Graffam v. Scott Paper Co., 870 F. Supp. 389 (D. Me. 1994)

*390 *391 Daniel W. Bates, James B. Haddow, Kenneth D. Keating, Petruccelli & Martin, Portland, ME, for plaintiffs.

William J. Kayatta, Jr., S. Mason Pratt, B. Simeon Goldstein, Pierce, Atwood, Scribner, Allen Smith & Lancaster, Portland, ME, for defendants.

Ray R. Pallas, Pallas & Waldron, Westbrook, ME, for Kathleen M. Woods.

GENE CARTER, Chief Judge.

This consolidated case arose from a reduction in force effective on March 13, 1991, in which a number of employees, including Plaintiffs, were permanently discharged from Scott Paper Company's S.D. Warren paper mill in Westbrook, Maine. Plaintiffs, eleven former salaried employees of the mill, allege age discrimination in violation of the Age Discrimination in Employment Act of 1967 ("ADEA"), 29 U.S.C. §§ 621-634 (1985 & Pamph.1994)^[1] (Count VI), and the Maine Human Rights Act ("MHRA"), 5 M.R.S.A. §§ 4551-4633 (1989 & Supp.1994)^[2] (Count I). The two remaining counts allege violations based on the "disparate impact" theory of *392 discrimination.^[3] The case was tried without a jury. Based on the testimony at trial and the exhibits submitted in evidence, the Court makes the following findings of fact and conclusions of law.

In October of 1990, S.D. Warren Company ("Warren") decided to reduce the number of salaried employees at the Westbrook mill. This decision was brought about, at least in part, by the company's decision to sell the mill and its desire to make the mill more attractive to potential buyers. Tr. at 341. Warren determined that it should be possible to increase its operating profit and thereby enhance its sale prospects by effecting a 20% reduction in the number of salaried employees at the mill. Christain Depo. at 34.

The selection process was devised and implemented with the help of the mill's department heads. All rated employees were placed in job groups for assessment. All Plaintiffs were at least fifty years of age as of March 13, 1991, when they were discharged from their employment at the mill as a result of the downsizing process. The downsizing process resulted in an overall rate of retention of 61% of employees age fifty and older, and an overall rate of retention of 91% of employees under age fifty. Tr. at 6.

In Wards Cove Packing Co. v. Atonio, 490 U.S. 642, 109 S. Ct. 2115, 104 L. Ed. 2d 733 (1989), the Supreme Court reformulated the law it had established almost twenty years earlier in Griggs v. Duke Power Co., 401 U.S. 424, 91 S. Ct. 849, 28 L. Ed. 2d 158 (1971), in a number of ways, including the allocation of the burden of persuasion for disparate impact claims. The Wards Cove Court declared that the burden of persuasion remains at all times with the plaintiff.^[4] Prior to Wards Cove, it was settled law that the burden of showing a business necessity shifts to the defendant once disparate impact of an employer's practice has been shown. See, e.g., Albemarle Paper Co. v. Moody, 422 U.S. 405, 425, 95 S. Ct. 2362, 2375, 45 L. Ed. 2d 280 (1975) (stating that the employer has the "burden of proving that its tests are `job related'"). The Civil Rights Act of 1991 ("CRA") legislatively overruled the burden of proof established in Wards Cove.^[5] The new law shifted the burden of demonstrating business necessity and job relatedness squarely back onto the defendant.^[6]

*393 The CRA became effective on November 21, 1991. Although these consolidated lawsuits were filed after the passage of the amendments to the CRA, the unlawful conduct alleged in this case occurred before the Act was passed. Therefore, the Court must determine whether to apply the amendments to the CRA retroactively to Plaintiffs' claims. Instead of clear direction, in the CRA Congress has provided courts with statutory language of compromise which creates an "ambiguity as to whether Congress intended the Act to be generally retroactive." Fray v. Omaha World Herald Co., 960 F.2d 1370, 1376 (8th Cir. 1992). Defendants maintain that the amendments to the CRA are inapplicable to this case because the statute's provisions are not retroactive^[7] and, in any event, the amendments do not apply to ADEA.

