Chang v. University of Rhode Island, 606 F. Supp. 1161 (D.R.I. 1985)
April 4, 1985
UNIVERSITY OF RHODE ISLAND, et al.
Diane R. SELEEN, et al.
BOARD OF REGENTS FOR HIGHER EDUCATION OF RHODE ISLAND, et al.
BOARD OF GOVERNORS FOR HIGHER EDUCATION OF RHODE ISLAND, et al.
BOARD OF GOVERNORS FOR HIGHER EDUCATION OF RHODE ISLAND, et al.
United States District Court, D. Rhode Island.
*1162 *1163 *1164 *1165 *1166 *1167 Abedon, Michaelson, Stanzler & Biener, Milton Stanzler, Julius C. Michaelson, Jordan Stanzler and Lynette Labinger, Providence, R.I., for plaintiffs.
Nicholas Trott Long, Gen. Counsel, University of Rhode Island, Kingston, R.I., Adler, Pollock & Sheehan, John F. Bomster, Edward L. Maggiacomo, Richard G. Galli, Michael Kelly, Providence, R.I., for defendants.
SELYA, District Judge.
While equality of opportunity was prominent in the collective minds of the Founding Fathers, it took nearly two centuries for the United States to recognize and act upon the existence of discrimination in the workplace based on race, religion, and sex. Congress attempted to move the nation closer to the accomplishment of our national goals by passage of the Equal Pay Act of 1963, Pub.L. No. 88-38, 77 Stat. 56 (codified as amended at 29 U.S.C. § 206(d) (1982)) (Equal Pay Act) and Title VII of the Civil Rights Act of 1964, Pub.L. No. 88-352, §§ 701-18, 78 Stat. 241 (codified as amended at 42 U.S.C. §§ 2000e-2000e-17) (1982) (Title VII). These laws were designed to reduce the discrimination which festered in American industry. Yet, Congress proceeded with understandable caution; the initial sweep of the statutes was not all-encompassing. Among those who were excluded from the prophylaxis of the laws were professional employees of educational institutions. It was not until 1972 that this gap was closed with the passage of the Equal Employment Opportunity Act of 1972, Pub.L. No. 92-261, 86 Stat. 103 (codified as amended at 42 U.S.C. §§ 2000e-2000e-17) (1982) (EEOA).
The cases at bar straddle this temporal span. They have their roots in a pattern of academic staffing which long antedated the enactment of the remedial legislation referred to above, but they have bloomed in the fertile soil of Title VII as enriched by the EEOA amendments. They require the court to strike a delicate balance between the rights of faculty members at the University of Rhode Island (URI or the University) and URI's fulfilment of its mission to educate and serve the people of Rhode Island. And, in calibrating this balance, the court must perforce tred in areas which, in simpler times, were deemed to be the sole province of educators and academicians. At bottom, these cases call upon the court to decide whether URI has engaged in a pattern and practice of discrimination in derogation of the rights of women faculty at the University.
I. THE PARTIES
A. The University
URI is the crown jewel in Rhode Island's coronet of state-supported higher education. Its main campus is situated in Kingston. The University also maintains a satellite campus located adjacent to the state capitol building in Providence (largely devoted to continuing adult education programs) and a unique facility on Narragansett Bay for its oceanography school. As the state's sole land-grant institution, URI operates various agricultural research centers as well as the state cooperative extension program.
As a state institution, the University is subject to the bureaucratic process. Its funding derives substantially from appropriations made by the state legislature (the Rhode Island General Assembly). The overall policies of the University, as well as its general management oversight, devolve from the Board of Governors for Higher Education. The Board consists of eleven members, eight of whom are appointed by the governor with the advice and consent of the state senate; the remaining members serve ex officio. R.I.Gen.Laws §§ 16-59-1, 16-59-2 (1981). The Board is empowered inter alia to establish broad policy necessary for the implementation of goals for higher education in Rhode Island, to formulate budget requests for the institutions of higher education, to appoint presidents of the universities, and to enforce all laws related to higher education. R.I.Gen. Laws §§ 16-59-4(1)-(10) (1981). The Board, however, is statutorily prohibited *1169 from operating or administering any subordinate college or university. Id. at § 4(3).
The day-to-day management of the University is vested in the school's administration. The operational hierarchy at URI is pyramidal in nature, with the president straddling the pinnacle of the obelisk. The president's duty is to ensure that the mission of the University, as defined by the Board, is carried out. During the currency of these events, there have been four chief executives at URI. Werner Baum served from prior to the passage of the EEOA until his resignation in 1973; William Ferrante, who was the vice-president for academic affairs (VPAA) at the time of Baum's resignation, was then designated as acting president and served in that capacity until August 1, 1974; Frank Newman became president at that time and continued in that capacity to August 1, 1983; and Ferrante, who had resumed his service as VPAA during Newman's reign, reemerged as acting president for a brief interval until Edward Eddy took office in 1984.
Historically, the president has been assisted by a number of vice-presidents. Among these plenipotentiaries, the VPAA is of paramount importance to this litigation. The VPAA is in charge of the scholastic programs at URI. His immediate subaltern is the assistant vice-president for academic affairs (AVPAA), whose major responsibilities include oversight of faculty hiring, setting salaries at hire, and allocation of teaching personnel to the various colleges and departments within the University.
Once overall policy has been formulated, its implementation is largely in the hands of the deans of the respective colleges. Sundry associate and assistant deans work under each of these administrators. The responsibility delegated to these officials varies from college to college, and depends to some extent on the particular dean.
The final level of administration is departmental. The department chair must implement administration policy within the particular academic enclave. This includes on-line budgetary responsibility, allocation of faculty, personnel management, curriculum development, and the like. In addition, the chair has first-strike responsibility for the evaluation of faculty members assigned to the department, and has considerable say in such momentous matters as promotion and tenure. The department chair has a foot in each of two sometimes opposing camps: the chair is both management and labor, both an administrator and a full-time faculty member. While other administrators hold faculty appointments, none are personally involved, in any appreciable way, in the day-to-day grind of teaching and research.
The plaintiffs have sued a melange of defendants; in fact, somewhat different defendants are named in each suit. These include URI, Board-R, Board-HE, the Rhode Island Department of Education, Carlotti, Dr. Richard Weeks (dean of the college of business administration), and Dr. Barbara Tate (dean of the college of nursing). All of the defendants are represented by the same battery of lawyers. For ease in reference, they will be grouped under the rubrics "URI" or "the University," except where the context plainly requires otherwise. No departmental chair has been sued as such.
B. The Plaintiffs
The named plaintiffs in these consolidated actions are or were faculty members at URI at various times from 1972 to the present. There are four separate actions consolidated for trial. Two of these are class actions (which overlap somewhat).
*1170 The seminal case (the Chang action) was filed in early 1977 by Lucy Peng-Fei Chang, who was terminated as an instructor in the college of business administration at the end of the 1972 academic year. She sought to represent a class of women who had been, are, or could have been, faculty at URI. A kindred class action (the Seleen action) was brought two years later by four women faculty members at URI. The quartet included Diane Seleen and Greta Cohen, both associate professors of physical education; Sharon Strom, a full professor of history; and Judith Anderson, a full professor in the department of speech communication. They also requested the court to allow them to represent a class of women who were employed at URI, are currently employed, or could have been employed after January, 1976.
In addition to these individuals, the American Association of University Professors-University of Rhode Island Chapter (AAUP) joined in the Chang and Seleen suits. The AAUP was designated as the exclusive bargaining agent for all faculty and professional library staff at URI in 1972, and has continued to act in that capacity. The AAUP negotiated a series of collective bargaining agreements, the first of which took effect in 1972; as we will see, these pacts had a substantial impact upon the administrative policies and personnel practices of the University.
In early 1983, almost four years after the Seleen action was commenced, Sandra Kraynek, an assistant professor of nursing, brought suit. She sought only individual relief for alleged violation of the Equal Pay Act. Soon thereafter, Wendy Roworth, an associate professor of art history, filed suit alleging violations of Title VII. She purported to sue both in her individual capacity and as a putative class representative (but, inasmuch as class certification was never sought, her action stands as a personal claim only). All of the plaintiffs share common counsel.
These, then, are the various major actors and actresses in this long-playing drama. With the main players in mind, the court turns its attention to the winding path so tortuously followed by this litigation.
While it can persuasively be argued that the events which led to this litigation began many decades ago, the stage was set for opening night when the defendant Weeks became the dean of the college of business administration in 1970. Weeks promptly terminated all faculty members not (i) holding doctoral degrees or (ii) then enrolled in doctoral programs. His professed aim was to upgrade the quality of the college. One of those affected by the new policy was Chang, then an instructor of mathematics and statistics. Following her termination, Chang sought relief from the Equal Employment Opportunity Commission (EEOC). The EEOC, as was and is its policy, see Black v. Brown University, 555 F. Supp. 880, 881 (D.R.I.1983), sent the complaint to the Rhode Island Human Rights Commission (RIHRC) for investigation. The RIHRC found no violation. The EEOC did, however, issue a right-to-sue letter. Chang, after receipt of this billet-doux, turned to the courts.
Chang's suit alleged that URI discriminated against her on the basis of gender both in terminating her contract and theretofore, in paying her under scale during her employment. She also alleged that URI's failure to rehire her at a time when the University, as she viewed it, engaged males with equal or lesser qualifications constituted discrimination in violation of Title VII. Her complaint contained allegations that her exposure to the pernicious effects of discrimination was but one example of a pattern and practice of disparate treatment which URI routinely utilized to *1171 the detriment of women faculty with regard to recruitment and hiring, rank at hire, pay at hire, promotion, annual compensation, tenure, and termination since March 24, 1972 (the effective date of the EEOA amendments and the date upon which Title VII became applicable to the Chang class).
The lawsuit filed by the Seleen litigants is strikingly similar to the Chang action in its pattern and practice averments. There are differences, of course, in the dates from which the litigation seeks class certification and in the claims of discrimination allegedly visited upon the putative class representatives. And the Seleen litigants only challenged URI's practices with respect to annual compensation, promotion, and tenure.
After the commencement of these suits, motions were filed seeking class certification pursuant to Fed.R.Civ.P. 23. In addition, the plaintiffs sought consolidation of the two actions pursuant to Fed.R.Civ.P. 42(a). In an unpublished order and memorandum dated September 2, 1980, Judge Pettine granted consolidation and certified the following class:
All women faculty members who are now employed at URI; who might become employed at URI; who were employed at URI on or after March 24, 1972; and all women applicants for faculty employment on or after March 24, 1972.
URI did not placidly acquiesce in class certification. The University repeatedly requested reconsideration and renewed its challenge to the timeliness of the Chang litigation. See note 4, ante. Judge Pettine rejected these initiatives on every such occasion. Subsequent to the rendition of the decisions of the Supreme Court in two significant Title VII cases, Delaware State College v. Ricks, 449 U.S. 250, 101 S. Ct. 498, 66 L. Ed. 2d 431 (1980) and General Telephone Co. v. Falcon, 457 U.S. 147, 102 S. Ct. 2364, 72 L. Ed. 2d 740 (1982), the judge agreed that reconsideration might be in order. URI also filed a motion for summary judgment under Fed.R.Civ.P. 56, based on these newly-emergent precedents. Before a hearing could be scheduled, however, the cases were coincidentally transferred from Judge Pettine's calendar.
Thereafter, in Chang v. URI, 554 F. Supp. 1203 (D.R.I.1983), this court denied the defense request for brevis disposition, holding that the defendants had failed to negate the existence of genuine issues of material fact. Id. at 1206. And, in an unpublished memorandum and order dated March 3, 1983, this court undertook reconsideration of the class certification question in the afterglow of Ricks and Falcon. Since the decision to terminate Chang occurred in December of 1971 (when Weeks decided not to renew Chang's contract) and was communicated to her at that point, Ricks barred Chang's Title VII claim for discriminatory treatment anent her separation from service despite the fact that the dismissal was not to take effect until after March 24, 1972, see Ricks, 449 U.S. at 259, 101 S. Ct. at 504; and this court so held, awarding partial summary judgment for the defendants vis-a-vis Chang's termination claim. Chang v. U.R.I., C.A. No. 77-0070S, at ___ (D.R.I. March 2, 1983). This court noted, however, that Chang's compensation claim was not, on its face, time-barred under Title VII inasmuch as she had continued to receive periodic salary payments after March of 1972; and that a nexus existed between compensation, promotion, tenure, and termination. Id. at 11-14. Given that linkage, Chang's claim was found to be sufficiently typical of the plaintiffs in the class and she was held to be a person who would adequately represent class interests in the litigation. Id. at 15. Thus, this court declined the invitation to tamper with the contours of the class as previously certified.
Both of the individual actions were commenced in early 1983. The plaintiffs were and are women faculty members, and each was and is a member of the Chang/Seleen class. Roworth's suit alleged that she was discriminated against on the basis of sex because less qualified or equally qualified *1172 males received early promotion and tenure while she did not. Kraynek's action alleged violations of the Equal Pay Act and Title VII with regard to the compensation which she received as an assistant professor of nursing. In mid-1983, the Roworth and Kraynek cases were consolidated with Chang and Seleen for trial.
The court thereafter bifurcated the issues, electing to try the liability issues in advance of any consideration of damages. The culpability phase of the consolidated cases was tried to the court for approximately eleven weeks. A complete trial transcript was subsequently obtained and a briefing schedule was arranged. The liability issues were taken under advisement in mid-1984. The court having heard the testimony, reviewed the transcript, studied the myriad exhibits, and ingested the amplitudinous briefs of the parties, this opinion constitutes the court's findings of fact and conclusions of law. Fed.R.Civ.P. 52(a). The protracted time span to which the litigation related and the voluminous evidence (both documentary and testimonial) which has been proffered, coupled with the complexity of the questions presented, renders it virtually obligatory that these findings and conclusions be woven together as part of the warp and woof of a unitary fabric. So, the court abjures the neat compartmentalization of factual and legal matters in favor of an endeavor to spin a seamless, homogenized web.
III. THE STATUTORY FRAMEWORK
A brief exegesis of the statutory mosaic is desirable in order to place the issues and the court's findings of fact into perspective.
Congress enacted the Equal Pay Act in an effort to eradicate the archaic notion that men, because of the posturings encouraged by societal role-playing, should be compensated more generously than women. Central to the legislative anodyne was the concept that industry must pay equal dollars for substantially the same work, irrespective of the workers' gender. Corning Glass Works v. Brennan, 417 U.S. 188, 195, 94 S. Ct. 2223, 2228, 41 L. Ed. 2d 1 (1974).
Though it was an important early step, the Equal Pay Act was not a panacea. Discrimination in the workplace was not limited to compensation disparaties between men and women. Unwilling to settle for a partial loaf in the face of the burgeoning demands of a fast-changing America, Congress proceeded to enact Title VII. Broader in its focus, Title VII banned discrimination in: "compensation, terms, conditions, or privileges of employment, because of such individual's race, color, religion, sex, or national origin." 42 U.S.C. § 2000e-2(a) (1) (1982).
While these legislative initiatives did not magically transform the employment land-scape overnight, they had their intended effect. Discrimination in employment began gradually to abate. In their initial incarnation, however, these acts were less than all-embracing. One notable exception to the reach of Title VII was the class comprised of the faculty of educational institutions. Powell v. Syracuse University, *1173 580 F.2d 1150, 1154 (2d Cir.), cert. denied, 439 U.S. 984, 99 S. Ct. 576, 58 L. Ed. 2d 656 (1978). And, to complicate the problem, the unamended version of the Equal Pay Act did not apply to professional employees of educational institutions. Compare 29 U.S.C. § 213(a) (1970) with 29 U.S.C. § 213(a) (1982).
For those concerned with secondary education, 1972 was a watershed year: Congress passed both the EEOA and the Education Amendments of 1972, P.L. No. 92-318, Tit. IX, 86 Stat. 235, 373 (Title IX). The former removed the exemption theretofore contained in Title VII for faculty members of educational institutions. Powell v. Syracuse University, 580 F.2d at 1154. And, Title IX left educational institutions subject to the dictates of the Equal Pay Act with respect to remunerative discrimination on the basis of sex. The regulations implementing Title IX required, inter alia, that each college and/or university conduct self-evaluation studies in order to root out discriminatory practices in all areas (including faculty employment). See 34 C.F.R. § 106.3 (1984). The effective date of the EEOA was, as noted above, March 24, 1972. Title IX became applicable to educational institutions on July 1 of that year.
Statutes rarely exist unembellished by subsequent interpretation, and the laws at issue here are no exception to this rule. Various executive orders (at both state and federal levels) and agency regulations have been promulgated to assist and insure compliance with Title VII and with the Equal Pay Act. These refinements play an important part in constructing the backdrop against which this litigation has been played out. Perhaps the most significant of these mandates was Executive Order 11375 (FEO), issued by President Johnson in 1967. It served to extend Title VII's coverage to all federal government contractors. Given the level of federal involvement in the funding of higher education and research, the FEO implicated many universities; URI wasand remainsa "federal contractor" within the reach of the FEO. Shortly after issuance of the FEO, the EEOC adopted a myriad of regulations, requiring federal contractors, among other things, to sponsor affirmative action plans and to designate affirmative action officers. See 29 C.F.R. §§ 1608.1-.12 (1984).
Rhode Island lagged behind. But, on May 16, 1972, then Governor Licht issued Executive Order No. 27 which prohibited discrimination on the basis of sex by any state agency. The governor further explicated this mandate on August 28, 1972 when he issued Executive Order No. 32, an edict which contemplated the formulation of equal employment plans and the nomination of compliance officers. Later that year, the Board adopted regulations requiring the appointment of a full-time affirmative action officer (AA officer) at each state institution of higher education.
This court's jurisdiction over the subject matter of the claims presented cannot be gainsaid. The bases for that jurisdiction need only be summarily stated.
These actions allege inter alia violations of Title VII, the Equal Pay Act, and 42 U.S.C. § 1983. 42 U.S.C. § 2000e-5(f) (3) provides for federal court jurisdiction over unlawful employment practice claims. This court also has jurisdiction over matters pertinent to the Equal Pay Act pursuant to 29 U.S.C. § 216(b). Actions brought under 42 U.S.C. § 1983 fall within the court's federal question jurisdiction as provided in 28 U.S.C. § 1331.
The defendants continue to dispute the extent of the court's jurisdiction, in that they persist in their objection to certification of a class which dates back to March 24, 1972. But, though their post-trial briefs preserve the point, the defendants do not offer any neoteric legal theories or newly-emergent caselaw. The court has disposed of this matter once, Chang v. URI, C.A. No. 77-0070S, slip op. (D.R.I. March 2, 1983), and sees no need to repastinate this tired ground. The court is satisfied that its jurisdiction at least coincides with the issues as framed by the cases in their present posture.
*1174 V. COMING OF AGE AT URI
The confluence of events which occurred within a telescoped period in calendar 1972, e.g., the enactment of the EEOA amendments and Title IX, the twin executive orders issued by Rhode Island's governor, and the promulgation of regulations by the Board-R, did not go unnoticed or unremarked at the University.
A. The Administration's Response
URI responded to the new laws and regulations in a punctual (if less than complete) manner. Joseph Rocha (about whom more will be told, e.g., text post at Part V(B)) was hired in 1973 to handle a potpourri of duties, including oversight of affirmative action at the University. (Though Rocha was not URI's first AA officer, see text post at Part V(B), his predecessor's involvement was so fleeting as to be insignificant.) Rocha's first order of business was promulgation of an affirmative action plan. The results of this labor, Pl. 3-1026, detailed the procedures to be followed in recruiting and hiring women and minorities into faculty and other professional positions.
To ensure compliance with the plan, Rocha claimed a right to approve the subsequent hiring of every new faculty member. While he was unable to object on the basis of qualifications or to demand that someone else be hired, he enjoyed (on paper, at least) a veto power as to whether the department and the administration had followed the procedures outlined in the plan. If Rocha failed to approve an appointment, URI was theoretically debarred from hiring the individual and from placing the person on the payroll. Rocha also gathered information on sundry hires to check URI's progress in meeting the goals of the affirmative action plan. This data helped Rocha compile affirmative action reports and reply to inquiries from adminstrators, faculty, and various state and federal agencies.
Coetaneous with Rocha's ascent to the affirmative action office, various members of the AAUP became vocal about perceived salary differential between male and female faculty members. During the negotiations which led up to the 1972 collective bargaining agreement, salary data had, for the first time, been made available to the faculty at large. Armed with these statistics, the AAUP formed a subcommittee to study the problem. Scott Desjardin, a professor of physics and president of the AAUP in 1972, headed the panel. Anderson was also a member. The group, which came to be known as the "Salary Inequities Subcommittee," compiled a report (SIS Report) which it made public on October 9, 1973. The SIS Report revealed that only 19% of the URI faculty, which then numbered 715 in all, were women; their average annual salary was approximately $4,000 less than their male colleagues.
The SIS Report was received by the administration of the University with considerable disdain. Those in authority pointed out, with indisputable accuracy, that it did not (i) distinguish among ranks in certain fields, (ii) include hires for the 1973-74 academic year, or (iii) treat calendar and *1175 academic year salaries on an equivalent basis. Despite these obvious flaws, the SIS Report caused a commotion among women AAUP members. Their demands for corrective action came to a head during the following year.
In 1974, the original collective bargaining agreement expired. Negotiations took on a fevered quality in an effort to reach agreement as to a new pact before the start of the 1974-75 academic year. A focal point of the bargaining was the SIS Report and the brewing storm over alleged inequities in distaff salaries. During the negotiations, the AAUP proposed the formation of an AAUP/URI joint committee to study the existence and extent of inequities and to devise solutions. The Board's negotiating team included the AAUP proposal in the tentative contract draft which was submitted to the Board in April of 1974. But, acting upon the advice of counsel, the Board rejected the joint committee concept.
In the counterproposal to the original 1974 contract draft, the Board suggested in substance that it undertake, in conjunction with URI, a study of putative salary inequities. The Board also promised to concoct a scheme to correct any imbalances revealed by the study. The AAUP acquiesced in this idea, although it was not memoralized in the signed collective bargaining agreement which ensued. The executed agreement did include a provision to adjust some specific salary inequities.
The task of carrying out the Board's assurances was delegated by Newman, in the first instance, to Rocha. In December of 1974, Rocha submitted his proposals to the administration. His core concept was the suggestion that URI construct a detailed "matched pairs" analysis analogous to the technique developed by the United States Department of Labor (U.S.D.O.L.) for the investigation of alleged violations of the Equal Pay Act. Cf. 29 C.F.R. §§ 800.114-.151 (1984). Rocha also recommended institution of a formal mechanism to unearth, adjust, and review inequities. Pl. 3-1065. The Rocha report generated little immediate interest after it was forwarded to Newman. Shortly thereafter, Rocha was discharged. His replacement was Barbara Brittingham.
At the time of her appointment, Brittingham was the assistant director of URI's Curriculum Research and Development Center. During her tenure as the University's affirmative action officer, she continued to hold her former position and to carry out her responsibilities at the Center. Since Brittingham was a part-time AA officer, her nomination to the post transgressed the Board's regulations. See text ante at Part III. Newman, a wily and experienced bureaucrat, did not directly confront the question of putative salary disparities. Instead, he commissioned yet another study: Brittingham was charged with the task of examining male/female compensation at URI. Together with Thomas Pezzullo, John Long, Glenworth A. Ramsay, Andrea Panciera and Roy Ageloff, she conducted the first sophisticated statistical analysis of faculty compensation at the University. The group's findings and conclusions were memorialized in a document which is familiarly known as the Brittingham/Pezzullo report (B/P Report), released on March 1, 1977. The B/P Report uncovered, but did not truly explain, some salient differences between the salaries of men and women. The authors were careful not to characterize these discrepancies as evidencing discrimination. Brittingham *1176 and her colleagues went no further than to state that there were differences and to suggest additional perscrutation of the relative productivity of male and female faculty members.
The B/P Report quickly attained much celebrity on campus. Many persons objected to its methodology and/or to its findings; others (mainly women) objected to its premises. Perhaps the chief accomplishment of the B/P Report was to convince the AAUP and many of the University's female faculty members that there was a lack of commitment on URI's part to recognize and correct gender-based inequities.
Three months passed and yet another survey of URI's employment practices surfaced. This study was one of a number of investigative analyses which trailed in the wake of the regulations implementing Title IX. Those regulations required URI to identify any discrimination in various areas of the University community and to devise methods for ameliorating any such unfairness. Pursuant thereto, the administration from time to time appointed several committees to investigate different aspects of URI's policies and practices. One of these committees focused on discrimination vis-a-vis employment of faculty. That committee was chaired by Bernice Lott. Lott, a full professor of psychology, had demonstrated her concern in this area by her service as a participant in the women's inequities committee of the AAUP, and as a member of the University Affirmative Action Committee.
After an extensive probe into faculty employment practices, the results of the Title IX survey (Title IX Report) were forwarded to Newman in June, 1977. Despite the fact that the Title IX Report contained an exhortation for its immediate release, the administration dawdled and did not disseminate it until November of 1977. The major finding of the committee was that the number of women faculty at URI was abysmally low and should be increased. The Title IX Report limned various actions which could be implemented to ensure more equal employment. Some of these recommendations were acted upon, some were delayed or deferred, and others were rejected or ignored. Lott's committee sought a meeting with Newman to discuss the Title IX Report, but the president never deigned to reply to this request.
While all of these in-house studies were in progress, URI was simultaneously under investigation by the Wage and Hour Division of the U.S.D.O.L. That agency, charged with the enforcement of the Equal Pay Act, centered its inquiry on disparities in faculty salaries. In a fashion characteristic both of government and of educational bureaucracy, the indagation generated an abundance of correspondence and kindred paper-chasing. Nevertheless, the U.S.D. O.L. investigation highlighted a number of problems. See Pl. 4-4020. URI realized that the federal authorities would not relent without a struggle, especially in light of the complementary ammunition furnished by the B/P Report and the Title IX Report. URI decided belatedly to act.
During the 1977-78 academic year, Newman asked Douglas Rosie, the incumbent AVPAA, to preside over yet another committee. Labelled the "Salary Review Committee" (but recognizable on campus almost exclusively as the "Rosie Committee"), that body was charged with investigating pay inequities among women faculty members and making recommendations for specific salary adjustments to cure any such shortcomings.
The Rosie Committee numbered among its membership the University's incumbent AA officer (Sylvia Feldman), Professor Ramsay (a veteran of the group which had authored the B/P Report), associate dean *1177 Margaret Robb of the college of arts and sciences, and the University's vice-president for personnel (Ron Snyder). The troupe reported back in the summer of 1978 with recommendations for backpay awards to 53 women totalling approximately $55,000. Newman endorsed these suggestions, but he was unable to convince the Board to go to the legislature to obtain incremental funding. As an alternative, Newman realigned the University's budget in order to effectuate the payments.
Rosie recognized that the outcome of his group's endeavors was less than utopian. Despite best efforts there could have been oversights or errors in judgment. Thus, Rosie advocated the formation of a salary appeals board. The panel would, as he envisioned it, entertain protests from faculty members who felt slighted (whether through exclusion or inadequacy of redress) by the recommendations of the Rosie Committee. The AAUP concurred in this idea, but suggested that women not participate in the appeals process.
Newman, recognizing that the Rosie Committee had barely scratched the surface, urged that URI and the AAUP establish a merged body to investigate the procedure by which URI set starting salaries for new hires. The joint committee was formed. Its report recommended the systemization of salary setting. The committee members apparently believed that a more rigorous procedure for initial salary recommendations from the deans and department chairs would ameliorate some of the effects of market forces and other pressures which counselled toward stipendiary inequality. But, this was another dead end; after issuing the report, the joint committee never heard from Newman and, from aught that appears of record, nothing came of the recommendations.
To sum up, the mid-seventies witnessed a cornucopia of studies anent employment practices at URI, but these studies were rarely translated into meaningful action. In the vernacular, Newman's overall strategy during this period seems to have been to attempt to talk the problem to death. Be that as it may, any reasoned assessment of the situation cannot be limited to the top echelons of the URI administration; it is equally important to focus on the operation of the University's affirmative action office.
B. Affirmative Action Officers
When President Johnson promulgated the FEO, URI, became obligated to appoint an AA officer. While the regulations implementing the FEO did not mandate that the AA officer be full-time unless the size or scope of the "business" of the federal contractor so required, e.g., 41 C.F.R. § 60-2.22(a) (1983), the Board commanded the appointment of a full-time AA officer at URI in 1972. Despite these directives, however, the University had no AA officer at all until the spring of 1973 when Baum appointed Monroe Moseley to the post. Moseley, who served on a part-time basis, had a very brief stay in the position. When Baum abandoned the presidency, Moseley also departed. Ferrante became acting president and hired Rocha as the University's labor relations and equal employment specialist. Rocha had no experience in equal employment matters prior to his arrival at URI. His previous position was a professorship in labor economics at Ball State University. Rocha handled various matters for URI with respect to the recently negotiated collective bargaining agreement. Notwithstanding that Rocha's title was that of special assistant to the president and though his job description involved oversight of labor relations matters as well as equal employment, he was the de facto AA officer. The special assistant designation effectively removed the holder of the office from the state's civil service system. Like Moseley, Rocha (because of his responsibilities in the labor arena) could not devote his undivided attention to the functions of an AA officer. Newman retained Rocha for a relatively short time before disenchantment set in. Rocha was fired in 1975.
As previously noted, Brittingham was next in line. Like her predecessors, Brittingham was hired as AA officer without a *1178 search and given the title of special assistant to the President. Brittingham's duties qua special assistant entailed only the oversight of affirmative action; yet, she, too, was part-time in view of her continued work at the Curriculum Research and Development Center. Despite her inability to work full-time as the AA officer, Brittingham developed a systematic procedure for use anent hiring, along with a compendium of accompanying forms. She, like Rocha, possessed the authority to prevent URI from engaging someone as a faculty member if she discovered that the procedures developed to insure non-discriminatory hiring had been breached.
Brittingham served as URI's AA officer to June 30, 1976. At that time, Feldman assumed office after a meticulous nation-wide search had been conducted. Unlike those who went before, Feldman had excellent professional credentials and demonstrable experience in the equal employment field. Prior to embarking on her stint at URI, Feldman had worked in a high-level position for the Equal Employment Opportunities Office of the New Jersey Department of Education. After four years and three AA officers, URI had finally bowed to the imperatives of the Board's regulations and had secured the services of a full-time AA officer after an acceptable search.
Feldman's tenure at URI was, by all accounts, a stormy one. An undercurrent of discord permeated her relationship with Newman: they disagreed as to many policies, and Feldman perceived obstacles in her path in the implementation of diverse equal employment and affirmative action initiatives. The flip side of the coin was Newman's belief that Feldman spent too much time in organizing conferences and in heightening awareness and not enough time in the nuts-and-bolts search for qualified minority faculty candidates (including females).
These undercurrents bubbled to the surface after Feldman planned, scheduled, and coordinated yet another conference on women in academia. In a press meeting after the convocation, Feldman aired some derogatory comments about Newman and his policies insofar as they pertained to affirmative action. URI's student newspaper reported the incident and printed quotations in this vein which it attributed to Feldman. The AA officer sensed trouble in the wind. She told Newman that the newspaper had taken her remarks out of context. Although the versions of this audience given by the protagonists vary, it is clear to the court that Newman voiced an opinion as to Feldman's inadequacy as an AA officer and that she became highly agitated and left the meeting in some haste. Feldman lost little time in tendering her resignation, and Newman lost even less in accepting it. When Feldman's passion cooled and she sought to withdraw her resignation, Newman demurred. He was unmoved by an outpouring of support for Feldman from the University's women faculty members.
Despite the fact that Newman accepted Feldman's resignation some eight months prior to her departure, the University's budget for the 1980-81 academic year made no provision for the position of AA officer. Thus, whether by his own doing or by the budgetary machinations of the Board or of the General Assembly, Newman *1179 felt that he was restricted in choosing Feldman's successor to in-house staff who could be spared on a part-time basis. Steven Hines, an employee of the University's personnel office, was appointed as a part-time AA officer on an interim basis.
On March 1, 1981, Hines was supplanted by Harold Wingfield. Wingfield wore three hats: he took on the AA officer's responsibilities in addition to his duties as the acting director of the University's black studies program and as an assistant professor of political science. Wingfield, despite the proliferation of his millinery adornment, had some credible background in affirmative action inasmuch as his doctoral dissertation had analyzed equal employment programs in Oregon colleges and universities.
The relationship between Wingfield and Newman was less than idyllic. Although Wingfield and Newman met frequently, and Newman was wont to intercede on Wingfield's behalf in certain matters, Newman often failed to respond to the AA officer's complaints about the hiring of people without proper regard for affirmative action protocol. While Wingfield's memory of events was faulty in certain respects, e.g., XXIX at 1-44, 45, other evidence corroborated the occurrence of such incidents and Newman's indifference thereto. These tensions and disagreements were, however, but the tip of the iceberg. A more fundamental conflict arose because Wingfield, according to Newman, sought both the authority to dictate the dimensions of the protected class from which URI should fill a given position and the power to select, from within that class, the individual to be hired. Newman believed, understandably, that Wingfield's approach, if sanctioned by the administration, would lift the lid from atop Pandora's jar. When this gantlet was thrown down in early 1981, Wingfield decided to return to teaching. Newman accepted the decision with no visible signs of regret.
Wingfield's exit led, in turn, to the appointment of Harold Smith as acting AA officer while the search continued for a permanent replacement. At the time of his engagement, Smith had been the director of minority affairs at URI. During the pendency of the search, Smith held both positions, each on a part-time basis. In March of 1982, Smith was chosen as the full-time AA officer. He became at that time only the second full-time AA officer at URI in the decade following the governor's edict (State Executive Order No. 32) and the Board's response.
Smith's term of office at URI was a learning experience both for him and for the faculty. He met with each department chair to discuss affirmative action. He continued to follow past precedent in approving advertisements for position vacancies and in reviewing the various forms that departments filled out in order to comply with affirmative action search procedures. He oversaw the mechanism by which offers were tendered. On the whole, Smith appeared to be conscientious in his job, especially given the limited amount of assistance provided by URI. He made an honest, diligent, good faith effort to fulfill the important responsibilities of the post and to achieve the legitimate objectives of affirmative action. Smith still serves as the AA officer at URI.
In fine, during the period covered by this litigation URI has had seven AA officers, only two of whom were employed full-time in the position. At best, this evinces an attitude that affirmative action officers were a necessary evil mandated from above, and suggests that URI had only grudgingly accepted the realities of Title VII and of kindred legislation. And, notwithstanding the continual and flagrant violation of federal, state, and Board mandates in the constitution of the University's affirmative action office, no sanctions of any kind have ever been levied on URI by the Board or by state officialdom. All indications are that the laissez-faire attitude of the University's administration was, if not a mirror image of a similar lassitude on the part of the Board, at least ratified by acquiescence.
*1180 C. Affirmative Action Efforts at URI: Reporting Requirements
The prime responsibility of URI's AA officer has been, as one might suspect, the oversight of the University's affirmative action efforts. Affirmative action has been defined as action to correct the effects of past discrimination and to prevent present or future discrimination without the ersatz motivation of litigation. 29 C.F.R. § 1608.1(c) (1984). To carry out affirmative action, employers are required to promulgate plans which will correct past discrimination and which tend to ensure equal employment in the present and future. The AA officer at URI was charged with collecting the data for the plan, designing the affirmative action program, and supervising its implementation.
Performance of these interrelated tasks necessitated the compilation of plethoric documentation: a summary of the basic data on all employees at URI, statistical delineation of labor pools for women and minorities, comparison of female and minority workers at URI with their availability within the relevant labor pools, and an analysis of men's and women's salaries and other benefits at the University. In the absence of this minimal information base, URI could not sensibly shape an affirmative action policy. And, even then, formulation of the policy alone was insufficient; URI also had an obligation to communicate the policy, design a plan to achieve the policy ends, and implement the resultant plan. Oversight of implementation required, at the very least, the installation of an internal audit system and the institution of some form of grievance procedure.
The attempt to achieve compliance with these requirements was a long and arduous process at URI. And, the difficulties inherent in the effort were exacerbated by highlevel footdragging along the way. To cite but one example, the University's affirmative action plan required the president to issue an equal employment statement. Newman was aware of this stipulation as early as 1975, but did not sign the statement until 1980.
The requirement that URI develop an affirmative action plan devolved as early as 1971. Neither Baum nor Ferrante took any discernible steps in this direction. Newman, shortly after his assumption of office, set the wheels in motion. Despite the defendants' not-altogether-unfounded intimation that the EEOC was partially at fault, the fact remains that the first formal submission of an affirmative action plan to any governmental agency was not accomplished until 1976. The 1976 plan was approved by the state, but rejected by the EEOC. Four more years of intermittent effort ensued. In 1980, nine years in arrears, URI finally submitted a plan which merited the approval of the federal government.
Nor was the failure to effectuate timely and appropriate submissions limited to affirmative action plans. After 1975, employers such as the University were required to submit various reports to the EEOC concerning, among other things, applicant pools and hirings. See 29 C.F.R. § 1602.50 (1984). This directive necessitated the systematic collection and retention of data anent hiring, promotion, tenure, demotion, transfer, layoff or other termination, and rate(s) of pay. 29 C.F.R. § 1602.49 (1984). In 1976, Brittingham requested assistance in categorizing data from the regional office of the federal Department of Health, Education, and Welfare. Two years later, Snyder, the head of the University's personnel office, notified Feldman that there were problems in collecting and organizing data for the 1978 report to the EEOC. Snyder told Feldman that a software package was being readied for operation and that this innovation would alleviate data collection problems for 1979 and succeeding years.
The 1979 report was filed with the aid of the computer system. But, this report, too, was suspect. For example, the document purported to supply data on applicant flows. Yet the testimony in this trial, e.g., XXII at 54, Pl. 5A-5540, demonstrated that such information was not kept. It is plain to the court that the deficiencies in reporting *1181 were not attributable merely to problems in organizing data, but also to the failure of URI to maintain records adequate to permit the required presentation to be made.
Shortcomings in URI's compliance with reporting obligations were still extant in 1981. Wingfield notified Newman that the data needed for completion of reports was unavailable. The state equal opportunity agency was lugubriously informed about the lack of adequate data concerning, inter alia, applicants. In point of fact, Wingfield's criticism of URI's disorganized recordkeeping continued unabated throughout his tenure as the University's AA officer. No credible evidence was adduced that the situation had at any time materially improved.
To sum up, both reporting and recordkeeping were lax in the affirmative action area. Data was not centrally collected, nor was serious effort expended to this end. An inordinate amount of time was taken in the design of affirmative action plans. URI never managed to satisfy completely the paper generation requirements of the EEOC, nor can its failure to do so be lightly excused. Yet, since there is more to affirmative action than recordkeeping, the court must scrutinize the substance of the University's operational efforts.
D. Compliance with Internal Affirmative Action Procedures
Though URI did not obtain approval of its affirmative action plan until 1980, it nevertheless undertook interim measures. The University, for instance, implemented search procedures suggested by Moseley, Rocha, and Brittingham. The starting place was the preparation and the filing of a job description for an emergent vacancy (e.g., Form A as devised by Brittingham). Once Form A was approved by the AA officer, a nationwide job search could begin. Under the school's guidelines, special emphasis was to be placed on affirmative action. The departmental search committee was required to delineate its search process (especially the attempt to attract minority and women candidates) on Form B. The third form in the sequence, Form C, compiled data on the pool of applicants. Once the applications had been studied, the search committee would choose the individuals to be interviewed. This information was supplied on Form D. Besides listing the candidates, that form contained information on the number of women and minority candidates and described the rationale by which the interviewees had been selected. (Expense factors dictated that only a comparatively few candidates be interviewed at this stage.) Following completion of the interviews, the finalists were ranked, and a nominee was tentatively designated. Form E was completed (describing, inter alia, the minority status of the finalists and the search committee's justification for the ranking). At this juncture, the University's AA officer was required to approve the search procedure before the nominee was offered the vacant position. But, the record reflects that this methodology was betimes honored only in the breach.
In 1973, during Moseley's reign as AA officer, the botany department initiated a search for a temporary faculty appointee. The department used the "old boy" network, contacting friends at other universities and totally ignoring the affirmative action channels suggested by Moseley. Robert Lepper (then dean of the college of arts and sciences) was informed of the department's failure to conduct a search within the guidelines prescribed by Moseley. The record is barren of any evidence that the situation was corrected or even that the department was rebuked.
In approximately the same time frame, the department of biochemistry also ran afoul of affirmative action. A number of courses in the department were being taught by a graduate student (Peter *1182 Smith). His part-time position terminated at the end of the 1972-73 academic year. In anticipation of the vacancy, the department advertised to fill the position for the next ensuing (fall, 1973) term. An apparently qualified female, Dr. Claude-Marie Janeway, applied. Dr. Janeway was never considered for the position, however, and Smith was appointed sub rosa, in the guise of "special lecturer," to teach the courses. The incumbent AA officer (Rocha) discovered the contretemps and complained to A.A. Michel, who was then acting as the VPAA (Ferrante having been pressed into service as interim president of the University). Michel and Ferrante jointly interceded and forced a reexamination of the affair. They told the department that Rocha would decide whether a supposed departmental promise to Smith, allegedly made at an earlier date, effectively nullified the vacancy. Rocha decided that Smith had no vested claim on the job and that Janeway's application should be considered. The department grudgingly invited Janeway to present a paper, telling her, however, that she should not assume that she was being considered for an opening. Janeway visited URI, presented her paper, and went away empty-handed. No further action was taken by the administration.
Rocha then turned his attention to the college of business administration's department of organizational management and industrial research. In the spring of 1975, Dr. Craig Overton, the department chair, offered assistant professorships to Robert Comerford and Dennis Callaghan. Rosie put them on payroll without bothering to secure Rocha's prior approval. Rocha complained to Ferrante, who beseeched Overton to cooperate with the AA officer. An audit ensued. Rocha thereafter found that the department's failure to comply with affirmative action procedures was inadvertent. But, no further mention was made of Rosie's disregard for the niceties of affirmative action in this situation.
The psychology department was also a source of discontent during the spring of 1975. The department chair, Allan Berman, was interested in hiring Wayne Velicer. Rocha became suspicious that the advertisements for the vacancy were tailored to fit Velicer alone. This level of concern rose dramatically when he discovered that Velicer had been posted on the payroll well before the expiration of the application deadline. Upon due inquiry, Rocha received an explanatory letter from Berman which, at bottom, explained very little. The application deadline was kept open. The air was apparently cleared by Barry Marks, dean of the college of arts and sciences, who convinced Rocha that the advertisements were kosher. Rocha subsequently exonerated the psychology department of any wrongdoing. No explanation was ever given, however, as to why or how Velicer had been placed on the payroll before the application deadline passed, nor was any remediation undertaken.
Such bevues were not limited to the departmental level. Affirmative action procedures were glossed over at the highest reaches of the URI administration. In 1977, Ferrante, who had returned to his position as VPAA, became involved in the hiring of an acting dean for URI's extension division. Under the affirmative action plan, advertising was necessary. Ferrante ignored the requirement. After the fact, he wrote to Feldman, apologized, and explained what he had done to attract minority candidates. And, when all was said and done, Ann Byrne, a female, was hired for the position.
Not every instance of Ferrante's absentmindedness with respect to affirmative action had as happy an ending. Some two years after the Byrne incident, Ferrante was involved with the search for a dean of the college of pharmacy. Instead of submitting the mandatory Form E to Feldman in advance of tendering a position, Ferrante offered the job to Louis Luzzi; only *1183 afterwards did he file Form E. Once again, Ferrante begged Feldman's indulgence after the deed was done, but no substantive corrective action was taken.
Ferrante and Rosie had no exclusive on such oversights. Leonard Kahn was hired in the physics department based on an ad hoc promise. Wingfield, the AA officer at the time, wrote to Dean Robb and questioned the administration's allocation of a slot for Kahn. Robb brushed off Wingfield's inquiry in a wholly perfunctory manner.
