United States v. City of Montgomery, Ala., 775 F. Supp. 1450 (M.D. Ala. 1991)

*1451 Marybeth Martin, Philip Eure, Employment Litigation Section, Civ. Rights Div., Dept. of Justice, Washington, D.C., for plaintiff.

Donald Watkins, Montgomery, Ala., for plaintiff-intervenor Williams.

Thomas M. Goggans, Montgomery, Ala., for white male police officers of the City of Montgomery, and Ledbetter, defendants-intervenors.

Kenneth L. Thomas, Massey, Means & Thomas, Montgomery, Ala., for plaintiffs-intervenors Williams, et al.

Vanzetta McPherson, Montgomery, Ala., for Timmi Alford re challenge to appointment.

J. Richard Cohen and M. Wayne Sabel, Argo, Enslen, Holloway & Sabel, Montgomery, Ala., for Pierce & Oyler.

Robert C. Black and Randall C. Morgan, Hill, Hill, Carter, Franco, Cole & Black, Montgomery, Ala., for all defendants except *1452 Wade L. Moss and Montgomery City-County Personnel Bd.

Robert D. Segall, Copeland, Franco, Screws & Gill, Montgomery, Ala., for Wade L. Moss and Montgomery City-County Personnel Bd.

George B. Azar, Azar & Azar, Montgomery, Ala., for defendants, Folmar, Wilson, McGilvray, Eckerman, Cooper & Duffee.

Julian McPhillips and Kenneth Shinbaum, McPhillips, DeBardelaben & Hawthorne, Montgomery, Ala., for Willie Davis re challenge to appointment.

Gary E. Atchison, Montgomery, Ala., for plaintiff-intervenor Boyd.

Theron Stokes, Montgomery, Ala., for Atty. Sabel re attorney's fees.

MYRON H. THOMPSON, Chief Judge.

Title VII of the Civil Rights Act of 1964, as amended, 42 U.S.C.A. §§ 2000e through 2000e-17, proscribes certain employment practices that have a "disparate impact" on employees that is, that have a discriminatory effect but for which there is no evidence of the employer's subjective intent to discriminate. Griggs v. Duke Power Co., 401 U.S. 424, 431, 91 S. Ct. 849, 853, 28 L. Ed. 2d 158 (1971). In these two cases challenging sex and race discrimination in the Police Department of the City of Montgomery, Alabama, the issue before the court is whether new permanent procedures developed by the police department for promotions to the rank of sergeant violate this proscription and thus may not be used. As explained below, the court holds, first, that certain parts of the new permanent procedures survive disparate impact analysis and thus may be used by the department for promotions to sergeant and, second, that the remaining parts of the procedures are not yet ripe for consideration by the court.

The relevant procedural events leading up to the current issue in these two consolidated cases are long but not complicated. In 1979 in United States v. City of Montgomery, civil action no. 3739-N, in response to a complaint-in-intervention brought by Sidney Williams, an African-American police officer, the court found that a sergeant's examination used by the Montgomery City Police Department had an impermissible disparate racial impact in violation of Title VII. 19 Empl.Prac.Dec. (CCH) ¶ 9239 (M.D.Ala.) (Johnson, J.). The court enjoined the police department from future use of that test, or any other test or procedure for promotion to sergeant which results in a disparate racial impact, unless and until the court found that the test or procedure met the requirements of Title VII. Id. at 7419. The department did not attempt to develop a new test but instead used subjective promotion procedures, which according to periodic reports the department made to the court did not have adverse racial impact.^[1] Later, however, in 1986 in Jordan v. Wilson, civil action no. 75-19-N, in response to complaint-in-intervention brought by Sandra Pierce-Hanna, a female officer, challenging the subjective promotion procedures, the court held that the police department had violated Title VII by systematically discriminating against female officers in promotions. 649 F. Supp. 1038 (M.D.Ala.) (Thompson, J.). The court ordered the department to develop both interim and permanent promotion procedures that would not violate Title VII. Id. at 1062. The department employed testing experts to develop the procedures. In 1987, as part of the relief in both United States v. City of Montgomery and Jordan v. Wilson, the court approved and ordered implemented an interim promotion plan for *1453 the police department. 667 F. Supp. 772 (M.D.Ala.1987) (Thompson, J.). The interim plan required, among other things, that promotions in the department not have adverse impact on either black or female officers. Id. at 777.^[2]