In general, substantive laws apply prospectively only. Procedural laws, on the other hand, apply both prospectively and retrospectively, unless there is legislative expression to the contrary. The change at issue here involves exclusively changes to the rules controlling the burden of proof which are not clearly substantive or procedural.^[8]

Traditionally, courts look to the legislative history of a statute in order to obtain guidance on the propriety of its retroactive application. In this case, the legislative history leaves the issue in a state of confusion. See 137 Cong.Rec.S. 15,483 (daily ed. Oct. 30, 1991) (statement of Senator Danforth, one of the bill's sponsors, that the Act was to apply only prospectively, while another sponsor, Senator Kennedy, expressed disagreement with that view.). Thus, the Court turns to the text of the 1991 CRA which states:

(a) In General. Except as otherwise specifically provided, this Act and the amendments made by this Act shall take effect upon enactment.

Section 402(a). One implication of this statement is that the amendments apply to all claims not yet decided at the time of enactment. However, the Supreme Court has recently observed that "there is no special reason to think that all the diverse provisions of the [CRA] must be treated uniformly" for purposes of retroactivity analysis. Landgraf v. USI Film Products, ___ U.S. ___, ___, 114 S. Ct. 1483, 1505, 128 L. Ed. 2d 229, 262 (1994) (holding section 102 of the CRA creating the power to recover damages and providing for jury trial did not apply to case pending on appeal on the date of enactment). The Landgraf court "understood the instruction that the provisions are to `take effect upon enactment' to mean that courts should *394 evaluate each provision of the Act in light of ordinary judicial principles concerning the application of new rules to pending cases and preenactment conduct." Id. Section 402(b) goes on to state:

(b) Certain Disparate Impact Cases.

Notwithstanding any other provision of this Act, nothing in this Act shall apply to any disparate impact case for which a complaint was filed before March 1, 1975, and for which an initial decision was rendered after October 30, 1983.

This provision was specifically designed to exempt Wards Cove, as it appears that virtually no other civil rights case fits the above description. If these amendments to the CRA applied only prospectively, section 402(b) of the Act, which carves out an exception for the Wards Cove case, would be unnecessary.^[9] As some district courts have agreed, the existence of subsection (b) supports the argument that the Act is retrospective.^[10]

Reading clauses (a) and (b) of section 402 together strongly indicates that the 1991 amendments to the CRA addressing the burden of proof in disparate impact cases apply to cases which were filed after, or which were pending as of, the CRA's effective date. The Supreme Court has stated on numerous occasions that "no provision [of a statute] should be construed to be entirely redundant." Kungys v. U.S., 485 U.S. 759, 778, 108 S. Ct. 1537, 1550, 99 L. Ed. 2d 839 (1988); Arcadia v. Ohio Power Co., 498 U.S. 73, 111 S. Ct. 415, 112 L. Ed. 2d 374 (1990). Along the same lines, the Supreme Court has also expressed a hesitancy "to adopt an interpretation of a congressional enactment which renders superfluous another portion of the same law." Mackey v. Lanier Collection Agency & Serv., Inc., 486 U.S. 825, 837, 108 S. Ct. 2182, 2189, 100 L. Ed. 2d 836 (1988) (citing cases). In this case, the only interpretation of Section 402 that does not make clause (b) meaningless is that the CRA is applicable to all cases which are filed after, or which are pending at the time of, its enactment, with the exception of Wards Cove.

Moreover, courts have held that where "Congress enacts [a] statute to clarify the Supreme Court's interpretation of previous legislation thereby returning the law to its previous posture," the act should be given retroactive effect. Ayers v. Allain, 893 F.2d 732, 754-55 (5th Cir.1990), reh'g granted, en banc 898 F.2d 1014 (1990). Accord Mrs. W. v. Tirozzi, 832 F.2d 748, 755 (2d Cir.1987) (holding that amendments to Education of the Handicapped Act should be applied retroactively since they codified "congressional purpose ... which Congress believes the Supreme Court had misinterpreted"); Lussier v. Dugger, 904 F.2d 661, 665-66 (11th Cir. 1990); DeGurules v. Immigration and Naturalization Service, 833 F.2d 861, 863 (9th Cir.1987). Were the Court to decide otherwise, Defendants would reap a windfall from the judicial misinterpretation of disparate impact doctrine. Therefore, this Court holds that the 1991 amendments to the CRA relating to the burden of proof requirements for disparate impact analysis apply to this case.