None of these incidents, standing alone, have any spectacular impact; but collectively, these episodesand others of like ilkspeak volumes about the attitude of URI's administrators toward affirmative action. AA officers and their efforts were tolerated, nothing more. Scant recognition or encouragement was given to successful affirmative action efforts. Similarly, sanctions were not meted out when guidelines were violated. Reduced to its barest essence, URI's institutional attitude toward affirmative action appears not to have been one of outright hostility, but one of studied indifference.
E. Conclusions Anent Affirmative Action
The court has examined the affirmative action mandates in place at URI, the relevant history vis-a-vis affirmative action and AA officers, the latters' efforts at instituting affirmative action principles, and the University's corporate response to those efforts.
Faced with repeated calls for better treatment of women, URI's antiphony was muted and larghetto. It took almost four years to hire the first full-time AA officer; when she left, the University neither brought in a permanent replacement nor even budgeted the position for nearly two years. Almost half a decade passed before a study of the allegations of salary discrimination was implemented and released. Even then, the University attempted grudgingly to apply only the mildest of palliatives. Moreover, despite ominous signs that something was out of pitch, available avenues were explored only partially, or at a snail's pace, or not at all. Rocha's 1974 matched pair proposal was shelved for several years; the joint commission's plan to systematize initial salary offers was ignored; Newman declined to meet with representatives of Lott's committee to discuss the Title IX report. URI's failure to file timely and adequate annual plans, its enduring neglect to maintain satisfactory records, and the repeated mishandling of paperwork in this sensitive area, were all of a piece with this pattern of behavior. Mistakes can and do happen; but the miscues here were too numerous, too frequent, and too one-sided simply to be shrugged off.
The University's response, taken as a whole, amounted to a sort of affirmative inaction, indicating a begrudging acceptance of the congressional mandate and a resistance to change. Certainly, the prevailing winds in the URI atmosphere were not hospitable to gender-neutral decisionmaking. And the Boardwhich, for example, never sanctioned URI for its patent and long-enduring failure to comply with a Board-imposed regulation mandating the presence of a full-time compliance officer exhibited no more discernible vigor in the pursuit of the goals and ideals of Title VII.
Although the failure of an employer to adopt or effectively to implement an affirmative action plan may be probative of discriminatory intent, Coser v. Moore, 739 F.2d 746, 751 (2d Cir. 1984); Craik v. Minnesota State University Bd., 731 F.2d 465, 472 (8th Cir.1984), the lack of an affirmative action plan or the poor implementation of one does not, by itself, prove classwide discrimination. Metrocare v. Washington Metropolitan Transit Authority, 679 F.2d 922, 929 (D.C.Cir.1982). Yet, given what has occurred in the past at URI, and an environment in which biased decisionmaking, while not condoned, was tolerated, the court must view the evidence on the issues at bar in a light shaded to some extent by skepticism. In other words, the University's overall record in this area is sufficiently blemished that heightened scrutiny is in *1184 order in an attempt to gauge whether or not the same vices infected other employment practices at URI. But, the burdens of proof do not change.
VI. LEGAL STANDARDS: AN OVERVIEW
Before venturing further into the thicket of URI's employment practices, it seems wise to address the allocation of the burden of proof and related mattersas all of the court's factual findings must ultimately be weighed and measured against this verdant backdrop.
A. Title VII
This litigation, as it now stands, is a stew composed of diverse ingredients. Chang and Seleen are certified as class actions, while Roworth and Kraynek are individual suits. The dissimilarities in the forms of action are significant with respect to the burdens and dimensions of proof.
The parties in the cases at bar recognized this overriding issue, yet attempted to slip it entirely within the integument of McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S. Ct. 1817, 36 L. Ed. 2d 668 (1973) as clarified in Texas Department of Community Affairs v. Burdine, 450 U.S. 248, 101 S. Ct. 1089, 67 L. Ed. 2d 207 (1981). The Burdine model, applicable to individual claims, first requires the plaintiff to establish a prima facie case by demonstrating the presence of four elements. Though precise definition may vary slightly depending upon the particular employment action which has come under scrutiny, the paradigm remains much the same. Thus, in a failure to hire case, for example, the quadrat of facts which an allegedly aggrieved plaintiff would have to prove would be essentially as follows: (i) membership in a protected class; (ii) suitable qualifications for the position in question; (iii) rejection notwithstanding possession of the necessary qualifications; and (iv) continued availability of the position and/or the hiring of a non-class member. McDonnell Douglas, 411 U.S. at 802, 93 S. Ct. at 1824. If proven, these facts give rise to an inference that the applicant was rejected for discriminatory reasons, Burdine, 450 U.S. at 253, 101 S. Ct. at 1093, creating a mandatory (but legally rebuttable) presumption that the employer unlawfully discriminated. Id. at 254 & n. 7, 101 S. Ct. at 1094 & n. 7.
The employer can rebut the presumption by producing evidence that "the plaintiff was rejected, or someone else was preferred, for a legitimate, nondiscriminatory reason." Id. The employer's burden is one of production only, that is, of adducing evidence sufficient to allow the factfinder reasonably to conclude that the employment decision was not motivated by discrimination. Id. at 257, 101 S. Ct. at 1095. "[I]f the trier of facts believes the plaintiff's evidence, and if the employer is silent in the face of the presumption," Burdine, 450 U.S. at 254, 101 S. Ct. at 1094, judgment must enter in favor of the claimant. If, however, the employer has articulated a legitimate, nondiscriminatory reason, the presumption of discrimination "drops from the case," id. at 255 n. 10, 101 S. Ct. at 1095 n. 10, and the claimant's burden of rebutting the proffered reason "merges with the ultimate burden of persuading the court that [the plaintiff] has been the victim of intentional discrimination." Id. at 256, 101 S. Ct. at 1095. The plaintiff can satisfy the devoir of persuasion at this point only by establishing that a discriminatory intention more likely than not motivated the employer (a task which is frequently accomplished by showing that the employer's proffered explanation is unworthy of credence). Id. The ultimate burden of proof remains with the plaintiff. White v. Vathally, 732 F.2d 1037, 1040 (1st Cir.1984).
Class action litigation implicating Title VII is, however, a fundamentally different breed of cat. The essence of a class action is an endeavor to prove that there was a systemwide pattern or practice of disparate treatment or that facially neutral policies had the effect of impacting one or more protected classes in a disparate and less favored way. International Brotherhood *1185 of Teamsters v. United States, 431 U.S. 324, 335 n. 15, 97 S. Ct. 1843, 1854 n. 15, 52 L. Ed. 2d 396 (1977). In contradistinction to individual discrimination claims, which are by their nature fact-intensive as to the details of a particular incident, class actions almost invariably seize upon statistical evidence as a tool to prove or to disprove discrimination. E.g., International Brotherhood of Teamsters, 431 U.S. at 339-42, 97 S.Ct. at 1856-58; EEOC v. American National Bank, 652 F.2d 1176, 1188 (4th Cir.1981), cert. denied, 459 U.S. 923, 103 S. Ct. 235, 74 L. Ed. 2d 186 (1982); Melani v. Board of Higher Education, 561 F. Supp. 769, 773, 781 (S.D.N.Y.1983). Although a sequential shifting of the burden of production is eminently suitable for individual discrimination claims and, with appropriate modification, can facilely be applied in an academic milieu, see generally Zahorik v. Cornell University, 729 F.2d 85, 93-94 (2d Cir.1984), the adaptability of the classic Burdine model to class actions is quite another question. "In a complex class action, utilizing statistical proof and counterproof, the value of the Burdine sequence to highlight the issues in contrast is about as relevant as a minuet is to a thermonuclear battle." Vuyanich v. Republic National Bank, 521 F. Supp. 656, 661 (N.D.Tex.1981), vacated on other grounds, 723 F.2d 1195 (5th Cir.1984). The Court's most recent discussion of the issue, in Cooper v. Federal Reserve Bank of Richmond, ___ U.S. ___, 104 S. Ct. 2794, 81 L. Ed. 2d 718 (1984), is revealing. In Cooper, a quartet of individuals intervened in an action brought by the EEOC against the bank. The action postulated racial bias. Subsequent to the intervention, the district court certified a class. The class members were duly notified of the pendency of the action and of their opt-out rights. Fed.R.Civ.P. 23(c) (2). Certain women never attempted to exclude themselves from the class (although they received the notice). Five of them later attempted to intervene in the relief phase of the Cooper litigation. Though intervention was denied, the district court permitted them to sue as individuals under 42 U.S.C. § 1981. The court of appeals reversed, holding that principles of res judicata effectively blocked the § 1981 claims. EEOC v. Federal Reserve Bank of Richmond, 698 F.2d 633, 674-75 (4th Cir.1983). The Court took certiorari and reversed.
The Court noted the fundamental differences in proof between class actions and individual actions, emphasizing that the main inquiry regarding the latter concerns a particular employment decision, whereas in a class action, the liability phase focuses not on individual decisions, but on the existence vel non of a pattern of discriminatory decisionmaking. Cooper, ___ U.S. at ___, 104 S. Ct. at 2800. Thus, it is entirely possible that a plaintiff will be unable to prove that a pattern and practice of discrimination exists "even though discrimination against one or two individuals has been proved." Id. ___ U.S. at ___, 104 S. Ct. at 2801. The Court reasoned, therefore, that the rejection of a class-based suit for discrimination did not obviate the possibility that an individual class member's claim was actionable: a judgment for the employer in a class action discrimination suit bars the class members from instituting another class action designed to traverse the ground covered in the initial litigation and estops the class members (save only those who, after receipt of notice, opted out) from further addressing the same pattern or practice issues, whether or not by way of individual claims. Id. ___ U.S. at ___, 104 S. Ct. at 2802. But, the Court rejected the bank's contention that subsequent separate actions would per se frustrate the purposes of Fed.R.Civ.P. 23. The Court noted that a rule preventing later individual actions, questioning idiocratic employment decisions, would require that every class members be permitted to intervene in a class action and to litigate the precise merits of his or her claim according to the Burdine model. Id. Although the court should confront the evidence concerning the personal claims of the class representatives, and should consider other anecdotal evidence generally insofar as it bears on the existence or nonexistence of the asserted *1186 pattern or practice, it need not adjudicate the myriad possible claims of each and every witness and/or class member. Id.
The import of Cooper is clear: inasmuch as the proof of class-oriented pattern or practice discrimination is on a different order than that requisite for the prosecution of a personal Title VII claim, the Burdine model, so carefully crafted to fit the demands of the latter, is of dubious utility in respect to the former. But, though Cooper teaches that class actions are a breed apart, it is less instructive as to the architecture of structuring a paradigm whereby Rule 23 actions may be judged.
That very issue was anticipated by the Eighth Circuit in Craik v. Minnesota State University Board, 731 F.2d 465 (8th Cir. 1984). Craik involved a suit maintained by four female faculty members at St. Cloud State University who alleged that a pattern or practice of discrimination existed with respect to a ragout of matters (e.g., selection of departmental chairs, rank, compensation, appointment to administrative positions) not unlike those here sub judice. In addition to the class claims, each of the four class representatives sought relief for one or more specific instances of discriminatory conduct allegedly visited upon her. The district court certified a class similar to the class at bar, viz.:
females who are or have been employed by St. Cloud University in a teaching capacity and who have been, continue to be, or may in the future be discriminated against because of their sex with respect to promotion, compensation ... and other conditions and privileges of employment.
Id. at 468.
For purposes of the instant litigation, the Eighth Circuit's analysis of the concept of a prima facie case within the purview of a class action is illuminating. The appeals court reiterated the class action model statistical proof and counterproofas set forth in International Brotherhood of Teamsters, 431 U.S. at 360-61, 97 S. Ct. at 1867, and concluded that the Burdine approach possessed little validity in a Rule 23 setting. Craik, 731 F.2d at 470. See also Cooper, ___ U.S. at ___, 104 S.Ct. at 2800-02; Perryman v. Johnson Products Co., 698 F.2d 1138, 1143 (11th Cir.1983).
Functionally, a class action operates on an entirely divergent principle: an effort is made, generally by the use of statistics, to prove disparities in, or arising out of, certain employment practices. It is incumbent upon the plaintiffs to do more than establish the occurrence of isolated or sporadic acts of discriminatory behavior. International Brotherhood of Teamsters, 431 U.S. at 336, 97 S. Ct. at 1854. The plaintiffs must prove, by a fair preponderance of the evidence, that sex discrimination, intentional in nature, was "the company's standard operating procedurethe regular rather than the unusual practice." Id. Normally, the statistical analysis employed by the plaintiffs will control for nondiscriminatory reasons which might justify the employment decisions, and thereby attempt to show "proof of the expected result of a regularly followed discriminatory policy." Id., 431 U.S. at 360 n. 46, 97 S. Ct. at 1867 n. 46. A counter-thrust is then mounted to realign the statistics. The employer can rebut the prima facie showing of discrimination by discrediting the plaintiffs' figures or by otherwise demonstrating the inaccuracy or insignificance of the plaintiffs' proof. If the plaintiffs' evidence is telling, and if the employer neither successfully rebuts it nor explains it away, then the court can conclude that the alleged violations occurred. Cf. International Brotherhood of Teamsters, 431 U.S. at 361, 97 S. Ct. at 1867. In effect, failure successfully to rebut the plaintiffs' evidence establishes a prima facie case that discrimination existed on a classwide basis. Craik, 731 F.2d at 470. And, if that is the case, then (logically) the class representatives also were, prima facie, victims of discrimination. Id. (citing International Brotherhood of Teamsters, 431 U.S. at 359, 97 S.Ct. at 1866). Moreover, once such a prima facie case has been established, the burden shifts to the employer to show that a specific employment decision was more *1187 likely the product of a lawful reason than of discrimination. International Brotherhood of Teamsters, 431 U.S. at 362, 97 S. Ct. at 1868; Franks v. Bowman, 424 U.S. 747, 773 n. 32 (1976); Craik, 731 F.2d at 470-71 & n. 9. Craik's ratio decidendi in this regard is consistent with, and flows smoothly from, International Brotherhood of Teamsters, in which the Court held that it was reasonable to require an employer to come forward and dispel such a prima facie case. 431 U.S. at 359, 97 S. Ct. at 1868. As the Court noted, proof of a pattern or practice of discrimination establishes a perceptibly greater probability that a decision affecting any given individual class member was part and parcel of the overall discrimination. Id. at 355 n. 41, 97 S. Ct. at 1866. The finding of a prohibited pattern or practice transmogrifies the employer from a presumptively innocent litigant whose liability, though alleged, is not yet established, to a "proved wrongdoer." See id. The Court plainly intended this language to have greater import than merely the imposition of a burden of production. See Craik, 731 F.2d at 471 n. 9. A divergent view of the law would fairly eviscerate the meaning and utility of a classwide finding of discrimination.
So postured, individual claims can be considered only after the class claims have been addressed. Craik, 731 F.2d at 471. The class representatives' personal claims, however, require a modified Burdine analysis. See Craik, 731 F.2d at 471. If an invidious pattern or practice has been established, then URI must prove that it did not discriminate. On the other hand, if the proof has fallen short of reflecting such a pattern or practice, those plaintiffs, as to their personal contentions, must be judged under a conventional Burdine analysis.
To close the circle here, Roworth's action, which is being maintained solely as an individual claim, requires treatment akin to that accorded to the personal claims of the class representatives. (Because Roworth is a member of the certified class, she would be entitled to the benefit of any favorable presumption arising from the class litigation.) The same holds true with regard to Kraynek, insofar as her suit is viewed as implicating Title VII rather than the Equal Pay Act.
B. Equal Pay Act
The Equal Pay Act provides that work requiring equal skill, effort, and responsibility, performed under comparable working conditions, must be rewarded with equal wages. Corning Glass Works v. Brennan, 417 U.S. at 195, 94 S. Ct. at 2228; Winkes v. Brown University, 747 F.2d 792, 793 (1st Cir.1984). The initial burden of proof rests on the plaintiff to establish a prima facie case: that he or she was paid less than one or more fellow employees of the opposite sex who were performing jobs requiring equal skill, effort, and responsibilities, under substantially the same working conditions. Corning, supra; Winkes, supra. If the plaintiff succeeds in so doing, the burden of proof shifts to the employer to show that any pay differential is a result of either a bona fide system of (i) seniority, (ii) merit, or (iii) productivity (quantitative or qualitative), or some other (permissible) differentiating factor besides sex. 29 U.S.C. § 206(d) (1) (1984). See Corning, 417 U.S. at 196-97, 94 S. Ct. at 2229.
This framework is a straightforward one, but it interlocks with the broader issues of sex discrimination in a suit which amalgamates Title VII and Equal Pay Act claims, inasmuch as 42 U.S.C. § 2000e-2(h) provides that:
It shall not be an unlawful employment practice under this subchapter for any employer to differentiate upon the basis of sex in determining the amount of the wages or compensation paid or to be paid to employees of such employer if such differentiation is authorized by the provisions of section 206(d) of Title 29.
Thus, an employer who proves that a controverted salary differential is the product of a valid exemption to the Equal Pay Act absolves itself of liability therefor under Title VII as well. And, the statutory rule *1188 must perforce apply with the same vitality to a class action as to an individual suit.
This enactment raises an interesting question: if the presence of a bona fide exception to the Equal Pay Act serves equally as a defense to Title VII liability, does Title VII somehow shift its burden of proof in rhythm with that prescribed under the Equal Pay Act? In grappling with this issue, the court notes at the outset that if a single claim alleges the existence of disparate pay for essentially equivalent work as being violative of both Title VII and the Equal Pay Act, the latter statute's standards (both in respect to a prima facie case and to burden-shifting) apply. Melanson v. Rantoul, 536 F. Supp. 271, 286 (D.R.I. 1982) (collecting cases). Even when such a claim is brought solely under Title VII, the modalities of the Equal Pay Act apply. Plemer v. Parsons-Gilbane, 713 F.2d 1127, 1136-37 (5th Cir.1983); Odomes v. Nucare, Inc., 653 F.2d 246, 250 (6th Cir.1981); Orahood v. Board of Trustees, 645 F.2d 651, 654 & n. 3 (8th Cir.1981); Stastny v. Southern Bell Telephone & Telegraph Co., 628 F.2d 267, 281 (4th Cir.1980); Taylor v. Phillips Industries, Inc., 593 F.2d 783, 785 (7th Cir.1979); Ammons v. Zia Co., 448 F.2d 117, 119-20 (10th Cir.1971). There is no logical basis upon which these authorities can be distinguished for purposes of this case.
Thus, the plaintiffs, to establish a prima facie case under Title VII with respect to their compensation claims, must prove that they performed equal work under substantially equivalent working conditions, with like responsibilities, skills, and effort, vis-a-vis their male counterparts. Plemer, 713 F.2d at 1136-37. If they fail to do so, the court need not address these contentions further. If they succeed, however, the defense has the obligation to invoke one of the four statutory exemptions. And a failure to do so, or to demonstrate the applicability thereof, allows the nisi prius court to find liability under Title VII. This scenario, of course, applies with the same force to all of the compensation allegations: the classwide determinations, the personal claims of the class representatives, and Kraynek's Equal Pay Act averment.
In the pages which follow, the court will apply these legal modalities to the facts as found.
VII. STATISTICAL EVIDENCE: A CAVEAT
Though the details, permutations, and effect of different types of intricate mathematical and quasi-mathematical analyses will be discussed with considerable meticulousness below, a general word of caution should be hung on the gatepost.
It is widely recognized that statistical evidence can be a valuable tool for proving or disproving employment discrimination. International Brotherhood of Teamsters, 431 U.S. at 339 & n. 20, 97 S. Ct. at 1856 n. 20. But, its worth can easily be overestimated. Statistics come in various shapes and sizes and are subject to the inadequacies of the database ("garbage in, garbage out" has come to be a cliche, but it is apodictic in the world of statistics), sophisticated manipulation, desiderative interpretation, and the bias of initial assumptions. Statistical evidence is far from irrefragable. As this court observed at a more embryonic stage of this very litigation:
While the progression of civilization from the quipu to the analog computer has added measurably to the store of available computational knowledge, even integrated microcircuitry and silicone chips know some bounds. Although statistical analyses serve an important role in employment discrimination cases, they are neither irrefutable nor necessarily definitive. Death and taxes, arguably, may be certain; but the colligation reached by application of the electronic dactylonomy of the statistical surveyor, and the conclusions suggested thereby, are not. Such analyses are, at best, sophisticated numerative generalizations, and they may, like other forms of generically-reliable evidence, be rebutted.
Chang v. University of Rhode Island, 554 F. Supp. at 1206 (citations omitted).
*1189 To the extent that the statistical evidence at bar comprises simple compilations of raw data, it is easy to understand and to manage. Analytical statistics, which in this litigation took the form of regression and multivariate analyses of the raw numbers, are of a genre as to which caution is most advisable. Cf. International Brotherhood of Teamsters, 431 U.S. at 340, 97 S. Ct. at 1856. These analyses, tredding the trail blazed by the Court in Castaneda v. Partida, 430 U.S. 482, 496 n. 17, 97 S. Ct. 1272, 1281 n. 17, 51 L. Ed. 2d 498 (1977) and Hazelwood School District v. United States, 433 U.S. 299, 309 n. 14, 97 S. Ct. 2736, 2742 n. 14, 53 L. Ed. 2d 768 (1977), calculate the fluctuation of a sample from some expected value (the standard deviation). The Court in Castaneda and in Hazelwood had indicated that a finding of two or three standard deviations would lead a prototypical social scientist to conclude that random variation did not explain an observed disparity. But, though the parties to these cases treat the standard deviation with a deference rarely seen since the heyday of the Oracle of Delphi, even a finding of two or more standard deviations does not prove the existence of discrimination; rather, a spread of that magnitude means only that chance can likely be excluded as an explanation for the result. Nor does the conversea standard deviation of less than two or threenecessarily exclude discrimination as a possible cause. Craik, 731 F.2d at 476 n. 13; EEOC v. American National Bank, 652 F.2d 1176, 1192-93 (4th Cir.1981), cert. denied, 459 U.S. 923, 103 S. Ct. 235, 74 L. Ed. 2d 186 (1982). The court categorically rejects the arguments (i) that a specified value in excess of two or more standard deviations is the only valid test of classwide discriminatory conduct, or (ii) that a specified value below that range is unrebuttable proof of the absence of such conduct.
The court will review a variety of other important statistical concepts and the assimilation and reliability of the databases utilized by the respective experts at diverse points in the discussions of specific issues which follow.
The time is ripe, then, to turn to an examination of those issues and the challenged employment practices which underlie them.
We start with a brief exegesis anent the hiring process at the University.
In preparing a budget request, the president asks each of the deans to list the faculty positions and the amount of funds needed in his or her college for the upcoming year. These proffers are massaged by the president and worked into a tentative fiscal plan which is submitted by the administration to the Board. The Board then revises the budget proposal, if and as it sees fit, and transmits a request to the General Assembly (which has the final say). Once the legislature has acted and the governor has approved (or in lieu thereof, his veto has been overridden), the URI administration receives the actual appropriation. But, adherence to the budget is not a mechanical exercise. The University has virtually complete autonomy over the appropriated funds so long as the administrators do not stray upwards from the bottom line. Thus, the use of money for faculty compensation and the inter-departmental allocation of teachers are left in large part to the wishes of URI's satraps (subject, however, to the constraints of the collective bargaining agreement). In reality, the power to apportion both money and faculty positions rested with, and was exercised by, a triumvirate consisting of the president, the VPAA, and the AVPAA.
Staff vacancies arise in two generic modes at URI. First, the budget may provide sufficient funds for new positions. Second, attrition occurs in a variety of guises: e.g., through sabbaticals, job-hopping, terminations (including non-renewals), retirements, or deaths. And, if retrenchment through budget cutting has not reduced the overall availability of funds, URI's filling of vacancies may often involve reassignments or transfers, thus creating corresponding *1190 openings in other positions. Once a vacancy arises, a search committee is formed. The committee is usually made up of a number of faculty members from the affected department, together with other representatives of the University community. The search itself is a process dictated, in part, by affirmative action guidelines.
Advertisements are placed in various publications in order to attract a wide group of applicants. Outreach efforts directed toward attracting women and minority applicants are set in motiona step which is made considerably more difficult by the fact that many fields of academic endeavor have relatively few women and minorities who have earned or are enrolled in doctoral programs. See text post. When the application deadline has arrived, the committee reviews the responses and selects a limited number for subsequent interviewing. (Legitimate fiscal constraints often play a role in holding down the number of candidates who reach this stage.) After these audiences have been completed, the search committee forwards its nomination to the dean of the college in question. If the dean approves of the choice, he or she sends the nomination (along with a salary recommendation) to the AVPAA. The latter's approbation, if coupled with the AA officer's endorsement of the fairness of the hiring process, almost always ensures that URI will tender the appointment.
B. Statistical Evidence (Hiring)
With this background of the hiring process in mind, the court turns next to the statistical evidence introduced by the parties concerning the issue of discriminatory hiring. The court begins its probe with the 1972-73 academic year, as this was the first year during which Title VII was fully effective in respect to University faculty.
In the 1972-73 academic year, there were 627 faculty members at URI. Of those, 112 (17.8%), spread among only 15 of URI's 60 departments, were female. By the end of the 1981-82 academic year, 134 women (19.5% of the faculty) were teaching in 33 of the 60 departments.
Most women faculty at URI have been sequestered in fields traditionally staffed by women (e.g., nursing, home economics, dental hygiene, library science, and education). According to Janet Wooley, quondam administrative assistant to the AAUP, the University in the 1973-74 school year hired 53 new faculty members, 18 of whom were women. Two-thirds of these women were in library science, nursing, or home economics. Although the next succeeding academic year reflected improvement in the percentage of women hired, fully three-fourths were relegated to fields dominated by women. The statistics for 1975-76 showed that while 50% of the women hired were in specialties other than nursing, home economics, or library science, the percentage of women joining the faculty dropped from nearly half to less than one-third. This overall percentage slid even further in 1976-77. In that year, slightly over 25% of the new hires were women; half of these were assigned to nursing, library science, or home economics. In both 1977-78 and 1978-79, the school-wide hiring percentage improved to the point where well over one-third of the new faculty recruits were women. And, the number of women lured into specialties which had traditionally been masculine preserves steadily increased: over half of those brought on board in 1978-79 went to departments other than nursing, home economics, and library science. Pl. I-3.
The raw data is notable in some respects. It plainly shows that females have not, as a class, historically comprised a large part of URI's faculty, despite the fact that their numbers are in the ascendancy. The naked *1191 and unadorned figures, though authentic, must be handled with care. There could well be a veritable host of reasons for URI's failure to hire more women or to hire women in specific fields. There was no credible data as to the number or gender either of overall applicants or of rejected offers. Obviously, the number of qualified women available and interesteda set of figures which cannot be divined from Wooley's statisticswould have a marked bearing on the overall equation. When all is said and done, the AAUP numbers are at best suggestive.
To shed light on the dilemma, a more sophisticated analysis was required by the plaintiffs. It was authored by their statistical expert, Dr. Harriet Zellner, an expert whose professional credentials are not open to legitimate dispute. Her report analyzed hires which took place from the commencement of the 1972-73 academic year through the end of the 1980-81 academic year. As a starting point, Zellner used the catalogued AAUP data described above, as well as information on the number of doctorates earned over time in various fields supplied by the National Center for Educational Statistics (NCES). The departmental structure at URI did not correspond exactly with the NCES format, so Zellner grouped departments to conform to the broad categories employed by NCES. The court finds this methodology to be acceptable.
Zellner's study assumed that URI was sex-neutral in its hiring process. By way of illustration, if thirty percent of the new doctorates in English were women in any given year, then Zellner hypothesized that thirty percent of the English professors hired by URI in that year should have been women. Zellner then compared the expected hires with URI's actual hires to form a predicate for her opinions. In making the expectancy calculation, Zellner did not utilize NCES statistics for each year and compare them with actual annual recruitment at the University. Instead, she chose 1976-77 as a midpoint and used NCES percentages for that year to calculate expected hires for all years in her study. Zellner described her rationale essentially as follows: she noted that the number of women enrolled in doctoral programs has been consistently on the rise since 1972; thus, any single year chosen as a benchmark would overestimate the availability of distaff doctorates in previous years and underestimate their availability in subsequent years. Zellner settled upon 1976-77 because she thought that, at this halfway point, overestimations and underestimations would cancel each other out. The court finds, however, that this assumption rested on rickety underpinnings, especially since it is clear that both the rate of growth in female doctorates and the raw number of doctoral graduates were much greater in the latter years than in the earlier years. And, the growth rate fluctuated widely according to field.
Once Zellner had calculated the expected hires, she used those figures, together with the raw actual data, to compute the odds of obtaining the observed difference. Probabilities were calculated to show the likelihood that detected differences were the result of random events. When the probability fell below a specified level, chance could, in her view, be ruled out as an explanation and the results deemed statistically significant. According to Zellner, that level is a probability (P value) of less than 5%. The court notes that the 5% level, while not uniformly accepted, is a common measure employed by social scientists. See D. Baldus & J. Cole, Statistical Proof of Discrimination § 9.02, at 291 (1980) (Baldus & Cole). Zellner studied hires for the ranks of instructor, assistant, associate, and full professor. She aggregated the data across disciplines for each rank and computed a probability for observing that number of actual hires. In addition, she computed totals across rank to obtain the probability for hiring that number of women over the period of the study. The data was analyzed in this manner *1192 with and without use of NCES figures anent health professionals.
An examination of Section A (including health professionals) of Table 1 of Zellner's report reveals the following:
Expected Actual Total Women Women Rank Hires Hires Hires Probability Instructor 56 18.6 42 1.00 Assistant 158 38 37 .421 Associate 31 7.67 6 .215 Professor 13 2.46 1 .14 ____ _____ ___ _____ Total: 258 66.8 86 .998
Zellner found no differences between actual and expected hires within a given field and rank which could be termed statistically significant, nor any which even approached statistical significance. Based on the results of her analysis, chance cannot be ruled out as an explanation for the paucity of women hired by URI.
An examination of Section B (excluding health professionals) of Table 1 reveals the following:
Expected Actual Total Women Women Rank Hires Hires Hires Probability Instructor 25 8.74 11 .848 Assistant 141 32.6 27 .111 Associate 29 7.03 5 .156 Professor 11 1.83 0 .061 ____ ____ __ ____ Total: 206 50.2 43 .098
As with the earlier table, there is a dearth of statistically significant differences in a comparison of hires within rank and discipline. Zellner did report that the probability of obtaining a result as low as zero "approached" statistical significance. She also testified that the probability of URI hiring only 43 women did come close to statistical significance. But, in neither instance was she prone to dismiss randomness of selection as an explanation for the result. And the court, given Zellner's statements anent the guidance to be derived from P values of five percent or less, cannot exclude stochastic variation as an explanation for the number of women hired on either paradigm.
In summary, whether or not health professionals are included in the mix, hiring patterns at the University were statistically consistent with the luck of the draw. To be sure, the number of women placed on the payroll tended almost universally to fall below the expected number of hires except at the instructor level. But, the results could well have occurred by chance or other random events.
This finding, however, is not dispositive of whether or not URI discriminated against females in its hiring practices. Zellner's study (the only statistical evidence proffered by any party on the point) only forecloses the statistical conclusion that chance did not play a part in the result. The court therefore proceeds to examine the relevant anecdotal evidence.
C. Anecdotal Evidence
1. Lucy Chang
Chang earned a master's degree in statistics from URI in 1968. While enrolled in the graduate program, she taught in the mathematics department. She was then hired as an instructor in the quantitative business analysis department (later to become the department of management science) of the college of business administration. Chang began teaching full-time in the fall of 1968. She claimed to have arranged with the dean of the college and the chairman of the department to teach 12 credit hours per semester and to forego any research. The chairman told her, she testified, that she would not be required to obtain a doctorate if she was content merely to toil at this level. Chang seemed pleased with this arrangement.
*1193 During her stint at URI, Chang taught basic statistics and mathematics. Her work was generally lauded by students and faculty. She received bonus compensation for dedicated work and obviously expected promotion in the usual course of events. These expectations, however, proved to be the merest of velleities.
In September of 1970, Weeks became the dean of the college. His mission, as he saw it, was to raise the standards of the college's undergraduate and graduate programs. When he came to URI, he found that the undergraduate program was barely accreditable; the graduate program was not only unaccredited, but was far from reaching that plateau. Weeks had been the head of the accreditation committee of the American Association of Collegiate Business Schools. The court unhesitatingly accepts his characterization of the college's quality (or lack thereof) at the commencement of his term of office. To raise these standards, Weeks believed it was imperative to strengthen the faculty. Such fortification logically required greater emphasis on doctoral status, and suggested replacement of teachers who did not have doctorates with those who had them (or who were currently enrolled in doctoral programs).
The new dean assembled all of the nontenured faculty members of the college in December of 1970. At the meeting, Weeks declaimed that he would not renew the contract of any faculty member who did not possess a doctorate or who, alternatively, was not actively pursuing one. This pronouncement was followed by letters to the eleven affected faculty members which reiterated the substance of the December monition. Chang was displeased. She visualized Weeks as reneging on the arrangement which she claimed to have made with his predecessors. In early 1971, she closeted herself with Weeks and told him of her special deal. According to Chang, Weeks replied that "there will be plenty of courses for you to teach." Weeks did not recall that remark, but remembered telling Chang that she would have a place at URI if she got her doctorate. He also offered to help her in any way possible to obtain the degree. Chang never availed herself of that volunteered offer. In fact, after that audience, Weeks never heard from Chang anent doctoral studies. Chang was a poor witness, and the court discounts her testimony in large measure. Weeks, on the other hand, was impressive in his candor.
In early 1971, Weeks began recruitment for the 1971-72 academic year. Time was short and he had difficulty in procuring the requisite number of faculty members. In August 1971, as a last resort, he renewed Chang's contract. The renewal was for one year only and required her to teach 12 credits per semester. At the conclusion of the 1971-72 academic year, her contract was terminated. Chang grieved the dismissal.
Chang learned of two vacancies within the college in the spring of 1972. Randolph Chen planned a two year leave and Wei Shih resigned from the faculty. Chang submitted applications for both positions. In her cover letter, she requested that the Shih position be kept open until the grievance which she had previously filed was decided. The University refused to abide by the request and hired Henry Parsons to replace Shih.
Nevertheless, the Chen position was still vacant. Chang's qualifications were adequate in the view of the department chair (Thomas Vollmann). But, instead of hiring Chang, the University hired a recent URI graduate, Tead Sosnowski. Vollmann testified by deposition that he was unable to hire Chang because he thought he would have to promote her and thus put her on a tenure track. Ferrante and Weeks testified that Chang could have been hired for a single year as an assistant professor without implicating tenure. Such a maneuver would, of course, have left URI with a vacancy in 1973-74 and with the concomitant need to undertake a new search for a second replacement for Chen. URI took an arguably easier option by hiring Sosnowski on an interim basis.
*1194 Chang, little daunted, did not abandon her quest for a position at URI. But, she made no effort to obtain a doctoral degree. In the spring of 1974, Chang saw an advertisement for a position in the management science department. The notice required that the applicant have a doctorate, be near completion of one, be enrolled in a course of study for one, or have other significant strengths. In addition, the advertisement noted that the applicant should have the potential to perform research. Despite her lack of the chief prerequisite needed for successful consideration, Chang applied for the position. She received a polite form letter from Weeks acknowledging receipt of her application and thanking her for her interest. She heard nothing further, so she made inquiry to Rocha. The AA officer informed her that Robert Pari and Ronald Norris had been hired on a temporary basis to fill in for absent faculty.
It is true that Chang applied for a position and that males were hired in her place. But, the record is devoid of any credible evidence which would show the "significant strengths" Chang possessed or how she was competent to perform the required tasks. The plaintiffs failed totally to prove as a factual matter that Chang was qualified for the position listed in 1974-75. Nor did the plaintiffs prove that either or both of the men hired in her stead were of lessor (or even equivalent) credentials.
2. Philosophy Department
A bumper crop of evidence anent hiring dealt with the philosophy department. The department was in the eye of the storm in the early 1970s, when unsettled conditions directly and indirectly impacted the recruitment of women.
The inclement weather began in the fall of 1973 when Oliver Martin announced his impending retirement. Michel, the interregnum VPAA, informed Dean Lepper that Martin's position could be filled at the assistant professor level, but only after an affirmative action search. During 1973, squalls had broken out in the philosophy department; by the time of Martin's resignation, a hurricane was building. In hopes of clearing the air, URI, shortly after the search for Martin's replacement began, authorized the department to recruit a new chair. This search impinged upon the department's quest to find a replacement for Martin. In order to accommodate the possibility of an outsider being hired into the department, Martin's position was left open until July 1, 1974. Thus, the philosophy department only had one position to fill. If a new chair came from within URI, then the department would have been empowered to hire an assistant professor in the philosophy department. Conversely, a chairperson brought in from outside URI would have debarred URI from replacing Martin's position with a new assistant professorship.
The search committee opened for business on February 7, 1974. Stephen Schwarz, a philosophy professor, served as its chairman. Other members were J. Morton Briggs (history), Stanford Cashdollar (languages), John Hanke, Yong Kim, William Young, and Donald Zeyl (all philosophy). Numerous applications were received for the departmental chair position, including those of Alice Koller and Hilda Hein.
Koller and Hein came up for consideration in May of 1974. By that time, the search committee had foreclosed the option of appointing an external candidate to chair the department. The search committee majority was composed of individuals banded together in one faction of the departmental civil war. These members of the philosophy faculty supported Fritz Wenisch in his candidacy for the chair; they did not want an outsider intruding into their fiefdom. Briggs and Hanke were appalled at the committee's behavior and complained to high administration officials. The latter interceded in the battle and demanded that the philosophy department consider outside candidates. Schwarz succumbed to the administration's exhortations, under protest, and reopened the search soon after the 1974-75 academic year began. Wenisch became the acting chair.
*1195 The search for an individual to replace Martin ignited in earnest when the search committee decided to review only internal candidates for the departmental chair (the committee being emboldened by the knowledge that this limitation freed up the single external availability for use in replacing Martin). The department received copious applications, including Koller's. Hein did not apply. By the time Koller's application was in, however, the department had settled on hiring John Holmes. Wenisch received departmental approval to go to Paris and interview Holmes and to hire him if he found Holmes to be acceptable. Nevertheless, Holmes' hiring was postponed until after the original July 1, 1974 cutoff date because of the administration's desire to reopen the search for a chair. Thus, the duad of positions which Koller sought again became intertwined. She was informed by Wenisch that the position opened by Martin's retirement was "frozen" until July 1 in order to allow external candidates to be considered for the chair. Wenisch's advisory was accurate as far as it wentbut it did not go far enough. He failed to inform Koller that reconvening the chairperson search before July 1 was extremely unlikely. In addition, Wenisch's letter neglected to inform her that the department had a pet candidate to fill Martin's position.
The frozen position suddenly thawed for one year in the middle of July. URI was unable to hire Holmes because of a delay in his thesis defense. With time running out and the start of the semester rapidly approaching, URI had to find someone to fill Martin's position on an interim basis. The department received permission to hire two graduate students on a part-time basis; Koller was not considered. On September 6, 1974, she was notified that the teaching position had been filled but that the search for a chair was continuing.
The travails of the department did not end with the hiring of the graduate student duo. The appointment of a departmental chair was still up for grabs and Schwarz was under administration orders to reconvene the search. He followed suit. While the search was underway, Newman (recently installed as president) instituted reviews of various units within the University. A budget task force completed its assessment of the philosophy department in December of 1974. The task force recommendations (which Newman endorsed) resulted in the philosophy department's loss not only of a faculty position but also of its graduate program. Thus, the temporary position split on an interim basis between the two graduate students was rendered obsolete at the conclusion of the 1974-75 academic year. In addition, the search for a new departmental chair was perforce restricted to internal candidates (since there was no longer an open position available to allow entry of an outsider). And, on December 30, 1974, Newman approved Wenisch's nomination as the new department chair.
This was not, however, the end of this convoluted saga. Almost four months later, Newman instructed the search committee to resume looking for an outside candidate. This turnabout was never coherently explained at the trial; it resulted either from Newman's failure to finalize the choice of Wenisch as chairman of the philosophy department or from the non-implementation of the budget cutbacks recommended by the task force.
The travails of the department continued when Damian Fedoryka announced in the spring of 1975 that he had elected to take an unpaid sabbatical for the 1975-76 academic year. Fedoryka's leave left both a vacancy and the wherewithal to fund it. Wenisch spotted the opening and expressed a desire to reassign the graduate students from Martin's position to Fedoryka's. Hanke objected, but to no avail; Wenisch got his way. Koller was neither informed of the opening nor considered for it.
Fedoryka's position again became available for the 1976-77 academic year due to his resignation. The department conducted an affirmative action search and sought external candidates. A request was made to allocate travel money to bring Galen Johnson and Edward Wierenga to URI for *1196 screening. Following those interviews, the search committee nominated Johnson. Wenisch communicated this decision to Brittingham (the resident AA officer); he also informed her that, if Johnson became unavailable, the department would not hire Wierenga immediately but would look into the credentials of Iris Young. The eventuality never came to pass as Johnson accepted URI's offer. In 1977-78, his position was converted from a one year position to a continuing tenure track position. No explanation was proffered for the delay in converting Johnson's position to an enduring one, although there was some suggestion that the dean of the college of arts and sciences was not certain whether Fedoryka's former position would be maintained or reassigned elsewhere within the college. There was no evidence that the department ever seriously considered Young for the post.
Throughout, the conduct of the philosophy department was petty. The majority of its members sought only to hire individuals known to them or referred through an informal "old boy" network. There is scant room to doubt that the department was not interested in hiring anyone, male or female, who did not emanate from within its self-circumscribed sphere. But, objectively viewed, the gender of Koller, Hein, and Young was less of an impediment than their lack of contacts in the bosom of the lodge.
3. Other Evidence
Networking, as a personnel practice, was not unique to the philosophy department. The department of biochemistry, for instance, employed such a fraternal approach on at least one occasion to find an instructor. The case of Claude-Marie Janeway is illustrative. The outlines of Janeway's story are already of record, see text ante at Part V (D), and do not bear reiteration. It is relevant under this rubric in that Smith was preferred over Janeway because the hierarchy of the department knew Smith and was comfortable with him. Yet, whether or not the selection demonstrated discrimination is an open question. There is insufficient evidence to compare Smith's and Janeway's qualifications or abilities. No resumes or independent evaluations of their respective plusses and minuses were produced. Accordingly, it would require outright speculation to hold that Janeway was a victim of discrimination instead of a victim of a closed-shop network which apparently operated to snub men and women equally.
Another vignette portraying allegedly discriminatory treatment in the hiring process involved Elaine Hagopian. In the spring of 1977, Hagopian was a candidate for the chair of the department of sociology and anthropology. The search committee recommended that she be hired. Richard Gelles (a male member of the committee) issued a minority report which gained some intra-departmental support. The disagreement over Hagopian's hiring reached the higher levels of the administration. Marks, the dean of the college of arts and sciences at the time, agreed with Gelles and so informed Newman. Marks' backing of Gelles' view ended any chance for Hagopian to prevail. Yet, no real evidence was introduced concerning Hagopian's qualifications; the support which she received from the search committee was lukewarm at best, and may have been more the result of the arduousness of the search rather than the credentials of the nominee. The search committee did not endorse the other finalist (Dr. Blumstock) and the committee was disbanded without filling the post. A new search was apparently commenced in the fall of 1977, but there is no evidence as to its outcome.
No other constitutive evidence was introduced which could be said to bear directly on discrimination in the hiring of faculty members by the University. The plaintiffs' reliance in this respect on the experiences of Jill Bonner and of Clarice Stasz as being probative of discriminatory hiring practices is unpersuasive. Bonner was hired. Stasz's case, if probative at all, smacks of discrimination in salary at hire rather than discrimination in hiring.