Both Officers Sidney Williams and Sandra Pierce-Hanna, along with the United States and the defendants (the City of Montgomery and its mayor and police chief), have continued to participate in this litigation. Williams currently represents a class of all black officers in the department and Pierce-Hanna currently represents a class of all female officers in the department. The court has also allowed Gordon M. Ledbetter and John D. Shumway to intervene in these two cases on behalf of a class of all white male officers in the department.

In late 1989, the police department submitted to the court a proposed new permanent plan for promotions to the rank of sergeant. In developing the procedures and in anticipation of a Title VII challenge to the "validity," or job-relatedness, of the procedures, the experts retained by the department followed a "content validity" strategy, one of three test validation strategies recognized by the Uniform Guidelines on Employee Selection Procedures, 29 C.F.R. § 1607 (1978) (hereinafter referred to as the "Uniform Guidelines"), and by the more recent Principles for the Validation and Use of Personnel Selection Procedures (3d ed. 1987) issued by the Society for Industrial and Organizational Psychology. The goal of the content-validity approach is to develop selection procedures "representative of important aspects of performance on the job." Uniform Guidelines, 29 C.F.R. § 1607.5(B). In other words "the test itself [must] closely approximate the tasks to be performed on the job. The classic example of a test having content validity is a typing test for a typist job position." B. Schlei & P. Grossman, Employment Discrimination Law 114 (2nd ed. 1983).^[3]

The department's experts used a two-step process to develop the new procedures under this approach: first, there was an analysis of the job of sergeant, and, second, there was the development of the promotion procedures themselves. The first-step job analysis defined the content of the sergeant's job by identifying both the work performed in the form of "work behaviors" and the "attributes" of the workers performing that work, that is, their knowledges, skills and abilities.^[4] In order to be included in the final definition of the job content, both work behaviors and worker attributes had to meet certain tests or "screens" to assure that they were important to the job and necessary at entry.^[5] The latter screens were designed to guard against including worker attributes which could be learned in a brief orientation period. The Uniform Guidelines provide that

In general, users should avoid making employment decisions on the basis of measures of knowledges, skills, or abilities which are normally learned in a brief *1454 orientation period, and which have an adverse impact.

29 C.F.R. § 1607.5(F). The worker attributes were also rated by incumbent police officers called "subject-matter experts" on the extent to which possession of each worker attribute would distinguish between superior and adequate overall job performance.^[6] All of the worker attributes included in the final definition of the job content met a minimum standard.^[7]

The department's experts then developed promotion procedures using both a written scored test and a structured oral interview to assess or measure the worker attributes that is, a worker's knowledges, skills and abilities identified by the job analysis. Both devices were used because, according to the department's experts,

While some [worker attributes] could be appropriately sampled by only one measurement strategy (either written examination or structured oral interview), an effort was made to distribute [worker attributes] across both strategies to enhance content validity and reduce potential adverse impact. Past experience in other law enforcement settings suggested that African American candidates would perform better on the [oral interview] than on the written test vis-a-vis white candidates. Whenever the sub-domain within a given [worker attribute] could be measured by the [oral interview], this method was preferred.^[8]

Weights "were computed for each measurement strategy to reflect the relative difference in the job content domain being sampled by each strategy."^[9] But, as a result of the efforts of the department's experts to minimize potential adverse impact by relying on the oral interview, the written test constituted only 40% of a candidate's final score.^[10]

After selecting appropriate measurement strategies, the department's experts developed questions for the written test as well as "critical incident and problem scenarios" for the oral interview. The police officers who served as subject-matter experts were selected and represented by race and sex and played an important role in both processes. First, they were used to insure that the test questions were job-related and free from bias.^[11] If a subject-matter expert considered a question biased, the question was eliminated.^[12] Second, they rated potential test questions on their power to distinguish between adequate and superior levels of performance.^[13] Only those questions that passed this hurdle were included on the test.^[14]