Defendants further argue that even if the amendments to the CRA apply retroactively, they are inapplicable to ADEA. Because of the similarity between the ADEA and Title VII of the Civil Rights Act, federal courts have historically applied the standards used for Title VII to ADEA. See, e.g., Western Air Lines, Inc. v. Criswell, 472 U.S. 400, 105 S. Ct. 2743, 86 L. Ed. 2d 321 (1985); LeBlanc v. Great American, Inc. Co., 6 F.3d 836, 843 (1st Cir.1993). Thus, although Defendants *395 strenuously argue that Congress never meant to change the burden of proof in the ADEA context, the Court is convinced that it is proper to apply the new burden of proof to this case. See Caron v. Scott Paper Co., 834 F. Supp. 33, 36 (D.Me.1993).

In order to establish a prima facie case of disparate impact, the plaintiff must: (1) identify the specific employment practices or selection criteria being challenged; (2) show disparate impact; and (3) prove causation. Watson v. Ft. Worth Bank and Trust, 487 U.S. 977, 108 S. Ct. 2777, 101 L. Ed. 2d 827 (1988). After the plaintiff has established a prima facie case of disparate impact, the burden of proof shifts to the defendant who may either discredit the plaintiff's statistics or proffer statistics of its own to show that no disparity exists. Id. at 996-97, 108 S. Ct. at 2790. The employer may also produce evidence that the practice is based on legitimate business necessity and that it is job related. Thereafter, it is up to the plaintiff to cast doubt on defendant's explanations. One way to do this is for the plaintiff to "`show that other tests or selection devices, without a similarly undesirable ... effect, would also serve the employer's legitimate interest in efficient and trustworthy workmanship.'" Id. at 998, 108 S. Ct. at 2790 (quoting Albemarle Paper, 422 U.S. at 425, 95 S.Ct. at 2375).

With respect to the first element of the prima facie case, Plaintiffs challenge the selection process as a whole. Defendants expand on the arguments previously made in their Motion for Partial Summary Judgment (Docket No. 18), explaining that the assessment process contains separable components which should not be lumped together for analysis. Specifically, Defendants point to the fact that the scoring, the draft, the corporate review, and the decision to make changes in several groups at the end of the process, were clearly distinct, separable components of the selection process. Defendants' Post-Trial Brief at 4-7. Throughout this litigation, Plaintiffs have challenged the assessment process as a whole. The Court has already ruled that Plaintiffs' challenge to the selection process should be scrutinized as one employment practice. Caron, 834 F. Supp. at 39.

Plaintiffs must identify a specific employer practice and show that the practice caused the disparate impact. Ordinarily this means that a plaintiff must point to a specific practice which is a component of the employer's decisional process, such as a test or an experience requirement, rather than to the decisional process as a whole. In Watson, however, the Supreme Court held an employer's facially neutral practice of committing employment decisions to the subjective discretion of supervisory employees was a specific employment practice properly subject to a disparate impact analysis. Upon revisiting this issue post-trial, the Court still concludes that the elements of the decisionmaking process at issue here are not capable of separation for analysis. Defendants' request that the Court limit the disparate impact model to a discrete component of the selection process would completely exempt the situation where the adverse impact is caused by the interaction of two or more components of the process. Accordingly, the Court is satisfied that in this case the entire subjective decisional process may be analyzed as one practice.^[11]

The Supreme Court has noted that the statistics which could be used to prove discrimination "come in infinite variety." International Brotherhood of Teamsters v. *396 United States, 431 U.S. 324, 340-41, 97 S. Ct. 1843, 1857-58, 52 L. Ed. 2d 396 (1977). Not surprisingly, the parties in this case have referred to different types of statistical analysis to support their positions. Plaintiffs employed a standard deviation analysis.^[12] Defendants' rebuttal consists of attacking the methodological integrity and explanatory value of Plaintiffs' statistics, as well as providing an alternative statistical analysis. The Court is satisfied that Plaintiffs' statistics show a significant disparate impact in the number of employees over age 50 discharged as a result of the reduction in force.

The Westbrook mill had 471 salaried employees prior to the reduction in force. Ex. 1. Of these, 143 employees were aged 50 and older, and 328 employees were under the age of 50. Id. Eighty-three employees were terminated due to the reduction in force. Id. Fifty-five of those terminated were 50 years old or older. Id. Twenty-eight of those terminated were under 50 years old. Id. Therefore, after the reduction in force, approximately 91% of Warren's salaried workers under age 50 retained their jobs, while only 61% of those age 50 or older were still employed. Tr. at 6.