*1197 D. Summary (Hiring)
1. Classwide Discrimination
The court has reviewed all of the evidence upon which the plaintiffs rely to establish discrimination in hiring. The statistical evidence reveals that women made up only about one-fifth of the faculty. That percentage remained relatively constant throughout the period. The scantiness of female representation within the population universe, however, is readily explicable in terms other than discriminatory animus on the part of the University. The pre-Title VII years, of course, created an opening imbalance. As the supply of qualified women increased, there was some perceptible improvement in hiring ratios. And, the statistical analysis presented by the plaintiffs was not probative as to the point at issue: although the data did not flatly rule out sex discrimination as an explanation for the numerical disproportion, neither did it rule out either randomness or the absence of discrimination.
Zellner's calculations in this wise rested on a faulty hypothesis; it would be pure guesswork to say that her overestimation of women Ph.Ds in early years was cancelled out by her underestimation for later years. And, she did not account in any sense for variation from discipline to discipline: other evidence showed, for example, that while the distaff percentage of doctorates in education increased as between 1976-77 and 1980-81 (from 35% to 46%), the percentage decreased in the physical sciences over the same period. Zellner's analysis was also flawed by her assumption that the labor pool should be restricted to doctorates awarded. This format underestimated the number of qualified candidates for positions at URI. The testimony throughout the trial showed that candidates without doctorates male and female were frequently recruited and hired. Indeed, in some fields (e.g., studio art, music, nursing, and theater arts), the hiring of those devoid of doctorates was the norm. The labor pool was not in any way limited to earned doctorates, and Zellner's data therefore underestimated the available number of potential candidates (male as well as female).
These defects are even more jarring when one looks not at the overall conclusion, but rather at the component parts. Zellner used the number of doctorates as an estimate of the instructor's labor pool; yet, her own rank at hire study, see text post at Part IX(C), revealed that comparatively few individuals with doctorates were assigned to the level of instructor at initial placement. Thus, the labor pool of instructors was unduly restricted by using the doctorate as a criterion for potential applicants. Similarly, Zellner utilized freshly minted doctorates as an estimate of the labor pool for associate and full professors. Faculty hired at these ranks had substantial experience beyond the doctorate or other terminal degree. To use the number of doctorates awarded in a given year as an estimate of the potential number of associate professors available in the same year was simply wrong; it entirely failed to account for the lag time needed to acquire the experience essential for appointment at these ranks. Zellner's calculations had the effect of markedly overstating the potential number of women available for hire at these levels.
Lastly, warts and all, Zellner's study does not on its face support a conclusion of bias in the hiring process. Zellner only found two results which were statistically significant at the 10% level. One of these values, the number of full professors hired after excluding health professionals, can be discounted due to the miniscule sample size and the inherent unreliability of small samples. See International Brotherhood of Teamsters, 431 U.S. at 340 n. 20, 97 S. Ct. at 1856 n. 20; Mayor of the City of Philadelphia v. Educational Equality League, 415 U.S. 605, 611, 94 S. Ct. 1323, 1328, 39 L. Ed. 2d 630 (1974); see also Baldus & Cole, op. cit. supra, § 9.12. The result for the across-the-board aggregation can likewise be discounted because it barely supports a finding of statistical significance even at the 10% level and does not justify such a finding at the 5% level favored by Zellner *1198 herself. Thus, the statistics proffered by the plaintiffs' expert do not exclude chance as a causative factor for the observed results.
Not only does the statistical trumpet sound an uncertain note, but the anecdotal evidence is also ambiguous. The particularized instances of purported discrimination cited by the plaintiffs are manifestly insufficient to sustain their burden. Chang's case (viewed by the plaintiffs as the centerpiece of their assault upon URI's hiring practices) falls far short of the mark. See text post at Part VIII(D) (2). And, while the University's disregard of its own affirmative action procedures was deplorable, that scenario alone does not carry the devoir of persuading the court that URI failed to hire women in proportion to their availability. In sum, the evidence proffered by the plaintiffs does not, even on a prima facie view, show a pattern or practice of excluding women from the University faculty.
URI has some sixty departments. Over the decade covered by this litigation, there were literally hundreds of employment decisions made by the University. Whatever problems existed in the philosophy and biochemistry departments appear to have been aberrations. Discrimination was not proven in either instance, let alone on a systemwide basis. The court is constrained to conclude that no discernible pattern or practice of sex-related discrimination has been proved with respect to the hiring of women faculty members at the University.
2. Class Representative Claim (Chang)
The court next addresses the issue of whether Lucy Chang was discriminated against when she applied for jobs in 1972-73, 1973-74, and 1974-75, an inquiry which the court will undertake assuming (though not deciding) that none of her claims are time-barred. Since no class-based discrimination existed in hiring, Chang is not entitled to any presumption that the defendants discriminated against her as well, Craik, 731 F.2d at 470, but must establish her claim independently, along the lines of the Burdine model. Id.
It is reasonable to suppose that Chang was qualified for the job which she sought in 1972-1973 since she had previously been terminated for a reason that had little to do with her ability to teach the courses assigned to her by the department (though there is legitimate doubt as to whether she was qualified to continue on in a regular faculty capacity, given the department's reasonable, gender-neutral change in direction).
URI's justification for not re-hiring Chang was, however, sound on several counts. She could, at the maximum, only have been reengaged for a single additional year; Sosnowski, her replacement, might well have been unavailable after that year. URI made a reasoned choice to hire an individual who did not implicate the problems of longevity that Chang presented to the department. The court finds that Chang's prima facie case, if proven at all, was flatly rebutted by the articulation of a legitimate business reason. And, the plaintiffs' efforts to demonstrate that the proffered reason was pretextual were wholly unavailing. Although Chang rails about Vollmann's allusion to the fact that URI could do nothing until her case was resolved, his statement (if made) was taken out of context and was not probative of any discriminatory animus. Her version of Weeks' remarks was incredible. The long and the short of it is that Chang's evidence did not measure up to the task of demonstrating either intent to discriminate or pretext. The court has scant difficulty in finding that Chang was not victimized by discrimination with respect to the job openings in 1972-73 and 1973-74.
Chang next contends that she was the object of discrimination with respect to a 1974-75 vacancy. This sortie need not long detain the court. Here, Chang never made out her prima facie case. The advertisement for the position required the prospective faculty member to possess a doctorate, be enrolled in a course of study leading to one, or have other strengths demonstrating *1199 a capacity to conduct research. Chang simply was not qualified under any rational application of this standard.
In sum, the court finds that URI is not liable to this class representative on her individual claim of discriminatory failure to hire.
IX. RANK AT HIRE
Though women were not denied entry to URI's faculty in any systematic fashion, that finding in no way foretells that the University did not discriminate against them once they were hired. The court therefore turns its attention to the remaining claims of the plaintiffs, addressing at the outset the issue of rank placement at hire.
A. Rank Placement: Process
URI, like most universities, has four basic faculty ranks. In ascending order, they are: instructor, assistant professor, associate professor, and full professor. Once brought on board, faculty are then placed on one of two tracks: tenure or nontenure. Neither the collective bargaining agreement nor the University's manual (Manual) provide criteria by which URI must determine the level at which an individual will start his or her University career. The determination is made on an ad hoc basis; in theory, it depends on such things as the money available, the advertised level of the position, the availability of open faculty slots on a temporary or permanent basis, and the applicant's credentials. The operative guidelines are similarly blurred as to when (if at all) a person will embark on a tenure path.
The tenure track is quite distinctive. Tenure track faculty are those members of the URI teaching staff who, after serving their apprenticeship for a contractually specified number of years, become eligible for tenure. If they are found worthy of retention on the faculty then tenure is granted thereby assuring the faculty member of a continuous appointment which cannot be terminated except for cause. If tenure is not earned within the stipulated span, the teacher is forced to depart. See, e.g., Pl. CC-4. Exactly the opposite is true of those placed on a nontenure track (although some such persons do eventually cross over to the tenure trail). Docents brought in to substitute for tenure track faculty members on temporary leave from the University are almost always kept off the tenure track. And, in rare cases, a visiting position (nontenured) may be created in order for URI to hire a qualified applicant. (If budgetary considerations later permit, the faculty member may then be transferred to the tenure track.) Faculty appointments outside of the tenure track are typically of a set duration, lasting anywhere from one semester to two years.
As indicated above, the allocation of positions (tenured and nontenured) depends largely upon the University's budget (and its prospects for future funding) and upon the number of slots made available by attrition of one sort or another (e.g., resignation, retirement, or leave on the part of tenured faculty) or by augmentation. The budget may de facto create new tenure track or nontenure track positions. On the other hand, the budget may prevent the replacement of faculty by withholding money from the University. Even if funds are available, URI must in the exercise of prudence consider the budgetary outlook for the future.
Once URI has resolved whether or not an opening will be created or filled, and if so, whether the position will or will not be on the tenure track, the University advertises for applicants. In some instances, URI may be quite specific about the rank of the position; but, on other occasions, the notices are noncommittal. In the latter case, rank is determined only after a fresh faculty member is selected. Rank placement is fixed by the AVPAA in consultation with the dean of the college in question. The decisionmakers take into account the budgetary constraints, if any, the prior experience and educational attainment of the individual, the discipline involved, and kindred factors.
Instructors are on the lowest rung of the academic ladder. Typically, they are people *1200 who do not have their terminal degrees in the particular field of study. Many of the instructors begin to teach while close to the end of the terminal degree pipeline; it is no rarity when, upon securing the terminal degree, such a person is moved up a rung to assistant professor.
Assistant professor is the most commonly used placement for new faculty. Such individuals, by and large, hold doctoral degrees. The colleges of business, pharmacy, engineering, and nursing, however, have historically (albeit not consistently) hired faculty without doctorates at the assistant professor level. No express rationale was articulated to explain this phenomenon. The court finds that such placements resulted not from any sinister motive, but from the inevitable inertia of academic habit and URI's perception of the market.
New entries at the associate professor level at the University were less frequent. But, when an individual with a terminal degree and significant prior experience (whether or not academic) was hired, placement at such a rank was sometimes considered. Though the combination of degree cum heavy experience did not guarantee such placement, it uniformly appeared that the absence of either one guaranteed assignment to a lower level.
The apex of the teaching pyramid is the full professorship. This is the highest rank accorded to a faculty member at URI and requires the most extensive experience and accomplishments. While the University rarely hires outsiders at the full professor level, such instances have occurred. On those occasions, the credentials of the individual were so outstanding that an award of immediate tenure was combined with the hiring. E.g., Grace Garner in nursing; offer to Eugene Stanley in physics (see text post). The decision to appoint at the full professor grade begins with the AVPAA, but requires the participation of both the AVPAA and the president. Tenure awards, of course, can be made formally only by the Board upon recommendation of the president.
The initial rank placement process leaves an enormous amount of discretion in the hands of the administrators. Subject only to fiscal constraints (and budgets are fair game for internal readjustment by the administration) and to the demands which must be met to lure desirable recruits, the University's top executives have an extremely free hand. The broad discretion inherent in rank at hire decisions led, in the period between the 1972-73 and 1980-81 academic years, to the following breakdown:
RANK AT HIRE MEN WOMEN Instructor 14 42 Assistant 121 37 Associate 25 6 Full 12 1
Pl. O(1), Table 2.
These compilations, in and of themselves, demonstrate little. Women may have been placed as instructors in greater proportion than men because fewer of them had terminal degrees. Or, the paucity of women hired at higher grades may reflect more women dropping out of the workforce for a time (and thus accumulating less overall experience than their male counterparts). If the raw figures are to be meaningful, a more sophisticated perlustration is obligatory. The oponents in this litigation have obliged. But, before scrutinizing the parties' conflicting analyses, the court must first take the measure of the somewhat dissimilar schemes used by the plaintiffs and by the defendants to assemble databases.
As the code name implies, a database is merely a collocation of raw information *1201 which is gathered, classified, and fed into a computer memory. In this case, the databases assembled by the rival litigants became twin wellsprings for their respective fountains of statistical truth.
1. Plaintiffs' Database
The plaintiffs gathered information from the AAUP's records and from the University. Oversight of the plaintiffs' data collection process was consigned to Ramsay, an active AAUP member and an associate professor of economics at URI. As a mathematical economist, Ramsay was familiar with the requisite statistical methodologies for assembling databases. Ramsay supervised the team of graduate students who actually gathered the data. To ensure continuity and consistency, Ramsay held an initial training session and thereafter periodically instructed newcomers to the project. The data was collected on computer coding forms. Despite Ramsay's efforts, many facets of the data gathering process were left to the discretion of the individual collectors. They had to decide, for instance, whether a faculty member had prior experience in teaching, research, and/or management; whether a publication was a journal article, monograph, or book; and like questions. Although judgment calls had to be made in the field, the collectors were graduate students; as such, they possessed some familiarity with the subjects committed to their discretion. The court does not consider that the delegation of the data gathering process to graduate students, without more, undermined the validity of the plaintiffs' database.
The next step was to have a commercial firm transfer the encoded data to computer-readable keypunch cards. The keypunch was then verified, whereupon Ramsay transferred the data to computer tape. At that point, Ramsay attempted some further verification by checking the tape for illogical entries and endeavoring to supply data that the graduate students had been unable to unearth. He then forwarded the tape to Zellner. Zellner "dumped" the data. Contrary to the first impressions of the uninitiated, this verb is descriptive of the process whereby a print-out of all the data in a readable form (proc print) is produced. Zellner then employed the proc print to determine what (if any) data was missing. She discovered several gaps. Zellner thereupon enlisted the aid of Nancy Murphy (an AAUP staffer), who spent a number of months in concert with Zellner attempting to fill in the missing data. The effort was partially successful.
Zellner had been given information as to the social security number, name, department, and sex of each faculty member. The proc print supplied her with data on starting salary, the year(s) in which degrees (including terminal degrees) had been obtained, year of hire by URI, year of highest degree obtained prior to hire by URI, amount of experience prior to hire, rank at hire, dates of various promotions, and annual salaries. While she also had data on publications, meetings attended, and committee service, Zellner did not analyze or attempt to synthesize that data in any of her studies. Ramsay testified that he was satisfied with the data collection process and, as a mathematical economist, would rely on the database to perform economic and statistical studies. Zellner agreed.
2. Defendants' Database
Like Zellner, the defendants' statistical expert, Bernard Siskin, was well credentialled. (The court, on the bare resumes of the two principal experts, gives neither a significant edge.) He relied upon data derived from the University's personnel files. The mechanics of gathering and transmitting the data were much different than the methodologies used by the plaintiffs. The *1202 defense's data collection process was overseen by the AVPAA (Rosie), a man with an excellent working knowledge of the records. Rosie retained a number of graduate students to search the files of URI's personnel office, review printouts from the budget office, and cross-check the University catalogues. From these sources, the students collected data on the year of mandatory retirement, the department, the year hired, the highest degree obtained, the year of highest degree, the institution from which the degree was received, the salary for each year (including merit awards and inequity adjustments), the year of any promotion, the year in which tenure was granted, and the year of separation from service. Special note was taken of whether the individual had a calendar year or academic year contract. The data was encoded under the aegis of Nancy Rieser, an official of URI's Computer Center. Rosie checked the coding forms for inconsistent or illogical results and then sent the sheets to be keypunched at a suitable facility maintained by URI's school of oceanography. The computer cards were then returned to the Center for transfer to magnetic tape. The product was eventually shipped to Siskin.
The first tape included all tenure track faculty hired or employed between July 1, 1972 and June 30, 1982. Data on the 1982-83 academic year was collected and sent to Siskin separately in July of 1983. Rosie also communicated with Siskin from time to time to rectify any after-discovered errors. Siskin plainly believed that the defense database was accurate and sufficient, in his opinion, to perform economic and statistical studies.
3. The Databases: A Comment
The court is satisfied that both databases were collected according to generally accepted principles of the implicated disciplines, and were adequate to perform the various studies proffered by the respective parties. This finding does not, however, blind the court to weaknesses inherent in each of the databases. But, the perceived flaws go less to the admissibility of the expert testimony than to the weight of the opinions and conclusions based thereon.
Generally speaking, the databases were used throughout by the experts, and served as the predicate for analyses on a myriad of fronts far surpassing rank placement. Unless specifically indicated to the contrary elsewhere in this rescript, the various analyses undertaken by Zellner and by Siskin on other issues sprang from these self-same databases.
C. Rank at Hire (Zellner's View)
Zellner first compiled total hires of men and women in each rank at URI in tabular form, see text ante at Part VIII (B), which showed a considerable disparity between the number of men hired at upper ranks in comparison to the number of women hired at those ranks. Zellner then tabulated her results according to the presence or absence of a doctorate at hire. She found the following.
I. DOCTORATES FEMALES MALES No. % No. % ALL RANKS 21 100.0 92 100.0 INSTRUCTOR 2 9.5 0 0.0 ASSISTANT 13 61.9 59 64.1 ASSOCIATE 5 23.8 21 22.8 PROFESSOR 1 4.8 12 13.0 II. NON-DOCTORATES FEMALES MALES No. % No. % ALL RANKS 65 100.0 80 100.0 INSTRUCTOR 40 61.5 14 17.5 ASSISTANT 24 36.9 62 77.5 ASSOCIATE 1 1.5 4 5.0 PROFESSOR 0 0.0 0 0.0
Perhaps the most intriguing of these figures were those reflecting placement at the instructor and assistant professor levels for individuals bereft of doctorates. Almost 62% of the women within this rubric were assigned the rank of instructor. In contrast, close to 78% of the men hired without doctorates were given the starting rank of assistant professor. These disparate results suggested the need for further investigation. And, Zellner's multivariate analysis of rank at hire attempted to determine whether the disparity in rank placement was brought about by chance or otherwise.
*1203 Zellner's attempt to isolate the effect of gender (if any) on an individual's rank involved the cross-matching of men and women who were comparably credentialled and as academically alike as possible in non-sex-related ways. To implement this goal, she constructed a statistical model. The model was "controlled" by the designer, i.e., it took into account, to the extent practicable, differences in all the factors other than the one being isolated. The underlying theoryshared by Zellner, Siskin and other statisticianswas that a properly controlled model permits the designer to compare individuals who are similar except for the particular factor to be studied. In this case, Zellner's model controlled for (i) the possession of a doctorate by the beginning of the academic year for which the faculty member had been hired, (ii) years since receipt of whatever degree the new faculty member possessed at the start of the relevant academic year, (iii) the amount and type of prior experience, and (iv) the departmental group or field. Zellner excluded any individual as to whom materially incomplete data was unavailable.
Zellner's hypothesis was, at bottom, a simple one: if sex impacted rank placement, an individual would probably be assigned a higher or lower rank than a comparably credentialled person of the opposite sex. Accordingly, Zellner constructed the model to facilitate comparisons between adjoining ranks. If sex discrimination existed and adversely affected females, then the statistics would reveal that, as between a higher or lower rank, the woman would more often be placed in the lower rank.
The results of Zellner's analysis can be found in table 3 of Pl. 0(1):
Number of Std. Net Deviations between Effect Net Effect of being and Probability Sample Female Zero of Effect Size Instructor vs. -1.473 -2.128 0.017 199 Assistant Assistant vs. -2.084 -1.961 0.025 176 Associate Associate vs. -2.787 -1.483 0.069 43 Professor
In each instance, Zellner found a negative effect in rank placement for females. But, the statistical significance of those findings is yet another tale. Some explication is desirable at this juncture.
A standard deviation gauges the amount of variability between a mean and observed results. By way of illustration, in this instance the standard deviation measures the amount of variability between neutrality (i.e., sex has no bearing on rank at hire, or put another way, a sex effect equal to zero) and a positive correlation (i.e., sex does have a bearing on rank at hire). Depending upon the level of significance and whether and/or how a "one-tailed" or "twotailed" *1204 test affects the statistical conclusion, inferences about randomness can be drawn with some assurance from certain resultant standard deviations.
All analysts agree that there are levels of significance below which a statistician can reliably aver that chance is not a likely factor in the outcome. But, these levels vary: each level has its statistical costs and benefits; each has a corresponding standard deviation; and the borderline levels produce, as among qualified statisticians, their proponents and detractors. To illuminate by a case in point, a 5% significance level corresponds to approximately 1.96 standard deviations (assuming that the sample size is large enough to generate a normal or bell-shaped distribution; sample sizes of 20 or 30 are sufficiently large to generate acceptable distributions). Levels of significance greater than 5% correspond to measurably lower standard deviations; lesser levels of significance correspond to higher standard deviations. See Baldus & Cole, op. cit. supra, § 9.03, at 297.
Since the lexicon of the statistician is idiocratic to the profession, this explanation produces the need for further definition. A standard deviation greater than 1.96 translates, as a general rule, to a finding of statistical significance at the 5% level of significance. Id. at 294-97. But, the true signification of the standard deviation analysis may vary, depending upon the type of test used. Tests may have either one tail or two tails. The tail concept is based on the existence of the type of bell-shaped distribution mentioned above.
The crest of the bell represents the mean of the distribution (in this instance, zero). A two-tailed test is more common, id. at 307; it seeks to examine whether there is variability above or below the crest of the bell (or the mean). Placed in the context of these facts, it is aimed at determining whether the discrimination favors or disfavors women. A one-tailed test probes only a single end of the bell. Since the bell curve is a mirror image of itself, any point above the mean will be exactly the same distance below the mean. Thus, for convenience, statisticians talk about absolute values of the standard deviations and ignore the sign of the deviation.
Zellner examined only the negative end of the bell. She assumed that, in a discrimination case, one would only look for an adverse impact fomented by an employment policy. Consequently, since she utilized a one-tailed test, statistical significance existed for a smaller tolerance of variation from the mean than would be so in the albedo of a two-tailed test. See, e.g., Brown v. Delta Air Lines, Inc., 522 F. Supp. 1218, 1229 n. 14 (S.D.Tex.1980). In terms of standard deviations, a one-tailed test at the 5% level of significance requires only 1.65 standard deviations in absolute terms to achieve some reliability, W. Curtis, Statistical Concepts for Attorneys 123 (1983), whereas under a two-tailed test, 1.96 standard deviations are necessary to exclude chance as an explanation of the result. Under Zellner's paradigm, any standard deviation with an absolute value greater than 1.65 signified that random events could not, as a matter of probability, explain the disparity from a sex neutral rank at hire policy. Zellner's results, computed in this manner, showed statistical significance for instructor versus assistant and for assistant versus associate, but not for the associate versus full professor comparison.
Zellner's use of a one-tailed test is open to question. Statisticians themselves disagree on the propriety of the one-tailed test in discrimination cases. Compare Harper, Statistics as Evidence of Age Discrimination, 32 Hastings L.J. 1347, 1355 & n. 65 (1981) with D. Kaye, The Numbers Game: Statistical Inference in Discrimination Cases, 80 Mich.L.Rev. 833, 841 (1982). The two-tailed test is more nearly equivalent to the two standard deviation measure which has most frequently been used by the courts. Baldus & Cole, op. cit. supra, § 9.2, at 308 n. 36. See also text ante at Part VII and cases cited therein. Baldus and Cole, however, approve the use of a one-tailed test as a means for ruling out as illogical the possibility that discrimination *1205 favors the claimant group. Baldus & Cole, op. cit. supra, § 9.2, at 100 (Supp.1983). More importantly, they note that any discord can be set to rest by the use of P values; and suggest that if a single tailed approach is used, it should be predicated upon P values. Id., § 9.03, at 297.
P values derive from a test statistic which is based on a ratio incorporating the standard deviation. See Gay v. Waiters' and Dairy Lunchmen's Union, Local No. 30, 489 F. Supp. 282, 308 n. 33 (N.D.Cal. 1980), aff'd, 694 F.2d 531 (9th Cir.1982). Once the value of the test statistic is ascertained, the probability of a test statistic of that value can be determined from mathematical tables. Id. This probability then can be compared with various levels of significance or the statistician can infer that the results are significant at the P value. In effect, use of the P values permits the analyst to evaluate statistical significance on a relative scale, comparing apples with apples and oranges with oranges. Zellner adopted this approach and computed the P values. She found that in the instructor/assistant and the assistant/associate brackets, the probabilities fell below the 5% significance level. And, since multiplying a one-tailed P value by two will yield the equivalent of a two-tailed P value, Baldus & Cole, op. cit. supra, § 9.2, at 102 (Supp.1983), the court can safely conclude that they fell below the 5% level of significance even under a two-tailed test.
Zellner concluded that chance was not likely to have been a factor in these findings. Given the presence of adequate controls in the model, she reasoned that sex discrimination had to be the causative factor.
D. Rank at Hire (Siskin's View)
To rebut Zellner's report in this respect, the defendants rely mainly on their expert, Dr. Siskin. Among other things, Siskin studied URI's rank placement at hire record. His rank at hire analysis differs somewhat from Zellner's in both scope and methodology. It includes any person hired since 1971 by URI. For each and all of these individuals, Siskin gathered data on degree year, degree held at hire, departmental group, and year of hire. He then applied this data in a logistic regression analysis in order to determine what effect, if any, gender had on rank placement at hire as between adjoining ranks. It should be noted that logistic regression is a statistical modality in which certain independent variables, e.g., year of hire, affect a dependent variable that does not take on continuous values. Thus, it is an appropriate technique for studying placement between ranks, hire or no-hire decisions, and pass/fail test results.
Siskin, in Table 7 of Df. JJJ-6, reported the following results from his analysis:
Coefficient Standard Statistical Sample For Sex Variable Deviations Significance Size Associate vs. -2.93 0.83 No 71 Full Assistant vs. -0.70 2.11 Yes 289 Associate Instructor vs. -.04 0.11 No 304 Assistant
Siskin found only the placement between assistant and associate professor to be statistically meaningful at the 5% level of significance (under the two standard deviation test). He reported that the difference in placement for the instructor versus assistant professor bracket depended in the main upon whether an individual held a *1206 terminal degree for the field. And, even though Siskin found a statistically significant difference at the assistant/associate professor level, he did not conclude that there was evidence of discrimination. Siskin's reasoning was three-pronged. First, he noted the unavailability of data on prior experience and proclaimed that such data, both quantitative and qualitative, was necessarily a salient factor in determining rank at hire. Second, Siskin maintained that the finding of statistical significance would have been materially different had eight more women been hired over the relevant temporal span. Therefore, Siskin's thesis ran, it was impossible to draw any earthshattering conclusion from the result. Third, a case-by-case study of the small number of women assigned to the rank of assistant professor would have revealed which faculty members had been misclassified. Thus, in Siskin's view, Zellner's study was fatally flawed by her attempt to draw conclusions from too meagre a group, which in turn undermined the reliability of her analysis. And, Siskin dispatched all other brackets summarily: after all, in controlling for year of degree, degree, department and hire year, there was no statistically significant result in rank placement in regard to either the instructor/assistant or associate/full professor comparisons.
E. Statistical Comparison (Rank at Hire)
The two studies demonstrate an axiom of the use of statistics in the law: statisticians can manipulate numbers to such a degree that sustenance can be found for an otherwise unsupportable position. Note, Judicial Refinement of Statistical Evidence in Title VII Cases, 13 Conn.L.Rev. 515, 525-26 (1981). The studies by Siskin and Zellner reach diametrically opposed results with virtually the same data. But, there are numerousand importantdifferences in the models.
Siskin's model excluded variables for prior experience, substituting a proxy (years since degree). Zellner, on the other hand, used the quantity of prior experience. Siskin criticized Zellner's study for failure to sift prior experience through any qualitative filter; yet, he conceded that prior experience was important and that his chosen proxy was a poor surrogate for an actual measurement of experience. To be sure, Zellner's use of the experience factor left something to be desired: her model did not take into account the quality of the experience. The court, however, must deal with probativeness, not with perfection; and Zellner's utilization of prior experience, though flawed, is superior to Siskin's use of an infelicitous proxy. Accordingly, the court must discount heavily the results obtained by Siskin anent this issue.
Siskin also assumed that rank placement was based on an individual's receipt of a terminal degree. That assumption overlooked substantial evidence that URI, in certain colleges, regularly granted the rank of assistant professor even though the faculty member did not have an appropriate terminal degree. Siskin never accounted for this phenomenon, nor was it adequately explained at trial. Thus, the court is loathe to accept Siskin's intimation that the determination of instructor or assistant professor rank depended solely upon the possession of a terminal degree.
Siskin also excluded individuals who obtained their terminal degrees after hire. By constructing his model in this fashion, he removed a number of individuals, both male and female, from the study. Siskin contended that this did not matter, since all *1207 of the individuals would be instructors; but, that assumption is a fallacious one, barren of record support. This, too, rendered his study of rank at hire suspect.
Siskin's explanation vis-a-vis the placement of women as assistant professors as opposed to associate professors skated on thin ice as well; in essence, it constituted an argument that the statistical result was nonetheless a random distribution because of the smallness of the sample size. The rationale of this thrust undermines the raison d'etre of statistics, which is to rule out the possibility of stochastic variation as an explanation for an observed result. If Siskin genuinely believed that the sparse sample was creating difficulties, he could easily have used acceptable statistical techniques in this instance to correct the problem. See Baldus & Cole, op. cit. supra, §§ 9.11-.12. His explanation is lame and appears to be nothing more than a feeble exercise in post hoc damage control.
For all of these reasons, the court would be hard pressed to accord any substantial weight to Siskin's findings on this issue. He excluded salient variables even though he knew of their importance, failed to include all individuals in his study by reason of incorrect assumptions, and compounded his sins by overreaching in a vain try to explain away results which did not support the defendants' theory of the case. On the rank at hire conundrum, Siskin's testimony was reminiscent of the monition of the Cheshire Cat: "It doesn't matter which way you go ... You're sure to [get somewhere] if you only walk long enough." L. Carroll, Alice's Adventures In Wonderland, 57 (Delacorte Press ed. 1966).
Zellner's study, to be sure, was also deficient in certain respects (over and above her neglect to account for qualitative experiential differences). Data was missing that could easily have been supplied. There was evidence of a modicum of improper coding. The study did not include all members of the faculty due to the slightly premature cutoff point: women hired in the 1981-82 and 1982-83 academic years were omitted. And, there was some multicollinearity. See note 25, ante.
Though recognizing these lapses, it is plain to the court that Zellner's model comprises a better, more useful, more reliable tool than Siskin's counterpart. The court finds that Zellner's conclusions as to rank placement at hire are sufficiently probative to be accorded substantial weight; and specifically, that her opinion that random events can in all probability be ruled out as an explanation for the disparity in placements at the instructor/assistant professor bracket and at the assistant/associate professor bracket is deserving of credence. Nevertheless, even giving due credit to Zellner, the disparity in placement between the associate/full professor levels was in all likelihood the result of random distribution.
F. Anecdotal Evidence
The anecdotal evidence on rank at hire, though scanty, tracks in the same direction. The parties agree that emergent faculty in the colleges of engineering, pharmacy, and business were assigned to the assistant professor rank despite the absence of the appropriate degree. This policy had the support of the University administration. Some nursing faculty also were appointed at this entry level without doctorates. While the defense attempted to explain these apparent aberrations on the basis of "market factors," that reasoning has a hollow ring. Any such problem could surely have been handled by salary adjustments without affecting rank placement. And in any event, the policy was not consistently applied. URI appears to have adopted a selective approach: business faculty without doctorates almost always entered at the assistant professor level, but only some nurses were so ranked at hire. The inconsistency was never explained. Both colleges were subject to the proffered market factor rationale, yet the male-dominated college of business enjoyed the benefices of the policy consistently, while the female-dominated college of nursing was similarly blessed only on a sporadic basis.
*1208 The evidence also showed that faculty in other colleges were placed at the instructor rank unless a terminal degree was extant. But, there were exceptions to this rule. For example, Robert Steinberg was hired as an assistant professor in the theatre arts department without such a degree. Again, the inconsistencies in application were never satisfactorily exposited.
There was no worthwhile anecdotal evidence probative of any placement discrimination whatever as between the associate professor and full professor ranks.
The court has reviewed the statistical and anecdotal proof amassed by the parties on this issue. The anecdotal evidence revealed a policy on rank at hire that was, at best, uneven and unexplained. The sketchy anecdotal evidence need not bear much of the weight of the court's conclusions on this point, however, as the statistical models of rank at hire are in this instance more compelling.
G. Summary (Rank at Hire)
1. Classwide Discrimination (Rank at Hire)
The rank at hire claim is a classic example of the proof and counter-proof model discussed in International Brotherhood of Teamsters, supra. The plaintiffs' case rests largely on the validity of Zellner's statistical analysis and whether it can weather Siskin's assault.
As noted above, see § (E) ante, the most troublesome aspect of Zellner's study was the failure qualitatively to assess prior experiencea factor which was critical in initial rank placement, but which was ill-suited for econometric modelling. Zellner's use of a quantitative measure, while not a flawless substitute, provided the court with a much more accurate gauge of the experiential element than Siskin's use of years since degree. Siskin's proxy had numerous defects. To cite but a few, it assumed that women would be steadily employed after receipt of their doctorates and that their ensuing opportunities would be equivalent to those enjoyed by males. Yet, there was no evidence to demonstrate that women stayed in the job market as long as men, and common sense dictates the contrary. Indeed, the defendants, in their brief, asserted that the opposite was true (thereby undermining their own statistician). And, if equality of opportunity was a given, then Congress was conjuring up spirits from the vasty deep when it enacted Title VII.
Zellner, after comparing similarly credentialled individuals, found that the net effect of being female on rank placement was negative at all ranks. Zellner also found that random events could be ruled out as an explanation for placement at the instructor/assistant professor level and at the assistant/associate level. Siskin's calculations buttress the latter conclusion, and his attempt to explain away the finding does not hold water.
Nor have the defendants successfully impeached Zellner's figures by casting aspersions on the size of her sample. There is no credible evidence whatsoever that the sample was so miniscule as to threaten the reliability of the analysis; and, as Siskin himself admitted, a statistician is ofttimes more comfortable with a smaller sample laden with appropriate values than with a larger sample denuded of those values. In the world of statistics, as elsewhere in life, the size of the package bears no necessary correlation to the worth of the contents. Only at the associate/full professor plateau does the small number of people involved in the sample cast serious doubt upon the reliability of Zellner's statistical model. See, e.g., International Brotherhood of Teamsters, 431 U.S. at 340 n. 20, 97 S. Ct. at 1856 n. 20.
Lastly, the defendants' attempt to gloss over the instructor/assistant professor discrepancies is unpersuasive. It is simple to say that the difference between the two placements related solely to whether or not the candidate in question had a terminal degree; but the point, once made, does not withstand analysis. To be sure, a large *1209 percentage of faculty members needed a doctorate to reach the assistant professor rank and comparatively few individuals were assigned as instructors if they possessed doctorates. But, this approach leaves wholly unexplained those individuals without terminal degrees who were made assistant professors. For reasons previously noted, the defendants' "market factors" thesis does not support the variation. And, the defendants' attempt to demonstrate that women bereft of the doctorate had the same chance to be placed at the assistant professor level as did their male colleagues was belied both by Zellner's statistical study and by Rosie's testimony anent URI's propensity to appoint men without doctorates to assistant professor at twice the rate as similarly situated women. Though defense counsel labored long and hard to "clarify" Rosie's testimony on this point, their efforts were entirely unconvincing. The court finds that Zellner's model was an accurate predictor of rank at hire, and that the endeavor to rebut her conclusions at the instructor/assistant and assistant/associate levels will not wash.
The court's task is incomplete, however, absent a determination as to whether Zellner's model sufficiently evidences sex discrimination. The court finds that chance can be excluded as an explanation for the disparities in placement at the two brackets in question, noting that by the use of Zellner's numbers, there is statistical significance at the 1.7% level and the 3.4% level, respectively, for one and two-tailed tests anent the instructor/assistant bracket. This gives strong support for an exclusion of chance. Cf. Baldus & Cole, op. cit. supra, § 9.221, at 101-02 (Supp.1983). At the assistant/associate level, Zellner found statistical significance at 2.5% and 5.0%, respectively, for the one and two-tailed tests, results which are also indicative of the absence of randomness. Based on Zellner's report, the court canand doesrule out random variation as an explanation for the observed disparities in placement at the instructor/assistant professor and the assistant/associate professor brackets. Finally, the court finds that the statistical evidence presented is such that the possibility of stochastic variation as the proximate cause of the disparity in rank placement between men and women at the associate/full professor level looms large.
Once the court has made these findings, it is permissible to infer that discrimination was the cause of the disparity at the two lower brackets if no other factor is evident as an explanation. Taylor v. Teletype Corp., 648 F.2d 1129, 1133 (8th Cir.), cert. denied, 454 U.S. 969, 102 S. Ct. 515, 70 L. Ed. 2d 386 (1981); see Hazelwood, 433 U.S. at 309 n. 14, 311 n. 17, 97 S. Ct. 2742 n. 14, 2743 n. 17; Craik, 731 F.2d at 475-78; Thompson v. Sawyer, 678 F.2d 257, 284 (D.C.Cir.1982). The variables relevant to URI's rank placement decisions were included in Zellner's model. There is no basis in the record for palming off the observed disparities in rank placement on factors such as degree or prior experience. The court is constrained to conclude that a practice and pattern of intentional sex discrimination was extant in rank placement at the two levels in question.
The court's conclusion is reinforced by the atmosphere at URI. Regard for the niceties of equal opportunity was lacking. Administrators involved in making and approving offers (including rank placements) were among the individuals who most frequently departed from the affirmative action principles. Given the ambience, the decisionmakers' relative disregard for rules and regulations, the anecdotal proof, and the statistical evidence, the court holds that URI from and after March 24, 1972 fostered a pattern and practice of discrimination based on sex with respect to rank placement as between adjoining levels below the associate/full professor bracket. The court holds further that no such pattern or practice has been proven with respect to decisions at the associate/full professor junction.
2. Class Representative Claims (Rank at Hire)
The sole class representative who has asserted an initial rank placement claim *1210 was Chang; and Chang's claim itself is necessarily unavailing, inasmuch as she was never hiredmuch less placed in ranksubsequent to the effective date of the operative Title VII amendments. Her rank placement prior thereto is, of course, time barred. Her claim fails.
X. SALARY AT HIRE
Having in mind the close relationship between initial rank placement and entry level remuneration, the foregoing discussion of rank at hire blends into an analysis of the plaintiffs' salary at hire contentions. Once again, we start with a review of the method and manner by which the University fixes original salaries.
The salary at hire procedure is somewhat akin to that which governs establishment of rank at hire. Salaries, like rank, have historically been determined by the AVPAA in conjunction with the appropriate dean. But, the constraints on salary decisions are considerably more formidable.
Starting salaries theoretically depended on a number of factors, the most important of which was the imperative of the collective bargaining agreement as it existed from time to time. Since 1972, the AAUP has negotiated minimum salaries for each rank, and these have been embodied in the union contracts. The minima escalate as the ranks rise. In the more recent collective bargaining pacts, there also are maxima applicable to the various ranks. Though the maximum for a lesser rank may be higher than the minimum for the next higher rank, it will always be above the maximum for the next lower rank.
Within each rank, the administration has broad discretion to set entry level salaries. Consideration is given to prior experience and its relevance to URI's needs, the perceived ability of the individual to perform teaching and research tasks, the quality of one's academic credentials (including degree attainment), and market demands.
Although each of these factors comprises a part of the decisional calculus for determining initial salary, some variables carry more weight than others. URI stresses, in particular, its institutional conception of market demand, prior experience, and degree possession. There is little room to doubt the appropriateness of prior experience and degree status in a reasoned determination of both rank and salary. But, the extent to which market demand is fairly apposite is a more vexing question.
A sound definition of the concept of market factors was supplied by Mary Reilly, an associate professor of sociology at URI (and an ally of the plaintiffs), who noted that,
certain areas should be able to demand higher salaries, because ... either they're more valuable, or there's more of a need, or there's less people, Ph.D.s for example, produced in certain areas, and so forth.
III at 139. Reilly then proceeded to explain why she felt that market demand was improperly utilized at URI:
I think that the whole issue of "market" has been arbitrary, there have been fields in which there are very few people with degrees, and yet they've not been able to command the kinds of salaries in the fields where other people have high degrees. There's been an attempt in certain fields to say that certain people should get what they can demand "in the market place," where that same thing has not been able to carry over into other disciplines, as well. It has not been a consistently-applied standard across all fields, and I might add it's been particularly found in those fields that have been dominated by men and not in fields in which you would find large numbers of *1211 women that have been "the marketable fields," or those that have been able to command the higher salaries.
Id. at 140. Reilly pointed out that a nurse, an engineer, or an accountant who held a doctorate was a rarity; yet, nurses with doctorates could not commandeer the same inflated salaries as engineers or accountants who were similarly endowed.
Newman and Rosie took issue with these comments. They testified in substance that market demand was a reality which simply had to be confronted for certain fields whether or not one accepted the inherent equity of the concept. Ferrante testified that, while the University could not hope to match salaries paid by private industry, it must nonetheless try to be reasonably competitive. As he saw it, to achieve this goal, URI had to do more than rely on the perceived benefits of an academic life. The administration felt that the University had to offer higher salaries to individuals who were in a position to command more money in the private sector. As a case in point, URI would have to pay chemical engineers more than most other comparably credentialled faculty members because of the pressure of bidding from the petroleum industry. And, URI had to set more munificent salaries not only to attract individuals who could fill these slots but also to secure the services of top quality personnel. Newman was especially vocal on this point: he recognized one of his primary missions, upon his installation as president of the University, as being the necessity to upgrade URI academically. To do so, an improved faculty was a high priority. He sought to fill this need by hiring better trained individuals and by inducing faculty in place to augment their credentials.
Rosie testified that determinations of market impact were not based wholly on some notion of supply and demand for certain fields. Rather, he relied on various surveys of salaries paid to beginning faculty elsewhere. A number of these surveys compiled data on average salaries of experienced professors. Chief among them was the annual report sponsored by Oklahoma State University, which agglomerated data on compensation schedules at land grant institutions vis-a-vis faculty at sundry ranks in a wide variety of fields. Rosie never explained with any precision the manner in which he integrated this data into salary offers, nor did he address the mechanism for resolution of the inevitable conflicts among the various surveys. On this record, it can only be said that URI sensed the existence of market factors and that an ill-defined effort was made somehow, some way, to take them into account.
Rosie described the salary reviews made by the Rosie Committee and criticized the comparisons ("matched pairs," so-called) between various professors relied on by the plaintiffs. Though he railed against the match-ups of docents in different fields, he was unable in certain instances (e.g., speech/English/languages) to articulate precisely what market factors made a difference.
The court agrees with the University that market factors exist, as indeed they must in a free, laissez-faire economy. The court unhesitatingly finds that certain fields and specialties command more generous pecuniary rewards than do others. Accordingly, the court perforce rejects the plaintiffs' notion that faculty are faculty, and that some type of comparable worth banner should be draped across all of the disciplines. The realities of the marketplace cannot be blithely disregarded; those teaching in dissimilar fields are different, inter alia, in the sense that the free economy values them at varying dollar levels. If this is discriminatory (in that male-dominated disciplines in some instances command more lavish remuneration than those fields in which women have more traditionally toiled), it is not the type of discrimination which Title VII interdicts. The answer lies in equalizing access to the higher-paying specialties, not in an artificial judge-made distortion of the interaction within an open marketplace.
An acknowledgement of the operation of market factors does not, however, sanctify *1212 the defendants' actions. The evidence is minimal insofar as it would support a determination as to which administration decisions were legitimately affected by market factors. The record is likewise unenlightening in explaining precisely how URI took those market factors into account when initial salary offers were tendered. At URI, there has been no coherent attempt to formulate a rational, sex-neutral scheme to account for such fluctuations. In many instances, market demand appears to have been employed as a concept of convenience, as a post hoc tool to rationalize results which decisionmakers, for other, entirely unrelated reasons, sought to reach. There is, simply put, inadequate evidentiary support for a blanket conclusion that market factors were the quintessential cause of salary disparities at URI.
The other criteria used by Rosie, Ferrante, and Newman to set salaries are easier to accept. The role of rank in fixing initial salaries is apparent. So is the validity of the tenet that the University must pay more lucrative salaries in order to induce higher quality applicants to accept URI appointments. Relevant, useful, and/or distinguished experience should be rewarded even where rank does not vary. It follows, therefore, that starting salaries among new faculty members at the University need not be uniform even though rank and duties may be identical.