As a further check against the use of unfair questions, a linguist was retained to review test questions to insure that none was written in a way that would be unfair to black test-takers.^[15] Finally, as an additional effort to create a level playing field, candidates were given a number of aids, including the following: a description of the examination, a list of study materials, and an opportunity to participate in tutorials to sharpen their test-taking skills conducted by a specialist in preparing minorities for written employment examinations.^[16]

After the written examination was administered, a sophisticated item analysis was conducted to reduce the number of test questions to be used in computing final scores.^[17] "Generally, items within a given content area having greater race and/or gender differences in difficulty were targeted for elimination. At the same time, the reliability of the test as a whole was an *1455 important consideration, as well as subtest reliability."^[18] Because of the small number of female test-takers, greater weight was placed on racial differences in identifying items to eliminate.^[19]

The oral interview was constructed through a similarly exacting process.^[20] Each interview panel consisted "of three interviewers, at least one of whom was an African-American and at least one of whom was female."^[21] The scores from the two components of the selection process were then combined, with the written test counting 40% and the oral interview 60%.^[22] The department also accepted its experts' recommendation that, rather than rank the candidates in strict numerical order, they should be grouped within five "bands" to guard against over interpreting small differences and to reflect that the selection procedures that they had designed like all selection procedures did not measure everything.^[23] Candidates within each band are considered equally qualified. Under this banding approach, each band is to be exhausted in turn, beginning with the band containing the highest scores. The total number of candidates within each band, from the highest scoring to the lowest scoring band, is as follows:

Band     Candidate Total   Black   White   Male   Female
A                8           0       8       7       1
B               25           3      22      24       1
C               44           9      35      41       3
D               39          12      27      37       2
E               14           6       8      14       0
Totals         130          30     100     123       7

In late 1990, while the parties were still preparing for a hearing on the sergeant's procedures, the police department asked to promote 20 officers immediately under the plan. The 20 officers to be promoted were as follows: the eight candidates in Band A (seven white males and one white female) and 12 of the 25 candidates in Band B (eight white males, one white female, and three black males). These 20 officers therefore consisted of 15 white males, two white females, and three black males. By order entered in January 1991, the court allowed the promotions to go forward over an objection filed by the class of black police officers. The court based its holding on the fact that the department desperately needed the promotions and the fact that the court believed that it retained the ability to provide the black officers with full relief should it later determine that the promotions violated Title VII. The court understood that the department intended to make 20 additional promotions; if the court rejected the procedures, the defendants could be required to include a sufficient number of blacks in the 20 additional officers to avoid adverse racial impact for the entire 40 promotions.

In July 1991, the court held a hearing on the proposed permanent procedures for promotion to sergeant to determine whether they would result in impermissible disparate impact in violation of Title VII. Counsel for all parties the United States, the black officers, the female officers, the white male officers, and the defendants participated.

Title VII prohibits two types of discrimination: "disparate treatment" and "disparate *1456 impact." With the former, an employee must prove intentional discrimination. Texas Department of Community Affairs v. Burdine, 450 U.S. 248, 253, 101 S. Ct. 1089, 1093, 67 L. Ed. 2d 207 (1981). With the latter, however, the employee challenges "practices that are fair in form, but discriminatory in operation," Griggs v. Duke Power Co., 401 U.S. 424, 431, 91 S. Ct. 849, 853, 28 L. Ed. 2d 158 (1971); the employee need not prove intentional discrimination. Id. at 430, 91 S. Ct. at 853. Here, the question for the court is whether the Montgomery Police Department's proposed permanent promotion procedures for the rank of sergeant would result in an impermissible disparate impact.