Plaintiffs presented a standard deviation analysis done by Rosemary Roberts, a statistics professor at Bowdoin College. In order to perform her analysis, Professor Roberts first determined what result would be expected if the selection for discharge had been made by a random selection process. Tr. at 14-15. Professor Roberts concluded that since 30% of the pre-downsizing workforce was over age 50, in a random selection process one would expect, with some degree of variation, that 30% of those selected for discharge would also be over age 50. Id. Therefore, the number of workers over age 50 that one would expect a random selection process to yield for discharge is 25. Id. The actual number of employees over age 50 terminated was 55, or approximately 66% of the discharged workforce. Tr. at 6; Ex. 1 at ¶ 3(e). Professor Roberts testified that the results of the mill's selection process differ from the result expected in a random selection by a magnitude of seven standard deviations. Tr. at 18. According to Professor Roberts, the likelihood of this departure occurring in a random selection process is less than one in a billion instances. Tr. at 19.

Defendants did not present any statistical evidence tending to show the absence of discrimination millwide. Rather, Defendants argue that Plaintiffs' statistics are flawed because they do not account for the breakdown of employees into job groups. The large standard deviation, Defendants suggest, results from the configuration of job groups. That is, the job groups that were to suffer the greatest number of job cuts had higher percentages of older workers in them. Defendants' expert, Dr. James Medoff, a professor of economics at Harvard University, presented a generalized critique of the suitability of Plaintiffs' methodology for measurement of discrimination in the present case. Dr. Medoff testified that it is important to analyze the data on a job-by-job basis rather than a millwide basis. Tr. at 758. He stressed, as is discussed in section II(B) (2) infra, that the placement of employees into job groups was an initial step in the downsizing process and under the control of Warren and Scott management. Dr. Medoff criticized Plaintiffs' statistical analysis stating that, in his opinion, millwide statistics could not be used to infer a correlation between age and termination because Plaintiffs had failed to account for the impact of the job groups on the selection process. Tr. at 733-34, 750, 770.

In an effort to control for, among other things, the placement of employees into job groups prior to selection for discharge, Dr. Medoff apparently performed a regression analysis on the data from Warren's reduction in force.^[13] Tr. at 733-34. The data, how *397 that data were arrived at, and the results of Dr. Medoff's completed study were never explained to the Court. The Court is not comfortable placing any significant weight on this testimony because there is no explanation of how the witness reached his conclusions. However, Dr. Medoff did explain, at least in part, what could be concluded from his work. Taking into account the composition of the job groups, in some unexplained way, Dr. Medoff determined that millwide, to a statistically significant degree, there was a greater likelihood that a worker over age 50 would be discharged as a result of the process used than that a worker under age 50 would be discharged as a result of the same process. Tr. 738-39, 741-42, 761. Dr. Medoff again questioned the correlation between age and termination in any of the individual job groups. Tr. at 750. Explaining that the millwide statistics could be a result of the economic decisions as to how many jobs were to be cut from each job group, Tr. at 762-63, Dr. Medoff did not look into any of the individual job groups to see if his explanation was in fact correct.^[14] Tr. at 760.

Although it may be helpful to know the information about age-related disparities in the individual job groups, it is not necessary for Plaintiffs to show that a disparity existed in every job group. And while the statistical disparity for any one group would most likely not be significant because of the sample size, it remains significant that the older employees were discharged to a disparate degree millwide. The selection process implemented a millwide reduction in force. Moreover, while Defendants' explanation, if believed, would diminish the probative value of Plaintiffs' statistical evidence, the Court does not believe that it vitiates its thrust significantly. The ultimate discharge decisions were made by persons who had control over all the job group employees. Therefore, the Court is persuaded that the proper focus should be on the aggregate members of the protected group. The suggested evidence of legitimate purpose the configuration of job groups does not discredit Plaintiffs' statistics.

From his regression analysis, Dr. Medoff also purported to draw a strong correlation between promotion over the last four years and the distribution of scores given in the assessment process. Tr. at 747-48, 755. That is, he says, employees who had been promoted within the last four years scored better in the assessment process than employees who did not receive a promotion in the last four years. Dr. Medoff also explained the time-in-grade phenomenon. That is, spending time on a job can do two things: (1) it can make a worker better at doing the job or (2) it can be an indicator that a worker has been passed over for promotion to the next job. Tr. at 757-58. Dr. Medoff's investigation revealed that younger workers at the Westbrook mill were promoted more frequently. Tr. at 757.