B. Salary at Hire (Zellner's View)
The variety of individuals, experience, and credentials renders simplistic salary comparisons a perilous venture. In the context of this litigation, a number of sexneutral elements might account for any apparent disparity. Zellner attempted to screen out such factors in order to correlate salary differentials with putative discrimination. She first computed the ratio of women's starting salaries to men's, converting stipends to current dollars so as to facilitate comparison of salaries for individuals hired in different years. Her findings are summarized below:
Academic Years Females Males Ratio 1972/73-1980/81 All Ranks $24,336 $29,749 81.8% Instructor $20,284 $23,380 86.8% Assistant $26,012 $27,232 95.5% Associate $35,262 $36,163 97.5% Professor $48,612 $48,337 100.6% 1972/73-1974/75 All Ranks $26,310 $30,638 85.9% Instructor $22,106 $23,857 92.7% Assistant $29,167 $29,276 99.6% Associate $37,297 $38,716 96.3% Professor None $59,589 N/A 1975/76-1977/78 All Ranks $23,375 $29,980 78.0% Instructor $17,279 $24,005 72.0% Assistant $23,253 $25,288 92.0% Associate $32,946 $34,491 95.5% Professor $48,612 $46,848 103.8%
*1213 Academic Years Females Males Ratio 1978/79-1980/81 All Ranks $21,391 $26,818 79.8% Instructor $18,697 $20,292 92.1% Assistant $24,138 $25,719 93.8% Associate $31,752 $35,457 89.6% Professor None $39,231 N/A
While this is useful information, it is inadequate, without more; the table fails to account for differences in prior experience, quality of academic background, or kindred variations.
Zellner then proceeded to reckon salaries at hire by sex, according to department, adjusting to current dollars in the same manner. This revealed:
Female Relative Females N Males N to Male DEPARTMENTAL GROUP $24,080 86 $29,755 172 80.9% Agricult & Natrl Res $26,964 2 $30,643 12 88.0% Biological Science $27,962 4 $28,120 17 99.4% Business & Mgmt $29,868 1 $32,138 27 92.9% Computer & Info Sci None $29,689 2 N/A Education $25,270 3 $30,750 7 82.2% Engineering None $34,084 15 N/A Fine & Applied Arts $22,658 6 $28,077 11 80.7% Foreign Languages $13,072 2 $21,530 3 60.7% Health Professions $22,698 43 $26,141 9 86.8% Home Economics $24,217 8 $35,685 3 67.9% Letters $22,429 5 $26,117 14 85.9% Library Science $25,775 2 $26,943 2 95.7% Mathematics None $32,358 4 N/A Physical Sciences $31,474 2 $32,992 17 95.4% Psychology $27,727 2 $28,139 9 98.5% Public Affairs $42,391 2 $28,847 2 147.0% Social Sciences $24,823 4 $26,573 18 93.4%
While it can thus be demonstrated that women were paid less than men in all but one field (public affairs), the results of this study fail to rule out a number of potential causative sex-neutral factors (e.g., length of time in the workforce, import of rank placement vis-a-vis salary, degree status, academic background). The table comprises, at best, only an indication that further analysis is warranted in order to determine whether sex discrimination was velivolant when URI set salaries at hire.
Zellner next computed the average salaries at hire (in current dollars) for men and women, considering educational attainment:
Female Females Males to Male Doctorates 27,591.10 30,281.50 91.1% Non-doctorates 22,063.90 28,084.90 78.5%
*1214 Years of prior experience or department were not featured in this tabulation. But, Zellner followed up by the performance of a regression analysis of initial salaries from 1972-73 through 1980-81. She testified that 258 individuals were hired by URI during that span. She excluded 32 individuals for whom the data was, for one variable or another, incomplete. In terms of model specification, she designated the following as independent variables: whether the doctorate had been obtained, years since receipt of the terminal degree, year of hire at URI, departmental grouping (as fashioned in her hiring pool analysis, see text ante at Part VIII), and years and type of prior experience. The resultant regression revealed:
Std. Deviations Net Effect between net Probability of of being female effect and zero this low an effect -0.057 -2.268 0.012
Under Zellner's model, gender was found to have a statistically significant impact on salaries at hire. She concluded that random occurrences could likely be ruled out as an explanation for observing a net effect so far below zero.
C. Salary at Hire (Siskin's View)
Siskin likewise carried out a multiple regression analysis to address this issue. His study included 400 individuals hired from 1971 through the 1982-83 academic year. The difference in sample size between the two studies resulted from Siskin's inclusion of 142 people hired in years which Zellner did not analyze, and from retention of the 32 "incompletes" which she had dropped. Siskin then specified a model in which the independent variables were: department (in his case, groupings of related fields), degree held, years with degree, year of hire, and rank at hire. He then ran the regression for each rank (except full professor) in order to determine whether there was a within-rank disparity in salaries at hire attributable to sex. The results of these regressions are shown below:
Effect of Standard Statistical Sex Deviations Significance R2 N Associate -135.9 0.15 No .85 49 Assistant -412.2 1.53 No .86 252 Instructor + 30.2 0.02 No .78 99
Siskin found no statistically significant disparity sufficient to give rise to an inference that gender impacted initial salaries. And, he reported that the model satisfactorily explained anywhere from 78% to 86% (the R2 value) of the variation in salary at hire.
Siskin then divided the ranks of instructor and assistant professor into two subsets. One included "inexperienced" individuals who had, by his definition, received their degrees fewer than two years before entry; the other, persons who had earned degrees two or more years before hire, *1215 were classifiedas one might have guessedas "experienced." The regression, run with the same independent variables as before, reflected the following:
Inexperienced New Hires Starting Coefficient for Standard Statistically R2 N Rank: Sex Variable Deviations Significant? ------------------------------------------------------------------------------------ Instructor + 11.3 0.02 No .78 84 Assistant + 92.7 0.28 No .89 150 Experienced New Hires Starting Coefficient for Standard Statistically R2 N Rank: Sex Variable Deviations Significant? ------------------------------------------------------------------------------------ Instructor +1037.3 0.65 No .67 15 Assistant - 454.5 0.94 No .84 102
Again, Siskin found no sex-based disparity at any statistically significant level. The model, by Siskin's lights, explained anywhere from 67% to 89% of the pay variation.
D. Statistical Comparison (Salary at Hire)
Once again, Zellner and Siskin land at opposite poles. Close perscrutation is necessary, therefore, as to the adequacy of the respective models.
Zellner's model excluded rank as an independent variable, based on her professed belief that comparably credentialled men and women deserved the same pay if they were in the same departmental grouping. Zellner also theorized that to use rank in the model masked the effects of discrimination on comparably credentialled faculty members because of her findings that rank placement was not itself neutrally aligned. Siskin, on the other hand, employed rank as an independent variable, contending that discrimination in rank placement (even if extant) would not affect salary at hire within a given rank. Siskin argued that his model actually examined whether there were sex-based disparities in rank given the actual placement decisions.
At first blush, neither asseveration as to the use of rank seems frivolous. On balance, however, the court is of the opinion that the exclusion of rank in the study of salary at hire is considerably more logical. If a female is misranked at hire by reason of discrimination, she will be over-qualified for the lower rank and her compensation will tend to be at the upper reaches for that rank. Though her salary will be equal to or higher than men at that level, such apparent equality will be doubly misleading: it will be fallacious in her case, and it will skew the true picture of salary at hire within the lesser rank. If enough women fall into this category, they will collectively disequilibrate the average compensation of women toward the high end of the scale in an artificial manner. Only if such misranked individuals are removed from the lower rank can an analysis within rank be appropriately performed to determine the existence of sex-based disparities in pay at hire. Since no such analysis is available, comparably credentialled comparisons are the most viable alternative. Siskin's model, of course, made no effort to outlaw individuals who should have been assigned to higher ranks. Given this court's finding that random chance can be ruled out as an explanation for most rank placement disparities and that the pernicious effects of discrimination were afoot in this area, see text ante at Part IX, Siskin's model is a farrago structured so as to conceal possible discrimination in salary at hire. To be sure, Zellner's model is not perfect. It *1216 fails to account for the importance of rank in salary determination. But, it has the offsetting virtue of pairing comparably credentialled individuals. It is a more plausible approach than Siskin's use of camouflage.
Moreover, the salience of prior experience in a reasoned determination of salary at hire cannot be gainsaid. Though Siskin acquiesced in this notion, he chose to forego experience in his model in favor of a proxy (years since degree). That choice defies logic; the relationship between the variable and the proxy is tenuous at best. The court finds that Zellner's inclusion of this variable per se in the regression analysis (even without qualitative differentiation) and Siskin's rejection of it in favor of an incongruous proxy weigh heavily in favor of Zellner's technique.
In addition, the equations used in Siskin's study were inappropriate. Siskin utilized real values for the independent variables. The court is inclined to the view, however, that it is more reliable to employ an equation with the natural log of earnings being a function of the independent variables (the years of experience should be squared). So, on the matter of functional deployment, Zellner's model seems superior; she used the appropriate form of equation in her multiple regression analysis. Furthermore, the inferior structure poses serious questions about Siskin's use of R2 as a statistical buttress for his results. Professor Fisher notes that a change in the functional form of an equation will alter the R2 values. Fisher, Multiple Regression in Legal Proceedings, 80 Colum.L.Rev. 702, 720 & n. 31 (1980). The R2 value is, under such circumstances, to be engaged with great caution. Id. The court concurs. Zellner's use of P values seems much more solid.
To sum up, while both models have faults, Zellner's is assuredly the more valid, and is more convictive on the salary at hire issue. The court is persuaded by its results. And, the court finds that the observed disparities in starting salaries were not the result of random events. Some other explanation must be inferred.
E. Anecdotal Evidence
Expert testimony was far from the only evidence introduced about entry level salaries. Much of the nonstatistical proof focused on the report of the Rosie Committee and on the study prepared by the Women's Inequity Committee (WIC) of the AAUP. Some specific case histories are also of interest.
1. The Rosie Committee Report
The origins and charge of the Rosie Committee have been cited, see text ante at Part V (A), and no useful purpose would be served by repastinating that ground. To carry out that emprise, the committee chose to use the matched pairs concept, originally advocated at URI by Rocha (in December of 1974). Matched pair analysis requires the investigator of an equal pay dispute to compare similarly situated employees. The underlying thesis, of course, is that such employees should earn the same amounts of money if they are performing equivalent work. If employees of one gender are consistently paid more for the same job than employees of the opposite gender, and if legitimate explanations for the divergence do not exist, then a salary inequity has been established (although not necessarily a violation of the Equal Pay Act).
The Rosie Committee adapted this theory to the URI setting. First, so-called "perfect matches" were sought, i.e., male/female pairs who, inter alia, were in the same department and enjoyed the same seniority at URI, and who had the same academic degrees and number of post-degree years in the workforce. If no such match existed within a department, an effort was made to replicate individuals as closely as possible, following the contours of this paradigm. In some departments gender was so one-sided that the possibility of finding a good match within the field was effectively foreclosed; the option of *1217 choice in such straitened circumstances was to seek matches in other departments within the same college, or if none such could be had, to look to other colleges. Rosie admitted in his trial testimony that some of his group's "comparables" were selected virtually out of desperation. Failing the discovery of any realistic match, the last resort was to utilize average salaries of a cross-section of faculty members for the comparison.
Ideal matches, to the extent they existed and could be identified, presented scant difficulty. The Rosie Committee examined 1977-78 compensation records and, if the salaries diverged, ascertained whether either individual in the comparison had received non-standard increases for merit, inequity, or market factors. The effects of these boosts were eliminated and the salaries were compared anew. If a gap still existed in favor of the male member of the pair, the Rosie Committee tentatively assumed that gender was a cause of the disparity and recommended a dollar award to rectify the situation. A similar process was undertaken both for the 1976-77 academic year and prospectively for the 1978-79 academic year (projecting compensation by means of the across-the-board percentage increase negotiated in the applicable collective bargaining agreement).
But, perfection is a goal rarely attainable in academia or elsewhere; and the committee had few ideal matches. Thus, in the majority of cases, the Rosie Committee was forced to reconstruct salaries. This process of restatement began with examination of average initial salaries of those persons hired at the appropriate rank in the year in which the female half of the pairing was initially employed. If the average salary was greater than the woman's starting salary, then the average became the jumping-off point. The across-the board percentage increase in compensation as per the union contracts for each year up to and including 1977-78 was applied, and merit increases (if any) actually received by the faculty member were added. The woman's reconstructed salary was then compared with her actual 1977-78 salary. If the former was greater than the latter, the existence of some type of sex bias was tentatively assumed, and a backpay award was suggested in order to ameliorate the situation.
To truncate a long and tedious tale, the Rosie Committee identified a total of 53 women who were arguably subjected to some form of pay-oriented discrimination and suggested lump-sum awards totalling $55,064 (an average of $1,038.94 per putative discriminatee). The appanages ranged from a low of $193 to a high of $1911 for specific faculty members. Most of the beneficiaries were women in home economics, nursing, and English.
WIC, as indicated above, was a group formed under the aegis of the AAUP after this litigation was underway. It had no official standing. Its main thrust was to conduct a matched pairs inquiry contrapuntal to that of the Rosie Committee. WIC's findings are set forth in a lengthy document variously known as Exhibit G or as Attachment A (WIC Report). The WIC Report assembled over sixty ostensible examples of women who had allegedly been treated unfairly in the context of salary vis-a-vis their male counterparts.
Reilly, who was intimately involved in the preparation of the WIC Report, explained the mechanics. WIC adopted, in essence, the matched pairs approach fathered by the Rosie Committee. Much of *1218 the data used by WIC came from the URI administration, including records of merit awards, inequity adjustments, and market factor increases. But, WIC devised some new wrinkles both in finding matches and in recasting salaries that varied materially from the earlier methodology.
WIC, like the Rosie Committee, began its labors with a search for the evanescent "perfect matches." Few existed. As an alternative, WIC then sought to compare individuals in the same department who had obtained degrees in the same year (even if the year of hire differed slightly). In cases where this alternative did not pan out, the WIC team sought to create pairings by bracketing similar disciplines or jobs with similar skills. Nursing faculty illustrates the point. Direct comparisons were extraordinarily difficult to come by. So, many members of the nursing faculty were mated with persons whose main function was instruction in studio-type courses (e.g., fine arts, music), on the rather dubious theory that both nurses and studio artists essentially taught in a practicum-oriented discipline which required equal skill and equivalent work. Despite this creative human engineering, WIC in certain instances had to go well outside the particular college to locate what it deemed to be an appropriate salary comparison. And, in the case of dental hygiene faculty, national salary data was used for analog purposes.
Once male counterparts were chosen (WIC was fond of using more than one male as a yardstick for each female), salaries at hire in 1972-73 and in 1978-79, after deduction of any merit increases or inequity adjustments, were compared. In cases where several men were used as comparators, their salaries were averaged for the appropriate year(s) and the distaff salary was compared to the average. Subtracting the woman's salary from the man's produced what WIC catalogued as an inequity. This difference was then compounded by the across-the-board percentage increases granted through collective bargaining, yielding a suggested backpay award and an addition to the 1978-79 base salary to correct the perceived inequity both retrospectively and prospectively.
3. Comparison (Rosie Committee/WIC)
Any reasoned comparison of the products manufactured by the Rosie Committee and by WIC, respectively, requires that the comparer note that the former reached only a three year timespan while the latter's grasp (from the effective date of the EEOA forward) was considerably more elongated. Moreover, the use of average salaries by the Rosie Committee in the reconstruction process would necessarily tend to reduce backpay awards if the average salaries of women were lower than those of men over the timespan, that is, the Rosie Committee's approach would mask any salary disparity by enshrouding it in an overall pay structure reduced by allegedly discriminatory salaries established for females. (WIC never referenced average salaries in the same sense.)
It also bears notation that the two groups differed significantly in their use of market factors. The Rosie Committee sought to employ such factors in its comparative analysis, but without much success. WIC, on the other hand, ignored market factors and instead concentrated on equal work and skill levels. Inherent in this paradigm is the assumption that all faculty personnel had to perform equal work. In one very general sense, the assumption was valid. All faculty at URI, when hired, were expected to teach, do research, and serve on University committees. But, the implementation of these expectations varied widely. A professor of nursing had to have certain teaching skills which were markedly different from his or her peers who taught only in the classroom. To like effect, the research expected of a studio artist had little in common with that anticipated from a chemist. WIC attempted to account for such variations by comparing those persons in clinical and/or practicumoriented *1219 fields with each other, despite farranging dissimilitudes between the specialties. And, the same broad gauge approach was reflected in the indiscriminate comparison of classroom-oriented faculty inter sese. Thus, WIC compared nurses to musicians and language professors to philosophers. It is readily apparent that the WIC plan suffered from a matched pair of drawbacks (over and above disregard of market factors). First, the court rejects out of hand the notion that, like Gertrude Stein's storied rose, a professor is a professor is a professor. And, the rather lame attempt to bisect the continuum at a few points to create two or three substrates does not cure the underlying malady. WIC's premise is simply wrong. Second, the integers which WIC chose to compare often bore no relation whatever to each other; WIC, for example, did not even try to substantiate the declamation that nurses and studio musicians required the same teaching skills (much less the same research skills).
The Rosie Committee report is also flawed. In addition to its narrow temporal focus and the inability of the group to explain or to apply market factor differences in any coherent manner, many of its comparables simply did not support the claimed analogy.
Neither the Rosie Committee report nor the WIC Report, as structured, are probative of salary discrimination at hire. The matched pair concept leaves too much to the discretion of the matchmakers and requires unwarranted assumptions of market value equivalency in various fields. The studies interfere, rather than assist, in the discernment of whether any pattern exists with respect to the treatment of men's versus women's salaries at hire at the University. If anything, these reports only established the difficulty inherent in effectuating a meaningful comparison of initial salaries. Neither was issue determinative as to salary at hire (or, for that matter, as to annual compensation). All in all, the value of these entries was extremely limited.
4. A Fresh Look
Although neither of the reports sheds a very bright light on the issue of entry level compensation, the contents do provide the court with an opportunity to examine pay at hire in a flesh and blood context, beyond the arcane world of regression analysis. The court (whilst using information from both studies) will, however, examine starting salaries in an idiocratic manner, different from the approaches which the two more partisan bodies have selected. The court, for instance, assumes at the outset (though recognizing the absence of evidentiary support for the premise) that the University's use of market factors was both well-defined and appropriate. Based upon the evidence, the court has identified the comparisons in which individuals were hired after March 24, 1972 and has reviewed each salary against the backdrop of the national and regional averages for new Instructors and assistant professors anent salaries at hire as reflected in the Oklahoma State surveys. The following tables result:
*1220 FEMALES NAME RANK YEAR START NAT'L REG'L NAT'L REG'L AT OF AVER AVER AVER AVER (DEPARTMENT) HIRE HIRE SAL (ASST) (ASST) (INSTR) (INSTR) Sussman Asst 1974 11,000 11,625 11,000 (Language) Morgan Instr 1973 12,000 12,712 14,202 10,843 11,206 (Nursing) Bockstael Asst 1976 16,200 16,419 15,133 (Res Econ) Brown Asst 1973 13,500 12,010 11,680 (Chem) Hufnagel Lect. 1973 11,109 13,574 N/A 10,840 (Microbio) Hufnagel Asst 1975 14,000 12,500 N/A (Microbio) Kass-Simon Asst 1973 12,500 12,545 11,413 (Zoology) Schach-Cook Asst 1974 11,100 12,208 14,500 (Hist) Murphy Asst 1977 16,000 14,223 13,857 (Poli Sci) Allred Asst 1974 15,000 14,044 12,932 (Bus Ed) DeCosta Asst 1976 13,000 13,713 13,325 (Nursing) Shea Asst 1975 14,000 13,079 12,483 (Sociology) Christner Instr 1974 10,000 13,753 N/A 10,947 10,641 (Home Mgmt) Higa Asst 1977 15,600 15,300 15,108 (Textiles) Roworth Asst 1976 12,500 13,000 N/A (Art Hist) Keller Asst 1975 12,000 12,351 12,133 (Studio Art) Kulberg Assoc 1974 15,000 15,850 16,234 (Psych)
*1221 MALES NAME RANK START NAT'L REG'L YEAR & AT AVER AVER OF (DEPARTMENT) HIRE SAL (ASST) (ASST) HIRE Arakelian Asst 13,000 13,355 12,845 1976 (English) Dempsey Asst 12,500 12,053 12,167 1973 (Music) LeBrun Asst 16,500 15,084 13,331 1977 (Plant Path) Laux Asst 14,500 13,574 13,574 1973 (Microbio) Levine Asst 14,000 12,581 13,371 1973 (Math) Heltshe Asst 14,000 13,934 13,667 1973 (Comp Sci) Frohlich Asst 14,000 12,526 13,500 1973 (Geol) Turnbaugh Asst 13,000 12,990 13,839 1974 (Sociology) Loy Asst 13,000 12,990 13,839 1974 (Sociology) March Asst 17,500 17,164 17,631 1977 (Mgmt Sci) Mangianelli Asst 17,500 17,164 17,631 1977 (Mgmt Sci) Brandon Asst 15,500 15,990 16,000 1973 (Accntg) Sanghvi Asst 15,000 14,693 15,574 1973 (Mgmt Sci) Dash Asst 16,000 15,097 15,000 1974 (Finance) Bracken Asst 18,000 17,608 17,000 1976 (Accntg) Rothstein Asst 15,000 13,617 13,115 1976 (Poli Sci) Glosson Asst 17,500 14,835 14,687 1978 (Theater) Onorato Asst 13,500 12,750 12,750 1977 (Art Hist)
*1222 The tables indicate that women and men were treated differently at URI in comparison to their peers elsewhere. Women were frequently paid less than their counterparts in the northeast or throughout the nation. Men, on the other hand, were almost invariably paid more (often, appreciably more) than either their regional or national counterparts. Although these homemade computations are not offered as significant in the scientific sense, they are reflective of a prevalent attitude. URI consistently treated men at least on a par with their landgrant colleagues vis-a-vis salary at hire, whilst just as consistently compensating women below that par. Not only is this evidence of differential treatment in starting salaries probative of the principal issue anent pay at hire, but it also demonstrates URI's unwholesome tendency to disregard market factors when such benign neglect suited the whims of the University administration.
The court's tabular presentation, though no model of sophistication, accepts market values and compares initial salaries of those hired with other new faculty in the same fields recruited elsewhere in the same year, thereby allowing a more insightful look into the existence vel non of a consistent gender-based pattern impacting the at-hire salary treatment of women and men at URI. And, the regular underpayment of women vis-a-vis their congenators in the same fields at other land-grant institutions is some evidence that the salary disparity between men and women was not purely the result of random events.
5. Specific Individuals
The uneven treatment of women at hire is not limited to disparate salaries in contrast to their counterparts at other institutions. URI also seems, on more than one occasion, to have sailed a different tack in initial salary negotiations with women as opposed to men. This practice led, not surprisingly, to inferior salaries for women. Four concrete illustrations should suffice.
a. Jill Bonner
Newman's charge from the Board was to upgrade URI in an institutional sense. To perform this task, he instituted studies of the various departments in order to discover strengths and weaknesses. An evaluation concerning the physics department led Newman to conclude that it was subpar. Improvement, in Newman's view, required some curtailment of the departmental mission. Accordingly, doctoral studies were restricted to areas of liquid state and neutron physics and a new department chair (Stanley Pickart) was lured to the campus. Pickart's strong suit was research. And, as part of his enlistment, he was promised an additional position within the physics department.
Pickart advertised to fill the new slot in early 1975. Anthony Nunes, an experimentalist, was found via a search and was hired as an associate professor at a salary of $17,000. His real recompense was greater, however, as URI obtained lucrative summer employment for him which brought initial compensation well over the national average. Nunes began teaching in the fall of 1976. Coetaneous with Nunes' hire, another vacancy sprang up by way of retirement. Pickart attempted to move Nunes into the retiree's position, thinking that he would still be able to fill the original open slot. But, for reasons never adequately explained, Nunes was deemed by the administration to have occupied the newly-created position.
Pickart was undaunted. He went ahead and advertised for a number of positions in the hope that URI would fulfil its original promise to up the complement of the department. The advertisement which Pickart placed in "Physics Today," a monthly *1223 publication of the profession, contained all of the customary boilerplate. But, the advertisement was somewhat atypical in stating that the ability to garner grant and research funds would be an important factor in choosing the successful candidate. The ad noted that the availability of any position was, as yet, subject to final budgetary approval.
Numerous responses ensued. The search committee found H. Eugene Stanley to be the best qualified applicant. Dr. Stanley had been a professor at Massachusetts Institute of Technology where he had obtained several hundred thousand dollars in grant money. Pickart requested (and received) a permanent position for Stanley from the administration, and Stanley was offered a full professorship with immediate tenure at an initial salary of $22,000. Pickart testified that URI expected at least part of Stanley's stipend to be paid by self-generated subvention, even though the offer was not made conditional upon such subsidization. Stanley demurred, taking a position elsewhere at more than twice URI's salary offer. And, the authorization for the position was revoked when Stanley declined. (No explanation was tendered at trial to explain why the slot vanished when Stanley withdrew.)
Pickart, still desirous of beefing up the department, turned to the next most suitable candidate, Jill Bonner. Dr. Bonner had received her Ph.D in 1969 from the University of London. She had acquired extensive teaching and lecturing experience during and after receipt of her doctorate. She eventually took a responsible position at Brookhaven National Laboratories, and was on staff there when she sent a letter of interest to URI. Pickart responded favorably and he began negotiating in earnest after the Stanley debacle. Pickart asked the administration for an enduring niche for Bonner, but to no avail. He then offered her a visiting assistant professorship at $14,500 per annum. Bonner categorically rejected the offer; indeed, she was offended by it. Pickart then raised the ante to a visiting associate professorship at $15,600 per annum, and told Bonner that was the best offer he could make. In addition, Pickart covenanted that he would continue his efforts to secure a permanent position for her. Bonner reluctantly agreed to enroll on this basis for the academic year 1976-77.
Pickart testified that it was University policy to attempt to fill temporary posts for the fewest possible salary dollars. The court credits this testimony. According to Pickart, however, a male physics professor was not likely to have contemplated such a picayune contract. The court infers (although Pickart did not expressly so state) that URI would not have extended such a miserly offer to a man. The record in the case leaves little doubt but that the physics department seemed to gain and lose positions depending upon the identity of the candidate for a possible vacancy. When a man was the prime candidate, a full-time continuing position existed; when the search focused on a woman, the full-time continuing position disappeared. And, in Bonner's case, an experienced, well-credentialled woman was compensated at a rate well below that paid to Nunes, (who was scarcely her equal), significantly less than her associate professor colleagues at other universities, and not much more than that earned by new assistant professors. Viewed in this light, Bonner's treatment further evidences that the starting salaries of women faculty members at the University were not equivalent with their regional and national counterparts. And, the shadows cast by Bonner are darker inasmuch as Nunes' salary was on a par with those of his peers. The narrative also exemplifies URI's disparate bargaining tack vis-a-vis female faculty.
*1224 b. Clarice Stasz
Clarice Stasz was recruited by the sociology and anthropology department to fill a senior level faculty position. During the negotiations, she specified certain conditions precedent to acceptance of the post. URI's resultant offer ignored many of her expectations. Dr. Stasz rejected the offer. Though the University reversed its field and enhanced the bid to bring it more in line with Stasz's original request, Stasz was thoroughly disgusted at URI's tactics and eschewed the new offer as well.
Nicholas Ordrey, on the other hand, was a doctoral student in engineering at Pennsylvania State University when URI offered him a position, conditional upon completion of his doctoral studies. Ordrey accepted. When he later found that, due to circumstances beyond his control, URI's doctoral deadline could not be met, the University promptly addressed and solved this problem. Ordrey was hired by URI notwithstanding the shortfall; though his rank was reduced to that of instructor, his remuneration was maintained at the assistant professor salary level. The department was so cooperative that it professed a willingness to have someone cover Ordrey's classes while he was finishing his doctorate.
The accommodation afforded an individual (male) who did not have a doctorate is in sharp contrast to the more intractable attitude manifested by the University toward Stasz, an individual (female) seeking to fill a senior level position. The court has little doubt that each would have fared differently but for the accident of gender.
c. Nancy Bockstael
Nancy Bockstael came to URI as a graduate student in resource economics. A teaching vacancy arose at about the same time she was completing her doctorate. Bockstael entered into negotiations with Darrell Hueth, the department chairperson, and they agreed on a salary of $17,000. Subsequently, but after the passage of time had severely diminished Bockstael's ability to find an alternate position, Hueth cut the anticipated wage to $16,200. Bockstael protested; Hueth was adamant. He told her that $16,200 was the final figure. Bockstael grudgingly acquiesced, although the pay cut put her below the salary norm for new assistant professors in the field.
Bockstael, not short of memory, renewed her complaint two years hence. Hueth brushed her off. She was eventually able to secure a salary upgrade and an "exceptional salary increase" from the administration, but as the result of a competing offer from the University of Maryland and not because of the shabby manner in which URI had treated her.
In contradistinction to Bockstael's situation, Dr. John Sutinen had no trouble in obtaining a salary on par with his colleagues. He received his doctorate from the University of Washington in 1973 and then joined URI as a research associate. After one year in that position, Sutinen was appointed to a visiting assistant professorship. He served in that post for two years when he was nominated to a full-time continuing position in 1976-77 as an assistant professor with a salary of $18,650. Bockstael was made an assistant professor the same year.
While some of the salary differential between Sutinen and Bockstael can be explained by the former's experience and seniority at URI, it does not explain all of it. Sutinen started at URI with a salary of *1225 $15,000 in 1973. Even after the application of standard collective bargaining increases, Sutinen's adjusted salary, without more, would not have approached the $18,650 salary which he in fact received in 1976-77. Though the defendants tried to distinguish the cases in terms of learned publications, the court rejects that suggestion as being pretextual and without adequate basis in the record. Significantly, Sutinen's starting salary in 1973 was higher than the national/regional average for new assistant professors in 1974-75. His 1976-77 salary was likewise more munificent than that year's average for assistant professors (or for associate professors, for that matter). Bockstael was hired for 1976-77 at a figure less than the national average. The only distinction which can, on this record, plausibly account for the discrepancy in the remuneration of these persons in comparison to each other and to their respective peers was sex.
d. Sandra Kraynek
Kraynek's saga really begins with the anointment of Tate as the dean of the college of nursing. The Board charged Tate with the task of upgrading the college to meet the standards required for accreditation of a master's degree program. To gain such stature, it was incumbent upon Tate to attract persons qualified to teach at the graduate level. This meant that Tate had to hire individuals with doctorates or those who were enrolled in (or willing to enter) doctoral programs upon hire.
Some years passed. Tate's efforts had borne fruit and URI had an accredited master's program in nursing. In June of 1981, a key faculty member, Brenda Bissell, abruptly quit. Tate, under pressure of time, needed to find a replacement who would maintain the standards of the college and specifically, an individual enrolled in, or ready, willing, and able to undertake, a doctoral program. She quite properly took the nearest way and drew upon the results of an ongoing search for another position in Bissell's specialty. That search had considered Joseph Champlin, a specialist in psychiatric nursing with a child psychiatry subspecialty. Champlin had a master's degree in nursing and had some experience as a co-supervisor of graduate students. Dean Tate made Champlin an offer of $19,000. Champlin demurred, and Tate raised the ante to $20,000 with the precondition that he enroll immediately in a doctoral program. Champlin accepted.
Kraynek had been hired as an assistant professor in the college of nursing for the 1981-82 academic year. Like Champlin, she had a master's degree and good relevant prior experience (supervising students, teaching on the undergraduate level, and in administration); she also possessed a specialty in psychiatric/mental health nursing. Her letter of appointment offered her a $17,000 salary with the condition that she obtain a doctorate on or before the mandatory tenure decision date (which occurs seven years after hire). Kraynek accepted and came to URI on this basis.
Tate labored prodigiously to explain the Champlin/Kraynek salary differential at trial. She testified that Champlin was required to secure his doctorate within three years of hire, while Kraynek had seven years within which to accomplish the same feat. But, there was no documentary support for the supposed Champlin precondition, and the court summarily rejects the proffered explanation as feigned. The court accepts Kraynek's testimony that it would require almost immediate enrollment in a doctoral program after hire for an individual to obtain the degree by the mandatory tenure decision date, bearing in mind that the person could work only part-time toward the doctorate while teaching. The court finds that Champlin was not required to complete his degree in three years; that the conditions of Kraynek's and Champlin's employment were the *1226 same; and that Kraynek was paid three thousand dollars less than Champlin. Moreover, the salary received by Kraynek was substantially less than the salary received by her peers regionally and/or nationally. The only distinction between these parties which could logically have accounted for so wide a gap in pay at hire was the difference in their gender.
F. Summary (Salary at Hire)
1. Classwide Discrimination
The court finds substantial evidence that the starting salary differentials between men and women at URI were not the result of random events. The court also finds that a pattern had developed at URI in which men were compensated on par with or better than their colleagues elsewhere, but women were not. And, the anecdotal evidence reinforces the statistically-suggested conclusion that the treatment of women vis-a-vis their congeners elsewhere in academia was no accident, but resulted in part from a negotiating stance that differed significantly between men and women. The University intentionally and sentiently applied a double standard. Given the manifest superiority of Zellner's model, see text ante at Part X (D), and the other available evidence, the plaintiffs have successfully demonstrated that, as a rule of practice, women were hired at URI at lesser salaries than comparably credentialled males.
One threshold question remains, however: were the positions "comparable with respect to skill, effort, responsibility, and working conditions?" Winkes, 747 F.2d at 793. On its face, the question strongly suggests an affirmative answer. The attendant "working conditions" were certainly, for Equal Pay Act and Title VII purposes, substantially the same for all faculty personnel. The variables accounted for by Dr. Zellner seem to insure rough equivalency in skills. The levels of required effort are not readily distinguishable. And, while some academic posts plainly impose more burdensome responsibilities than do others, the record does not bear out any differentiation along male/female lines in this wise.
Nonetheless, confronted with an alleged case of discrimination between a female in one department and a male in another department, the defendants invariably claimed that equivalency was wanting. This contention has been made so insistentlyand it so blatantly misconstrues the concept of equivalency under the Equal Pay Actthat the court will attempt to set it to rest once and for all.
In Melanson v. Rantoul, 536 F. Supp. 271, 286 (D.R.I.1982), Judge Pettine noted the following with respect to equal pay:
As a general proposition, the term "equal" is one of "substantial equality." Comparison of jobs requires that the duties be "closely related" or "very much alike," and require substantially equal skill, effort, and responsibility under similar working conditions. In making these determinations one looks to the "actual job performance and contentnot job titles, classifications or descriptions." (citations omitted).
Unlike Melanson, this case is a class action in which all equivalencies are called into effect. Thus, the court's focus must be on blanket comparisons, comparisons which (while undeniably supportable in the record) blur departmental distinctions to some extent. And, it must be remembered at all times that the Equal Pay Act is a broad charter for women's rights in matters of economics, which should be construed and applied in an expansively remedial fashion. E.g., Bence v. Detroit Health Corp., 712 F.2d 1024, 1029 (6th Cir.1983), cert. denied, ___ U.S. ___, 104 S. Ct. 1282, 79 L. Ed. 2d 685 (1984); Shultz v. Wheaton Glass Co., 421 F.2d 259, 265 (3d Cir.), cert. denied, 398 U.S. 905, 90 S. Ct. 1696, 26 L. Ed. 2d 64 (1970).
*1227 As previously noted, all URI faculty members are expected to build on all three legs of the academic stool: teaching, research, and service. The trilogy itself establishes, at least on one level, a general equivalency among faculty members. But a detailed analysis makes the point more powerfully for Equal Pay Act purposes.
Most departments at the University require faculty to teach three courses per semester and simultaneously to engage in research. Though there are some slight variations among departments and colleges, the courseload is nearly uniform, and teaching obligations are much the same. In this respect, the majority of the faculty have equivalent jobs. To be sure, presiding over an introductory history course composed of 200 freshmen is different than teaching neutron physics to doctoral candidates; but for purposes of these cases, the distinction is more apparent than real. Cf. Brennan v. Prince William Hospital, 503 F.2d 282, 288 (4th Cir.1974) (frequent performance of catheterizations and other procedures not a sufficiently meaningful added duty so as to justify wage differential for male orderlies inasmuch as female aides performed the same work), cert. denied, 420 U.S. 972, 95 S. Ct. 1392, 43 L. Ed. 2d 652 (1975). Whereas the Melanson defendant segregated faculty into rigidly demarcated lines of work, no such pattern of assignment was evident at the University. With rare exceptions (e.g., community planning), no URI faculty members were limited by writ of hire to the teaching of graduate degree courses. At bottom, scant difference in job responsibility results from the fact that URI has undergraduate and graduate degree programs. Most faculty teach both; and most are hired with the intention that they be able to do so. The University's line-up card is peopled, within a given field, with all-around players rather than designated hitters. Examples abound: Bonner taught at both levels; Weeks sought Chang's dismissal because she could not teach graduate students. In short, the faculty members have similar responsibilities for the most part (at least with respect to lecture-type courses).
It is true, of course, that teaching outside the environs of the lecture hall (e.g., directing a laboratory course or a clinical seminar in the performing arts) involves somewhat divergent skills. It would, at first sight, appear inappropriate to compare an individual in the chemistry department with, say, a person in the philosophy or history department. But, appearances are apt to be deceiving. It is rare that a laboratory course does not have a lecture component. Further, almost every professor who presides at a clinical course also teaches one or more lecture courses. See, e.g., Pl. UU(3). And URI is prone to hire adjunct or specialized clinical faculty in fields (e.g., nursing, dental hygiene) laden with heavy clinical components. See id. Thus, the seeming disparity is more a hollow than an abyss.
In fine, the court finds that the responsibility imposed, the skills required, the effort expended, and the attendant working conditions are substantially equal for most faculty, male and female alike. The sprinkling of exceptions (which, in any event, are not by and large gender-dependent) do not materially detract from the inherent classwide equivalency.
The court need not tarry long in its consideration of the second leg of the stool. Little variation exists anent research expectations: with the exception of a few departments (e.g., those specializing in the performing arts and in applied fine arts), research is a sine qua non, and publications and/or presentations at various scholarly colloquia are obligatory. With only minor divagations, all University faculty must carry on research and announce the results. The skills needed are much the same despite the diversity in disciplines: all teaching professionals must meet basic requirements of originality of thought, expertise, perseverance, organization, and command of language skills sufficient to make the fruits of one's research meaningful and comprehensible. Research responsibilities, skills, and effort are, in the court's view, *1228 relatively uniform across department lines. And, a like conclusion can be drawn concerning the artistic criteria which apply to those engaged in less traditionally intellectual endeavors.
Service, too, is indistinguishable along gender-based lines for Equal Pay Act purposes. The chief components of such service entail sitting on various University committees and organizing sundry activities and programs at URI. The record is ample to show that no faculty members are, as such, exempt from such a commitment. And, all faculty serve URI under substantially similar conditions, demanding equivalent time, skill, and effort.
Taken in their ensemble, the rigors of the three legs of the stool impose on the faculty as a whole substantially equal responsibilities; and the fulfilment of those duties demands roughly equal talent, prowess, and effort. The court finds that the faculty within the salary at hire subclass (i.e., women hired by the University subsequent to March 24, 1972) as a group performed comparable work to the male faculty similarly so hired, within the ambit of the Equal Pay Act. The evidence presented a compelling case that women with credentials similar to men were offered and paid less than men at hire for performing comparable work. The plaintiffs have met their initial burden and established their prima facie case. And, the burden thus shifts to the defendants to prove the applicability of one of the four Equal Pay Act exemptions. Corning, 417 U.S. at 196-97, 94 S. Ct. at 2229; Winkes, 747 F.2d at 793.
The perceived differentials in salary at hire do not relate to a "seniority system;" by definition, all of the members of the subclass and their male congenators were starting afresh in URI's employ. The "prior experience" component is different than seniority; and, in any event, it has already been factored into the mix, see text ante, without materially improving the defendants' position. Nor can the defendants rightfully claim that initial pay differentials existed due to differences based on productivity. Since the persons in question had yet to produce anything at URI when first hired, it would be an incredible feat of vaticination to attempt to justify disparities on the basis of this statutory exemption. "Productivity," as that term is used in the Equal Pay Act, implies stratification of compensation based on a "quantity or quality system." See 29 U.S.C. § 206(d) (1) (1982). There is no evidence here sufficient to bring the University within that integument.
There has been no merit system at URI in the strict sense of the phrase. And, the First Circuit has plainly indicated that the merit system exemption is inapposite in the traditional university setting. Winkes, 747 F.2d at 793. Assuming arguendo that an Equal Pay Act "merit system" is not limited in application to current employees, cf. EEOC v. Aetna Insurance Co., 616 F.2d 719, 725-26 (4th Cir.1980) (open question whether merit system limited to current employees or can be utilized to differentiate starting salaries), the only merit system arguably available to URI was one which differentiated initial salaries on the basis of quality, quantity, and type of prior experience and the highest degree held. And, Zellner's model demonstrated a statistically significant salary differential for individuals in the same department, with the same degrees, and possessing similar amounts and types of prior experience. The defense has not established the existence of merit-based salary differentials within the meaning of the Equal Pay Act.
The final exemption to be considered is a general one, that is to say, "a differential based upon any factor other than sex." 29 U.S.C. § 206(d) (1) (1982). The statutory language is, however, not entirely permissive. See generally Note, Not Just "Any Factor Other Than Sex": An Analysis of the Fourth Affirmative Defense of the Equal Pay Act, 52 Geo.Wash.L.Rev. 318 *1229 (1984). Courts have routinely declined to accept such a defense unless the employer can show that its justification for the salary spread relates to the bona fide requirements of the job or to the person's performance of the work. E.g., Brennan v. Prince William Hospital, supra; Shultz v. First Victoria Nat'l Bank, 420 F.2d 648, 656-57 (5th Cir.1969).
The defendants contend that market factors and rank, singly and in combination, accounted for the seeming salary disparities. The court demurs. While market factors, appropriately used, can be a valid distinguishing element in some cases, see Lamphere v. Brown University, 685 F.2d 743, 750 (1st Cir.1982), they are of scant utility in circumstances in which URI administrators are unable either rationally to explain their use in the University's salary-setting process or to account for wide variations in their haphazard utilization in essentially kindred fields. Moreover, the court's finding that the University's treatment of women with respect to their male and female colleagues nationwide was inferior to that accorded to men undercuts URI's professed reliance on market demand in this context.
Rank is certainly an important element in the even-handed determination of starting salary. In fact, the salience of rank is documented by inclusion in the collective bargaining pacts of rank-based minima for each rung of the academic ladder. But, URI discriminated against women with respect to rank placement at hire. See text ante at Part IX. Inclusion of rank as a possible defense to an equal pay claim would be equivalent to permitting URI to profit from its own wrongdoing. Such a result cannot be countenanced, as it would do violence to abecedarian principles of English and American jurisprudence. E.g., Glus v. Brooklyn Eastern District Terminal, 359 U.S. 231, 232-33, 79 S. Ct. 760, 761-62, 3 L. Ed. 2d 770 (1959); Deitrick v. Greaney, 309 U.S. 190, 196, 60 S. Ct. 480, 483, 84 L. Ed. 694 (1940); Holman v. Johnson, 98 Eng.Rep. 1120, 1121 (1775) (Mansfield, L). And further, the anecdotal evidence supports a finding that, even as between equally-ranked men and women, the University impermissibly preferred the former in fixing pay at hire. As with the market factors defense, URI's plea that rank placement was a nondiscriminatory differentiator vis-a-vis starting salary is bogus.
The defendants have failed to prove, by a preponderance of the evidence, that any of the four Equal Pay Act exemptions apply to the subclass. No legitimate management reasons underlay URI's penchant for disparate treatment in this regard. The court finds that the defendants had and exercised a pattern and practice of discriminating against women with respect to starting salary at hire from and after March 24, 1972. Their continuing fealty to gender-based discrimination after the advent of the EEOA necessarily imports classwide liability.