In assessing whether there is impermissible disparate impact, a court should engage in a three-step process. First, the employee must identify the specific employment practice challenged and, further, must show that the challenged practice falls significantly more harshly on one group than another, that is, that the practice under attack has created "adverse impact." Wards Cove Packing Co., Inc. v. Atonio, 490 U.S. 642, 656, 109 S. Ct. 2115, 2124, 104 L. Ed. 2d 733 (1989).^[24] If this showing is made, the burden then shifts to the employer to produce evidence of job relatedness for the employment practice. The employer's burden is one of production not persuasion, for the burden of persuasion always remains with the employee. Finally, if the employer satisfies its burden, the employee may prevail only if he or she shows that the employer's justification for the practice has no basis in fact or that another practice, without a similarly undesirable adverse effect, would also serve the employer's legitimate employment interests. Id. at 660, 109 S. Ct. at 2126. "Of course, any alternative practices which [an employee may] offer up in this respect must be equally effective as [the employer's] chosen ... procedures in achieving [the employer's] legitimate employment goals." Id. at 661, 109 S. Ct. at 2127. Furthermore, because "Courts are generally less competent than employers to restructure business practices," id. (quoting Furnco Construction Corp. v. Waters, 438 U.S. 567, 578, 98 S. Ct. 2943, 2950, 57 L. Ed. 2d 957 (1978)), "the judiciary should proceed with care before mandating that an employer must adopt a plaintiff's alternative selection ... practice in response to a Title VII suit." Id.^[25]

All parties agree that the Montgomery Police Department's proposed sergeant's procedures have a significant adverse racial impact on African-American candidates, and no one questions the department's evidence of an employment justification for the procedures. Still, the parties disagree on three points. First, the black police officers claim that the court should not approve the proposed promotion procedures because there is an alternative selection procedure available that would serve the department's interest in an effective police force equally as well but with significantly less adverse racial impact. Second, the white male police officers contend that promotions on the basis of the procedures should be made in rank order rather than from bands. Third and finally, both the United States and the white male officers argue that the department should be required to develop procedures for making selections from within bands. The court will consider each of these contentions in turn.

The black police officers claim that the written test should be used as a "pass/ *1457 fail" device with the oral interview scores used to order candidates. The officers argue that this alternative procedure would have less adverse racial impact than the police department's procedures and yet would serve the department's interest in an effective police force equally as well. To support this argument, the officers maintain that, because most of the candidates who did fairly well on the oral interview also did fairly well on the written test, their alternative approach would not yield people decidedly less able to meet the demands of the sergeant's position.^[26] The officers further maintain that the defendants' experts always intended to use the written test solely as a pass/fail procedure. Indeed, according to the black officers' expert, this approach "carries the strong endorsement" of the Uniform Guidelines.

Although the black officers' position has a certain surface appeal, the court is not persuaded that their alternative would be equally effective in serving the police department's legitimate interest. First, the manner in which the job analysis was carried out and the test questions prepared reflects that the written test was intended to be used to distinguish among levels of job performance. The Uniform Guidelines recognize that, in such a situation, ranking may be appropriate. They provide that

If the user can show, by a job analysis or otherwise, that a higher score on a content valid selection procedure is likely to result in better job performance, the results may be used to rank persons who score above minimum levels.

29 C.F.R. § 1607.14(C) (9).^[27] Second, although most of the candidates who did fairly well on the interview also scored 90 or better on the written test, the fact remains that some scored below 90. Similarly and more importantly, scores from 90 to 100 on the written test actually cover a broad range in statistical terms, four standard errors of measurement, and, in lay terms, a range sufficiently broad to allow for meaningful distinctions among the candidates who scored in the range. To use the test simply on a pass/fail basis would treat the scores within the range as if they were the same and would ignore the distinctions among the candidates that the differences in the scores reflect. The approach would, as a result, undermine the reliability and validity of the proposed promotion procedures and thereby sacrifice, rather than further, the legitimate interests of the police department.

Anticipating this conclusion, the black officers argue that any shortcomings in those promoted could be overcome after a brief period of service at the sergeants' rank. But accepting this argument would place the court in the position of second-guessing, without any basis, the judgment of incumbent police officers. As explained earlier, the proposed promotions procedures measured only those worker attributes that is, knowledges, skills, and abilities that police officers rated as necessary at entry to the sergeant's rank and which could not be learned in a brief orientation period. In addition, adopting this argument would place the court in the position of assuming, without evidence, that differences among candidates in their performance on the test would not translate into differences in their performance on the job once they were promoted. With the public safety of the people of the City of Montgomery at stake, the court cannot make such a critical but unsubstantiated assumption.