Although Dr. Medoff looked at slices, job groups, in the corporate pyramid, Tr. at 764, the promotion rationale does not satisfactorily explain the problem with this part of the evidence. That is, it does not take into consideration the opportunities for promotion. Tr. at 757-58. Given the meritocratic promotion system at the mill, there are generally fewer opportunities for promotion of longer service workers. Tr. at 761-62. The Court does not think that, standing alone, this testimony has much probative value. Moreover, the evidence of recent promotion does nothing to cast doubt on the probative value of Plaintiffs' statistical showing of disparate impact discrimination.

Plaintiff must show that the disparity was caused by the particular employment *398 practice being challenged. Defendants argue that Plaintiffs' prima facie case is flawed because they failed to show that the statistical disparity was caused by the challenged subjective selection process. Plaintiffs respond that the statistical evidence that they have presented is sufficient to create the inference of discriminatory cause. The Court finds that the requisite causal inference has been shown in this case.

In Watson, the Supreme Court upheld the applicability of disparate impact analysis to subjective employment decisions. Later, in Wards Cove, the Court confirmed by reference "the use of `subjective decisionmaking;'" in a list of practices that it said plaintiffs would have to connect causally to a disparity in order to prevail. Wards Cove, 490 U.S. at 657, 109 S. Ct. at 2125. "[T]he plaintiff must offer statistical evidence of the kind and degree sufficient to show that the practice in question has caused the exclusion of [employees] because of their membership in a protected group." Watson, 487 U.S. at 994, 108 S. Ct. at 2788. The statistical disparities must be "sufficiently substantial that they raise such an inference of causation." Id. at 994-95, 108 S. Ct. at 2788-90. The significance or substantiality of numerical disparities is judged on a case-by-case basis. Id. at 995-96 n. 3, 108 S. Ct. at 2789-90 n. 3.

When a plaintiff's statistics indicate a disproportionate discharge rate for a protected group, there are three possible explanations for the discrepancy. D. Baldus and J. Cole, Statistical Proof of Discrimination, § 1.23 at 45 and § 9.02 at 291 (1980). First, the disparity may be the result of unlawful discriminatory bias. Second, the disparity may have a legitimate and nondiscriminatory cause. Third, the disparity may be the product of chance. When a plaintiff demonstrates a significant statistical disparity in the discharge rate, he or she has provided strong evidence that chance alone is not the cause of the discharge pattern. Here, Plaintiffs' expert has shown that the statistical results depart significantly from those that chance alone would select. What the standard deviation statistics do not determine is whether the more likely cause is a bias in the selection process or legitimate selection criteria.

Here again, Defendants contend that the nondiscriminatory reason for the disparate impact was that the job groups selected to reduce the highest number of employees contained the greatest number of older employees. The job grouping, Defendants assert, in combination with job shedding having one's job function eliminated altogether caused the disproportionate discharge rate for older employees.^[15] Defendants question the statistics, arguing that the statistics simply assumed every worker had exactly the same chance of being terminated and that this was not the case. Although Plaintiffs still carry the burden at this stage of the analysis, if Defendants demonstrate that the specific employment practice did not cause the disparate impact, Defendant will not be required to demonstrate that such practice is required by business necessity and is job related. But the Defendants have not made that showing in this case.

The Court's response to Defendants' argument is two-fold. Most important, as discussed in II(B) (1) (b) supra, Defendants' reliance on the numerical discrepancies of older employees found in job groups have no factual support in the record. Second, the alternative cause should be something in the Plaintiffs' background instead of a reason embodied in the assessment process itself. It makes no sense for Defendants to point to an alternative part of the challenged selection process as the cause since the process still results in a disparate impact for workers over age 50. It was, after all, Defendants who selected in which job group(s) Plaintiffs would be placed in the first instance. Thus, the Court finds these alleged nondiscriminatory reasons job shedding and job groupingcannot *399 be used to account for the statistical disparity.

The Court assumes that in a large-scale reduction in force, as the instant case presents, deviations in the selection rates of younger and older employees may well result. But the statistical disparity is too significant in this case to ignore.^[16] Because the statistical disparity is substantial, the Court finds that Plaintiffs have established a prima facie case without the need of providing additional evidence that bias is the more likely cause of the discharge than are legitimate nondiscriminatory explanations.^[17] Accordingly, the Court believes that the Plaintiffs' statistical proof, on its face, establishes a prima facie case of discrimination.