2. Sandra Kraynek
Kraynek has sued in her own behalf, claiming transgression of the Equal Pay Act as to starting salary. The parties have not briefed the impact of certification of the class on Kraynek's solo flight. But, because (i) her case was consolidated for trial with the class actions, (ii) since the defendants assented to a consolidated trial and never sought to stay the prosecution of Kraynek's action, and (iii) Cooper erases any questions of res judicata as to an individual claim of disparate treatment, ___ U.S. at ___, 104 S. Ct. at 2801, Kraynek's suit appears to be properly before the court.
*1230 The facts surrounding the hiring of both Kraynek and Champlin (a male) as assistant professors in the college of nursing have heretofore been set out in some detail. See text ante at Part X(E) (5) (d). Champlin was hired in the same rank and same field. Both were required to teach undergraduates. The jobs were performed under identical working conditions. The positions entailed comparable talents, effort, and responsibility. The court finds that Kraynek and Champlin performed substantially equal jobs which required utilization of essentially the same skills and energies. Kraynek, accordingly, has established her prima facie case.
The defendants' spavined attempt to rebut Kraynek's proffer rested principally on two grounds. They contended that Kraynek's contract of employment contained less onerous conditions than Champlin's because he was given a shorter time within which to obtain his doctorate. That averment was pretextual: the documentary evidence did not bear it out, and Tate's unsubstantiated testimony had the hollow ring of a strategically-devised afterthought. And, even if Champlin was expected to enroll immediately in a doctoral program, there was no meaningful difference between that requirement and the requirement (imposed on Kraynek) that a doctorate be obtained prior to the mandatory tenure decision date. The court credits Kraynek's testimony that fulfilment of this latter condition would necessitate almost immediate enrollment in a doctoral program. In no way did doctoral preconditions warrant the creation of a situation where Kraynek's initial salary was only eighty-five percent of Champlin's.
The defendants' effort to justify the disparity on experiential grounds was likewise unpersuasive. Champlin's prior experience was, in a limited sense, marginally superior to Kraynek's: he had shouldered some limited supervisory duties with respect to graduate students, an undertaking which Kraynek had never essayed. But, this experience was largely irrelevant in that Champlin was hired to teach undergraduates. In all other respects, Kraynek's credentials at hire were as good or better than Champlin's. While some minute difference in salary might arguably have been justified, the size of the spread belied Tate's assertions.
The court is convinced that URI failed dismally in its endeavor to prove that any of the exemptions specified in the Equal Pay Act apply to this case. The University's explanations appear to be gerry-built after the fact, and are not deserving of credence. The court finds that URI, despite its post-event rationalizations, discriminated against Kraynek with respect to her initial salary.
XI. ANNUAL COMPENSATION
Salary at hire is, of course, only the acorn from which the tree of annual faculty compensation, with its variegated branches, takes root. The previous discussion of initial rank placements and starting salaries comprises a springboard for perlustration of the plaintiffs' claim that the University has impermissibly repressed the yearly pay of female faculty members in comparison to their male colleagues. We begin our journey through this minefield by peeking behind the scenes to visualize the methodology employed at URI in establishing faculty salaries on an ongoing basis.
Prior to 1972, URI's personnel practices were governed almost exclusively by the Manual. Salaries were determined on a case-by-case basis. The compensation of each faculty member was reviewed yearly by the department chair and a report prepared. The chairperson would recommend whether an individual deserved a merit raise or not. If suggested by the department chair, such a boost was committed entirely to the administrative discretion of *1231 the University. At times, URI in a noblesse oblige fashion, ceded across-the-board increases to all faculty.
A new era in faculty/University relations dawned with the certification of the AAUP as the exclusive bargaining agent for the faculty and with the execution, in 1972, of the initial collective bargaining agreement. These master contracts were thereafter successively revised and renewed at divers intervals, invariably including negotiated across-the-board increases. Because these pay raises were customarily crafted on an absolute percentage basis unaffected by rank, salary level, or merit, both URI and the AAUP recognized that some further mechanism was desirable. Three separate bridges were ultimately erected to span the perceived gap: the inequity adjustment, the merit raise, and the exceptional salary increase. See ante at n. 30. There is also some evidence that "administrative supplements" were allotted to various faculty members on occasion, but these were dehors the collective bargaining agreements and appear to be immaterial to the questions presently at bar.
In 1974, URI/AAUP negotiations resulted in the establishment of a fund (slightly less than $60,000) designed to alleviate market factor inequities. Oversight of the fund was vested in a joint committee on which both the AAUP and the Board were represented. The intent was to distribute monies in the "most equitable manner so as to reduce, if not completely eliminate, the inequities which may presently exist in the faculty salary structure." The fund was to last until June 30, 1975. Men and women were equally eligible for the benefices of the pelf.
The same contract also inaugurated a collectively-bargained merit raise system, which endured through the end of the 1978-79 academic year. A special annual fund was set aside for the stated purpose, either a specific dollar amount or an amount equal to a percentage of the yearly salary base, that is, the total salaries paid in that year to all staff covered by the collective bargaining agreement. The setaside was then allocated to individual colleges in proportion to the number of full-time equivalent faculty (FTE) in each college. (The acronym "FTE" connotes a faculty member who teaches full-time, or some combination of faculty equivalent to one full-time member. For example, two half-time instructors would be counted as a single FTE at the instructor level.)
The Manual obligated each department chair to prepare an annual review for each member of his or her department, detailing achievements in teaching, research, and service. The merit raise system required the department chair, predicated on the annual review, to notify the appropriate dean of the identity of those faculty members who were deserving of merit increases and to recommend in each such instance a specific dollar award. The amount could vary from $50 to $750. Under the initial collectively-bargained version of merit hikes, the awards were made as a one-time proposition and were not added to the individual's salary base. In subsequent years, however, merit awards for 1974-75 and all future merit raises were added to the salary base. The system did not, on its face, differentiate between males and females, and there was no proof that the system was operated in a sex-partial manner. The record is tenebrous as to the criteria used by chairs and deans in the decisionmaking process.
The exceptional salary increase was introduced via the collective bargaining pact signed in 1975, and has endured since that date. As its nomenclature indicates, the exceptional salary increase was not commonly employed. Its function has been as a device to fend off overtures from other institutions to highly valued personnel. The University is, by contract, severely restricted both as to the number (one percent of the faculty per year) and amount ($3500 maximum per person) of such premiums.
With this background in place, the court turns to the statistical evidence anent annual compensation.
*1232 B. Compensation (Zellner's View)
Zellner began this inquiry by assembling data on the average salaries for men and women by rank and field. While the parties dispute her mathematics (and therefore, the size of the gap), it is clear that she found women to have been paid less than men in each and all disciplines save one. This data, by itself, does not prove discrimination. Obviously, salary differentials may result from perfectly unassailable causes: seniority, greater merit, swifter (deserved) ascension up the rank ladder, and the like. To carry her studies further, Zellner next performed a multiple regression analysis to determine the effect of sex on annual compensation. As independent variables for her study, she chose: department, attainment of the doctorate, years since terminal degree, years and type of prior experience, and longevity at URI. She did not control for rank for reasons already cited. See text ante at Part X(B). She did not isolate promotion as a separate variable. Her findings revealed:
Probability Net Effect Number of of Net Effect Academic of Standard This Far Sample Year Being Female Deviations Below Zero Size -------- ------------ ---------- ------------- ------ 1972/73 -0.040 -1.752 0.040 504 1973/74 -0.071 -2.471 0.007 554 1974/75 -0.051 -2.080 0.019 611 1975/76 -0.042 -2.329 0.010 605 1976/77 -0.051 -2.758 0.003 620 1977/78 -0.091 -3.663 0.000 625 1978/79 -0.041 -2.473 0.007 631 1979/80 -0.040 -2.279 0.012 621 1980/81 -0.063 -3.884 0.000 608
In her view, these results were statistically significant under either the standard deviation or P value test, without regard to whether a one or two-tailed test was used. From this data, Zellner concluded that salary disparities of this magnitude could not occur by chance and that sex was a likely explanation for the outcome.
The results of her study and the conclusions drawn therefrom are, however, fraught with uncertainty. Zellner wholly neglected to analyze the effect of collective bargaining on the salaries of those hired before 1972. Similarly, she omitted a separate analysis of the union contracts' aggregate effect on the compensation of those enlisted thereafter. Zellner did not analyze merit awards, inequity raises, exceptional salary increases, or administrative supplements; nor did she account for the impact of any of these. The court finds that Zellner's model failed to address many of the main decisional criteria used in determining annual compensation levels at URI. It follows that her conclusions in respect to this issue are highly suspect.
C. Compensation (Siskin's View)
Siskin also endeavored to study the impact (if any) of sex on annual compensation. He controlled for department, highest degree (at time studied), years since *1233 highest degree, longevity at URI, current rank, and rank at hire (or in 1971, if hired prior thereto). He reported the following results:
Coefficient Standard Statistically Year for Sex Variable Deviations Significant? R2 N --------------------------------------------------------------------------------------- 1971/72 -480.7 2.03 Yes .77 597 1972/73 -462.4 1.83 No .77 610 1973/74 -613.0 2.34 Yes .76 650 1974/75 -514.8 2.01 Yes .77 655 1975/76 -502.3 1.83 No .76 651 1976/77 -686.4 2.45 Yes .77 663 1977/78 -311.6 1.08 No .79 666 1978/79 -151.5 0.49 No .77 665 1979/80 - 85.3 0.27 No .78 676 1980/81 -218.7 0.66 No .77 663 1981/82 -180.9 0.53 No .77 666
Siskin found some evidence that sex was arguably a statistically significant factor in compensation in four of the eleven years studied. But, it was his opinion that the study evidenced no sex discrimination inasmuch as the analysis commingled hires antedating and postdating Title VII.
To isolate the effect of gender in post-1972 salary decisions vis-a-vis pre-Title VII recruits required a more sophisticated technique. Siskin therefore performed two additional regressions. In each, he used 1970-71 as the base year and controlled for sex, department, highest degree (at time studied), years since highest degree, longevity at URI, current rank, and rank in 1970-71. One such regression analyzed the dollar difference between salary in a given year and in the base year. The other examined the ratio of a given year's salary to salary in the base year. In essence, both of these regressions compared compensation decisions in each year rather than actual salaries.
Siskin found no statistically significant difference in the growth rate of men's salaries as contrasted with women's salaries between 1970-71 and 1981-82. And, somewhat surprisingly, he discovered that the ratio of current salary to 1970-71 compensation was statistically significant and favored women in four of the years studied (1974-75, 1978-79, 1980-81, and 1979-80).
No similar study of compensation decisions was made by either expert for individuals hired after 1971. But, Siskin did perform a separate review of annual compensation for all faculty recruited since 1971. That analysis was identical to the one which he conducted for all faculty (except that the sample size was smaller). Siskin found no statistically significant effect of sex on yearly compensation. Df's JJJ (6), Table 13. Like Zellner, Siskin failed to study in isolation the impact of the collective bargaining process on annual compensation, and never separately analyzed the awards for inequity or merit. So, while his conclusions are more soundly based than *1234 Zellner's, they, too, are less than shatter-proof.
D. Comments on the Compensation Statistics
While the statistical evidence tends to be not as helpful as a beleaguered court might wish, it is nevertheless possible to sort through the detritus. The key to the compensation question rests with the recognition that the collective bargaining agreements first locked in, and thereafter perpetuated, salary discrepancies. For example, if two individuals were hired at different salaries in 1975, their recompense in the following year, ceteris paribus, will remain different because the across-the-board percentage increase for all salaries will not tend to close the gap (to the contrary, the divergence will expand). It necessarily follows, then, that any discrimination in salary at hire will be riveted into future across-the-board pay increases, unless and until elsewise ameliorated. And, the court has already found that the starting salaries for women at URI, commencing with the 1972-73 academic year, were lower than those for comparably credentialled males. See text ante at Part X(F).
The situation for pre-Title VII hires is considerably more complex, especially because of the change in the decisional process. Before Title VII (and the concomitant emergence of collective bargaining agreements), URI's discretion in determining annual compensation was virtually unbridled. Starting salary, therefore, was not as critical as in the post-Title VII era, inasmuch as compensation discrepancies were not routinely perpetuated by subsequent across-the-board percentage increases. The introduction of collective bargaining, however, had the same effect on salaries for the 1971-72 academic year that it had on salaries of people hired thereafter. Disparities in salary in the 1971-72 academic year would be reflected in future years due to the ensuing across-the-board percentage increases. Thus, proof of differentials in salaries for 1971, be they at hire or otherwise, would tend to establish the existence of some disparity in the post-Title VII time frame. But, since valid statistical evidence on this point is wanting, the court must look to the anecdotal evidence for what it may reveal concerning salary differentials in the pre-1972 period.
E. Anecdotal Evidence
1. The WIC Report
The plaintiffs place major reliance on the WIC Report to substantiate their claim of discrimination in annual compensation. But, this document has marked flaws, and is not probative, by itself, of salary discrimination.
As noted previously, the WIC Report matches selected men and women faculty members, comparing, inter alia, salaries at hire in 1972-73 and in 1978-79. Yet, the force of the WIC Report is inextricably intertwined with the validity (or invalidity, as the case may be) of the handpicked comparisons. It should be borne in mind that the WIC Report was prepared by members of the plaintiff class under guidelines which they, themselves, formulated. There was a total lack of independent oversight of the work. The potential for self-serving choices was virtually limitless, and the signposts are clear that this potential was duly exploited. Moreover, the validity of the comparisons rested on gross assumptions never independently verified. By way of illustration, the WIC Report assumed that salary disparities existed because of discriminationan unscientific, unsubstantiated, post hoc ergo propter hoc rationalization. In like manner, comparison of salaries in 1972-73 assumed the presence of discrimination at that point. And, the WIC Report blithely ignored the realities of the salary decision-making process, *1235 especially for years prior to the onset of collective bargaining.
The court finds that, on this issue, the WIC Report lacks probative force. The court credits Rosie's testimony that the many of the pairings were not apt. (Though the court questions Rosie's reasoning in some instances, the accuracy of his overall conclusion is plain.) The plaintiffs have not demonstrated that any consistent and independently verifiable guidelines were used to prepare the study, nor have they shown that the comparisons contained therein were the product of supportable assumptions and reasoning. The exhibit is nothing more than a handy compilation of eclectic data; it cannot be considered as proof of any kind of discrimination.
2. Lucy Chang
Chang's background, history, and scope of service at URI has already been explored, see text ante at Part VIII (C) (1), (D) (2), and it would be pleonastic to repeat that material anew. Her salary in the 1971-72 academic year was $8,874 ($4,437 per semester). To support the claim of salary discrimination, the plaintiffs pointed to Tead Sosnowski. Sosnowski had a bachelor's degree in engineering and had finished his course work toward a master's degree in business administration. The record is obscure as to whether Sosnowski had his master's degree in hand when he undertook the instructorship at URI, but it seems that the degree was not officially awarded until June of 1972. He taught a total of twelve credits in production management and computer science. Dean Weeks testified that Sosnowski was paid $4,500 during the spring, 1972 term in order to meet competition in the marketplace.
Both Chang and Sosnowski were instructors in the college of business administration. Chang had a master's degree in an appropriate field and Sosnowski had completed the requirements for the degree. Chang had four years of teaching experience while Sosnowski had none. Chang earned $63 less than Sosnowski for the semester in question. The chief difference between the two was that Chang was an instructor in basic mathematics and statistics courses (and was not qualified to teach more erudite subjects), whereas Sosnowski taught courses at the next highest level. The variation in course work implicated the amount of skill and time needed to teach the course. Professor Vollmann, a creditable witness, testified that the knowledge and effort required to teach twelve credits of lower-level or upper-level undergraduate courses was materially different. This evidence, in turn, undermines the contention that the tiny difference in salary between Chang and Sosnowski was discriminatory or unexplained. Indeed, the variance in course levels would have supported a larger gap. Though Chang's seniority should have counted in her favor, this was offset by her exemption from research responsibilities (a benefice which, on this record, Sosnowski did not enjoy). The court finds Chang's claim of compensation discrimination to be utterly meritless.
3. Judith Anderson
Dr. Anderson joined the URI faculty in 1970 as an instructor in speech communication with only the defense of her dissertation remaining to secure her doctorate. This hurdle was cleared in September of 1970; she simultaneously received an automatic promotion to the rank of assistant professor. Her salary for the 1970-71 academic year was $10,400. It was increased to $11,569 by 1972-73. In 1975 she was promoted to associate professor. In 1978-79, the Rosie Committee granted her an adjustment in pay.
During her tenure at URI, Anderson taught nine credit hours per semester. Such a courseload was about the norm in the college of arts and sciences. It should be noted, perhaps parenthetically, that the number of courses and/or credit hours frequently varied from the norm for myriad reasons, e.g., the specifics of the course-work, involvement in administrative duties, acceptance of student advisory or counselling functions. Individuals who had no research *1236 responsibilities customarily taught four courses (twelve credits per semester). In addition to her teaching duties, Anderson performed research which led to publications in professional journals, and also fulfilled her obligation to the University community by serving on various bodies (including the URI affirmative action committee).
Anderson claimed that she was paid less than men who were comparably credentialled and performing jobs that required the same skill, effort, and responsibility. Specifically, she contended that three analogous males (Anthony Allen, Robert Jirsa, and Richard Roth) were paid more.
Allen was hired by the department of education in 1969 at a salary of $10,833. A year later he earned his doctorate. In addition to teaching, research, and University service, Allen headed the Curriculum Research and Development Center, established coetaneously with his arrival at URI. In 1972-73, Allen was paid $13,135. Of course, Allen was on a calendar year salary, thereby reducing the apparent spread between his 1972-73 wage and Anderson's. More importantly, Allen's duties as the director of a research center involved significant administrative responsibilities which Anderson did not share. These added responsibilities were more than sufficient to account for the salary disparity.
Dr. Jirsa, a speech pathologist and audiologist, was hired in 1970. He obtained his doctoral degree that same year; his starting salary was $13,000. URI procured the funding necessary to cover his remuneration from a United States Department of Health, Education, and Welfare grant rather than from the University's legislative appropriation (as was true in Anderson's case). Given Jirsa's peculiar circumstances, the University understandably lacked the incentive to approach the question of his compensation for the spotlighted year with its accustomed parsimony. In taking advantage of the available federal largesse, URI "was within a permissible area of choice, and made a reasonable decision." Winkes, 747 F.2d at 797. The plaintiffs' effort to contrast Anderson with Jirsa seeks to compare budgeted prunes with federally-sponsored plums. The comparison proves nothing.
Roth, Anderson's last match, was hired to teach in the speech department in 1966 at a salary of $7,200. In 1972-73, Roth (who still had not obtained his doctorate) earned $931 more than Anderson. Anderson asseverated that she taught speech communication in the same environment as Roth; that each party performed equal work under identical conditions; that her doctorate rendered her a more skilled individual; and that her recompense should have been equal to, or greater than, his.
This argument, despite its superficial appeal, is doomed to failure. Roth had been employed at URI for several more years than Anderson, and this built-in seniority enabled him to reap salary increases during the time Anderson was sowing the seeds for her doctorate. And, the two did not perform the same function at URI. Roth, unlike Anderson, had operose duties outside the classroom as a part of his position, especially as the coach of the URI debating team (a post which required him to train students for the rigors of intercollegiate debate and to travel with the team as an advisor). He frequently had to do his job off-campus and out of a classroom setting. Roth's employment required skills different from those of most other faculty in the speech department (Anderson included), and his working conditions were atypical. These differences were sufficiently substantial to destroy the validity of the match.
4. Sharon Strom
Dr. Strom received her Ph.D in 1969, the same year she was hired by URI. At that time, she became the sole woman member of the history department. Her salary was $9,500 ($11,519 by 1972-73). Her pay was adjusted upward by the Rosie Committee in 1978-79. Strom became a full professor in 1982. The other half of Strom's matched pair was Dr. Anthony Bryan. Bryan likewise received his doctorate in 1969 and *1237 joined the URI faculty that fall (as an assistant professor with a starting salary of $9,500). Both individuals taught nine hours per semester. Their compensation remained in lockstep until the 1972-73 academic year, when Bryan received an increase of almost $3,400 (bringing his salary to $14,000).
Rosie tried to justify Bryan's sudden rise in pay by pointing out that he was prized by a number of other institutions. As an inducement to persuade Bryan to remain at URI, Rosie's explanation ran, the University awarded him an exceptional salary increase (which Rosie, with an oxymoronic flair, described as "routine"). Rosie failed, however, credibly to demonstrate any concrete offers to Bryan from other schools, or to make a case that Strom was not similarly in demand. And, once the disparity was in place, the natural operation of the collective bargaining agreement thereafter widened the gap. The court has greeted Rosie's professed reason with some skepticism, as the evidence failed to show any contractual provision expressly authorizing exceptional salary increases in 1972-73. Nor did URI introduce any evidence of the specific market factors extant anent black studies at the time. On the record as it stands, Strom and Bryan were comparable; they had kindred skills, duties, seniority and the like; and the University has not demonstrated any plausible reason to explain the favoritism extended to Bryan and withheld from Strom.
5. Greta Cohen
Greta Cohen was hired in 1966 as an instructor in physical education at an annual stipend of $6900. She taught a wide variety of courses, coached the fencing team, and transformed a ragtag terpsichorean group into a dance company. By 1969, Cohen was earning $8,800. In that year, John Norris was hired by the physical education department at a salary of $12,000. Norris had a master's degree, taught a multitude of courses, and served as head coach of the varsity baseball team. As a coach, release time was ceded to Norris during the baseball season. Gaza Henni was likewise recruited by the physical education department in 1969, earning $13,000. He possessed a master's degree, carried a teaching load, and coached the varsity soccer team (for which release time was granted during the soccer season). By 1972-73, Henni was earning upwards of $14,000; Norris' salary was $14,417; and Cohen's was $10,592.
Coaching was the foremost responsibility of both Norris and Henni, whereas Cohen was primarily a teacher. Cohen's duties anent the fencers never warranted release time. In fact, by 1974 Cohen no longer had any coaching duties at all. The positions held by the two men, on the one hand, and by Cohen, on the second hand, entailed different skills, were performed in disparate environments, and were not fairly comparable. And, the trio came to URI with vastly different backgrounds (though all three held master's degrees). Cohen had but two years of high school teaching experience; Norris had spent five years of teaching cum coaching at Norwich University; Henni's credentials included international soccer stardom in Europe and nineteen years of coaching at various levels (including professional and international soccer competition). The broadscale disparities in background and experience, without more, would suffice to explain the salary differentials both in 1969 and thereafter. The plaintiffs' effort to pair Cohen with either or both of these men was bootless.
Cohen also used her dance group as the foundation to essay a comparison with other (male) performing artists. Specifically, she targeted Robert Steinberg (theatre) and John Dempsey (music) in this regard. *1238 Steinberg was hired in 1971 to help URI develop a program directed at obtaining accreditation to award the degree of master-of-fine-arts in theatre. While he did not hold such a degree himself, he possessed substantial professional experience in stage design. In 1972-73, he was paid $11,000. Dempsey was engaged in 1973 to teach music theory and violin performance. He had earned a master's degree, was partway on the road to his doctorate, and was a noted concert performer. URI paid him $12,500 for the academic year 1973-74.
Again, the attempted comparisons flounder. Cohen's main responsibility was her coursework (devoted chiefly to dance). While Dempsey's paramount function was the teaching of violin performance, he (unlike Cohen) was required to perform in various University-sponsored musical organizations. Steinberg's lot was even more alien: though he taught various seminars and practicums in scenic design, he also served as an intermediary between URI and the O'Neill Theater, a Connecticut-based repertory company which was to serve as a training ground for the students in the MFA program.
The evidence was clear that the physical education department, albeit impressed with Cohen's choreographic progress, did not view her work as integral to the departmental enterprise. In stark contrast, both Steinberg's and Dempsey's nonteaching responsibilities were vital parts of their jobs and significant to the advancement of the missions of their respective departments. And, Dempsey's performance responsibilities were part and parcel of his job. Similarly, Steinberg was hired not so much because of his academic or professional credits but because of his contacts in the professional theatre. To compare Cohen with either of these two men is simply unfair and unrealistic.
6. Diane Seleen
Dr. Diane Seleen is currently an associate professor of physical education. She came to the department as a graduate student in 1969. When a temporary position became available for the 1971-72 academic year, Seleen (who had earned her master's degree that year) was recommended for the position by the department chair, M. Dorothy Massey. Seleen was appointed at a salary of $8600 per annum. The following year, Carol Plunkett resigned. Seleen was moved into Plunkett's spot on a continuing basis, with a $400 hike in pay. Seleen earned her education doctorate in 1981 from Boston University. She initially managed 12 teaching credits per semester, coached the women's varsity swim and volleyball teams, and served with considerable distinction as a student advisor (a role which was, at bottom, optional on her part). In 1974, female coaches were given the choice of coaching in addition to their regular teaching duties, or not coaching. Seleen relinquished her palaestral duties. By 1978-79, Seleen had become an assistant professor at a salary of $15,345 (exclusive of any merit increments).
Seleen proffered three matches: Norris, Douglas Johnson (a community planner), and Charles Latos (an economist). None of these men had doctorates when hired. The attempted comparison with Norris need not long detain the court. Norris' chief function was coaching intercollegiate sports; Seleen's primary undertaking was teaching. The tasks were different, the responsibilities were different, the skills were different, the working environment was different, the seniority of the two was different, and their prior experience was dissimilar. They are less a match than an odd couple. The court turns, therefore, to the two remaining analogues.
Johnson matriculated at URI in 1969, as a graduate student in a program leading to a master's degree in community planning. While so enrolled, he worked part-time for the Rhode Island Department of Community Affairs and served as a summer intern for the Metropolitan Area Coordinating Council in Wilmington, Delaware. In July of 1971, Johnson simultaneously donned two University hats: he became an instructor in community planning and the director of the Urban Research and Planning Center *1239 of the Graduate Curriculum in Community Planning and Area Development, at a salary of $10,900. A year later, his pay for the dual jobs rose to $11,610. Johnson was promoted to assistant professor in 1974. By 1976-77, he was earning $16,741. He left the University at the end of the 1977-78 academic year.
Whereas Johnson taught students in a graduate professional degree program, Seleen taught primarily undergraduates (many of whom were not majoring in her field). The differences in skill and responsibility requisite for the two positions cannot be gainsaid. And, the jobs were performed under materially different conditions.
Latos came to URI as an undergraduate economics major. He obtained his bachelor's degree in 1968 and his master's degree in 1970. As the bestowal of his master's degree approached, Latos was appointed to a special part-time instructorship at URI's extension division. In parallel with this assignment, Latos commenced work at Brown University toward his doctorate. He labored in this dual capacity from the fall of 1970 through the spring of 1972. He received a full-time instructorship at URI at a salary of $9,500 for the 1972-73 academic year. He continued to work on his Ph.D. His salary increased to $10,450 for 1973-74. During 1974-75, Latos became a so-called special instructor. From 1975 until January, 1977, he also coordinated the bachelor's degree program in URI's extension division. Latos earned his doctorate in late 1976 and was made an assistant professor in January 1977. He was assigned to the extension division at a wage rate of $14,000 per annum. By the 1978-79 academic year, his salary had risen to $16,151 (with merit, $16,551).
Basically, any salary differential between Latos and Seleen resulted from their starting salaries and the effect of the collective bargaining agreement on those salaries. Both were hired on a continuing basis in the fall of 1972. Latos' salary at that time was $500 greater than Seleen's. Even though both were instructors and neither had doctorates, Latos was enrolled in a doctoral program and Seleen was not. This difference alone could reasonably have accounted for the slender initial salary differential; and once the disparity was set in place, the collective bargaining agreements ensured that the gap would in the ordinary course be maintained (indeed, widened slightly). And, there were sufficient differences in the duties of Seleen and Latos, respectively, to justify the disparity in their salaries wholly apart from their divergent doctoral paths.
When all is said and done, the specific case histories, taken individually or as a whole, do not bear the weight of the plaintiffs' asseveration that URI discriminated against women in the fixing of ongoing pay rates.
F. Summary (Annual Compensation)
The issue of annual compensation requires the court to undertake a somewhat different analysis than that utilized for salary at hire. The need for such a detour arises from the combined impact upon the wage structure of the court's classwide findings anent initial salary and rank placement at hire, and the confluence, in 1972 and thereafter, of Title VII and collective bargaining. These factors require the court to treat post-Title VII hires differently than pre-Title VII hires. Only then need the court scrutinize any individual claims.
1. Annual Compensation: Faculty Hired After March 24, 1972
The issue of the fairness of annual compensation for women hired after the effective date of Title VII depends, to a large extent, on the interaction of starting salaries and the collective bargaining agreements. To backtrack briefly, it must be borne in mind that two individuals who started with different paychecks at, say, the beginning of academic year 1973-74, would have those differences maintained and swelled throughout their careers at URI by reason of across-the-board percentage raises dictated by the Union pacts (at least in the absence of special wage adjustments or intervening promotions). Since *1240 the court has found classwide discrimination both in salary at hire and in rank placement at hire postdating March 24, 1972 (which in turn impacted starting salary), the effects of that discrimination can be presumed, classwide, to have permeated the ongoing wage scales. The defendants, therefore, have the burden of producing evidence sufficient to show that intervening actions on their part ameliorated the inferentially adverse effects of these builtin disparities.
It is true that the University, on divers occasions, ceded a variety of payments to women in an effort to offset such inequities in remuneration. But, URI has failed miserably to show that the interim palliatives which it undertook eliminated the disparity between men's and women's starting salaries. Rather than effecting any overall cure, URI fretfully and immethodically applied a bandaid or two. Accordingly, annual compensation must be viewed as subject to the self-same pattern and practice of discrimination anent post-Title VII hires. The court finds that those women hired at discriminatory starting salaries after March 24, 1972 and who commenced employment after the effective date of the first collective bargaining agreements were, as a class, within the penumbras of the University's impermissible conduct in regard to annual compensation.
With respect to the damages portion of this case, URI will have to prove either that the initial salary of a particular claimant was fair or that the University, through inequity adjustments, the Rosie Committee awards, or exceptional salary increases, successfully eradicated an impermissible differential. There is, however, no basis in the evidence for a suggestion that merit raises were similarly tainted; and individual claimants will be foreclosed, on a pattern and practice predicate, from contesting any putative failure on URI's part to award them merit increases. The computation of damages in respect to this subclass should be limited to the difference in initial salary as perpetuated and augmented by across-the-board percentage increments thereafter (except to the extent offset or eliminated by interim awards of the type and kind previously mentioned).
2. Annual Compensation: Faculty Hired Before March 24, 1972
To the extent that the claim of discrimination for pre-Title VII hires is grounded upon the concept of a continuing violation, it has not been proven. Before March 24, 1972, URI was not prohibited from discriminating against women with respect to pay or any other employment practice. See Lamphere v. Brown University, 685 F.2d 743, 747 (1st Cir.1982); Powell v. Syracuse University, 580 F.2d 1150, 1154 (2d Cir.), cert. denied, 439 U.S. 984, 99 S. Ct. 576, 58 L. Ed. 2d 656 (1978). After that date, it became unlawful to discriminate with respect to the terms of faculty employment (including compensation). See, e.g., 42 U.S.C. § 2000e-2(a) (1) (1982). There is absolutely no evidence, however, to suggest that the across-the-board raises contained in the collective bargaining agreements were distributed in a biased manner. Thus, this subclass of plaintiffs can establish liability only by showing that discrimination existed prior to the implementation of the pacts, and that it was intentionally perpetuated thereafter. This raises a new question concerning the validity of the continuing violation theory based on present effects of past (lawful) discrimination.
In United Air Lines, Inc. v. Evans, 431 U.S. 553, 97 S. Ct. 1885, 52 L. Ed. 2d 571 (1977), the Supreme Court held that a Title VII continuing violation claim must do more than allege simply that an injury persists in the wake of prior discrimination. Id. at 558, 97 S. Ct. at 1889. To pass muster as a continuing violation, the Court required the showing of a presentthat is, ongoingbreach of Title VII's commands. Id. See also Velazquez v. Chardon, 736 F.2d 831, 833 (1st Cir.1984); Jones v. Somerville, 735 F.2d 5, 7 (1st Cir.1984).
A number of courts have considered the viability of a continuing violation theory in the albedo of Evans. Most have dismissed *1241 such claims with great celerity. E.g., Velazquez, 736 F.2d at 833-84; Goldman v. Sears, Roebuck and Co., 607 F.2d 1014, 1018 (1st Cir.1979), cert. denied, 445 U.S. 929, 100 S. Ct. 1317, 63 L. Ed. 2d 762 (1980); Masco v. United Air Lines, Inc., 574 F.2d 1127, 1130 (3rd Cir.1978); Fricker v. Town of Foster, 596 F. Supp. 1353, 1359 (D.R.I. 1984); see generally B. Schlei & P. Grossman, Employment Discrimination Law 1053-54 (2d ed. 1983) (Schlei & Grossman). When an individual claims discrimination with respect to pay, however, each paycheck in which the individual receives less money than another employee may constitute a separate violation of Title VII. Bartelt v. Berlitz School, 698 F.2d 1003, 1004 (9th Cir.1983), cert. denied, ___ U.S. ___, 104 S. Ct. 277, 78 L. Ed. 2d 257 (1984); Hall v. Ledex, Inc., 669 F.2d 397, 398 (6th Cir.1982); Satz v. ITT Financial Corp., 619 F.2d 738, 743 (8th Cir.1980). This rationale holds even if the only discriminatory act was the threshold decision to hire the individual at a discriminatorily low salary. Lamphere, 685 F.2d at 747; Jenkins v. Home Insurance Co., 635 F.2d 310, 311-12 (4th Cir.1980). Thus, the plaintiffs' present theory of liability remains, on its face, viable notwithstanding Evans.
The plaintiffs attempted to show that past discrimination existed by demonstrating current disparities in salary. They reasoned that any existing gulf could only have resulted from unevenness in pay at some time prior to the institution of the collective bargaining pact. Their logic is flawed.
First, there was no evidence to support the assumption that faculty personnel generally received equal compensation increases in the pre-1972 era. In fact, the evidence was overwhelmingly to the contrary. In those bygone days, the salaries of individuals were determined on a case-by-case basis as required by the Manual.
Second, there was little or no proof which tended to indicate that pre-1972 salary decisions were infected by discriminatory animus. The statistical model formulated by Zellner examined salaries in the postcollective bargaining period alone. Her paradigm revealed nothing about the decisionmaking process before 1972. A large number of valid, non-discriminatory explanations might account for differentials in salary prior to the effective date of the EEOA amendments. The burden was on the plaintiffs to show that their pay was set in a discriminatory manner. They offered no evidence from which this court can rule out valid non-discriminatory explanations for salary distinctions. Having failed entirely to show that either the setting or the maintenance of salaries was conducted in a discriminatory fashion prior to 1972, the plaintiffs cannot establish post-1972 salary discrimination as to those who were on board when the Title VII era dawned on campus. Any other result would require the court to find, as a matter of fact and law, that URI discriminated in the pre-Title VII era simply because the defendants were permitted to do so and because URI discriminated in one or more respects post-Title VII. While the argument may be tantalizing, it is based on no more than unbridled speculation. The court has a duty to base its findings on facts and on reasonable inferences drawn from those facts. And, there is no basis in the record sufficient soundly to bottom the plaintiffs' thesis.
Inasmuch as the plaintiffs have not established the existence of discrimination in the period prior to 1972, their claim of continuing violations postdating the EEOA must fail.
The claimants are on even shakier ground in their endeavor to attack the University's overall annual compensation scheme, insofar as it impacted persons hired before March of 1972, on any basis not dependent upon developments which transpired before the EEOA took effect.
In the first instance, it should be noted that the plaintiffs are unable to rest their argument on salary decisions made after the inception of, and pursuant to, the collective bargaining agreements. The pacts removed most of the University's discretion in determining annual salaries subsequent *1242 to initial hire. Discretionary raises were limited to merit hikes, inequity and market factor adjustments, and exceptional salary increases. Insufficient evidence was introduced about the award of these discretionary raises upon which a finding of disparate treatment or disparate impact could be premised. There is little evidence of the criteria for these awards, and no evidentiary predicate for a finding that male faculty members received them disproportionately (either more frequently or in larger amounts) to women. And, the plaintiffs have never claimed that these discretionary raises were meted out in a biased manner. In fact, merit raises were eliminated when salary calculations for the WIC Report were compiled. Apparently, the class members were satisfied with the distribution process on an ongoing basis, and there is no warrant in attempting to second-guess this view on so meagre a record.
Finally, the plaintiffs attempt to build a case based on the WIC Report. But, this frayed document raised fully as many questions as it answered. For individuals hired before 1972-73, the pairings matched starting salaries, salaries in 1972-73, and salaries in 1978-79. From this data, the class would have the court infer discrimination. The only way in which the court could reach such a destination would be to assume, in the absence of any adequate evidentiary basis, that salary differentials at hire (before 1972) and during the first year of collective bargaining were the result of discrimination. The court declines this open-ended invitation to hinge its decision on the gossamer strands of conjecture.
Nor does the record reveal any classwide discrimination in post-hire salaries of those persons recruited antecedent to the onset of collective bargaining at URI. Without such evidence, the court, for much the same reasons, cannot conclude that the existence of salary differentials was the functional expression of rampant discrimination.
The documentary evidence introduced in a rather lame attempt to bolster the WIC Report fails to fill the void. The court has been given no meaningful guidance as to how faculty credentials were evaluated anent compensation decisions before Rosie became AVPAA. (The record does not reveal how starting salaries were determined or what factors went into the decisional calculus in that era.) Thus, the court has no rational method for assessing whether discrimination in starting salaries existed in those bygone days, and if so, whether the discrimination in starting salaries was continued in post-hire salaries. Nor have the plaintiffs, given the presumably important impetus of pre-1972 decision-making, sufficiently refuted the notion that valid, nondiscriminatory salary decisions led to the differentials reflected in the WIC Report. As heretofore pointed out, see text ante at Part XI(E) (1), the pairings were so maladroit and self-serving that heavy reliance on the WIC Report would, in any event, be a perilous undertaking. The plaintiffs have not proven discrimination in annual compensation as to this subclass.
Because of this failure, the court need not address in this proceeding the viability of the defendants' claim of entitlement to Equal Pay Act exemptions.
3. Individual Claims
a. Lucy Chang
Chang's claim rested upon the comparison between her and Sosnowski. Chang, however, did not establish that her teaching load required the same skill as Sosnowski's. See text ante at Part XI(E) (2). Without proof that the jobs required the same prowess, Chang has stumbled in her effort to make out a prima facie case. Her claim must fall; and the court need not reach the defendants' alternative contention that Chang's compensation claim was, because of its temporal incidents, beyond the scope of the Equal Pay Act.
b. Judith Anderson
Anderson based her compensation claim upon a comparison with three other professors (Allen, Roth, and Jirsa); only the last two of these individuals were in the *1243 speech department. Anderson was unable to establish that she performed tasks equivalent to those handled by Allen or Roth. See text ante at Part XI(E) (3). She had no administrative or coaching duties at URI, and her position (a straight teaching assignment) therefore required different skills and responsibilities. Cf. Hein v. Oregon College of Education, 718 F.2d 910, 914 (9th Cir.1983) (athletic coaching jobs required different skills and responsibilities than noncoaching jobs). Though Jirsa presented a closer question, Anderson falls short on her final match as well. Jirsa's field was not rhetoric and oratory (as was Anderson's) but the pathology and audiology of speech. Both were, however, in the same department. The court assumes arguendo that Anderson established a prima facie case by Equal Pay Act standards. Both speech professors were required to perform much the same work under much the same conditions with much the same responsibilities. And, Jirsa earned substantially more.
But, even if this assumption is made, it avails Anderson naught. The University is nevertheless exculpated so long as it has proven the applicability in her case of an Equal Pay Act exemption. Corning Glass Works, 417 U.S. at 196-97, 94 S. Ct. at 2229; Winkes, 747 F.2d at 793. The evidence is persuasive that, in this instance, the salary differential arose because Jirsa's remuneration was paid through a federal grant. The subsidy permitted the Universityat no cost to itto set Jirsa's pay at any level within the confines of the grant exchequer. The process by which his salary was determined had little to do with the ordinary decisionmaking channels which governed faculty compensation at URI. The court finds that Jirsa's rate of pay was established with respect to the availability of federal subsidization and not with respect to his sex, thus satisfying the fourth exemption as set out in 29 U.S.C. § 206(d) (1). URI is not liable for the salary disparity between Jirsa and Anderson; and Anderson is not entitled to prevail in her individual claim.
c. Sharon Strom
Strom's compensation claim revolved around a comparison with a colleague in history, Anthony Bryan. Strom made out her prima facie case: she established that they performed the same work, involving identic talents, under the same conditions and with the same responsibilities, but were compensated differently. See text ante at Part XI(E) (4). The defendants' explanation of the disparity (which arose initially in 1972-73) defined logic: Rosie attributed it to a "routine" exceptional salary increase. Nowhere in the 1972-73 collective bargaining agreement was there a provision for exceptional salary increases. (Rosie testified that the offer and adjustment occurred before the institution of the collective bargaining agreement. This explanation did not obviate the problem with the permissibility of such an adjustment. The adjustment took effect after the first union pact became operative and after Title VII was fully effective. The salary control postings did not show an increase in Bryan's salary until after the effective date of both. No waiver was granted by the AAUP. The court must conclude that the adjustment was not permitted by the agreement.) Nor was any evidence introduced concerning exceptional salary increases prior to their incorporation at a subsequent date into a successor version of the collective bargaining agreement.
Winkes, 747 F.2d at 793-97, stands as authority in this circuit for the proposition that upgrading an individual's pay to match or forestall an outside offer can comprise a valid defense under the fourth exception to the Equal Pay Act. But, to obtain the benefits of this holding, there must at least be some evidence that an offer existed and that the University had a policy of trumping external offers in that fashion. The only evidence germane to the first half of this standard was an assertion by Rosie that Bryan received an outside offer. He did not remember how he became aware of the supposed competition, whether Bryan approached him on the subject, the source or amount of the alleged offer, or any *1244 other helpful details. Not a scrap of documentary evidence was addressed to the point. Bryan did not testify.
URI had the burden of proof to establish that one of the four Equal Pay Act exemptions applied. The defendants' vague assertions concerning a totally hypothetical competitive bid were unpersuasive. Rosie's garbled attempt to pass off the exceptional salary increase in Bryan's case as "routine," and his unconvincing semi-denial that gender was a factor in the spread between the pay levels of Bryan and Anderson, e.g., XXVII at 46-49, drive the final nail into this particular coffin. The University's burden has not been met; Strom has proved that URI is liable on her claim of unequal pay for equal work. Damages will be based on the across-the-board increases as they affect the disparity in the 1972-73 salaries of Strom and Bryan, reduced by the net effect of any voluntary intramural award made to Strom for the purpose of lessening the salary disparity between her and Bryan, or by any subsequent increase in Bryan's favor attributable to increased responsibilities or the like (not shared by Strom).
d. Greta Cohen
Cohen's compensation dispute arose out of salary differentials between her, on the one hand, and a male quartet (two coaches and two artists) on the other hand. The underlying facts have been set forth in Part XI(E) (5), ante.
The comparison with the two coaches failed for two reasons. Cohen's primary responsibility was teaching; Norris and Henni were, first and foremost, coaches. As the Ninth Circuit has noted, coaching requires different skills and responsibilities than teaching. Hein, 718 F.2d at 914. Cohen, who was largely uninvolved in agonistic pursuits, failed to prove a prima facie case in this regard. And, even if the jobs were comparable, the fourth exemption of the Equal Pay Act would hold URI harmless. Both Henni and Norris had a wealth of significant experience. Norris was a college coach for five years; Henni's acquaintance with soccer was vast and ran the gamut from the "Magnificent Magyars" (the eponym for Hungary's national soccer team) to coaching at the international and professional level in the United States and abroad. Cohen's experience paled by comparison; it was limited to high school coaching. The defendants have proven that the sizable gulf in experience accounted for any salary differences. Sex was not a factor.