Before leaving the topic of the black officers' proposed alternative selection procedure, it is necessary for the court to address one final point. The black officers have noted that the black test-takers' correlation between the written test and the oral interview appears to be different from the white test takers' correlation 0.23 versus *1458 0.01.^[28] The officers point to this difference as the primary source of the adverse impact resulting from the department's selection procedures and as a justification for adopting their proposed alternative.^[29]

The court's response to this observation is simple. The black officers appear to be confused by the requirement that they identify the "specific or particular employment practice that has created the disparate impact under attack." See Wards Cove, 490 U.S. at 656-57, 109 S. Ct. at 2124-25. Admittedly, "Such a showing is an integral part of the plaintiff's prima facie case in a disparate-impact suit under Title VII." Id. However, there is no question in this case but that the department's proposed promotion procedures are the source of the adverse impact against black candidates. Given this point, it is not necessary for the black officers to offer a more elaborate statistical description of the manner in which the adverse impact occurs. Even if the black officers had not called the court's attention to the apparent difference in correlations, it would have been incumbent on the court to consider their proposed alternative. In any event, the difference to which the black officers refer is more apparent than real. The evidence reflects that, given the small number of black test-takers, the difference between the two correlations, when compared directly, is not statistically significant. Admittedly, the black officers' expert claimed that the two correlations should not be compared directly. According to their expert, both correlations should first be compared to zero. When this procedure is followed, the correlation for white test-takers, 0.23, is significant, but that for black test-takers, 0.01, is not. The expert then concludes that, because one correlation is statistically significant and one is not, the two are different. But this approach cannot withstand closer analysis. Even if the correlation for black test-takers had been 0.23, the same as that for white test-takers, it would still not have been statistically significant when compared to zero given the small number of black candidates with the illogical result that the two correlations would have been different even if they had been identical. Moreover, the efforts of the black officers' expert to salvage his approach with the argument that statistical tests do not apply because the figures were not drawn from a sample, does not wash. Commentators have recognized that "statistical tests can be a useful aid ... in assessing the plausibility of chance patterns" in such circumstances. D. Baldus & J. Cole, Statistical Proof of Discrimination § 9.32 at 316 (1980) (discussing "whole universe" problem). Indeed, the black officers' expert himself performed statistical tests when comparing the correlations to zero.^[30]

The white male officers contend that the scores from the department's proposed procedures should be used in rank order rather than in bands. According to the officers, rank ordering would maximize organizational efficiency in the long run even if the difference between two scores, say 92 and 93, were not statistically significant. The officers contend that, because the female officer and two of the three black officers promoted from Band B had numerical scores that were lower than those of some of the white males in Band B who were not promoted, a ruling that a rank-ordered use of the procedures was required might entitle the white male officers to relief.^[31] Actually, the white male officers make their argument in only a half-hearted fashion. Their expert candidly acknowledged that he is "not opposed to banding," that the procedure is "an attractive device in terms of administration," and that the theoretical increase in efficiency from rank ordering is premised on the repeated, long-term use of the same procedures, a premise that is unlikely to hold true here. Moreover, other courts have endorsed a banding approach. See, e.g., Bridgeport Guardians, Inc. v. City of Bridgeport, 735 F. Supp. 1126, 1136-37 (D.Conn.1990) (Daly, J.), aff'd, 933 F.2d 1140 (2nd Cir. 1991). It appears as if the white male officers' real objection concerns the absence of a formalized procedure for making selections within bands, and it is to this objection that the court now turns.

The white male officers and the United States complain that the police department has no formalized procedure for making selections from within bands. They contend that the problem will be especially critical if the department reaches Band C, the band that contains 44 candidates. The lack of formalized procedures, according to them, invites arbitrariness. The police department responds that it has asked its experts to develop a training program to help guide the department in making selections from within bands. The training will be designed to help the department avoid common rating biases and focus its attention on those abilities that are important for sergeants but were not capable of being measured by the written test and the oral interview. Furthermore, because these additional guidelines developed for the department will necessarily be subjective ones that would be difficult if not impossible to validate, the department has agreed to avoid adverse impact for selections from within bands. Cf. Local 28 of the Sheet Metal Workers' International Association v. EEOC, 478 U.S. 421, 450-51, 106 S. Ct. 3019, 3036-37, 92 L. Ed. 2d 344 (1986) (numerical goals acceptable when promotions are necessary and nondiscriminatory procedures are not available).