Section 3 of the CRA identifies the statute's purposes. These include, inter alia, restoring disparate impact doctrine to its pre-Wards Cove state and providing guidelines for the application of disparate impact analysis. Section 3 provides, in pertinent part:

The purposes of this Act are

(2) to codify the concepts of "business necessity" and "job related" enunciated by the Supreme Court in Griggs v. Duke Power Co., 401 U.S. 424, 91 S. Ct. 849, 28 L. Ed. 2d 158 (1971), and in the other Supreme Court decisions prior to Wards Cove Packing Co. v. Atonio, 490 U.S. 642, 109 S. Ct. 2115, 104 L. Ed. 2d 733 (1989);

(3) to confirm statutory authority and provide statutory guidelines for the adjudication of disparate impact suits under Title VII of the Civil Rights Act of 1964....

The restoration of the pre-Wards Cove evidentiary standard with regard to what is meant by "business necessity" and "job related" has not, in fact, fully clarified the standard because the pre-Wards Cove cases offered a variety of definitions for these terms.

Starting with Griggs, the seminal disparate impact case, the Supreme Court confusingly formulated the required elements of a defendant's response to plaintiff's prima facie case.^[18] Subsequent decisions by the Court added further confusion to the task of defining defendant's response.^[19] Then in *400 Wards Cove the Court eliminated the rigorous burden of justification altogether.^[20] Coming full circle, the 1991 amendments to the CRA, for the avowed purpose of putting back the pre-Wards Cove disparate impact standards, reinstalled the pre-Wards Cove state of confusion with one added twist: defendants are now required to show that the employment practice is both a business necessity and job related. 42 U.S.C. § 2000e-2(k) (1) (A) (i) (for text see supra note 6). In an attempt to capture the essence of these two terms, this Court has made a logical definitional distinction between them. That is, business necessity inquires whether the job criteria arise out of a manifest business need and the job related standard inquires whether there is a correlation between the criteria used and successful job performance.

Plaintiffs do not question that Defendants discharged a large number of employees due to a business necessity. Instead, Plaintiffs challenge whether the selection process was job related. Here the fact that the mill formulated a rating system in which subjective decisions would be made does not mean that the practice is immune to challenge for illicit discrimination. Disparate impact theory arises in large part from the notion that facially neutral practices might be "the functional equivalent of a pretext for discriminatory treatment." Watson, 487 U.S. at 998, 108 S. Ct. at 2791. See also Wards Cove, 490 U.S. at 660-61, 109 S. Ct. at 2126-27 (holding that a refusal to adopt equally effective alternative defeats disparate impact justification because "such a refusal would belie a claim by [the employers] that their incumbent practices are being employed for nondiscriminatory reasons."). The question, therefore, is whether the selection criteria are correlated to successful job performance and whether the overall process was a consistently effective way for the mill to achieve the necessary reduction in force. As the Court previously discussed, the burden for this element of the case is on Defendants.

Defendants' approach at this stage of the analysis is two-pronged. First, Defendants presented testimonial evidence explaining the importance of the selection criteria to the mill's managerial jobs and, second, they challenged on cross-examination Plaintiffs' accounts of work performance and proficiency. Using this strategy, Defendants have set forth legitimate, nondiscriminatory reasons for discharging each of the eleven employees. According to Defendants, each individual Plaintiff was chosen for discharge because his or her position was judged to be less essential to the company's continued operations and because other individuals in the same position, if any, were judged to be better qualified or more proficient. The Court has reviewed the testimony and the exhibits regarding the selection process to determine if the process was job related and if it effectively reached the business necessity of reducing the work force.

In order to develop and organize the selection process, the Mill Leadership Team ("MLT"), which was made up of the heads of several departments and the mill manager, met for a number of days in January of 1991. Tr. at 270-71. Each department identified job functions and job positions that were to be eliminated in the downsizing. Tr. at 274-76, 279, 293-96, 1107-08. In this manner, approximately 20% of the jobs at the mill were eliminated. Tr. at 198, 1107. The department *401 heads divided the jobs in their respective departmentsincluding those to be eliminatedinto "job groups." As a result, the entire rated work force was divided into job groups, with each job group sharing comparable skills. Tr. at 74, 305, 358, 360-61; Ex. 8 (W10004); Ex. 33. This step included placing into job groups employees whose jobs had been eliminated.^[21] Tr. at 356-57.

During the same time frame, the MLT members collectively defined the selection criteria. Tr. at 313-14, 332-33. As their starting point the MLT modified a downsizing assessment instrument that Scott Paper had previously used. Tr. at 307-314. The MLT members reviewed drafts of the selection process, sought to understand exactly what each criterion was, and discussed its applicability to the Westbrook mill. Tr. at 207-208, 312-14. The assessment tool designed by the MLT was tested before being finalized. Tr. at 207-208, 210, 219-20, 332-34, 1019-22, 1109-10. The MLT did one or two test runs consisting of a mock assessment of a person known to the team who was not in fact being assessed for discharge, to confirm its understanding and use of the instrument. Tr. 319, 1024-25, 1055-56, 1113, 1130.