The same holds true when Cohen is compared to Dempsey and Steinberg. They were featured individual performers; Cohen was not. Their posts required them to participate in performance-based specialties; Cohen's job had measurably divergent specifications. Her oversight of the dance company was not even a requirement (until 1978 or thereafter, when she received release time). Cohen's position had different responsibilities, invoked variant talents, and demanded different working conditions than Dempsey's or Steinberg's. Cohen has fallen short of establishing her prima facie case. And, had she done so, her claim would nevertheless falter. A close examination of her salary in 1978-79 (as compared to Dempsey's and Steinberg's) revealed nothing amiss. Exclusive of merit, the salaries reflected largely the effect of the collective bargaining agreement. Steinberg, whose rate of pay was (justifiably) higher than Cohen's in 1972-73, had basically no more than a proportionately larger salary in 1978-79. And, her pay was higher than Dempsey's at that time (reflecting her greater seniority). In any event, the unique talents and contacts of the two men render the proffered analogies worthless.
URI is not liable to Cohen vis-a-vis annual compensation in manner and form as alleged.
e. Diane Seleen
Seleen's claim of compensation rested upon comparisons with Norris, Latos, and Johnson. See text ante at Part XI(E) (6).
The court must summarily dismiss the pairing with Norris. Seleen failed to make *1245 out a prima facie case with respect to the Norris analogy; she was principally a teacher, he was principally a coach. Hein held that the skills attendant upon the two jobs bore little resemblance. 718 F.2d at 914. This court concurs. And, passing that point, the marked discrepancy in prior experience between them itself accounts for the salary differential.
Johnson was responsible for teaching graduate students in a professional degree program. In addition, he had administrative functions. His job required notably different skills and responsibilities than Seleen's; and their work environment was dissimilar. Seleen did not establish a prima facie case via her comparison with Johnson.
Finally, the endeavor to spotlight Latos as a congener was equally unavailing. The compensation variance between the two resulted from a difference in starting salaries and the effect of the collective bargaining agreements on those initial stipends. In turn, the difference in starting salaries was the result of Latos' enrollment in a doctoral program. The salary differential resulted from a factor other than sex.
The University has no liability to Seleen in regard to her personal claim for discrimination in annual compensation.
At URI, as elsewhere in the academic world, promotion confers not only added prestige and the comfort of better job security, but also affects compensation. An individual who is earning a salary less than the minimum for his or her new rank is automatically lifted to at least that mesa at the time of promotion. Additionally, the later collective bargaining agreements provided lump-sum monetary awards in escalating amounts ancillary to promotion. Thus, discrimination in promotion, if it existed, also would work a more indirect form of discrimination with respect to compensation.
Once more, the logical incunabulum is an exegesis of the promotion process.
The promotion process is essentially consigned to the tender ministrations of the URI hierarchy. The collective bargaining agreements only provide a wide canvas upon which individual promotion decisions must be painted. The detailed brush strokes are in every substantial respect left to the artfulness of the University administration.
Prior to 1972, the promotion process was articulated only in the Manual. It provided in substance that proposals for advancement (or, for that matter, for retention in rank, separation from the University, or bestowal of tenure) be initiated by the department chair. In theory, the annual review (the yearly report anent each faculty member's accomplishments in teaching, research, and service to the URI community) was the catalyst for a promotion recommendation. If such a recommendation eventuated, it was tendered to the dean of the college along with the annual review. If the dean acquiesced, the wish became an actuality; if the dean demurred, the file would go forward to the VPAA. In such instances, the decision would be made at the vice-presidential level.
The Manual also cited, in general terms, criteria for elevation. Research was the most important factor in the promotion calculus. Knowledge uncommunicated was knowledge lost; thus, a demonstrated record of imparting wisdom gained through research (or conversely, the absence of such a record) played a pivotal role. Teaching was understandably another salient element in the decision. The third leg of the stoolthe faculty member's history *1246 of service to the University community (or the lack thereof)was also a factor. And, the onset of collective bargaining did not materially alter either the criteria used in the promotion process or the subjectivity necessarily inherent in evaluating and applying those criteria.
The collective bargaining pacts did, however, formalize somewhat the procedures and lines of authority. The responsibility for the initiation of a promotion decision stayed with the department chair (although the dean of each college was given the authority to inaugurate the promotion process with respect to a faculty member if the department chair was disinclined to do so). The annual review remained a crucial document. In preparing it, the chair was expected to solicit, assess, and crystallize comments from the faculty member's peers. The essence of these critiques, along with the chairperson's own views and the faculty member's dossier, combined to make up the annual report. The department chair next formulated the written recommendation itself, which necessarily included a statement of the faculty member's duties, an assessment of professional achievement, and an evaluation of pedagogical skills. When finished, the appraisal is sent to both the dean and the faculty member. The latter may respond; if this right is exercised, the rejoinder likewise goes to the dean.
The dean reviews the material, confers with the department chair, and then prepares a written report pro or con. If the dean is opposed to promotion, his or her reasons must be specifically delineated. The dean's recommendations are then forwarded to the president (through the intermediary of the VPAA) and to the council of deans. (Prior to collective bargaining, the council, which is composed of all the URI deans, sat as a decisional body; the council, since 1972, acts only in an advisory capacity.) As a general rule, the president meets with both the VPAA and the AVPAA in order to solicit their views and those of the council on the particular application. And, the president then passes judgment. If his decision concurs with one upon which the department chair and the dean had previously agreed, that decision (whether positive or negative) goes into effect. But, if the president vetoes such a mutual agreement, he is required to state in writing the reasons for his overriding determination (and, in such an event, his decision can be grieved to Rhode Island's commissioner of education and thereafter at binding arbitration). To square the circle, it should be noted that if the dean and the department chair have split in their views, a presidential decision determines the matter.
As is evident from the foregoing, the procedure by which promotion decisions are reached at URI is highly subjective in nature. In and of itself, this is no indication that nefarious forces are at work: the nature of the beast lends itself to a subjective approach. Accordingly, the system vests administrators at each level with fairly broad discretion. And, it is the way in which this discretion has historically been exercised which gives rise to the instant protestations. As the complainants see it, the discretion has not only been abused, but it has been meted out unfavorably to women in a systematically discriminatory manner.
B. Promotion (Zellner's View)
Zellner calculated the expected number of promotions for women, assuming that a non-biased promotion process would elevate females proportionate to their representation within each rank. She attempted to refine her study by both including and excluding those faculty members who left URI. She hoped, in this way, to take account *1247 of the reason why a faculty member departed. Her reasoning was that if a teacher left due to causes wholly unrelated to the promotion process, that departure in effect removed the promotion decision from URI's institutional hands. When a person sought greener pastures in anticipation of a (presumably adverse) promotion decision, however, then that result should be reflected in the statistical analysis. But, since Zellner never collected data on the causes underlying faculty departures from URI, she could not exclude those who left the University for reasons unrelated to the promotion process. She adopted what she perceived as the next best alternative: exclusion as a group of all who had left URI by the end of the 1979-80 academic year, regardless of motivation. Zellner's findings follow:
I. Including Those Who Terminated Before 1980/81 Actual Relative to Actual Expected Expected Rank to Which Promoted Number Number (Percent) ---------------------- ------ -------- ----------- 1972/73-1980/81 Assistant 29 35.7 81.2 Associate 38 51.4 73.9 Professor 8 15.2 52.6 1972/73-1976/77 Assistant 26 29.8 87.2 Associate 17 33.8 50.3 Professor 8 11.3 70.8 1976/77-1980/81 Assistant 7 9.2 76.1 Associate 24 28.8 83.3 Professor 2 10.1 19.8 II. Excluding Those Who Terminated Before 1980/81 Actual Relative to Actual Expected Expected Rank to Which Promoted Number Number (Percent) ---------------------- ------ -------- ----------- 1972/73-1980/81 Assistant 21 24.4 86.1 Associate 36 44.4 81.1 Professor 5 12.9 38.8 1972/73-1976/77 Assistant 22 23.4 94.0 Associate 17 29.9 56.9 Professor 8 11.2 71.4 1976/80-1980/81 Assistant 7 8.9 78.7 Associate 24 27.9 86.0 Professor 2 9.9 20.2
As can be seen from the above, women at no time and in no rank managed to achieve promotion in the expected proportions. But, the raw compilations, standing alone, *1248 prove little. Thus, Zellner performed a multivariate statistical analysis of promotion. She controlled for whether or not an individual left URI before 1980-81, whether a terminal degree was held at inception and/or had been obtained subsequent to that time but prior to the end of the period (or was held when the individual exited, if termination occurred prior to 1980-81), years since receipt of terminal degree, years and type of prior experience, years in rank from which individual was or was not promoted, years in the lower rank, and department. She reported the following results:
III. Including Those Who Terminated Before 1980/81 Number of Net Standard Effect Deviations Probability of between of Net Effect Being Net Effect This Far Sample Period Female and Zero Below Zero Size ------ ------ ---------- ------------- ------ 1972/73-1980/81 -0.920 -2.333 0.010 567 1972/73-1976/77 -0.887 -2.202 0.014 479 1976/77-1980/81 -1.199 -2.759 0.003 458 IV. Excluding Those Who Terminated Before 1980/81 Number of Net Standard Effect Deviations Probability of between of Net Effect Being Net Effect This Far Sample Period Female and Zero Below Zero Size ------ ------ ---------- ------------- ------ 1972/73-1980/81 -1.214 -2.471 0.007 462 1972/73-1976/77 -0.567 -1.451 0.073 404 1976/77-1980/81 -0.983 -2.167 0.015 404
Zellner found that the effect of being female on promotion was negative and that it was statistically significant for all but one of the subsets in the analysis, viz., femininity was not statistically significant if one analyzed the faculty from 1972 to mid-1977 and excluded those who had left URI before the 1980-81 academic year.
Zellner, who termed her estimates "conservative," stated that the analysis only revealed differences in the actual number of women or men chosen for promotion. If an equal number of men and women had been promoted by the end of 1981, her study would not show any discrimination. Zellner pointed out that a woman who remained in rank longer than a man solely due to sex was a victim of sex discrimination, but that this slower rate of progress would not be observable through her analysis. Nor would Zellner's study have revealed any discrimination on the basis of the number of promotions received in a given time period. In short, Zellner's study was limited to an examination of whether or not comparably credentialled individuals would have been promoted once during the applicable period.
Zellner's study, however, is fatally debilitated. It in no way takes into account the time needed to achieve promotion. Inasmuch as Zellner did not exclude any individuals from her study of promotion, assistant professors hired, say, for the 1977-78 academic year were factored into the mix although it was extremely unlikely that those persons (male or female) had sufficient time to gain promotion prior to the cut-off *1249 date of Zellner's promotion analysis. And, this defect is particularly jarring in that the comparative rate of female hires was on the upswing during the late 1970s. See text ante at Part VIII(B).
More importantly still, Zellner entirely neglected to focus on the most critical elements involved in promotion decisions. She compared individuals who were consubstantially credentialled by prior experience, a factor important to initial placement but of marginal relevance to advancement at URI. The promotion process is not based on antecedent accomplishment, but on achievements after arrival at the University. Implicit in Zellner's report was the assumption that individuals with comparable prior credentials were equally productive once hired by URI. No evidence in this case supports that questionable premise, and common sense is offended by it. The court rejects Zellner's supposition out of hand.
The idea of measuring post-hire accomplishments, and of predicating promotion decisions on such progress, is fundamentally sound. The University has specific missions, missions which its faculty must implement. A primary incentive which URI can offer to induce faculty to extend best efforts in teaching, research, and service is the carrot of promotionpromotion which hinges to some appreciable degree on how well individual teachers have done at URI and on their contribution to the accomplishment of the University's goals. Zellner's failure to examine post-hire accomplishments (other than her formalistic recognition of receipt of the doctorate or other terminal degree) renders her study virtually nugatory. Zellner's findings fail in any way to address the key question: whether the standards used for promotion were applied differently to women.
C. Promotion (Siskin's View)
Siskin recognized the importance of achievement during University employment in the URI promotion scheme. Since directly quantifying this variable was not feasible, he selected a proxy: denials of promotion recommendations at levels above that of the department chair. At bottom, Siskin used the chair recommendation as the best available substitute for hard data on the pivotal point of post-hire success. Siskin found that only 10.3% of these men and 8.5% of these women were denied promotion to associate professor. A denial of promotion in this context meant that an individual was begrudged elevation by decree of the dean or the president, and was not promoted to that rank by 1981. The percentage differences were not statistically significant. At the full professor level, Siskin observed a refusal rate of 16.0% for men and 14.3% for women, likewise statistically insignificant.
Siskin also performed a companion survey which disregarded peer group evaluations. He compared comparably credentialled faculty members in order to examine promotion rates to associate professor and to full professor. The study was similar to Zellner's except in one important respect: Siskin bifurcated the promotion process into two parts, viz., assistant-to-associate decisions and associate-to-full decisions. He found no statistically significant difference in promotion frequencies between men and women.
Of the available statistical analyses, Siskin's first model is the most telling. It monitored actions taken after one group of decisionmakers concluded that the faculty member's achievements at URI warranted promotion. As such, it provided some evidence of how the promotion standards were applied. It is true that Siskin's experiential proxy can be questioned: its use assumes, for example, that peer level review was not discriminatory; it did not purport to analyze the standards employed in the deanery and in the presidential suite in reversing departmental recommendations; and denials were not broken down by discrete years. The court has examined each of *1250 these arguments, and others raised by the plaintiffs. They diminish the weight of Siskin's results. Yet, all things considered, his chosen proxy is a sensible one.
The plaintiffs also point out, correctly, that Siskin never studied promotions from instructor to assistant professor. But, in the majority of cases, such promotions were triggered reflexively by the obtaining of a terminal degree. The court does not believe that Siskin's putative omission in this respect invalidated his results. The court accepts Siskin's study as competent proof of the lack of differential application of promotion standards.
D. Evidence re Promotion Standards
The articulation and refinement of standards applicable to advancement received abundant attention at the University during the decade at issue.
Newman, from and after his installation, felt that the decisional process was fair, but constantly sought better ways to address the question of appropriate criteria for promotion. At his instigation, Ferrante and Rosie drafted a statement of faculty expectations in 1977. Pl. GG (1-3). The statement was disseminated to a number of new faculty members. It was eventually withdrawn after the AAUP protested. Dean Knauss (of URI's school of oceanography) suggested to the council of deans that the individual deans make certain that the departmental chairs within their respective jurisdictions were aware of the standards for promotion. In his continuing quest to upgrade the faculty, Newman admittedly sought to impose stricter requirements for promotion (as well as for tenure). Newman expressed considerable doubt as to whether he had succeeded in this endeavor; and Ferrante, at least, felt that the standards for promotion had not changed in ten years. Though the plaintiffs perfervidly urged that correspondence between Marks and Ferrante had the effect of changing the rules in mid-stream, the court finds that this was no more than an attempt to clarify the standards and procedures for promotion. It did not modify the criteria.
The somewhat uncertain understanding of the rules resulted primarily from the subjective nature of the topic and variances from department to department. The problem was exacerbated by the seeming inability of the powers-that-were to draft a coherent set of guidelines for faculty review. In this wise, the court credits particularly the testimony of Professor Briggs, who was chair of the history department in 1974-75 and again in 1976-77. Briggs testified that the standards and methods of evaluation preliminary to promotion varied by department and by college. The court is, however, less ready to accept Briggs' view that the standards for evaluation became more rigorous as Newman's tenure progressed.
Despite the pervasive amphibology, however, certain facts come through with reasonable clarity. Newman's unceasing effort to upgrade the institution was well-known throughout the University community. Thus, though the criteria for promotion did not materially change, it is not surprising that some perception arose that the yardstick for faculty advancement had become tougher. To the extent that problems were endemic in the process, they were principally problems of communication. With this overview of the prevailing administrative winds, the court turns to the comparisons offered by the complainants in their attempt to demonstrate that promotion standards were applied less harshly to men than to women.
E. Anecdotal Evidence
1. Judith Swift
Judith Swift came to URI in 1968 as a graduate student. She received her master's degree in English in 1971 and was immediately hired as an instructor in the theatre arts department. Swift taught three courses per semester. The offerings *1251 were laboratory/practicum and required substantial contact time with students. Whereas a normal burden of three courses might involve nine hours of classroom time per week, Swift's courseload required her to be in the classroom or on stage roughly eighteen hours per week. She also oversaw one or more theatrical productions. After three years, and without an appropriate terminal degree, Swift was promoted to assistant professor. In 1976, she received tenure.
Professor Flannery, chairman of the department in 1977, recommended her for promotion. Marks agreed, but Newman did not. Newman proffered no satisfactory explanation of his rebuff, either contemporaneously to Swift or at the trial. She was, however, encouraged to "reapply." But, during the following year, Swift and Flannery had a falling out. A series of petty incidents and comments grew into a chasmal rift. During the next round of promotion recommendations, Flannery penned a most unflattering annual review. The report questioned, for the first time, her teaching competency, the value of her research, her credentials to teach drama, and her attitude. Marks, not surprisingly in light of Flannery's diatribe, decided against promotion. Swift grieved the adverse action, initially without success. She went to New York and directed an off-Broadway production which received critical acclaim. Marks was unswayed. Finally, an outside evaluator was brought in, recommended promotion, and convinced Marks to expunge Flannery's review of Swift. She was thereupon promoted.
The evidence does not support any credible defense of Swift's treatment at the hands of Newman, Marks, and/or Flannery. But obnoxious behavior, without more, is not a violation of Title VII. The plaintiffs introduced only negligible evidence relating to the standards used by the department in evaluating promotions. The court is unable to find that the standards (whatever they may have been) were applied differently in Swift's case than in the case of men. She did not sustain her burden of proving that she was qualified for promotion when the subject initially surfaced, especially given her lack of a terminal degree. And, her case is even more flaccid as to the second go-round. There is every reason for the court to believe that a male, similarly embroiled in such internecine warfare, would have been treated with no less vituperation. Swift's comeuppance seems to have been the result of a personality clash with her immediate superior, totally unrelated to her gender.
2. Carol Avery
Carol Avery came to URI in 1970 as an instructor in the department of textiles and clothing. In 1974, she was promoted to assistant professor despite the lack of a terminal degree. In 1977 she received her doctorate; a year later, she was awarded tenure, but not further promoted. No evidence was introduced whereby the calibre of Avery's teaching, research, or service could legitimately be gauged. The plaintiffs relied, in this instance, principally on the award of tenure without concomitant promotion. It is true that the two often went hand-in-hand at the University; but the proof demonstrates that the grant of tenure sans promotion was by no means unprecedented. And, no probative evidence was introduced about Avery's record, the department's standards, the college's standards, or kindred factors. It has not been shown either that Avery was entitled to promotion or that similarly situated males were accorded greater largesse.
3. Karen Stein
Karen Stein joined URI in 1968 as a half-time instructor in the English department, teaching only three courses per year. *1252 Newman promoted her to assistant professor, over Marks' objection, in 1976. Stein did not receive her doctorate until 1982. The department recommended Stein for further promotion. Marks agreed, but Newman rejected the suggestion.
The plaintiffs introduced no meaningful evidence of Stein's accomplishments after 1976. There is little in the record to intimate that Newman's action was arbitrary, capricious, or even ill-advised; and there is absolutely nothing from which the court can conclude that she was a victim of differential treatment in promotion.
4. Frances Chin
Frances Chin obtained a Ph.D in bacteriology in 1941. In a rather curious switch of disciplines, she received a master's degree in library science from the University of Kentucky in 1962. She then worked as a cataloguer for various institutions of higher education before she was hired at URI in 1965 as an associate professor of library science. Three years later she was tenured. Chin was never promoted. In her case as well, no real evidence was introduced concerning her post-hire accomplishments.
The complainants took a matched pair approach at trial, and claimed that Daniel Bergin was promoted to full professor even though his credentials and record were not superior to Chin's. This allegation of disparate treatment stands without substantial evidentiary foundation, however, as the plaintiffs never furnished competent proof whereby a valid comparison of Chin and Bergin could be conducted. The plaintiffs' position contains a hint that length of service, in and of itself, adequately defines accomplishment for preferment purposes. The court summarily rejects any such insinuation. The focus must be on achievement, not mere longevity. Chin's claim of biased treatment vis-a-vis promotion is, on this record, unsupported.
5. Greta Cohen
Cohen is, of course, a named plaintiff and class representative. Her background and overall situation has been summarized earlier. See text ante at Part XI(E) (5). After receipt of her doctorate in 1981, Cohen approached the department chair, Polidoro, to discuss promotion prospects. Her overtures were rejected; she was told to devote more time to research if she desired promotion to a full professorship. Cohen claimed that this was the first she had heard of any supposed insufficiency in research (but, the 1978 annual review suggested as much; and Brittingham had reiterated the point to Cohen in 1979). Cohen asserted that her artistic achievement should serve as an adequate surrogate for research, but Polidoro disagreed.
Cohen's claim in this respect is predicated largely upon a belief that a different standard, one requiring more extensive research, was applied to her than to her male colleagues. The record, however, does not sustain this contention. It is clear that Cohen assumed (erroneously, as matters turned out) that her artistic accomplishments were a suitable substitute for research. The University lulled her into no such false sense of security. The record is bereft of any proof that men in her situation were allowed to trade off virtu for intellectual effort. Furthermore, Cohen introduced no meaningful evidence about the standards used by the physical education department in evaluating candidates for promotion to the august rank of full professor; nor did she demonstrate that comparably situated men attained that rank without the requisite productivity in research.
Moreover, her claim is little bolstered by analogy to other disciplines. To be sure, artistry is sometimes treated as a substitute for research in other fields. But, Cohen's department (physical education) was not compellingly similar to those in which such a transmogrification was commonly *1253 effected: drama, music, or fine arts. Certainly, to the lay eye, there are marked differencesboth extrinsic and intrinsic. Thus, the court is unable to find that the departmental standards in physical education were irrational, or that they prejudiced Cohen in any unlawful way vis-a-vis her colleagues in other departments, or that the standards were unevenly applied in her case.
F. Summary (Promotion)
1. Legal Standard
The burden of proof with respect to promotion is somewhat different than that normally associated with Title VII cases. Promotion requires an evaluation of a faculty member's credentials, productivity, and teaching ability. Universities are far better equipped to perform this function than the courts. Faro v. New York University, 502 F.2d 1229, 1231-32 (2d Cir. 1974). But, Title VII nonetheless requires the judiciary to intrude when sex is a factor in promotion decisions. See Sweeney v. Board of Trustees, 604 F.2d 106, 112 (1st Cir.1979), cert. denied, 444 U.S. 1045, 100 S. Ct. 733, 62 L. Ed. 2d 731 (1980). Such a dichotomy must carefully be reconciled.
A solution was proposed in Zahorik v. Cornell University, 729 F.2d 85 (2d Cir. 1984). There the Second Circuit held that an individual does not make out a prima facie case for tenure unless the faculty member can show "that some significant portion of the departmental faculty, referrants or other scholars in the particular field, hold a favorable view of the question." Id. at 94; cf. Banerjee v. Board of Trustees, 648 F.2d 61, 63 (1st Cir.) (plaintiff failed to prove that articulated reasons for denial of tenure were pretextual), cert. denied, 454 U.S. 1098, 102 S. Ct. 671, 70 L. Ed. 2d 639 (1981). Although the Zahorik rule was devised for judge-wrought oversight of tenure decisions, it can be adapted for use with respect to individual promotion claims without missing a beat. Exactly the same spectrum of considerations which demand that the judicial bull tread gingerly in the academic china shop impact tenure and promotion alike. Nevertheless, a literal application of Zahorik to classwide claims is unwieldy, absent some specially crafted adaptations.
Zahorik stands for the proposition that peer review and support for a candidate is an excellent benchmark for the court's judgment anent the qualifications of candidates for promotion or tenure. In a similar vein, classwide analysis of discrimination in respect to such decisions must begin with peer review. If the plaintiffs have proved by a fair preponderance of the evidence that bias existed in peer review and was not corrected by upper-level administrators, then an inference will lie that URI tolerated discrimination. Alternatively, if the plaintiffs have shown that the administration ignored peer review and made decisions which favored men as opposed to women, then the same inference will lie. And, either proposition can be established circumstantially. It is unrealistic to restrict the plaintiffs to direct evidence of discrimination (which will, in the ordinary course of events, be difficult to come by); the odds of finding a smoking gun are remote.
One proxy for direct evidence of bias is the likelihood of promotion of comparably productive individuals. Since research is the prime consideration in the evaluation of promotion candidates, it seems fair to assume that faculty with equally impressive research credentials and the same seniority at URI will enjoy an equal probability of promotion. If equally productive candidates do not garner comparable advancement in rank, the promotion scenario becomes suspect. Yet, the proposition is easier stated than implemented. An examination of faculty research credentials requires the compilation of articles published in refereed and non-refereed journals, books published or edited, and papers presented at various conferences. Once this data has been assembled, some formula must be devised to give weight to the different types of publication. This, in itself, is no easy task. Not only are there differences between classifications (a book carries different weight than a journal article) *1254 but there are subtle shadings within classifications (some journals are more selective and/or prestigious than others). And, publications in journals may, as a generic matter, be more impressive in some fields than in others. E.g., Langland v. Vanderbilt University, 589 F. Supp. 995, 1011 (M.D.Tenn.1984) (contrasting journal articles in the sciences and in the humanities). If this data can be meaningfully assembled, an analysis of it will then show whether equally productive researches are being treated comparably by the institution.
This model has a number of advantages. First, it provides some outward indication of quality of research: a book will usually not be published if it is meritless; acceptance of an article by a refereed journal is a tangible indicator of some intrinsic worth; it is unlikely that an individual will be requested to present a paper at a recognized scholarly conference if certain standards have not been met. Because academic research, to be worthy of the appellation, is by its very nature subjected to outside review at some point, a model pivoting off this fact avoids the possibility that internal peer review may be biased against one group or another. And, the model adapts to judgments at the department, dean, or president level of review. Similarly productive individuals should be treated comparably at all stages. If they are not, and if the inequality works consistently to the disadvantage of an identifiable class, then evidence of discrimination exists.
Admittedly, this model has some drawbacks. It excludes analysis of teaching, performance oriented specialities, and service commitments. Furthermore, it is expensive to gather the data and to analyze it. But, these obstacles can be overcome. Teaching and service are subject to analysis in a variety of ways; or, in lieu thereof, to the use of meaningful proxies. And, the instant plaintiffs did gather information anent publications and research (though they never used it).
Another approach involves the examination of the peer review process. At URI, the apex of the peer review pyramid is topped off when the departmental chair compiles the annual review, summarizing the consensus of a faculty member's colleagues. If the annual review process is not gender-neutral, then some proof of peer discrimination exists. Because of the difficulties inherent in attempting to ascertain the existence of such bias, a more reasonable substitute might involve an examination of faculty votes on promotion and a comparison of those votes to the ensuing recommendations of the departmental chairs. If the chairperson routinely rejects the outcome of the faculty vote when it favors a woman and does not do so when it favors a man, some evidence of discrimination is detectable. And, if this paradigm is extended to include an analysis on a university-wide basis, then a showing of class discrimination can be discovered.
This method, too, has advantages and disadvantages. It closely follows the conceptual pattern in Zahorik and takes account of all three legs of the stool. And, it is not prohibitively expensive; it requires significantly less data gathering than the research modelling technique. But, the methodological vices are also significant. The approach is built upon the assumption that the faculty themselves are not biased in any particular manner. Moreover, the method requires the ascertainment of faculty votes even when no formal votes (or recorded votes) exist. Lastly, it does not account for discrimination at higher reaches, condonation of peer review discrimination through reflexive acquiescence in departmental recommendations, or correction of such discrimination by upper level administrators.
The logic of the peer review analysis technique requires one further step. The plaintiffs must examine what happes to the chair recommendations at the dean and president levels. If there is a consistent pattern of rejecting the favorable outcome of the annual review when women are involved, and a different or less consistent pattern vis-a-vis men, then further evidence *1255 of discrimination is established. On the other hand, a consistent pattern of rejecting all unfavorable outcomes would not evince discrimination at the upper levels. Finally, if the upper level administrators accept the chairs' recommendations and there is evidence of peer review bias, then this will establish condonation of discrimination (and thus, discrimination itself).
2. Discrimination In Promotion (Classwide)
The statistical model formulated by Zellner examined the promotion prospects of individuals hired with the same credentials and who held the same degrees. The underlying hypothesis of Zellner's study was that comparably experienced and degreed individuals were likely to enjoy like productivity at URI. If true, this assumption obviated the need to employ any of the more sophisticated techniques previously outlined by the court. But, the assumption, like the daring young man on the flying trapeze, has no visible means of support; it cannot be countenanced. The very reason that peer review was needed was because all faculty did not produce equally once they arrived on campus. Initial hire may be analogized to a professional football team's draft choices and peer review to a subsequent performance-oriented assessment of who deserves all-pro honors. As any sports fan can attest, the highest draft choices do not always prove to be the most accomplished players. And, once the balloon of Zellner's underlying assumption is punctured, her model collapses in shambles around it. The court finds that Zellner's handiwork failed in this instance to establish a prima facie case on behalf of the plaintiff class.
The anecdotal evidence fares no better. Out of some eighty promotions which Zellner studied, the plaintiffs primarily relied on only five. That handpicked quintet proved manifestly insufficient to demonstrate the presence of a pattern or practice of discrimination. Cf. Cooper, ___ U.S. at ___ _ ___ n. 11, 104 S. Ct. at 2801-02 n. 11. This is especially true given that only two of the five cases revealed any evidence of meaningful peer support for the derailed promotion. In respect to three of their exemplars (Avery, Chin, and Stein), the plaintiffs introduced no evidence concerning faculty or other referrants' support for promotion. Nor was any probative evidence introduced from which the court could make an independent judgment on the relative qualifications of these persons for promotion. Thus, even if it is assumed for the purpose of argument that both of the two remaining instances were illegal in contemplation of Title VII, but see post at Part XII(F) (3), they could not alone constitute a pattern or practice of discrimination. See Cooper, ___ U.S. at ___ _ ___ n. 11, 104 S. Ct. at 2801-02 n. 11; Presseisen v. Swarthmore College, 442 F. Supp. 593, 607-08 (E.D.Pa.1977), aff'd mem., 582 F.2d 1275 (3rd Cir.1978). One swallow does not a summer make; even two will not suffice, when dealing with numbers of large magnitude, to thaw the icescape of the plaintiffs' burden.
It follows inexorably that classwide relief must be denied, whichever legal standard is applied. The peer review process was not shown to be slanted; promotion criteria were neither fashioned so as to hamper the advancement of females nor applied to women, as a class, in a factitious manner; examination of referrants' recommendations at the department chair, dean, and presidential level have not been cast into gender-based disrepute; the use of disparate evaluative techniques was not demonstrated; and no effort was made to prove the classwide claim by means of a research modelling approach.
3. Discrimination In Promotion (Class Representative Claim)
Failure to prove a class claim does not foreclose consideration of individual claims. Cooper, ___ U.S. at ___, 104 S. Ct. at 2802. The individual allegations will not, of course, benefit from the Craik presumption. In addition, the court need consider only those individuals who are named as plaintiffs either as pleading-identified *1256 class representatives or as intervenors. See Clark v. Chrysler Corp., 673 F.2d 921, 929 (7th Cir.1982); Dickerson v. United States Steel Corp., 582 F.2d 827, 832 (3rd Cir.1978); McDowell v. Safeway Stores, Inc., 575 F. Supp. 1007, 1067 (E.D. Ark.1983); Harrison v. Lewis, 559 F. Supp. 943, 947 (D.D.C.1983); cf. Cooper, ___ U.S. at ___ _ ___, 104 S. Ct. at 2802-03. Accordingly, only Greta Cohen's personal claim anent animus in the promotion process need be adjudicated here.
Cohen's asseveration was based upon the alleged failure of the physical education department to inform her of the standards required for elevation to full professor. From this, Cohen inferred that URI sought to apply a different standard of promotion to her than to other (presumably male) individuals. For several reasons, the remonstrance comprises little more than aimless whistling past the graveyard of frustrated academic aspirations. First, contrary to her assertions, Cohen was informed of the research requirement. She received a letter from Brittingham in which the latter requested that Cohen devote more time to research, and similar sentiments were expressed in at least one of Cohen's annual reviews. Cohen knew of this deficiency but chose to ignore it for purposes both of her subsequent discussions with Polidoro and of this litigation. If she was ignorant of the requirement, her ignorance was self-induced, and therefore unavailing. Cf. Emery-Waterhouse Co. v. Rhode Island Hospital Trust National Bank, 757 F.2d 399, at 402 (1st Cir.1985).
Nor did Cohen in any way establish that she was deserving of promotion. The departmental majority felt that Cohen should devote more time to research to qualify for the rank of full professor. Simply put, her peers, by and large, did not believe that she was qualified. Cohen did not attempt to introduce the opinions of other scholars to support her quest for promotion. Instead, Cohen argued that her work with the dance company should have been given more credit towards the fulfillment of the research requirement. But, this contention overlooks that peer review, by its nature, demands the exercise of judgment by those most qualified to give it. The physical education department is the best arbiter of how much weight to accord to performance as opposed to published research. Unless Cohen can show that the department gave more credit to a man's performance work than it gave hers, her claim of disparate treatment must fail. She made no such showing at trial. And, her litany of comparisons with coaches, musicians, and studio artists are nothing more than red herrings. Coaches require different skills, have different responsibilities, and work in a special environment. Moreover, there is an objective evaluation scheme in the coaching domain: the won-lost record. The comparison with musicians and artists fails as well, for reasons previously stated. See text ante.
Cohen's argument, at bottom, is a plea that each faculty member should have the leeway to weigh his or her own credentials and to construct an individual balance. Such a thesis would lead to a special kind of institutional chaos. The court will not place its imprimatur upon such an evisceration of the peer review system. Cohen fell far short of the mark. She did not make out a prima facie case under Zahorik. Her claim for discrimination in promotion is baseless.
At URI, as at most universities, a faculty member, after a certain number of years, is considered for tenure. Once tenure is granted, he or she has the right, with exceptions that are not relevant to this litigation, to remain on the URI faculty until retirement. Those who are passed over for tenure are ordinarily required to leave the University's employ in short order.
The tenure process is set out in some detail in the collective bargaining agreement. The annual reviews comprise the foundation upon which tenure decisions are built. Aided by these assessments, faculty *1257 peers, the department chair, the appropriate dean, the VPAA, and the president sequentially examine the credentials of a faculty member and determine whether tenure is merited. Each untenured faculty member is reviewed annually, and each has an outside time limit within which to achieve tenure (the mandatory tenure decision date). The length of the period varies depending upon rank at hire; assistant professors have the most generous span and full professors have the most niggardly. In exceptional cases, a faculty member may be awarded tenure prior to the mandatory decision date, consistent with the collective bargaining agreement. Early tenure requires that a faculty member possess a stipulated minimum number of years in service at the University. And, as with the mandatory decision date, eligibility for early tenure depends upon rank at hire.
From and after the date that an individual becomes eligible for early tenure, he or she is assayed annually vis-a-vis tenure until tenure is granted or the mandatory decision date has expired (whichever first occurs). A faculty member has the right to solicit postponement of tenure consideration up to the arrival of the mandatory decision date. That date is an absolute; it cannot be extended. At the assistant professor level, the mandatory decision date is eight years from the date of hire. If tenure is not bestowed prior thereto, the faculty member must leave the University. In the typical case, a teacher is informed no later than the end of year seven as to whether or not tenure is in the offing. Although the temporal spans differ, the procedure is much the same at other ranks. Thus, discussion of this issue will focus at the assistant professor level unless the context indicates to the contrary.
Under the terms of the collective bargaining agreement, recommendations for or against tenure must originate with the departmental chair (whose pivotal role in the process is reminiscent of the chair's involvement in URI's promotion scheme). Such recommendations are delivered to the dean no less than nineteen months before the end of the candidate's eighth year at URI. The timing necessitates that the tenure decision focus on the aspirant's accomplishments during the initial six and one-half years of his or her career at URI (though the faculty member has the right to amend his or her dossier to reflect chickens which come home to roost, e.g., neoteric publications or newly-funded grants, after the report has gone forward to the dean). The dean and the chairperson then confer. If the dean denies tenure, the process dead-ends. The affected individual may protest the rejection only by filing a grievance. But, if both agree that tenure is appropriate, the joint recommendation is transmitted to the VPAA, and eventually, to the president. The council of deans performs the same advisory function with respect to tenure decisions as in connection with the promotion process. See text ante at Part XII(A).
Presidential review is similarly an either or proposition. If the president overrides the recommendation and vetoes tenure, he must set forth his reasons in a writing forwarded to the faculty member in question by the end of the seventh year after hire. Once again, such a negative decision may be contested only in a grievance proceeding. If the president favors the grant of tenure, there is yet another step. Unlike promotions, the president does not have the authority to award tenure. That power rests nominally in the Board (though there was no evidence that the president's tenure recommendations were not universally honored). If the Board chose to disagree, however, its decision would be final, and no grievance or other recourse would be available to the candidate.
The standards for tenure are not set out in any document. The Manual is wonderfully general; it proclaims only that tenure should not be ceded to an individual who *1258 has not adapted to the standards of the University. Apart from a ritualistic incantation of the trilogy (teaching, research, service), the Manual does not specify the applicable criteria. And, the collective bargaining agreement is equally silent on this point. The various departments handle the question in different ways: some carry out faculty polls as to whether a candidate deserves tenure; others do not. In some fields, the departmental faculty have the option to abstain from voting on aspirants for tenure; other departments operate differently. The rules fluctuate widely.
At bottom, the procedures which inseminate the initial tenure recommendation are left to the whim of the individual departments (and, in many cases, to each department chair). Aside from an abortive effort by Knauss to formulate meaningful guidelines (an effort whose product never received any official imprimatur), it is plain that the benchmarks for tenure are and have been subjective and that the shadings of those chiaroscuro "standards" vary somewhat from department to department and from college to college. And, procedures for evaluation are likewise nonuniform. Some departments, such as history, have a detailed and well organized methodology for assessing, say, a faculty member's teaching prowess; others take an ad hoc approach. Such variances are not, in and of themselves, unusual in academic circles; after all, "(t)he procedures prescribed for making the tenure decision ... plainly contemplate a subjective evaluative decisional process by academic professionals rather than an objective fact-finding process by tribunals adapted to that quite different purpose." Siu v. Johnson, 748 F.2d 238, 244 (4th Cir.1984).
URI apparently relied upon the higher levels of review in order to bring uniformity to the tenure process. But, this reliance overlooked both the key role of the departments in putting the ball into play and the fact that the upper level scrutinization apparently operated on a wavelength some kilocycles apart from that to which many of the department chairs were tuned. Knauss, for instance, complained bitterly about the failure of the deans to inform lower-level administrators of the standards to be applied to tenure decisions (and to promotion recommendations, for that matter). See XXIII at 106. There was some suggestion that the criteria for tenure ebbed and flowed over time. Newman testified that the standards became increasingly more rigorous. Marks supported higher tenure standards, as did Brittingham and Rosie; at one point, they drafted a letter (never implemented, due to an AAUP protest) limning and advocating such requirements. Ferrante, on the other hand, testified that tenure criteria remained constant. Other testimony equivocated on the subject. Compare XI at 68-69 with XI at 93-95. The court finds that URI did gradually begin to apply progressively tougher standards vis-a-vis tenure decisions during the relevant period and that the University did not communicate the changes in a manner designed adequately to inform the faculty.
There is a definite linkage between tenure and promotion. Ordinarily, tenure will not be given if promotion to the rank of associate professor is not warranted; conversely, such a promotion will not normally be effectuated in the absence of a concomitant grant of tenure. In certain circumstances, however, the University may award tenure sans promotion, e.g., text ante at Part XII(E) (2) (Carol Avery), or may upgrade an individual from assistant to associate professor without awarding tenure (and the collective bargaining agreement does not flatly interdict these practices). Thus, cases of failure to earn promotion are at times interwoven with problems anent tenure.
With this background in mind, the court turns its attention to the evidence concerning alleged gender-based discrimination in URI's allocation of tenure.
B. Tenure (Zellner's View)
Zellner computed the expected number of females and compared that to the number of females actually tenured. She deleted *1259 individuals who left URI before 1980-81 for the same reasons described earlier. See text ante at Part XII(B). Zellner found the following:
I. Including Those Who Terminated Before 1980/81 Actual Relative to Actual Expected Expected Period Number Number (Percent) ______ ______ ________ ___________ 1972/73-1980/81 55 83.8 65.6 1972/73-1976/77 29 52.4 55.3 1976/77-1980/81 26 36.7 70.8 II. Excluding Those Who Terminated Before 1980/81 Actual Relative to Actual Expected Expected Period Number Number (Percent) ______ ______ ________ ___________ 1972/83-1980/81 52 68.6 75.8 1972/73-1976/77 29 45.5 63.9 1976/77-1980/81 25 34.0 73.5
At no time did the number of females actually tenured reach the level mathematically to be expected if there had been complete sex neutrality in the adjudication of tenure decisions. But, the raw data is itself inconclusive.
Zellner next conducted a logistic regression study which followed the format of her comparison between actual and expected tenures. She excluded any faculty members who began teaching after the fall of 1977, on the theorem that such persons would not have had sufficient time to attain eligibility for tenure. Zellner controlled for department, whether the doctorate had been obtained by 1972-73 or by the time of hire (depending on the year of hire), whether the doctorate had been secured during the time period studied, years since receipt of terminal degree, years and type of prior experience, and year of hire at URI. She reported the following findings:
III. Including Those Who Terminated Before 1980/81 Number of Net Standard Effect Deviations Probability of between of Net Effect Being Net Effect This Far Sample Period Female and Zero Below Zero Size ______ _______ __________ ______________ _______ 1972/73-1980/81 -0.666 -1.590 0.056 365 1972/73-1976/77 -1.177 -2.292 0.011 293 1976/77-1980/81 -0.753 -1.120 0.131 167
*1260 IV. Excluding Those Who Terminated Before 1980/81 Number of Net Standard Effect Deviations Probability of between of Net Effect Being Net Effect This Far Sample Period Female and Zero Below Zero Size ______ ______ __________ _____________ _______ 1972/73-1980/81 -0.908 -1.026 0.152 289 1972/73-1976/77 -1.074 -1.767 0.039 253 1976/77-1980/81 -0.697 -0.753 0.226 142
By Zellner's own admission, the results lacked statistical significance except in the period 1972-73 to 1976-77. In the court's view, the conclusions which Zellner sought to draw are invalid for other reasons as well.
Tenure decisions, like promotion decisions, clearly implicated the productivity of individuals while at URI. Zellner's study did not attempt to examine the research or teaching output of any faculty member. Without evidence of factors that were relevant to the decisionmakers, the court is unable to treat her computations as a probative base for a finding that sex played a role in the tenure process.
Zellner's calculations are also open to other criticisms. The second set of figures (Table II, ante) analyzed tenure for those who had not left URI before 1980-81. But, inasmuch as the collective bargaining agreement mandated the up or out policy described above, see text ante at Part XIII(A), Zellner's extrapolation in effect studied only (i) those individuals who had not yet reached the mandatory tenure date and (ii) those individuals who had successfully scaled the tenure mountain. Those for whom tenure had been eschewed were, in reality, not a part of the examination. Almost by definition, despite its sophisticated trappings, the analysis did not study whether women deserving tenure were denied it due to their sex; it did notand could notanalyze whether women were disproportionately denied tenure. It is a sheep in wolf's clothing.
Moreover, Zellner appears to have double-counted, in that her "early" subperiod ended in 1976-77 and her "later" subperiod began in the same academic year. There was no plausible explanation of the possible effects of this overlap on the findings. Zellner's paradigm was also marred by persistent overinclusion and underinclusion. E.g., note 50, ante. And, notwithstanding these deficiencies, her figures in most cases failed to exclude the likelihood of chance as an explanation for disparities in the conferment of tenure.