The court is of the view that the issue of within-band selections is not ripe for resolution. Because the police department has yet to develop these additional guidelines, neither the parties nor the court can evaluate them. In addition, the department may never make promotions from Band C, the band with 44 remaining candidates. Although it appeared earlier, at the time the court approved the initial promotion of 20 candidates, that the department would promote 20 more persons during the two-year life of the rankings under the proposed promotion procedures, it now appears from the evidence presented at the hearing on the proposed procedures that unanticipated budgetary problems may make it impossible to do so. Given that only white males remain in Band B, the only band from which further promotions are likely, it is unlikely that an issue about promotions from within the bands will arise in the context of this employment discrimination lawsuit. Accordingly, the defendants in this litigation the City of Montgomery, *1460 Alabama and its mayor and police chief will be allowed to promote from Band B, but not Band C without further court approval.^[32]

In a year, this litigation will be 20-years old, and if it continues at the same pace it will be well into the next century before it comes to an end. Indeed, the court today could give only partial approval to the sergeant's promotion procedures and has yet even to consider promotion procedures for the other ranks. It is therefore evident that the parties and the court must investigate and consider other means for resolution of this litigation. Otherwise, these two cases will give a literal meaning to the figurative comment this court made a number of years ago, in another lawsuit, that "Unlike old soldiers, cases such as the one now before the Court not only never die, they never fade away." United States v. Frazier, 14 Empl.Prac.Dec. (CCH) ¶ 7599 at 4929 (M.D.Ala.1976) (Johnson, J.).

In accordance with the memorandum opinion entered this date, it is the ORDER, JUDGMENT, and DECREE of the court:

(1) That the objections to the sergeant's promotion procedures for the Police Department of the City of Montgomery, made by the United States, the Williams intervenors on behalf of all black police officers, and the Ledbetter intervenors on behalf of all white male police officers, be and they are hereby overruled;

(2) That it be and is hereby DECLARED that the sergeant's promotion procedures for the Police Department of the City of Montgomery are approved and may be used by the defendants that is, the City of Montgomery and its mayor and police chief to the extent that they seek to make promotions from ranks A and B; and

(3) That the defendants, their officers, agents, servants and employees and those persons in active concert or participation with them who receive actual notice of this order, be and they are each hereby ENJOINED and RESTRAINED from making any promotions from ranks other than ranks A and B without first obtaining approval from the court.

The clerk of the court is DIRECTED to issue a writ of injunction.

[1] For a more detailed history of United States v. City of Montgomery and Williams's participation in it, see Sims v. Montgomery County Commission, 686 F. Supp. 878 (M.D.Ala.1988) (Thompson, J.).

[2] For further discussion of the interim plan, see United States v. City of Montgomery, 731 F. Supp. 436 (M.D.Ala.1989) (Thompson, J.) (promotion of black officer to rank of lieutenant under interim plan); United States v. City of Montgomery, 744 F. Supp. 1089 (M.D.Ala.1990) (Thompson, J.) (promotion of female officer to rank of captain under interim plan).

[3] The other methods of validation are "criterion-related" validation, which is established when "there is a significant positive correlation between comparative success on the test and comparative success on some measure of job performance," B. Schlei & P. Grossman, Employment Discrimination Law 114 (2nd ed. 1983); and "construct" validation, which is established when "there is a significant relationship between the test and the identification of some trait, such as `intelligence' or `leadership,' which is required in the performance of the job." Id.

[4] Auburn University at Montgomery Job Analysis for the Ranks of Sergeant, Lieutenant, Captain, and Major in the Montgomery Police Department (1989) (hereinafter referred to as "Job Analysis Report").

[5] Auburn University at Montgomery Content Validation Report Police Sergeant Promotional Procedures for the Montgomery, Alabama Police Department at 5 (1991) (hereinafter referred to as "Content Validity Report").