The final assessment instrument allocated a total of 100 points to seven criteria. Ex. 9A (W11003). Namely, technical job skills (20 points), leading change skills (15 points), interpersonal skills (15 points), self management (10 points), performance (10 points), length of service (10 points), and versatility (20 points). Ex. 9A. Five of the criteria leading change skills, interpersonal skills, self management, performance, and versatility were defined identically for use in all job groups. Tr. at 314; Ex. 9A (W11004-W11008, W11010). The remaining criterion, technical job skills, consisted of an assessment of each worker against a list of technical skills which differed from job group to job group. Tr. at 314-15, 1023, Ex. 9 (W11012-45). The technical skills list, developed by the MLT member for the job groups in his department, contained skills deemed to be important and needed for each job. Tr. at 314-16, 1021, 1022, 1110-11. All the criteria, except length of service, were subjective in nature.

Plaintiffs contend that Defendants do not meet their burden of job relatedness because they failed to present testimony that the Warren selection process accurately predicted or was significantly correlated with the skills necessary for the performance of any job, and that no witness with personal knowledge testified that any Plaintiff's job skills were accurately measured and compared with the job skills of the people against whom he or she competed for the purpose of selection. Plaintiffs' Post Trial Brief (Docket No. 99) at 15. Plaintiffs' expert psychologist, Dr. James Mahoney, testified that Warren's selection process was not valid. Tr. at 838-39. This was, he said, because the selection criteria used were not correlated with actual performance of any job for which the selections were being made. Tr. at 835. Dr. Mahoney testified that of the three types of validity, content validity was the most important in this case. Tr. at 839-41. Content validity mirrors a job skill thought to be important to the job. Tr. at 839. It is not necessary, however, for the mill to introduce a validity study. See Watson, 487 U.S. at 998, 108 S. Ct. at 2790.

Moreover, Dr. Mahoney testified that the assessment process employed by Warren was invalid in that it did not accurately measure the presence of important job skills or correlate them with actual performance of any job for which selections were being made. Tr. at 838-39, 860-61. He stated that the mill's selection procedure was problematic because of the use of six identical rating categories for all job groups, as well as the use of identical weights for all seven categories across all job groups despite the variations from job to job within the salaried work force at the mill. Tr. at 851-853. The varying degrees to which those skills were required, Dr. Mahoney testified, meant that the selection process could never serve to accurately *402 measure whether, and in what level of concentration, employees possessed the skills necessary to perform the jobs for which they were being selected. Tr. at 376-77. He opined that these generic rating categories are contrary to the concept of job specificity that underlies the validity concept. Tr. at 852-54.

Dr. Mahoney testified that he had no direct experience with paper mills. Thus, he was unable to name any jobs at the mill in respect to which the selection criteria did not represent important aspects of job performance. Tr. at 847-48. Unlike the evaluation process here, all the tests that Dr. Mahoney was professionally involved with validating were paper and pencil tests. Tr. at 865. Dr. Mahoney never actually performed a statistical analysis to determine the actual validity of the mill's process. Tr. at 870. Dr. Mahoney testified about the manner in which he would design and test a selection instrument.^[22] Tr. at 839-846. However, Dr. Mahoney did not offer any previously validated methodologies that he would recommend Defendants should have used. Tr. at 880.

The Court is persuaded that the criteria Warren used were job related. The mill's expert, Dr. Richard S. Barrett, a consultant in the field of industrial psychology, testified that based on his personal knowledge, all the criteria used, with the exception of the specific technical skills, described job behaviors required in managerial and technical jobs in a manufacturing facility such as a mill. Tr. at 1166-68. For example, Barrett testified that interpersonal skills are important at all levels of the corporate institution and every job. Tr. at 1168.