C. Tenure (Siskin's View)
Siskin also performed a study of tenure, but it was not an elaborate one. Beneath the rococo facade of his handiwork, he simply computed the number of men and women who obtained tenure and those who did not. He found that there was no statistically significant difference in the rejection rates for men and women. This study is, of course, not skewed, as in Zellner's; but it is fundamentally unenlightening. Siskin made no attempt to find a proxy for the trilogy (as he had done in his promotion analysis), nor to examine tenure decisions at various administrative levels, nor to focus on individuals who departed URI for presumably greener pastures because their tenure chances were slim, nor to analyze why rejections eventuated.
Siskin's study is arguably some evidence that gender-based discrimination did not permeate the tenure process, but it is not a very sturdy reed. Despite Siskin's adroit *1261 number crunching, the court is constrained to peek beneath the statistical coverlet.
D. Anecdotal Evidence
1. Katherine Schach-Cook
Katherine Schach-Cook was a doctoral candidate in history at the University of Nebraska in the early 1970s; during two years of this period, she doubled in brass as an assistant professor at Huron College in Huron, South Dakota. She was awarded her doctorate, with a specialization in European diplomatic history, in May of 1974. The following September, she settled in as an assistant professor of history at URI.
The collective bargaining agreement permitted URI in its discretion to grant credit (not exceeding three years) towards tenure for time in service at another institution. Schach-Cook was offered the credit for her Huron days, and took it. Her acceptance was apparently predicated on her incorrect belief that she would not be reviewed vis-a-vis tenure for five years. But, tenure eligibility (and therefore, review) begins in the fifth credit year, irrespective of the fact that some of the tenure-credited time was spent elsewhere. To the extent that Schach-Cook thought otherwise, it was due to her own misreading of the tea leaves and not to any deceptive conduct on the part of the URI power structure. The court finds that the rule existed and was applied to Schach-Cook in a regular, non-discriminatory fashion.
At URI, Schach-Cook taught three courses per semester, the norm for history department faculty. She toiled on a number of committees and was elected to the faculty senate. In addition to her instructional activities, Schach-Cook developed new courses in European women's history and served as an advisor to prospective history majors. She began to research an exotic topic in Russian foreign policy, but was thwarted in her efforts when an article on the same subject was published just as she was poised to complete her work. She selected another niche and started afresh. At this juncture (1977), she became eligible for tenure review. Aware that she had not been able to do five years worth of creditable research, Schach-Cook asked Ferrante for a waiver of the years previously credited toward tenure. She argued that, due to the lack of facilities at Huron, she had been unable to produce significant research during those two years. Ferrante did not respond, and tenure consideration went forward.
As part of the history department's deliberations, a vote was taken. Schach-Cook received six ayes and three nays. There were three abstentions. Schach-Cook felt that the vote revealed only lukewarm support within the department and asked that her name be withdrawn. In accordance with the applicable provision of the collective bargaining agreement, her request was granted. Schach-Cook thereafter intensified her scholarly efforts. She had published an article in the Modern Encyclopedia of Russian and Soviet History. A second piece was requested by the publishers; it was subsequently authored and printed. And, she began working on an anthology of the writings of nineteenth century European diplomats.
Her tenure review recommenced in the fall of 1979. The only additional arrows in her quiver were the second encyclopedia article and a broader exposure to the history faculty. The departmental vote favoring her tenure was, on this occasion, nearly unanimous. The departmental chair made an affirmative recommendation to Dean Marks.
Schach-Cook then met with Dean Robb to discuss her quest for tenure. Robb characterized her record as marginal and in need of improvement. Robb (not herself an historian) described diplomatic history as a dying field and suggested that a change in Schach-Cook's research focus (to women's studies) might be helpful. Schach-Cook, taken aback, replied that she would think about it. Robb, however, perceived this temporizing as a rejection of the idea. The session culminated with an accord to seek outside evaluations of Schach-Cook's work. While the external assessment was in progress, the International History Review *1262 informed Schach-Cook that it was going to publish a learned piece which she had submitted. Elated, she lost no time in notifying the deanery. Marks was obviously not impressed; soon after, he recommended that tenure be denied. The president concurred. At the end of the 1981 academic year, despite the fact that her mandatory decision date had not yet arrived, Schach-Cook departed.
The crux of Schach-Cook's contention was that men with records no stronger than hers were tenured. But, with a single exception, all of the men to whom she alluded were tenured under the more complaisant standards which prevailed in the era before Newman and Marks took command. Those cases are, therefore, not truly comparable. The exception was Dr. Michael Honhart, who came to URI as an instructor in 1971, was promoted to assistant professor in 1972 (upon receipt of his doctorate), and became eligible for tenure review in the 1976-77 academic year.
A comparison of the records of the two, in terms of the three legs of the stool, is instructive. The primary indeed, the sole focus of Honhart's research appears to have been the reworking of his doctoral thesis into a publishable manuscript. Yet, publication was, at initial tenure review, a mere velleity. And, Honhart had published nothing else. (Though his publication record was clearly inferior to Schach-Cook's, there was considerable debate over whose research was superior). Both Honhart and Schach-Cook received copious praise for teaching skills. In terms of service to the University, Honhart stood head-and-shoulders above Schach-Cook (and most other faculty members). Honhart had chaired the curriculum affairs committee, which oversaw the development of new courses, majors, and degree programs, and modification of existing ones. Honhart's work in this capacity was extraordinary and his contribution to the University was extremely valuable. To be sure, Schach-Cook had a perfectly adequate repertory of service to URI, but she was, in this respect, simply not in Honhart's class. Even the plaintiffs have acknowledged that his service record was scintillating.
Honhart's search for tenure was supported by members of the department and by the department chair. Marks, however, objected. Newman overruled the dean and recommended that the Board grant Honhart early tenure. The Board acceded to the president's request. The plaintiffs harp incessantly that Honhart's publication record was abysmal; yet, there is little room to doubt that Honhart was a unique case, rewarded for his stellar service record despite patent deficiencies in publication. At the presidential level, superb service, excellent teaching, and adequate research outweighed the absence of an acceptable bibliography. So viewed, Honhart was an aberration. Nothing in Schach-Cook's record distinguished her in the same manner. She came to the tenure decision equipped with fine teaching, satisfactory service, and a less-than-average rating for research and publication. Unlike in Honhart's case, there was no reason to go the extra mile. The court finds that Schach-Cook was not denied tenure on the basis of being female, nor was Honhart preferred on the basis of being male.
2. Janet Kulberg
Dr. Janet Kulberg, an assistant professor in the department of psychology, was recommended for early tenure in 1975-76. Marks rejected the suggestion. He specifically held that there were no exceptional circumstances which warranted a grant of early tenure in her case. This assessment *1263 was not contradicted by any probative evidence introduced at trial.
The collective bargaining agreement in force at the time required tenure review with respect to any individual who had achieved the five year minimum eligibility. The same contract also provided that early tenure was not to be granted except in unusual circumstances. Marks relied on this criterion. It has not been shown that he acted capriciously as to Kulberg, or that similarly credentialled men were allowed to slip through when she was not.
3. Linda Shamoon
Linda Shamoon was offered by the plaintiffs as another example of a woman who was discriminated against in the award of tenure. The plaintiffs have made an argument in this regard so convoluted as to defy rational analysis, centering about Marks' alleged alteration of an agreement concerning Shamoon. The record reflects, however, that Shamoon was in fact tenured prior to the date (1980) when Marks is alleged to have indulged in the chicanery in question; both databases support that fact. Absent some drastic change in the general theory of relativity, the court acknowledges that it cannot follow the anfractuous stream of consciousness which has prompted the question; and simply finds that Shamoon's case in no way, shape, or form tends to prove the existence of gender-based discrimination at URI vis-a-vis tenure decisions (nor anything else, for that matter).
4. Linda Hufnagel
Dr. Linda Hufnagel, after earning her doctorate at the University of Pennsylvania in 1967, held a number of research fellowships and teaching positions. In 1973, she was appointed a special lecturer at URI. This position was not covered by the collective bargaining pact; it required her to operate URI's electron microscope, and to do some teaching. Two years later, she was appointed an assistant professor. In 1976-77, she was initially considered for tenure. Her department chair, as well as the majority of her faculty peers in the department, thought that Hufnagel was not ready for tenure. She was denied tenure at that stage, but received it in 1979. There was nothing either remarkable or suspicious about her case. It certainly fails to carry the burden of showing either personal or class-based discrimination.
5. Sharon Strom
The circumstances of the University's hiring of Strom (a named plaintiff and class representative) and of her early service have already been cited. See text ante at Part XI(E) (4), F(3) (c). In the 1972-73 annual review, the chair of the history department recommended that she be retained in rank (as an assistant professor). Thereafter Strom became eligible for tenure consideration in 1974. Though the history department was split, a clear majority supported her quest. Findlay (then the chairman) concurred and recommended tenure.
Ferrante, however, rejected the proffer; he found no circumstances sufficiently exceptional to warrant early tenure. He did not mention that tenure, absent promotion, violated the University's general practice, e.g., text ante at Part XIII(A), but that fact must have weighed heavily. He refused to award tenure to Strom in 1974. She was again reviewed in 1975. The department vote was consentient in her favor. Findlay at that point recommended both tenure and a promotion. There were no dissenters. Strom was ceded tenure and promoted to associate professor on July 1, 1975.
Strom contrasts her case with that of Anthony Bryan, a match which she also employed anent annual compensation. Bryan's antecedents are likewise set out ante at Part XI(E) (4), F(3) (c). Bryan was first reviewed for tenure in 1974. The department chair recommended both promotion and an award of tenure effective July 1, 1974. The suggestion was greeted favorably and blossomed into reality.
No evidence was introduced about Strom's progress toward associate professor, nor is there record support for a contention that the decision to retain her in *1264 rank was unfair or discriminatory. Indeed, she has asserted no such claim (and on the proof, none would lie). Her first departmental tenure recommendation, unlike Bryan's, made no mention of promotion. No probative evidence suggests an anti-female bias as to tenure decisions at the departmental level in history. And, when Strom received the dual encomium the following year, the administration readily embraced the recommendations. Since URI's general practice was to link tenure and promotion, it is easy to see why Bryan received a more hospitable administration response in 1974. The reasons previously discussed in the court's finding that Bryan's higher pay was not justified, id., do not ipso facto support the conclusion that the earlier tenure grant in his case was likewise infirm. In the one case (compensation), the University made a radical and unexplained departure from policy and precedent in the man's favor; in the other (tenure), the University adhered to established practice by declining to grant tenure in the absence of simultaneous promotion. Strom was treated equitably and in a non-discriminatory way vis-a-vis tenure.
6. Wendy Roworth
Wendy Roworth, the named plaintiff in the most recent of the cases at bar, see text ante at Part I(B), came to URI in 1976 as an instructor of art history with a specialty in seventeenth century baroque art. After successful defense of her doctoral thesis, she was promoted to assistant professor and received a raise. Roworth became eligible for tenure consideration during the 1980-81 academic year. By December of 1980, she had amassed an enviable record of publications: her thesis had been published in book form, to the acclaim of art historians internationally; a brace of articles had appeared in referred journals; a trio of refereed papers had been presented (along with perhaps ten others); some abstracts and minor pieces had been published; and she was hard at work on her second book. Except for a one-time courseload reduction in lieu of guaranteed maternity leave, Roworth carried the usual teaching burden (and received plaudits for her work in the classroom). She had earned some grants and fellowships. Her credentials and abilities were well recognized within the department, and she was recommended for early tenure and concomitant promotion, effective July 1, 1981. Marks demurred; he believed her record would be stronger still in another year.
Roworth, miffed by this rebuff, met with Marks and Robb. Marks held fast to his position that tenure was, in her case, too early. Roworth felt that she had not received an adequate response to her inquiry, and suspected that the forces of discrimination were at work. And, her suspicions were reinforced when she learned that a number of male faculty members had been granted early tenure and promotion during the same review period. In 1980-81, fifteen people were promoted to associate professor (fourteen men, one woman). Eleven of those so promoted and tenured were hired with Roworth or after her. Of the eleven individuals, at least five had not reached their respective mandatory tenure decision dates. Roworth claimed that her claim of discrimination was proven five-fold by the elevation of Paul Arakelian, Galen Johnson, Nelson Hairston, Jr., Charles Collyer, and Robert Manteiga. *1265 The court has examined each aspect of this contention.
Arakelian, who had received his doctorate in 1975, had been an assistant professor at URI since 1976. By December of 1980, Arakelian had written and published seven articles and a book review. Arakelian originated and oversaw development of the New England Studies Program at URI. He was widely regarded as a superb teacher. With the blessing of all parties in interest, he was granted early tenure and promotion, effective July 1, 1981. Arakelian's case does not bolster Roworth's claim. His record was markedly superior to hers in University service, and probably better (but at least equal) in research and teaching. He had been more frequently published. And, there was an extenuative circumstance in Arakalian's case: URI was fearful of losing him entirely to another institution if tenure and promotion were not granted. The court finds that this threat was genuine, and notes that no such concern was expressed by the art history department in respect to Roworth.
Dr. Johnson was recruited by URI in 1976 (the same year in which he earned his Ph.D) as an assistant professor of philosophy. He was first reviewed for tenure in 1980 in itself an oddity, as the assessment took place a year before he was tenure-eligible. This deviation from the norm was never explained during the trial. Be that as it may, the evaluation showed that Johnson had submitted a manuscript to a number of university presses (two of which had sought to secure exclusive pre-publication options) and that three of his referred articles had been printed. Johnson's classroom performance drew rave reviews; he was nominated for an excellence in teaching award in 1978. No particular attention was ever called to his service record, and the court assumes that it was run-of-the-mine. On the whole, Johnson's qualifications were on a par with Roworth's; there was little to choose between them.
Collyer received a Ph.D in psychology from Princeton University in 1976, whereupon he became an assistant professor at URI, specializing in experimental psychology. After four years at URI, Collyer had established a perception laboratory which was highly regarded. His first tenure review occurred in 1980-81. An outside evaluator termed the few papers Collyer had written as excellent (but they were based on his doctoral thesis). Collyer's teaching record drew superlatives. He received tenure during his first year of eligibility. Again, Roworth's record appears to have been Collyer's equal. Yet, Marks wrote that delaying Collyer's tenure for another year would not be utile; in Roworth's case, he suggested that she wait another twelve months. The proof reveals no rational basis for such a distinction.
Dr. Hairston received his Ph.D in 1977 and joined the URI faculty that year as an assistant professor of zoology. According to Marks, Hairston had been at URI for only three years; the record reflects no basis for any credits toward tenure eligibility. Despite this, he was considered for tenure and awarded it in his fourth year at URI. Marks opined that:
Dr. Hairston has compiled a very impressive record. His teaching is excellent, he has established his research lab, he is advising graduate students, and he has published.
Pl. P (13).
Ferrante wrote Hairston that he had been awarded tenure for:
your fine record as a teacher and major professor, your growing scholarship record and research support, and your growing reputation as a biologist.
Yet, beneath these high-minded platitudes, there were no accomplishments which, fairly viewed, singled Hairston out for preferment over Roworth.
Dr. Manteiga also was rewarded with the gift of early tenure for satisfying the demands of the trilogy. Manteiga, who had been an instructor at URI for one and a half years, received his doctorate in 1977 and became an assistant professor at that time. Thus, giving due deference to his *1266 prior service, he became tenure-eligible in the 1980-81 academic year. His instructional performance earned him good to excellent ratings. He had extensively revised a part of his doctoral thesis, and it had been published in book form. He had also authored three articles on essentially the same subject. He received (and accepted) invitations to present papers at a number of scholarly conferences. His research record was varied and noteworthy. And, Manteiga's service to the University was exemplary: it included stints as section head of the Spanish area within the language department and as a foreign studies advisor. He received tenure a year before his mandatory decision date.
The Roworth/Manteiga call is a close one. As an isolated comparison, given Manteiga's slight edge in seniority and his superior record of service to the University, it would not carry the day for Roworth. But, viewed as part of the Johnson/Collyer/Hairston mise-en-scene, it suffices, at the least, to heighten the suspicions of an impartial observer. The overall similarity of the records would lead one to conclude that both individuals had met URI's legitimate expectations in a roughly proportionate fashion.
The totality of the circumstances anent the 1980-81 tenure considerations tends, on balance, to indicate that it was more likely than not that Roworth was put off because of her gender. Certainly, at least three and perhaps four males no more deserving than Roworth were treated more favorably.
The Roworth incident is very different from the general class-based claims of discrimination. Roworth sued as an individual, claiming that early tenure was denied in her case because of her sex. In the same year, a number of men were tenured early. Many had records no better than hers. The defendants' proof did not demonstrate that these males satisfied the "unusual circumstances" standard which the collective bargaining agreement attached as a condition precedent to the grant of early tenure (at least to any greater extent than did Roworth herself). So viewed, the evidence pertaining to Roworth's situation shows disparate treatment in her case; but, fairly weighed and balanced with all of the other credible evidence on the issue, it falls far short of establishing that a sex-biased pattern and practice of discrimination infected tenure decisions generally at URI.
The analytic modalities previously postulated with respect to promotion, see text ante at Part XII(F), apply equally to tenure decisions. With this framework in mind, the court turns its attention to the class claims, the claims of the named class representatives, and to Roworth's suit.
E. Summary (Tenure)
1. Classwide Discrimination
a. Mandatory Tenure
The main weaponry for the plaintiffs' assertions of tenure discrimination came from Zellner's statistical report. But, as the court has previously noted, see text ante at Part XIII(B), Zellner's study was out of plumb. She neglected to take into account the productivity of individuals after they were hired. Viewed in the most charitable light, Zellner's calculations revealed no more than that women with prior credentials and seniority equivalent to those possessed by men arguably failed to receive tenure on an equal basis with such males during the period from 1972 to 1976. In the utter absence of any proof that a necessary (or even a logical) correlation existed between pre-URI credentials and post-hire accomplishments, however, Zellner was shooting blanks. There is no reliable statistical evidence in the case which is *1267 probative of discriminatory decisionmaking in the tenure area; Zellner's study, by itself, does not support a prima facie case of classwide tenure discrimination.
Nor does the anecdotal evidence sufficiently bridge the gap. Though the atmosphere at URI was subtly scented with potentially poisonous vapors, there is little to show that such negative attitudes infected the tenure process. The plaintiffs' non-statistical proof on the point is too slight to carry the weight of their argument, especially since isolated incidents, without more, do not constitute a pattern or practice. See Cooper, ___ U.S. at ___ & n. 11, 104 S. Ct. at 2801 & n. 11; General Telephone Co. v. Falcon, 457 U.S. 147, 159, 102 S. Ct. 2364, 2372, 72 L. Ed. 2d 740 (1982).
The credible evidence adduced at trial fails to make out the existence of a prima facie case. The court must conclude that the plaintiffs' contention that URI discriminated against women with respect to denial of mandatory tenure has miscarried.
b. Early Tenure
The plaintiffs' evidence with respect to early tenure is a trifle stronger. Although the statistical proof is arguably even weaker (Zellner's model, after all, was designed to study only whether tenure was received, not when it was granted), the anecdotal evidence tends to be somewhat more substantial. The plaintiffs produced evidence of four decisions that implicated early tenure.
Of course, even if all four calls were vagariously made as the plaintiffs suggest, the quartet of incidents would not necessarily constitute a pattern or practice of discrimination on the facts at bar. See Cooper, ___ U.S. at ___, 104 S. Ct. at 2801; Falcon, 457 U.S. at 159, 102 S. Ct. at 2372; Goff v. Continental Oil Co., 678 F.2d 593, 597 (5th Cir.1982). Cf. Chisholm v. United States Postal Service, 665 F.2d 482, 495 (4th Cir.1981). The court, however, need not wrestle with quantification on this order of magnitude, for the majority of the cited examples cannot withstand close perscrutation. Neither Hufnagel nor Strom was able to state a prima facie case; and the dean's basis for his rejection of the Kulberg recommendation must, on this record, be labelled as legitimate. Roworth, standing alone, cannot carry the entire class on her back.
Classwide, on this record, the plaintiffs have not established that early tenure was parcelled out in a discriminatory way.
2. Class Representative Claim
Strom, a named plaintiff, has already been discussed in some detail. E.g., text ante at Part XIII(D) (5). She did not make out a prima facie case. Upon the initial tenure nomination, her peers did not simultaneously support her for promotion. URI's policy was to cede tenure, generally, coincident with promotion to associate professor. By failing to show that she was then qualified for a higher rank, Strom undercut her contention that she was ripe for tenure. Moreover, even assuming arguendo that Strom was qualified for tenure on this occasion, the University articulated a valid nondiscriminatory motive for denying her at that time. Strom introduced no evidence to overcome URI's professed reason. She did not show, for example, that men were tenured without promotion in the year she sought early tenure; nor even that men were from time to time allowed that privilege under congenerous circumstances. Without evidence of pretext, the court cannot find that Strom overcame URI's production of a genuine ground for denial of tenure. And, since Strom's second tenure review (concomitant with a recommendation for promotion) turned out favorably, no claim can be premised on that series of events.
Her case fails.
3. Roworth Suit
Roworth demonstrated that she was qualified for tenure and promotion to *1268 associate professor effective July 1, 1981. She also established that she was rejected in her quest. Under the Zahorik model, Roworth plainly set out a prima facie case. URI attempted to articulate a nondiscriminatory rationale for denying her tenure: Marks told her that she had not fully matured as a scholar and she should linger yet awhile in order to develop a better record. This showing shifted the burden back to Roworth to demonstrate that the commentary was pretextual and that discrimination was the real reason behind the decision.
Roworth proved that all but one of a series of matched males, each of whom received anachronous tenure, had records no better than her own. The evidence also showed that, at least in 1980-81, a consistent pattern of early tenures occurred; many of those decisions involved men with whom she could fairly be compared. And, scholarly maturity did not debar any of the latter.
In the court's view, Roworth established that URI, in 1980-81, construed the provisions in the collective bargaining agreement requiring unusual circumstances for early tenure awards in such a way that the decisions invariably favored men. Roworth showed that in over one-third of these cases she had a record that at least equalled those of males who received early tenure. To the extent (if at all) that Marks' reasoning had merit, it should have applied equally to several of the male aspirants (but it did not). These facts, when scrutinized in the light of an omnipresent atmosphere at URI in which little regard was paid to eliminating subtle (or unsubtle) forms of sex discrimination, suffice to establish liability. Roworth overcame URI's rebuttal to her prima facie case; she demonstrated that the decision with respect to her denial of tenure resulted from the fact that URI, at least in her instance, was unduly loathe to award early tenure to women. No rational explanation other than sex discrimination can explain the adverse employment decision in Roworth's case.
To summarize, though the court has not found any pattern or practice of discrimination with respect to tenure (early or mandatory) over the time frame of this litigation, Roworth not only satisfied the devoir of persuasion with respect to a prima facie case, but also overcame URI's articulation of a professed nondiscriminatory reason for the denial of her tenure. She showed that the University's explanation was not deserving of credence and that intentional discrimination was, more likely than not, the sticking point in this instance. The defendants are liable.
XIV. RECAPITULATION OF FINDINGS AND CONCLUSIONS
URI, as an institution, exhibited scant respect over time for the niceties of affirmative action programs; compliance was at best a quondam thing. And, this light regard for affirmative action mandates set the mood, in a very real sense, for this mammoth litigation.
Yet, these cases fundamentally deal not with heightened sensitivity, but with patterns and practices. The court's inquiry is, of necessity, fact-intensive. Attitudes become important only insofar as they have been translated into acts or omissions, insofar as they created a frame of reference within which the court may lucidly choose among conflicting inferences, insofar as they shed light upon amphibolous behavior.
The scope and intricacy of the network of issues which have been confronted and resolved suggest that, before proceeding to consider the ramifications of the court's findings and conclusions, a pause should ensue, and an attempt should be made to sum up the state of the record to this point.
*1269 There is no solid evidence that URI discriminated in deciding whether or not to hire women faculty members. There were, to be sure, isolated incidents where the University fell short; but it cannot be said that the plaintiffs have proven the existence of a policy disfavoring the engagement of females or preferring recruitment of males. While there may well have been an aberrant department or two, the practices of a small, retrenched bastion cannot, on this record, be imputed to the University as a whole. Furthermore, the allegations made early in the game to the effect that the Ph.D requirement which Weeks imposed for the college of business administration (i) adversely and unfairly impacted the opportunities for women to be hired and/or (ii) was applied heavyhandedly to females and winked at for males, see, e.g., Chang v. URI, C.A. No. 77-0070, slip op. at 8-11 (D.R.I. March 2, 1983), were virtually abandoned by the plaintiffs at and after the trial. In any event, there was a notable dearth of proof to support either such contention. Discrimination in hiring remains, on this record, unproven.
The assignment of rank at hire stands on a different footing. Reliable numerological evidence demonstrated that there was a statistically significant variance between the rank placement of men and women with similar backgrounds. And the anecdotal evidence, while scanty, seemed to point to unexplained differences in rank placement differences which habitually favored males and disfavored females. The record supports a finding of purposeful discrimination in this regard. Classwide liability exists as to rank placement at hire, with respect to those women who were (i) enlisted subsequent to March 24, 1972, (ii) at the rank of assistant professor or below.
A similar note was sounded by the evidence anent salary at hire. The statistical evidence revealed an unexplained differential as between men and women. Chance was sufficiently excluded as a basis for this asymmetry. The defendants' attempt to explicate the disparity in terms of market factors was an empty exercise in casuistry. At hire, men at URI were invariably paid better in relation to the national and regional going rates than women. A pattern and practice existed. Once again, the court has found classwide liability in respect to starting pay as to female faculty members who began their institutional service at the University after March 24, 1972.
The question of annual compensation was perhaps the most difficult to resolve. The statistical evidence was not issue-determinative. And, the parties were singularly unhelpful to the court in attempting to screen out the warp which was created by preferment in salary at hire in order to avoid double-counting, or in endeavoring to factor the award of inequity adjustments into the mix so as to permit an accurate overview of the situation as a whole. Nor did any party accurately gauge the impact of collective bargaining on pre-1972 and/or post-1972 annual compensation. The most that can authoritatively be said, given this tangled web, is that, with respect to yearly pay, gender has not been proven to be a substantial classwide determinant. There is, however, one salient exception to this overall conclusion. The exception relates to persons hired after the effective date of Title VII, and runs to the extent that discrimination in salary and/or rank placement *1270 at hire may have created genderbased disparities which were thereafter distended by the routine application of across-the-board percentage increases under the collective bargaining pacts.
With regard to both promotion and tenure, the record is considerably less tenebrous. Such decisions were largely judgmental in nature, and were implemented against a backdrop of poorly articulated criteria. The results, accordingly, were not letter perfect, and it is possible to nitpick here and there. Nonetheless, in neither area did the evidence create any substantial foundation for a charge of classwide sex discrimination. The plaintiffs have failed to heft the devoir of persuasion.
Although Chang, as pled, arguably raises questions implicating URI's policy and practice in the disparate application of termination standards to female faculty members, there was no proof of such at the trial; Chang's own claim in this regard was never proven (Weeks' applied his doctoral degree requirement, which was itself fair and neutral, evenhandedly to all faculty members, irrespective of gender); and the plaintiffs neither briefed nor argued this contention classwide. The claim was, therefore, waived. See O'Brien v. Sky Chefs, Inc., 670 F.2d 864, 868 (9th Cir. 1982); Volyrakis v. M/V Isabelle, 668 F.2d 863, 865 n. 1 (5th Cir.1982) (plaintiff waives error in dismissal of claim due to failure to prosecute it on appeal); cf. Larkin v. United Association of Journeymen, 338 F.2d 335, 336 (1st Cir.1964), (points not argued are waived), cert. denied, 380 U.S. 975, 85 S. Ct. 1337, 14 L. Ed. 2d 270 (1965).
The claims of the class representatives sort out as follows. Chang proved neither her claims of discriminatory termination, hiring, nonhiring, or setting of pay levels, nor her paper charge that the defendants discriminated against her in the promulgation and/or the implementation of the business college's doctoral requirement. Seleen, Cohen, and Anderson were uniformly unsuccessful on their respective initiatives. Strom fought to a draw with the University in that she failed to prove intentional discrimination pertaining to tenure or promotion in her case, but succeeded in doing so vis-a-vis compensation.
Finally, both of the individual suitors have established liability: Kraynek with regard to violation of the Equal Pay Act and Roworth with respect to URI's denial of her attempt to secure early tenure.
XV. LIABILITY OF THE DEFENDANTS
As indicated previously, see text ante at Part I(A), these various suits were prosecuted against a hodgepodge of diverse defendants. Having ascertained liability as a general matter, it becomes necessary to sort out the status and responsibilities of the sundry defendants.
A. Chang Class Action
Chang's suit was instituted against three defendants: URI, Board-R, and Dean Weeks. Insofar as Chang's individual claim is concerned, judgment must enter for all defendants. As to the class claims, judgment must be ceded to all defendants as to (i) recruitment and hiring, (ii) promotion, (iii) tenure, and (iv) termination. As to (i) rank at hire, (ii) pay at hire, and (iii) annual compensation, liability, to the extent found by the court and indicated in the text hereof, will run classwide against URI. Board-R has passed into oblivion and has been supplanted by Board-HE, which has assumed all of its rights and responsibilities. See text ante at note 1. See also R.I.Gen.Laws § 16-59-1 (1981). The court hereby substitutes Board-HE as a defendant in the place and stead of Board-R, cf. Fed.R.Civ.P. 25(d), and orders that the liability findings on the successful class claims be entered as against Board-HE. No individual liability has been shown with respect to Weeks, however, and judgment should enter in his favor on all claims.
B. Seleen Class Action
The Seleen case was instituted against URI, Board-R, the Rhode Island Department of Education, and Carlotti (in his capacity as chairman of Board-R, and not individually). Board-HE must replace *1271 Board-R, see text ante at Part XV(A), and that substitution is hereby effectuated. As to all defendants, judgment shall enter in their favor on the unsuccessful classwide claims regarding promotion and tenure. The state Department of Education is not a proper party defendant, and shares no responsibility for the matters at bar; ergo, judgment shall enter in its favor on all claims. To the extent liability has been found vis-a-vis annual compensation, that finding runs against URI, Board-HE and Carlotti in his representative capacity. (He remains today as the chair of Board-HE.)
Insofar as liability has been imposed in favor of Strom (annual compensation), the finding implicates URI and Board-HE. As to all other personal claims of the class representatives, judgment shall enter in favor of the defendants, collectively.
C. Kraynek Suit
Sandra Kraynek's suit was brought against Board-HE, Carlotti (in his official capacity as the chair of Board-HE) and Dean Tate, individually and as dean of the college of nursing. The favorable liability finding which Kraynek has obtained indisputably runs against Board-HE and Carlotti (in his representative capacity). Kraynek has also proved her case against Tate, both individually and in her official capacity. Tate was undeniably acting in the interest of URI in setting initial salary levels for newly-hired faculty members, yet her discriminatory actions in rigging the Champlin/Tate pay differential went beyond the routine and evenhanded implementation of University policy. Tate, therefore, is properly held liable as an individual defendant in the suit, as well as in her official role. See 29 U.S.C. § 203(d) (1982). E.g., Weiss v. Marsh, 543 F. Supp. 1115, 1118 (M.D.Ala. 1981).
D. Roworth Suit
Wendy Roworth prosecuted her action against Board-HE and against Carlotti (as its chair). The liability finding limited above has force and effect against both defendants.
A. Classwide Claims
Insofar as the court has held against the defendants on certain of the classwide claims, viz., rank placement at hire (exclusive, however of those originally enlisted at a rank of associate professor or higher), salary at hire, and annual pay (to the limited extent that this finding implicates the ongoing compensation of persons hired after March 24, 1972 by virtue of the application of global percentage increases to impermissibly-fixed starting salaries), further proceedings are requisite. The very nature of a bifurcated trial demands such follow-up in order to determine which members of the class deserve relief. See Teamsters, 431 U.S. at 361, 97 S. Ct. at 1867; Craik, 731 F.2d at 484. Any woman who was originally hired by URI after the effective date of the EEOA (March 24, 1972) may participate in the second phase proceedings. The court's findings mandate the establishment of three successful subclasses. As a prelude to further action, the defendants shall examine the University's records and shall file with the court, within sixty days next following the date hereof, a roster for each of the three subclasses comprising all known persons who are or may be eligible for membership therein, together with the last known address of each. The plaintiffs shall, within thirty days thereafter, supplement those proffers by an incremental filing of information within their possession indicating *1272 omissions (if any) from the membership lists prepared by URI.
As soon as practicable after completion of these rosters, the court will confer with counsel as to an appropriate form of notice to be mailed (at the defendants' expense) to each subclass member, see Fed.R.Civ.P. 23(d) (2), and as to a satisfactory method of notice by publication further to insure protection of the rights of the members of the subclasses and the continued fair conduct of the Chang and Seleen class actions. Id.
The court, or the special master(s), see text post at Part XVII(B), shall, with the assistance of the parties, then formulate a final and definitive roster of the membership of each subclass. To the extent (if at all) that the parties disagree as to the inclusion of any individual(s), the defendants shall bear the burden of proving, by a preponderance of the evidence, that the woman in question is not a member of the particular subclass as defined by the court.
The determination of eligibility will raise a presumption that each particular claimant was a victim of the type(s) of discrimination suffered generically by her subclass. The defendants can then introduce evidence to prove, by a fair preponderance, that the particular individual was not herself subjected to an impermissible employment decision. (This showing is, of course, different than the preliminary showing that a woman is not eligible for participation in these proceedings, i.e., not a bona fide member of one or more of the subclasses.) If the defendants are unable to prove that the discriminatory pattern and practices avoided a given claimant, that individual will be entitled to damages. Consistent with the prophylaxis hereof, members of Subclass 2 who are shown to have received appropriate entry level wages will, notwithstanding their initial inclusion on the Subclass 3 roster, thereby be precluded from participation in the incremental treasure hunt contemplated for the benefit of the last-mentioned subclass.
Each individual claimant whose claim survives the second stage determination must then shoulder the initial burden of establishing the extent of her damages. See Teamsters, 431 U.S. at 361, 97 S. Ct. at 1867. And, the methodology for this third stage determination will vary based on subclass membership and the genre of discrimination. Thus, in determining damages stemming from initial rank placement, a third stage claimant will have to show the rank to which she was entitled at hire and the adverse economic effects proximately attributable to her original misranking. Likewise, in assessing damages flowing from inadequacy of salary at hire, a third stage claimant will have to prove her proper initial salary level, and then trace the likely course and consequences of that discrimination.
Assessment of suitable ranks and salaries at hire should not, however, be attempted on a University-wide basis. The evidence in the case reflects permissible, non-gender-based variations from department to department. Zellner used a form of departmental breakdown in her salary at hire analysis; and the court, in examining salary at hire policies, also used a departmental *1273 analysis. Parity, for Title VII or Equal Pay Act purposes, in no way requires that, say, a male assistant professor in the humanities earn the same as a female assistant professor in engineering.
Such methodologic distinctions demand appropriate accounting for discipline differences as well. To the extent feasible, the focus should be on comparably credentialled individuals within the same department and discipline (e.g., Kraynek vis-a-vis Champlin). If a valid match is nonexistent, then departmental averages may come into play. For example, on the salary at hire issue, the starting stipends of men hired in the department at the rank and in the year in question can be averaged and that product used as a benchmark. If insufficient numbers exist so as to make such averaging utile, broaderbut logically related groupings can be sought to arrive at the benchmark salary.
There are a few divisions in which departmental agglomeration is inadvisable. Nursing and dental hygiene are the prime exemplars. In addition, there may be departmental groupings in which there are insufficient hires in a given year for averaging to be fair and productive. Should that occur, the calculation of damages in these instances could proceed on a multi-level averaging process. The first level of the computation would involve averaging the starting salaries of all male hires within a given rank and year at URI. For the same rank and year, the average starting salaries of all individuals in land-grant institutions can be ascertained (for example, from information compiled in the Oklahoma State University surveys). From such data, a ratio of national and/or regional averages to URI averages can be established. Then the individual plaintiff's salary can likewise be compared to the regional and/or national average for the appropriate rank and department. By comparing the overall URI ratio to this specific departmental ratio, a rational foundation for calculating backpay can emerge. To illustrate: if the ratio of URI to national average is 120% and the woman's salary is only 95% of the national average for the department, then the woman's salary should be increased to 120% of the national average.
Once a prejudice-free starting salary for a third stage claimant is fixed, the justifiable wage for any later year can be computed by factoring in the various across-the-board percentage raises and increases received as a result of promotion. This procedure will yield an unbiased salary for the terminal (or current) year as well as the total backpay award which the woman deserves to make her financially whole for the discrimination suffered.
Issues of offset and prejudgment interest must also be addressed. In addition to authorizing backpay, 42 U.S.C. § 2000e-5(g) provides:
[i]nterim earnings or amounts earnable with reasonable diligence by the person or persons discriminated against shall operate to reduce the backpay otherwise allowable.
The court must resolve whether this language requires, in respect to the computation of damages, the offset of any award by the amount of any inequity adjustment *1274 (under the aegis of the Rosie Committee or otherwise).
The purpose of Title VII is to end discrimination and to make the affected individual(s) whole. Albemarle Paper Co. v. Moody, 422 U.S. 405, 418, 95 S. Ct. 2362, 2372, 45 L. Ed. 2d 280 (1975). If the court does not permit the offset of backpay by voluntary payments designed to ameliorate discrimination, then the purpose of Title VII would be thwarted. Cf. Ford Motor Co. v. EEOC, 458 U.S. 219, 228-30, 102 S. Ct. 3057, 3063-65, 73 L. Ed. 2d 721 (1982) (unconditional job offer to Title VII claimant tolls further accrual of backpay liability, thereby giving employer incentive to end discrimination through voluntary compliance). The law requires an individual to mitigate damages by accepting another comparable job. Id. at 231-33, 102 S. Ct. at 3065-66. Such a principle of mitigation should apply equally to a person who retains her employment. If the individual accepts a monetary payment designed to alleviate discrimination, then there is a mitigation of damages.
Another tenet of Title VII requires that individuals be compensated solely for economic injuries. Albemarle Paper Co., 422 U.S. at 418-22, 95 S.Ct. at 2372-73; Merriweather v. Hercules, Inc., 631 F.2d 1161, 1167 (5th Cir.1980); Pearson v. Western Electric Co., 542 F.2d 1150, 1152 (10th Cir.1976). Any award of damages that purports to do more is prohibited. See DeGrace v. Rumsfeld, 614 F.2d 796, 808 (1st Cir.1980). The voluntary payments made by the University scumbled some of the economic harm stemming from the discrimination. If the backpay award is not offset to the extent of the volunteered reparation, the claimant would receive a trouvaille, that is, remuneration beyond the parameters of the economic injury suffered. Such a windfall is forbidden by the spirit of Title VII and by prudential precepts consistently enunciated by the courts of this circuit. E.g., Kolb v. Goldring, Inc., 694 F.2d 869, 872 (1st Cir.1982) (ADEA). Therefore, the offset supports the purposes of the Act.
For these reasons, any backpay award should be cushioned by the amount of Rosie Committee award or other inequity adjustment (except that the backpay award shall not be trimmed by the inequity adjustment mentioned in the 1974-75 collective bargaining agreement unless the defendants prove by a preponderance of the evidence that such an award was made to a particular claimant in order to correct salary discrimination). The failure of any person to receive such an inequity adjustment shall have no relevance in the determination of backpay.
The award of prejudgment interest is within the sound discretion of the trial court. Segal v. Gilbert Color Systems, Inc., 746 F.2d 78, 82-84 (1st Cir.1984); Domingo v. New England Fish Co., 727 F.2d 1429, 1446 (9th Cir.1984); EEOC v. Wooster Brush Co. Employees Relief Association, 727 F.2d 566, 578-579 (6th Cir. 1984); Donovan v. Freeway Construction Co., 551 F. Supp. 869, 880-81 (D.R.I.1982). Prejudgment interest often is awarded as an element of damages in Title VII actions. See Grossman & Schlei, op. cit. supra, at 1451 & n. 149. Cf. Earnhardt v. Puerto Rico, 744 F.2d 1, 3 (1st Cir.1984).
The First Circuit has recently reviewed the generic considerations counselling for and against an award of interest in a federal question case. Segal, 746 F.2d at 82-83. See also Furtado v. Bishop, 604 F.2d 80, 97-98 (1st Cir.1979), cert. denied, 444 U.S. 1035, 100 S. Ct. 710, 62 L. Ed. 2d 672 (1980); Citizens Savings Bank v. Bell, 605 F. Supp. 1033, 1047 (D.R.I.1985). The test is, as Segal indicates, constructed by a balance of the equities. 746 F.2d at 83. Justice Black noted in Rodgers v. United States, 332 U.S. 371, 373, 68 S. Ct. 5, 6, 92 *1275 L.Ed. 3 (1947), that "the failure to mention interest in statutes which create obligations has not been interpreted by the [the] Court as manifesting an unequivocal congressional purpose that the obligation shall not bear interest." Rather,
[A] persuasive consideration in determining whether such obligations shall bear interest is the relative equities between the beneficiaries of the obligation and those upon whom it has been imposed. And this Court has generally weighed these relative equities in accordance with the historic judicial principle that one for whose financial advantage an obligation was assumed or imposed, and who has suffered actual money damages by another's breach of that obligation, should be fairly compensated for the loss thereby sustained.
In Segal, an employment discrimination case brought under 28 U.S.C. § 1875, the court of appeals recognized that
[I]f an aggrieved employee is to be made whole in the face of his employer's insensitivity ..., such an award may be an invaluable remedial component. We hold, therefore, that prejudgment interest can be recovered under the Act [28 U.S.C. § 1875]. The "relative equities' demand no less.
746 F.2d at 83 (quoting Rodgers, supra).
The same principle holds true in Title VII litigation. The societal value of gender-blind employment decisionmaking is on a par with that of encouraging federal jury service. The purpose of Title VII is remedial and restorative: to make victims of discrimination whole. Presumably, some of the money which should have been paid to them, but was not, would itself have earned interest. This case, unlike Farnhardt, supra, presents no compelling reasons to eschew an interest award. If the salubrious policies of Title VII are to be furthered, those harmed by URI's invidious practices should receive prejudgment interest. The "relative equities" demand no less.
The rate of interest to be applied is within the trial court's sound discretion. Grendel's Den v. Larkin, 749 F.2d 945, 950 (1st Cir.1984); Tosco Corp. v. Federal Deposit Ins. Corp., 723 F.2d 1242, 1249 (6th Cir.1983); Citizens Savings Bank, at 1047; Freeway Construction Co., 551 F. Supp. at 880-81. Cf. Segal, 746 F.2d at 84. The court is mindful of the wide swing of interest rates over the lifespan of the events sub judice. Compare, e.g., R.I. Pub. Laws 1970 ch. 184 § 1 with R.I. Pub. Laws 1981 ch. 54 § 1 (prejudgment interest rate increased over time from 8% to 12%). The court, having weighed the many and sometimes competing considerations which impact this choice, directs that prejudgment interest be calculated at the rate of 9% per annum, simple interest (not compounded).
B. Equitable Relief
Title VII authorizes the court to "enjoin the respondent from engaging in such unlawful employment practice, and order such affirmative action as may be appropriate ..., or any other equitable relief as the court deems appropriate." 42 U.S.C. § 2000e-5(g) (1982). The plaintiffs seek such injunctive relief here.
It is apodictic that "[c]ourts, after all, do not enjoin parties from violating the law without proof of a real likelihood that such will happen." Derwin v. General Dynamics Corp., 719 F.2d 484, 491-92 (1st Cir.1983). Yet,
When defendants are shown to have settled into a continuing practice ..., courts will not assume that it has been abandoned without clear proof.... It is the duty of the courts to beware of efforts to defeat injunctive relief by protestations of repentance and reform, ...