[6] Job Analysis Report at 22.

[7] Content Validation Report at 8.

[8] Id. at 12.

[9] Id.

[10] Id. at 12-13, 55.

[11] Id. at 23, 27-29.

[12] Id. at 28-29.

[13] Id. at 27-28.

[14] Id.

[15] Id. at 24.

[16] Id. at 32.

[17] Id. at 34 & App.K.

[18] Id. at 34.

[19] Id., App.K (last page).

[20] Id. at 38-54.

[21] Id. at 49.

[22] Id. at 55.

[23] Id.

[24] For a discussion of the various methods of establishing significant disparate impact, see Richardson v. Lamar County Bd. of Educ., 729 F. Supp. 806, 816 (M.D.Ala.1989) (Thompson, J.), aff'd, 935 F.2d 1240 (11th Cir.1991).

[25] The Uniform Guidelines similarly provide:

Consideration of suitable alternative selection procedures. Where two or more selection procedures are available which serve the user's legitimate interest in efficient and trust-worthy workmanship, and which are substantially equally valid for a given purpose, the user should use the procedure which has been demonstrated to have the lesser adverse impact.

29 C.F.R. § 1607.3(B).

[26] The black police officers' trial brief at 14.

[27] See also Equal Employment Opportunity Commission: Adoption of Questions and Answers to Clarify and Provide a Common Interpretation of the Uniform Guidelines on Employee Selection Procedures, Q. & A. 62, 44 Fed.Reg. 11,996, 12,005 (1979) (hereinafter referred to as the "EEOC Questions and Answers").

[28] According to the black officers' expert, the correlation for black test-takers was 0.10, and, according to one of the police department's experts, it was 0.01. Because the use of the 0.01 figure would actually make the black officers' argument stronger by making the difference between the correlations appear to be greater, the court has adopted the department's figure.

[29] The black officers' expert originally referred to the difference in correlations as an example of "differential validity." At the start of the July 1991 hearing, however, he acknowledged that his use of the term had been too loose. Differential validity refers to a situation that may occur only in the context of a criterion-related validity study, not a content validity model. See EEOC Questions and Answers, Q. & A. 71. The phenomenon of a difference in correlations for two groups on two different selection devices developed following a content validity model is not a situation that is covered in the Uniform Guidelines or in other relevant professional standards in the field of industrial/organizational psychology.

[30] In choosing an example to illustrate the difference in correlations, the black officers' expert made an error that resulted in exaggerating the apparent difference. According to their expert, there were four blacks among the top 30 scorers on the written test, and there were five blacks among the top 30 on the oral interview; yet there were only two blacks among the top 30 on the combined results. The expert maintains that the fact that the number of black high scorers is lower on the combined figures than on either component of the process is anomalous. In fact, however, the actual figures are 2.4, five, and two. The fraction results from the fact that there were many tie scores on the written test. Moreover, had the black officers' expert chosen to focus on the top 35 candidates instead of the top 30, the analogous figures would have been 3.5, six, and four. The expert admitted that figures such as 2.4, five, and two or 3.5, six, and four would not have been troublesome.

[31] Content Validity Report, App. T (last page)

[32] Although it argues that the issue is not ripe for resolution, the class of female police officers has stated that it opposes any method of selection from within bands that will leave the police department with any discretion in light of the department's history of retaliation. See United States v. City of Montgomery, 755 F. Supp. 1522 (M.D.Ala.1990) (Thompson, J.) (promotion of Alford to the rank of captain), aff'd, 934 F.2d 1265 (11th Cir.1991) (table); United States v. City of Montgomery, 744 F. Supp. 1074 (M.D.Ala. 1989) (Thompson, J.) (selection of deputy chief), aff'd, 911 F.2d 741 (11th Cir.1990) (table); Jordan v. Wilson, 667 F. Supp. 772 (M.D.Ala.1987) (Thompson, J.) (development of interim plan to address retaliation); Jordan v. Wilson, 649 F. Supp. 1038 (M.D.Ala.1986) (Thompson, J.) (promotion of Pierce-Hanna).