United States v. Oregon State Medical Society, 343 U.S. 326, 333, 72 S. Ct. 690, 695-96, 96 L. Ed. 978 (1952) (citation omitted). Even the voluntary cessation of illegal conduct does not moot an action for injunction "if there is a possibility of recurrence, since otherwise the defendants `would be free to return to [their] old ways.'" Allee v. Medrano, 416 U.S. 802, *1276 810-11, 94 S. Ct. 2191, 2198, 40 L. Ed. 2d 566 (1974) (quoting Gray v. Sanders, 372 U.S. 368, 376, 83 S. Ct. 801, 806, 9 L. Ed. 2d 821 (1963)). Accord Los Angeles v. Davis, 440 U.S. 625, 631, 99 S. Ct. 1379, 1383, 59 L. Ed. 2d 642 (1979). As Justice Brennan has cautioned, United States v. Parke, Davis & Co., 362 U.S. 29, 48, 80 S. Ct. 503, 514, 4 L. Ed. 2d 505 (1960), courts must be chary of "deny[ing] relief altogether by lightly inferring an abandonment of the unlawful activities."
In determining whether to extend injunctive redress, the caselaw demands a realistic appraisal of the facts of a particular suit against the framework constructed by the Court in United States v. W.T. Grant Co., 345 U.S. 629, 633, 73 S. Ct. 894, 897, 97 L. Ed. 1303 (1953). In the instant cases, the character and duration of the past violations, coupled with the manifest reluctance of the University to face up to the legal demands of the twentieth century and the perils of relying solely on voluntary cessation of the interdicted practices, are such that this court perceives "some cognizable danger of recurrent violation," id., if a restraining order does not issue. At least four perdurable facts conduce to such a conclusion.
First, there is no credible evidence that the defendants have ceased paying women less than men at hire, or that they have begun assigning women to rank in an unbiased fashion. Second, many of the players who fostered antipathy toward women and, in certain instances, made discriminatory decisions, are still active in the University hierarchy. Third, the kind of choices that are made at URI require the application of subjective factors under which the sinister influences of discrimination can all too easily be masked. Fourth, the atmosphere of discrimination, e.g., text ante at Part V (E), has not wholly abated. Given these factors, and the recognition that old habits die hard, the court concludes that injunctive relief is warranted on this record. There is a reasonable likelihood of future transgressions sufficient to warrant the engagement of the court's Rule 65 powers. See United States v. W.T. Grant Co., 345 U.S. at 633-34, 73 S.Ct. at 897-98; United States v. Oregon State Medical Society, 343 U.S. at 333-34, 72 S. Ct. at 695-96.
Such an injunction must, of course, be drawn precisely and in close adherence to the tenets of Fed.R.Civ.P. 65(d). In general, it should require that the two chief defendants, URI and the Board-HE, use their collective best efforts (i) to prohibit and prevent discrimination against women with respect to initial salary offers and rank assignments at hire, and (ii) to comply with the University's affirmative action plan as the same may from time to time lawfully be framed, and (iii) to systematize the policies and procedures for making salary determinations and rank placements at hire. The terms of the permanent restraining order should further assure adequate documentation of each ensuing initial rank and salary offer in accordance with the guidelines promulgated herein. And, the system and procedures so developed must be effectively disseminated throughout the URI community.
The defendants must also refrain from making any initial salary offer and/or rank assignment until the affirmative action officer has approved the same and certified that, to the best of his or her belief and knowledge, it does not violate Title VII, the Equal Pay Act, or the dictates of this court's order.
The order shall not be applicable to, inter alia, (i) tenure decisions and/or promotions; (ii) initial rank placements at the level of associate professor or higher; (iii) hirings or continued employment funded in their entirety out of sources external to the University's appropriations from the General Assembly (e.g., certain endowed chairs; certain grant-subsidized positions); and (iv) administrative personnel, non-teaching personnel, and faculty (if any) not subject to the URI/AAUP collective bargaining agreement(s). Salary documentation will, however, be required with respect to categories (i) and (ii) ante, notwithstanding this paragraph.
*1277 The injunction must also require, of course, that each subclass member who remains currently employed at the University and who is ultimately found to have been misranked because of URI's discriminatory bent, see text ante at Part XVI (A), be placed in the appropriate adjusted rank, with all regular time-in-rank credits and other appurtenant benefits (if any); and further, that those found to be entitled to backpay, see text id., and who are still in the service of the University, receive such adjustments of current base wages as to afford them, on an ongoing basis, the "unbiased salary," see text id., demanded by this court.
Each party shall submit to the court, within thirty days of the date hereof, a proposed form of permanent injunction designed to accomplish these objectives, together with such embellishments (as to periodic monitoring, intramural appeals, or otherwise) as the parties, respectively, may deem desirable in connection therewith.
C. Class Representative Claims
In accordance with the findings and conclusions herein contained, and as noted above, see text ante at Part XV (A), (B), judgment may enter in favor of the defendants in respect to the individual claims of Chang, Seleen, Cohen, Anderson, and Strom (except that Strom is entitled to recover on her Equal Pay Act/compensation claim). Strom may prove her damages in this wise in due course to the court.
D. Individual Suits
1. Sandra Kraynek
Kraynek, having prevailed with regard to her Equal Pay Act claim, may prove her damages to the court. Kraynek's entitlement to damages raises, however, yet another issue: the availability of liquidated damages pursuant to the Fair Labor Standards Act, 29 U.S.C. § 216 (1982). The statute's anodyne, incorporated by reference as a remedy available under the Equal Pay Act, provides for liability not only for backpay, but also an equal amount in liquidated damages.
Liquidated damages should be awarded unless URI has "show[n] to the satisfaction of the court that the act or omission giving rise to such action was in good faith and that [it] had reasonable grounds for believing that [its] act or omission was not a violation of the ... Act." 29 U.S.C. § 260 (1982). Absent this showing of good faith or reasonable grounds to believe that a violation of the Equal Pay Act had not occurred, the award of liquidated damages is mandatory. Doty v. Elias, 733 F.2d 720, 725-26 (10th Cir.1984); Laffey v. Northwest Airlines, Inc., 567 F.2d 429, 464 (D.C.Cir.1976), cert. denied, 434 U.S. 1086, 98 S. Ct. 1281, 55 L. Ed. 2d 792 (1978); Melanson, 536 F. Supp. at 292. And, the discretion to award liquidated damages exists even if good faith inheres. Marshall v. Brunner, 668 F.2d 748, 753 (3rd Cir.1982); Melanson, 536 F. Supp. at 292.
Good faith in the context of this litigation depends on whether URI had an honest intention to uncover the requirements of the Equal Pay Act and to comply with those imperatives. Marshall v. Brunner, 668 F.2d at 753; Melanson, 536 F. Supp. at 292. In Kraynek's case, URI knew she was started at a lower salary than Champlin and made some (manifestly insufficient) attempt to correct the disparity. The differential still remained. Melanson stands *1278 as persuasive authority that such circumstances cannot constitute good faith.
In addition, the actions of URI do not reflect a reasonable belief on the part of the University that it was in compliance with the Act. Kraynek brought the disparity to the attention of the dean, to no avail. It was no secret that both individuals were required to perform the same work under almost identical conditions. Given all of the facts, URI seemingly had no irreproachable basis on which it could reasonably have relied in refusing to bring Kraynek's pay to the proper level.
While this evidence bodes ill for URI on the issue, it would be premature to hold the University liable for liquidated damages. The trial was bifurcated; the question of damages was not before the court. Fundamental fairness necessitates that the defendants be afforded the opportunity to demonstrate that their actions were not beyond the pale of 29 U.S.C. § 260. Accordingly, the court reserves decision on the liquidated damages question in Kraynek's case pending the damage phase trial.
2. Wendy Roworth
Roworth, having prevailed with respect to her claim that URI impermissibly denied her early tenure, may prove her damages to the court in due course.
XVII. FURTHER PROCEEDINGS
A. Retention of Jurisdiction
The court will retain jurisdiction over these cases to the extent reasonably requisite to insure the implementation of its orders and to provide complete relief in the premises. Damage phase trials anent Strom's annual compensation entitlement and in regard to the Kraynek and Roworth cases will be scheduled in the immediate future. Counsel should plan accordingly.
B. Special Master(s)
It is evident that the identification of the members of the three prevailing subclasses, the determination of the defendants' liability for individual claims from within the membership of those subclasses, the correction of misrankings and of salary levels, and the computation of a damage award for each successful claimant, will in the aggregate require an extra-ordinary expenditure of time, effort, and detail work. This is precisely the sort of litigation where the existence of an "exceptional condition," Fed.R.Civ.P. 53(b), counsels reference to a special master.
The court, therefore, determines that one or more special masters should forthwith be appointed, Fed.R.Civ.P. 53, to conduct hearings, to take and receive evidence, and to report his/her/their findings and recommendations to the court. The scope of the reference shall embrace all questions anent (i) the identity of individual members of each of the three subclasses as to which classwide liability has been found, (ii) the notice to be afforded to such subclass members, see text ante at Part XVI (A), (iii) the existence of discrimination vel non as to each such subclass member, and (iv) the relief to which each such subclass member is entitled (quantified in dollars where appropriate). In making his/her/their recommended findings of fact and conclusions of law, the special master(s) shall be guided by the findings, principles, and holdings enunciated in this opinion, to the extent applicable.
The master(s) shall have and exercise, without limitation, all of the powers prescribed in Fed.R.Civ.P. 53(c), (d). The master(s) may, from time to time, seek guidance from the court by written inquiry (which inquiry, and any response thereto, shall be made part of the record of these proceedings). The master(s) shall proceed *1279 with all reasonable diligence; shall submit written status reports to the court at monthly intervals; and shall render his/her/their final report(s), together with a transcript of all proceedings undertaken and the exhibits introduced thereat, see Fed.R.Civ.P. 53(e) (1), within fifteen months from the date hereof (unless further extended by order of the court, for cause shown). The master(s) shall be entitled to reasonable compensation for his/her/their services, and to reimbursement for all reasonable expenses incurred, to be borne by the defendants (unless the court, by subsequent order, determines that some part of such costs should equitably be allocated otherwise).
The reference to the master(s) shall not include C.A. No. 83-0044 (the Kraynek case) or C.A. No. 83-0099 (the Roworth case); nor shall such reference extend to the class representative claims decided herein; nor shall such reference extend to the damage phase trial of Strom's compensation claim (that being reserved to the court).
The parties shall each submit to the court, within twenty days from the date hereof, proposed nominee(s) for the office of special master(s) and a suggested form of order of reference, consonant with the foregoing.
C. Intermediate Appeals
Inasmuch as this opinion is rendered at the conclusion of the first (liability) stage of a bifurcated trial, the order(s) entered in pursuance hereof will not ripen automatically into final judgments, and will not be immediately appealable as of right. But, 28 U.S.C. § 1292(b) provides that:
When a district judge, in making in a civil action an order not otherwise appealable under this section, shall be of the opinion that such order involves a controlling question of law as to which there is substantial ground for difference of opinion and that an immediate appeal from the order may materially advance the ultimate termination of the litigation, he shall so state in writing in such order. The Court of Appeals may thereupon, in its discretion, permit an appeal to be taken from such order, if application is made to it within ten days after the entry of the order: Provided, however, that application for an appeal hereunder shall not stay proceedings in the district court unless the district judge or the Court of Appeals or a judge thereof shall so order.
The court is mindful that intermediate appeals are not favored as a rule, and that resort to interlocutory certification under 28 U.S.C. § 1292(b) should be employed only in extraordinary circumstances. McGillicuddy v. Clements, 746 F.2d 76, 76 n. 1 (1st Cir.1984); In re Heddendorf, 263 F.2d 887, 888-89 (1st Cir.1959). This litigation, however, has a symptomatology which fairly cries out for the balm of the statute: it presents an interleaved series of difficult and pivotal questions of law as to which there is a dearth of controlling precedent and as to which there is appreciable room for differences of opinion. An immediate appeal from the classwide orders contemplated hereby would have the salutary effect of resolving some of these critical questions with a greater degree of finality. And, such a process would in this court's judgment both materially advance, and reduce the cost of, the ultimate termination of the legal battle. To permit the second and third stage proceedings to run their course, at enormous expense to the parties and to the judicial system, with the grey eminence of appellate review lurking in the wings, would run a thoroughly unacceptable risk of prodigal wastefulness.
Should the parties (jointly and/or severally) wish to essay at this time one or more interlocutory appeals in pursuance of § 1292(b), the court stands disposed, upon timely presentment of an order to that effect within twenty days of the date hereof, to issue the requisite certificate (limited, however, to the Chang and Seleen cases).*1280 If such a request is forthcoming, the court will then and there hear the parties as to whether further class action proceedings hereunder ought to be stayed while the application for leave so to appeal is pending before the First Circuit (and thereafter, for the duration of appellate proceedings, should the application be granted).
D. Attorneys' Fees and Costs
Plaintiffs, to the extent that they are the prevailing parties on certain of the issues raised in this litigation, will be entitled to have the defendants pay reasonable counsel fees and costs. 42 U.S.C. § 2000e-5(k) (1982); Hensley v. Eckerhart, 461 U.S. 424, 434-37, 103 S. Ct. 1933, 1940-41, 76 L. Ed. 2d 40 (1983); Nadeau v. Helgemoe, 581 F.2d 275, 278-79 (1st Cir.1978); Coble v. Texas Dept. of Corrections, 568 F. Supp. 410, 411-12 (S.D.Tex.1983). So much time, effort, energy, and expense have heretofore been consumed that the court will entertain an interim fee application from the plaintiffs if they choose to file one. E.g., Yakowicz v. Pennsylvania, 683 F.2d 778, 781-82 & n. 4 (3d Cir.1982); James v. Stockholm Valve & Fittings Co., 559 F.2d 310, 358-59 (5th Cir.1977), cert. denied, 434 U.S. 1034, 98 S. Ct. 767, 54 L. Ed. 2d 781 (1978). Westfall v. Board of Commissioners, 477 F. Supp. 862, 867-68 (N.D.Ga.1979). Any such application must, of course, be suitably detailed and documented. See generally Grendel's Den, Inc. v. Larkin, 749 F.2d at 950-51. If filed, the application should cover all services rendered through and including the day preceding the date hereof. It should sort out and identify, to the maximum extent practicable, time spent on issues upon which the plaintiffs prevailed and time spent in the unsuccessful or idle pursuit of other issues.
If no such application is filed within sixty days from the date hereof, this invitation shall be considered withdrawn, and the question of fee entitlement can be taken up as a whole at or after the time of entry of final judgment. Should any such fee application be filed, the defendants shall have thirty days thereafter within which to file their opposition or other responses thereto.
Though the court has found for the defendants on several of the classwide and class representative claims, the defendants have a substantially heavier burden to meet in any attempt to coerce payment of counsel fees and costs by the plaintiffs. Christiansburg Garment Co. v. EEOC, 434 U.S. 412, 421-22, 98 S. Ct. 694, 700, 54 L. Ed. 2d 648 (1978). The court is of the opinion that those claims, though they proved to be unmeritorious, were not frivolous, nor were they sued out unreasonably or wholly without foundation. This court is cognizant that "post hoc reasoning" and "hindsight logic" cannot be applied in making such an assessment. Id. See also Hughes v. Rowe, 449 U.S. 5, 14-15, 101 S. Ct. 173, 178, 66 L. Ed. 2d 163 (1980) (per curiam); White v. South Park Independent School District, 693 F.2d 1163, 1170 (5th Cir.1982). Since the plaintiffs had colorable grounds upon which to premise the full panoply of their asseverations when they were brought, those claims, albeit unsuccessful as matters now stand, were not meritless in the Christiansburg sense. The court, accordingly, sees no reason to entertain cross applications from the defendants in this respect.
*1281 XVIII. EPILOGUE
Litigation of this genre, conducted on so grand a scale, inevitably reveals much about the anatomy of a university. As discovery and trial progressed, layer after layer of skin was peeled back until the ossature of URI was laid bare. The resultant picture is not a pretty one in all respects. Yet, in most cases, reportage of the view from the bench is clouded both by the need to avoid conjecture and speculation, and by the requirement that facts, once found, must be filtered through the seine of legal principle. The court's juridical task having been accomplished in the preceding pages, an eschatocol of sorts seems appropriatea postscript which, though not a necessary adjunct of the decision, attempts to place matters into a functional perspective.
It is plain to the court that URI, whatever its shortcomings are or may have been in affirmative action and in its dealings with women faculty members, is a great institution ... and has the potential for achieving even greater eminence. Neither its contribution to higher learning nor its enrichment of life and culture in Rhode Island can (or should) be gainsaid. URI is a principled institution, not a lawless renegade. But it is because of the ideals, values, and accomplishments which the University symbolizes that it must be held strictly accountable to the high standards of decency and fair play which underlie Title VII and the Equal Pay Act. If the seeds of discrimination can be allowed to take root and to sprout in the very citadel of enlightenment, then an horrific harvest awaits us all.
The path which URI followed in setting ranks and salaries at hire detoured around the permissible boundaries demarcated by the Congress in order to insure equality of opportunity; and this detour, in turn, rerouted the handling of annual compensation decisions in an unfair manner. Even in those arease.g., hiring, discharge, tenure, promotionwhere the University has not been proven guilty of sex discrimination, the conduct of its affairs has been measurably less than one would hope to have found. The inability of the plaintiffs to prove intentional classwide discrimination on a given issue is not the equivalent of ceding a clean bill of health to URI on that issue; the fact that the University prevailed on a particular point is neither an endorsement of its policies and practices nor a signal that there is no room for improvement. Quite the contrary is true.
The court recognizes that University administrators confront an increasingly complex array of tasks in these modern times. It is difficult to strike the requisite balances between excellence in education and the constraints of a parsimonious budget; it is difficult to reconcile the competing demands of faculty and students. And, this equilibration is especially difficult for a land-grant institution, which must also pay obeisance to the heavy burdens carried by the taxpayers. Yet, difficult or not, the juggling act must stay within the law; if the balls can be kept in the air only by trampling upon the inalienable rights of women (or blacks, or Roman Catholics, or whomever), then it is time to bring in a more dextrous troupe of jugglers.
Let there be no mistake: the court does not regard the URI hierarchy as being comprised of blackguards and dolts, of evil or incompetent men and women. With but a few exceptions, those administrators who testified in these proceedings appeared well-intentioned, dedicated, and professional. But a certain quality of sensitivity has obviously been lacking, nourished by a struthionoid tendency to see, hear, and speak no evil. To get along, one goes along. In the court's view, a reordering of administration priorities would be a healthy thing.
Nor are the problems which have festered at the University and which underlay this litigation entirely attributable to the *1282 defendants. The plaintiffs, too, must share some of the responsibility. Faculty members undeniably have rights, and must be free to assert those rights. But, the AAUP and its members seem to have developed an unfortunate penchant for confrontation, an uncanny ability to see ghosts in the most austere of closets, a self-annihilating eagerness to tackle the administration for the sake of the battleirrespective of the importance or insignificance of the issue or, sometimes, the merits of the administration's position. And, there are signs that faculty self-interest is, on occasion, much too far to the forefront. The faculty, too, would benefit from a reasoned reappraisal of its collective priorities.
The implementation of this court's orders willas the parties were forewarned doubtless be expensive, time-consuming, and to some extent disruptive. Yet, that is a price which must be paid when the judiciary is forced to intrude in the groves of Academe in order to rectify demonstrable wrongs. It is a consummation devoutly to be wished that the protagonists in this epoch will focus not on the nightmarish aspects of what has transpired, but on their shared hopes, dreams, and aspirations for the dawning of a better, more productive day. There are people of intelligence and good will on both sides of this dispute Carlotti, Pezzullo, Knauss, Ramsay, Lott, Anderson, Strom, to name but a few. A new president, uninvolved in the obumbrations of the past, has been installed. There is no plausible reason why, despite the travail, URI cannot fully achieve its mission. All that is required is a synergistic willingness on the part of the administration and the faculty to walk hand-in-hand, proudly, into a just and equitable future.
In a very real sense, therefore, this epilogue is no epilogue at all; the authentic epilogue to this litigation is yet to be written. The court can, by judicial fiat, insist that equality of opportunity become a reality in the University workplace and that the vestiges of discrimination be eradicated. But, the extent to which the bright promise of URI's future is to be fulfilled rests largely in the hands of the parties.
Each and all of the orders hereinabove contained are so mandated.NOTES
 In 1981, the responsibility of governing URI was transferred from the Board of Regents (Board-R) to the newly-constituted Board of Governors for Higher Education (Board-HE). R.I.Gen.Laws § 16-59-1 (1981). The court will refer to these commissions collectively as the "Board" and will distinguish between them only when necessary. During the times most directly relevant to this litigation, the defendant Albert Carlotti was chairman of the Board.
 The predecessor Board-R was similarly excluded from the daily administration or operation of colleges or universities. R.I.Gen.Laws § 16-49-4(3) (1956) (repealed 1981).
 URI has a number of colleges in which allied fields of study are grouped and administered. The colleges include: Arts & Sciences, Business Administration, Engineering, Graduate, Human Services, Nursing, Pharmacy, and Resource Development.
 The timeliness of Chang's suit was contested by the defendants. In an unpublished opinion dated May 28, 1980, Judge Pettine of this court rebuffed the challenge. The same opinion also rejected URI's contention that the court lacked subject matter jurisdiction because Chang was terminated before the effective date of Title VII's applicability to university faculty. The latter issue was raised anew later in the litigation. See text post.
 For purposes of this opinion, transcript testimony will be designated by volume number in roman numerals. Plaintiff's exhibits will be labelled with the prefix "Pl." and defense exhibits will be designated as "Df."
 § 3 of the Equal Pay Act, now codified at 29 U.S.C. § 206(d) (1) (1982), provides as follows:
No employer having employees subject to any provisions of this section shall discriminate, within any establishment in which such employees are employed, between employees on the basis of sex by paying wages to employees in such establishment at a rate less than the rate at which he pays wages to employees of the opposite sex in such establishment for equal work on jobs the performance of which requires equal skill, effort, and responsibility, and which are performed under similar working conditions, except where such payment is made pursuant to (i) a seniority system; (ii) a merit system; (iii) a system which measures earnings by quantity or quality of production; or (iv) a differential based on any other factor other than sex. Provided, That an employer who is paying a wage rate differential in violation of this subsection shall not, in order to comply with the provisions of this subsection, reduce the wage rate of any employee.
 URI has two types of employees: classified and non-classified. Classified employees are hired in accordance with the imperatives of the state civil service system pursuant to R.I.Gen. Laws §§ 36-4-1 et seq. All state employees are classified except those specifically immunized pursuant to R.I.Gen.Laws § 36-4-2. Among the positions exempted are the faculty slots at URI. While the hiring and recruiting of women and minorities in the classified system was subject to affirmative action, the posts so encompassed are not the subject of this litigation. These cases focus exclusively on those in unclassified service. Thus, Rocha's doings with respect to classified employees will not be discussed.
 The court recognizes that there are numerous possible explanations for these findings. For example, men had undoubtedly been at URI longer and in greater numbers and were probably in higher ranks than women. Seniority is, in and of itself, a consequential economic indicator in academe; the higher one's rank, the greater one's compensation. Given the social, socio-economic and historical predicates, the range of possible explanations for the phenomena revealed by the SIS Report is great. No useful purpose would be served by further speculation in this regard. It suffices to say that the SIS Report neither proved nor disproved the existence of sex discrimination.
 URI utilizes a duad of basic time/money salary formats. Some employees are paid on a calendar year basis (either January 1 to December 31 or September 1 to August 31). Others are compensated on an academic year hypothesis, from the beginning of the school year (traditionally, September 1) to the end of the school year (in June or thereabouts) next following. While the latter formulation is used almost exclusively for faculty, the former is used with some regularity for non-faculty employees, and some faculty members (selected, apparently, for a rather haphazard variety of reasons) enjoy the same benefit. To convert a calendar year salary to its academic year equivalent, URI multiplies the calendar year salary by 0.80. No claim has been made in this litigation that the University discriminated against women faculty in bestowing calendar year contracts.
 In addition to an AA officer, URI had in place numerous affirmative action committees. Each department had its own committee designed to wrestle with affirmative action. The umbrella under which these departmental groups toiled was the University-wide committee on which Lott served. Its responsibilities included oversight of affirmative action within each of the departments as well as oversight in respect to areas which fell between, across, or outside of the individual jurisdiction of the departments and colleges at URI.
 Board regulations required URI to conduct a search anent vacancies in the AA officer position. Whether the regulations applied to temporary assignments, such as Brittingham's, is tenebrous. The court, though noting that no search occurred, draws no conclusion from URI's failure to conduct one.
 These events occurred in late 1979. Feldman, in writing her resignation, settled upon an effective date of August 1, 1980. She remained in office until that time.
 The record is murky as to the precise cause of this lapse. Several rationales are possible: Newman did not budget the position when he went to the Board, or the Board axed the position, or the legislature refused to appropriate the funds when URI's budget request was approved. In any event, the University had sufficient flexibility to reallocate monies in order to fund the position, had Newman so desired or had the Board so insisted. The failure to do so was not explained.
 It should be borne in mind that the full panoply of safeguards described in the preceding paragraph was not set in place until Brittingham grasped the reins. Thus, Moseley (and Rocha, after him) operated in a somewhat different paperwork milieu.
 In the jargon of the University this process was termed "posting" because it required the AVPAA to list the individual on a USP-2 (a personnel form used to record payroll information vis-a-vis URI employees).
 There may, of course, be exceptions for part-time or plainly temporary posts.
 The search committee need not be unanimous. It is sufficient if a candidate has majority support within the committee. In such cases, members of the body who disagree with the nomination have the option of filing a dissenting report with the dean.
 Women dominated the number of hires among the health professionals at URI. NCES statistics, which included fields not taught at URI (such as medicine and dentistry), revealed that only 32% of the health professional doctorates were held by women. If this statistic was used, the expected number of women hires would be underestimated. To correct the problem, Zellner would have had to determine precisely how many of the 32% were in specialties offered by URI (e.g., nursing), and substitute the resultant figure. Inasmuch as that data was not readily available, Zellner performed her calculations with and without reference to health professionals.
 Most full-time faculty taught 9 hours/semester. Research was de rigeur.
 "Terminal degree" is academic argot for the highest degree that is regularly required to teach a particular subject on the University level. In most fields, the terminal degree is a doctorate. But, in some areas (e.g., studio art or music) a master's degree is considered to be the terminal degree (whether or not it is possible to obtain a doctorate). And in other specialties (e.g., accounting) a doctorate may be inapposite. There is no entirely objective yardstick; sometimes, denomination of a degree as terminal vel non is susceptible to debate.
 Although the court will not repeat this overall discussion as part and parcel of its analysis of the statistical evidence anent other practices at the University, Zellner used the plaintiffs' database as the touchstone for her studies generally. Similarly, the key defense expert, Siskin, made relatively constant reference to the defendants' database. On particular issues, these databases were supplemented as indicated specifically in the relevant text.
 The elements for which Zellner controlled seem indeed to be the most relevant items vis-a-vis rank at hire.
 Due to the small number of associate and full professors hired laterally by URI, Zellner had to exclude any control for years since degree and possession of the doctorate at hire. In addition, she had to combine certain departmental groupings at these ranks. Although she asserted that these problems were of a minor, technical nature, the court is not so sanguine.
 In fact, if eight more women had been assigned to the associate rank, the placement rate would have been exactly equal between men and women.
 This conclusion is not affected by another problem in Zellner's model. She used two variables, years since degree and prior experience, which are arguably correlated with each other. That is, the value of years of prior experience will change, at least to some extent, with the years since degree. This phenomenon is referred to as multicollinearity. While less multicollinearity is better than more, the problem is not severe. See Fisher, Multiple Regression in Legal Proceedings, 80 Colum.L.Rev. 702, 713 & n. 21 (1980).
 Special treatment in this regard was not, however, limited to men. Karen Stein, for example, was promoted to assistant professor of English without possessing her terminal degree.
 Prior to the negotiation of the initial collective bargaining agreement, University salaries were set in accordance with URI's Manual. The Manual correlated salary with a "unit" and specified the number of units necessary to achieve a given rank. The annual salary of an individual faculty member was determined by multiplying the salary per unit by the number of units assigned to the appropriate rank. In this fashion, a minimum was fixed for each rank. While this paradigm constructed a series of floors, there were no visible ceilings.
 Multiple regression is a statistical technique used to predict the effect of independent variables on some dependent variables. In its most basic form, regression analysis fits a line between observations such that the variation (in squared form for mathematical purposes) from the line to the observations is minimized ("least squares regression"). Multiple regression expands the concept from two dimensions (a single dependent variable and a single independent variable) to multiple dimensions, with numerous independent variables impacting the dependent variable.
In all statistical analyses, the independent variable(s) may not fully explain all of the variation in the dependent variable. While probability testing speaks to statistical significance and the elimination of random events as possible explanations of a phenomenon, multiple regression speaks in term of unaccounted-for variation. This is denoted by an error term in the statistical model. The value of multiple regression depends upon the validity of certain assumptions about the behavior of the error term. The error term must (i) not be related to changes in any independent variable, (ii) not have any discernible tendencies, and (iii) follow a normal (bell curve) distribution. Zellner's analysis did not transgress these rules.
If these assumptions are met and the study is otherwise valid, the statistician can use the technique to do two types of analyses: (i) prediction of the outcome of a dependent variable in the future from past observations; and (ii) derivation of the extent to which each independent variable affects the dependent variable. Zellner focused upon the latter.
 Nursing had only a scattering of men, dental hygiene had none. Conversely, many of the engineering and business departments had few, if any, women.
 The AAUP from time to time had negotiated both across-the-board raises for the faculty, and certain other, more irregular increases. The latter included augmentations set aside for merit, inequity adjustments, and market factors. While these will be discussed in more detail post, it is advisable at this point merely to note the practice in order better to comprehend the workings of the Rosie Committee. Such special increases were sporadic and were not uniformly included in successive collective bargaining agreements. Compare Pl. CC (1) with Pl. CC (4).
 If persons so identified were hired before 1972-73, salaries were also compared at year of hire; the wages paid latecomers were likewise scrutinized in the year of hire.
 The survey information does not trace back prior to the fall of 1974. Thus, for individuals hired before that time (but after the effective date of Title VII), the court has used 1974-75 figures as a proxy. Surely, this substitution leaves something to be desired, but the record presented no viable alternative. The surveys are less than all-encompassing in other respects as well; for example, a few of the samples are pitifully small. Moreover, the survey data does not specify a starting salary for new associate professors but only the average salaries of associate professors irrespective of experience level.
 The court notes, with respect to Hufnagel, that the lecturer position is not one covered by the collective bargaining agreement. The evidence disclosed that the position was created in this instance as an inducement for Hufnagel's husband to come to URI. She shifted to an assistant professorship in 1975. As such, it would be unfair to compare her to a newly-minted assistant professor since she had two years of teaching and research experience at URI as well as other relevant prior experience, e.g., research at Yale for two years, teaching in a community college, and research at Wayne State University. In gross dollars, Hufnagel's 1975-76 salary was high for a beginning assistant professor but low for an experienced assistant professor.
 Exceptional salary increases were often given by URI to induce selected faculty to remain at the University when they were coveted by other institutions. The practice is not uncommon in academic circles. See, e.g., Winkes v. Brown University, supra.
 The court notes, with a small bow in the direction of irony, that Champlin was eventually terminated by the University because he failed to enroll in a doctoral program. The assertion to the contrary in the defendants' post-trial brief has no record support; the testimony to which they allude refers not to Champlin but to Robert Jacques.
 Though research per se may not be required of faculty in, say, the performing arts, they are expected to undertake roughly equivalent creative endeavors in order to boost their stock at the University.
 An interesting question hovers on the periphery in consequence of this ruling. Kraynek's case falls squarely within the parameters of the certified class; yet, knowing that classwide litigation was afoot, she chose to sue individually and to prosecute her cause without awaiting a determination on the classwide issues. Under such circumstances, can she now clothe her case in the mantle of the court's finding of a classwide pattern and practice of discriminatory conduct in the setting of pay levels at hire at URI? Despite the intriguing nature of the problem thus presented, it is, in Kraynek's case, academic (in every sense of the word); viewed without benefit of any classwide presumptions, she has nevertheless clearly and convincingly proven her claim. The court, therefore, assumes for purposes of argument that she derives no succor from the classwide finding, but expresses no opinion on the question noted.
 Promotion plays a prominent role in salary determination. If an individual is promoted but his or her salary is less than the minimum for the new rank, an automatic raise would result. Since 1979, the collective bargaining agreements have provided for a lump sum payment upon promotion (the size of the award being dependent upon the rank to which the faculty member was elevated). So, Zellner reasoned, if promotion was itself manipulated in a gender-biased manner, factoring the results of the promotion process into the annual compensation mix would simply spread the infection.
 The court notes that Siskin's R2 values in these computations are puny. In his absolute growth study, for example, R2 never reached 70%; in the proportionate growth study, R2 never reached 50%. Thus, much of the variation, especially in proportionate growth, was left unexplained. Inequity adjustment awards may hold the key to unlocking this puzzle. Some such were made in 1974-75; if females received the bulk of the awards, then the growth in their current salaries would perforce be higher than males. Similarly, in 1978-79, the Rosie Committee made awards to 53 women (41% of the female faculty at URI). These awards necessarily bumped up the base salary of women in a manner from which men did not benefit. It is possible (although not likely) that the effect on proportionate growth in salary may have carried forward to 1979-80 and beyond.
 Siskin's study did exhibit statistically significant differences in salaries in 1971-72. But, the analysis failed to account for the case-by-case salary determination in vogue prior to and during that academic year. Without a study for the period focused upon post-hire productivity and the like, the court cannot make the quantum leap urged by the plaintiffs so as to attribute the origins of this differential to faculty gender.
 Release time constitutes a reduction in a faculty member's teaching load so that he or she may perform other, usually non-academic, duties for the good and welfare of the University. In discussing coaches, it is well to bear in mind that oversight of varsity teams was, as a rule, vested in URI's department of athletics. The policies of that department are not a subject of this litigation.
 Despite the "annual review" nomenclature, it should not be assumed that each member of the University faculty was evaluated in this manner every year. It would be the epitome of futility, for example, to review at each 12-month interval a tenured full professor; such an individual has both job security and no higher rank to attain at URI. Nevertheless, the University jargon speaks in terms of "annual reviews," and the court adopts the custom. The evidence did not reveal any instance of an in-house promotion where an annual review had not been performed on a reasonably current basis.
 The dossier is assembled by the faculty member in question. It recites his or her accomplishments (including, but not necessarily limited to, a demonstration of excellence in teaching and/or advising, publications authored or coauthored, presentations made at conferences, membership in professional societies, research conducted, grants received, service to URI and to the community at large, and academic, professional, and/or artistic achievements of divers kinds), usually from the flattering perspective of enlightened self-interest.
 It should be noted, when one focuses on promotion from assistant to associate, that neither Siskin nor Zellner adjusted for the effect (if any) of more strenuous tenure criteria. See text post at Part XIII(A).
 There were, however, a number of examples to the contrary, i.e., promotions to assistant professor (or above) absent terminal degrees. E.g., Karen Stein (post at Part XII(E) (3))); Greta Cohen (ante at Part XI(E) (5)); Judith Swift (post at Part XII(E) (1)).
 The appropriate terminal degree would have been in drama (either an M.F.A. or a Ph.D), not an M.A. in English.
 Implicit in the Avery argument was thesis that women were regularly tenured without being promoted, in derogation of URI's policy. The evidence is much too sparse to permit such a conclusion.
 The Manual provided shortened pre-tenure periods for those individuals initially hired at associate or full professor rank. In the case of associate professors, the minimum period was three years.
 If a faculty member initiates a grievance (whether in consequence of adverse action emanating from the deanery or from the president's office) and loses, separation from URI occurs at the end of the eighth year of employment.
 This is both underinclusive and overinclusive. Initially, it fails to acknowledge that faculty hired at the higher ranks need less time to qualify for tenure consideration. Such persons would have had ample seniority at URI to be tenure-eligible. And, while an academician who joined URI as an assistant professor in the fall of 1977 would have been eligible (albeit barely) for tenure consideration as of the effective date of Zellner's study, the mandatory decision date for such an individual would be well beyond the reach of her analysis.
 The court specifically rejects the notion that Schach-Cook was somehow punished (by denial of tenure) for her perceived unwillingness to accede to Robb's imprecation that she shift her research focus. The court notes, however, that even if such retaliatory conduct could be inferred from the evidence, the thrust of the point would nevertheless fail to implicate concerns germane to sex discrimination. The plaintiffs presented no evidence that similar exhortations were not made to males when circumstances warranted or that men were freer than women to shrug off such administrative advice.
 Certain scholarly journals (refereed journals) submit articles received for publication to academicians for pre-screening. These interlocutors comment on the quality, originality, and importance of the scholarship and the editorial style. Pieces are accepted for publication only after successfully running this gauntlet. A similar screening process is sometimes employed as a pre-condition to the presentment of learned papers at conventions, colloquia, and other professional meetings. Hence, the parallel terminology "refereed papers." Work product which has not been submitted to such testing, or which has been submitted and rejected (unrefereed work), may nevertheless be published or presented under less exacting auspices.
 While Robb attempted to disparage Roworth's accomplishments, her criticisms were both uninformed and unintelligent. Marks, to his credit, placed no apparent reliance on this sort of drivel; and it played no discernible part in his decision.
 There is at least one difference, in that tenure decisions unlike promotion decisions require concurrence by the Board. Thus, in theory, discrimination can occur at this rarefied level as well, that is, the plaintiffs could make out a case on this claim either by showing that the Board itself discriminated in tenure awards or failed to correct discrimination which routinely transpired at lower levels. But, this point is more theoretical than real; the plaintiffs have not argued that it was the Board which was the instigator of any disparate treatment in connection with the granting of tenure.
 Other evidence revealed at least two instances where the University did opt to tenure females without simultaneously promoting them. See text ante at Part XII(E) (3) (Karen Stein); Part XIII(D) (3) (Linda Shamoon).
 The defendants claimed that there is a difference between a finding that Roworth ought to have been promoted and a finding of intentional sex discrimination. That may be true, but the observation begs the issue. If Roworth ought to have been promoted and tenured and was not, then the reasons proffered by the defendants were unworthy of credence. If the reasons were unworthy of credence, then Burdine dictates that the trier of fact must find the existence of sex discrimination. 450 U.S. at 256, 101 S. Ct. at 1095.
 The court is unmoved by the defendants' portentous submission (in a belated, eleventh-hour attempt to discredit the plaintiffs' statistical evidence) of a recent decision of the United States District Court for the District of Oregon, Penk v. Oregon State Board of Higher Education, No. 80-436, slip op. (D.Ore. Feb. 15, 1985), addressing the issue of sex discrimination in a university setting. The Penk court's recognition of the flaws inherent in adapting statistical models of proof to show discrimination in an academic setting is mirrored throughout this court's opinion. See text passim. But here, unlike in the Oregon case, the ancedotal evidence substantially enhanced the probative value of the statistical evidence anent rank at hire and salary at hire; the statistical proof evinced greater indicia of reliability; and the aura of negative attitudes, so striking in the case at bar, does not permeate the pages of Penk. Furthermore, this court, with respect, does not share Judge Frye's pristine view of the defendants' expert testimony. Compare, e.g., text ante at Part IX(E) with Penk, slip op. at 374-75.
 The subclasses are as follows:
I. Rank at Hire Subclass (Subclass 1)All women faculty members initially hired at URI at the rank of assistant professor or below on or after March 24, 1972.
II. Salary at Hire Subclass (Subclass 2)All women faculty members initially hired at URI on or after March 24, 1972.
III. Annual Compensation Subclass (Subclass 3)All members of Subclass 2 who continued in service at URI for more than one year after initial hire and who received across-the-board percentage increases in compensation by reason of the application of the global provisions of one or more URI/AAUP collective bargaining agreements.
 Subclass 3 presents no special problems. It is composed exclusively of members of Subclass 2, and, once liability has been adjudicated in a given case, it would appear that the damage computation for disparate salary at hire and annual compensation can be performed in tandem. No claimant should be entitled to a double recovery or windfall; provided however, that this statement is not intended as a prevenient pronouncement of judgment vis-a-vis liquidated damages (if the same are otherwise appropriate) in any given instance. And, if entry level salary was condign, no incremental damages for annual compensation would inhere. See text ante.
In point of fact, the special master(s) may decide to treat all damage calculations in a unitary fashion. A prevailing member of Subclass 1 will in most instances be entitled, along with correction of her misranking at hire, to a corresponding retrospective adjustment of her entry level salary, cf. note 60 post, and thus to the corollary remediation of annual compensation adjustments commensurate with her stay at URI. So, in the majority of cases where infelicitous rank placement has occurred, all three concepts of classwide relief will interlock.
 This should be carried out only after all necessary rank corrections have been made. It is anticipated that, in the third stage proceedings, members of subclass 1 who have been found to have been victimized by discrimination in the form of misranking will be entitled to receive whatever adjustments in rank can be shown to be reasonable in order to palliate the effects of the initial discrimination in rank placement. On this third stage issue, the burden of proof reverts to the plaintiffs.
 No evidence was adduced that merit awards were allocated in a sex-slanted manner. The damage computation should not, therefore, include any attempt to determine whether merit awards should have been given to those who did not receive them. Merit awards should, however, be considered in the salary computations in those instances where they were in fact granted. The calculation of backpay should track the collective bargaining agreements vis-a-vis the inclusion of such awards vel non in the salary base. But, merit premiums should not be utilized as an offset. See text post.
 Inasmuch as there was no evidence that any putative recipient of an inequity adjustment premium refused to accept the hike, the court need not decide in this litigation whether the rejection of such a gratuity would terminate backpay liability to the extent of the proffered olive branch. Cf. Ford Motor Co. v. EEOC, supra.
 The court notes that the findings against Chang on her individual claims do not affect her status as an appropriate class representative. The issue of Chang's adequacy as a class representative was addressed by this court several times after class certification first took place. E.g., Chang v. U.R.I., C.A. No. 17-0070S, slip op. at 8 (D.R.I. March 2, 1983). The mere fact that the sole named plaintiff in the Chang action has failed to prove her case does not now require the court either to decertify the class or otherwise to nullify the claims of the class members generally. See East Texas Motor Freight v. Rodriquez, 431 U.S. 395, 406 n. 12, 97 S. Ct. 1891, 1898 n. 12, 52 L. Ed. 2d 453 (1977); Scott v. City of Anniston, 682 F.2d 1353, 1356-59 (11th Cir. 1982). The Court has made it abundantly clear that a class action may continue although the class representative's personal claim has become moot. Sosna v. Iowa, 419 U.S. 393, 399, 95 S. Ct. 553, 557, 42 L. Ed. 2d 532 (1975).
 If liquidated damages are eventually awarded, prejudgment interest would not accrue. See Doty, 733 F.2d at 726; Hodgson v. Miller Brewing Co., 457 F.2d 221, 229 (7th Cir.1972); cf. Gibson v. Mohawk Rubber Co., 695 F.2d 1093, 1102 & n. 9 (8th Cir.1982) (ADEA plaintiff not entitled to both liquidated damages and prejudgment interest absent exceptional circumstances); Kolb, 694 F.2d at 875 (prejudgment interest unallowable in age discrimination case if liquidated damages granted).
 Kraynek and Roworth seem to the court to stand in a different posture. They are individual actions, implicating a relatively well-settled framework of legal concepts. Prevenient review would, in these instances, needlessly promote piecemeal justice and place an unwarranted burden on the court of appeals. The cases can be concluded in short order. There is, as to this brace of suits, no sound justification for deferring the day of reckoning or for splintering the regular judicial process.
 For much the same reasons previously mentioned, see note 65 ante, the grant of permission for the filing of an intermediate fee application applies only to the class actions (Chang and Seleen), and does not reach either Kraynek or Roworth. Time productively spent on those cases, and costs incurred, should play no part in any interim impetration. Those plaintiffs may, upon the entry of final judgments in their cases, petition for appropriate fee-shifting orders in the ordinary course.
 In so ruling, the court acknowledges that the state Department of Education might perhaps be viewed in a more charitable light. There was no ostensible reason or need to sue the department. But, all of the defendants shared common counsel and the state's coffers fueled the engines of the joint defense. No operose work-load was imposed thereby, nor did the nominal presence of this defendant create extra expense. (The parties, by and large, simply ignored the fact that the department had been linked to the litigation.) Thus, no harm was done by the improvident joinder of the Department of Education.