Three department heads, Gary Parafinczuk, William Foley, and Richard Matlack, each testified that the common criteria were fair representations of important skills needed in their departments. Tr. at 335, 340, 342-43, 344, 348, 1021-22, 1110. Further, they each testified that they developed the technical skills lists for the job groups in their departments, and that those also contained important skills needed in the jobs in their departments. Tr. at 314-16, 1021, 1110-11. Parafinczuk, the Human Resources Director at the mill, testified regarding the relationship between the assessment criteria and the needs of the mill. Tr. at 335-49. He gave specific examples of how the criteria applied to the demands of the jobs, and how the mill, prior to downsizing, had devoted money and effort to teaching and encouraging the very skills that were assessed. Tr. 335-44. Dr. Mahoney acknowledged that all the criteria, except leading change skills, were important aspects of job performance in some jobs. Tr. at 883-86.

The possession of the necessary technical and managerial skills was important to all the rated jobs at the mill. Tr. at 314, 1021, 1068, 1076, 1088, 1110-11, 1168; Ex. 9A (defining all the assessment criteria). Specifically, versatility was valued at the mill because of the nature of mill operation.^[23] Tr. at 335, 1021-22, 1068, 1088-89. That is, the interrelatedness of the production process, Tr. at 1115, the need for round-the-clock shifts, Tr. at 275, and the prospect of trying to staff the mill with 20% fewer employees, all combined to provide a basis to value a worker's ability to perform several different job assignments. Interpersonal skills, Dr. Barrett testified, are always important at *403 both the peer level and supervisory level. Tr. at 1168. Leading change skills, although more amorphous in nature, was something Warren had been attempting to encourage the development of by sending employees to a total quality management course. Tr. at 338-40. Self-management was also key to any salaried position at Warren and a quality which the mill management had encouraged the development of and provided employee training for that purpose. Tr. at 524.

Generally, the qualities rated in the Warren selection process are important qualities for any employee to possess. The Court finds that the assessment criteria used, identified necessary job skills in an industrial mill. The Court concludes that, through the scoring of the selection criteria for each employee, the overall process correlates with successful job performance. This conclusion is supported by the testimony of Dr. Barrett. The method chosen by Warren to implement its reduction in force is clearly not the only way to assess salaried employees. Nor was this Dr. Mahoney's ideal selection process. However, the Court is satisfied that it was job related for these salaried mill positions.

After the assessment instrument was developed, Plaintiffs were evaluated by a team of at least three coworkers using that instrument. Each assessor was given a text explaining the procedure and criteria, Tr. at 203, 314; Ex. 9A, and a list of technical job skills applicable to the particular job group within which assessments were to be made. Tr. at 315-16, 319. Each assessment team had at least one person who had substantial first-hand knowledge regarding the skills and past performance of the employee being assessed. Tr. at 321, 436, 486-87, 625, 723-26, 786-87, 1026-28, 1114-15. The assessment team also included a member of the MLT who had participated in discussing and developing the written criteria. Tr. at 206-208, 1024, 1113; Ex. 9A (W11002). No single assessor or group of assessors rated all employees. A final rating for each employee was developed through consensus by members of the evaluation team. With a few exceptions, the employees selected for termination were those with the lowest scores in their job group.^[24]

Plaintiffs also challenge the method in which the assessment criteria were applied. Dr. Mahoney's testimony supported Plaintiffs' criticism. Dr. Mahoney testified that the written assessment criteria were poorly defined and ineffectively "cued." Tr. at 850. Along the same lines, Plaintiffs also question the adequacy of the training of the raters. The essence of the objections is that the assessors did not understand the criteria or how to apply the criteria consistently.^[25] Dr. Barrett did not share this view. The Court heard testimony from six individuals who did assessments for Warren employees, including one Plaintiff, all of whom indicated that they understood the criteria and how to apply the criteria. See Tr. at 457-58 (Whitten); Tr. at 383 (Parafinczuk); Tr. at 1134 (Matlack); Tr. at 1067 (Reed); Tr. at 1053 (Foley); Tr. at 1087 (Boulanger). Tr. at 319-22, 332-35, 340, 342-43, 344, 348, 457-60, 1020-25, 1067-69, 1087-89, 1110. All six assessors who testified at trial felt that they had received adequate training in how to assess individuals. Tr. at 316-48, 457-58, 1019-25, 1068-69, 1087-88, 1109-13.

The Court heard no evidence that raters did not know what they were doing in making the assessments. There were written instructions in each text for assessors to use. Ex. 9A. Each assessment team contained a MLT member who had spent a significant period of time reviewing and revising the criteria and procedures. The assessment format called for discussion and consensus *404 which usually uncovered any team member's misunderstanding or confusion about the criteria or process. Tr. at 319-20, 1024-25. Dr. Barrett testified that the attempt to reach a consensus was an effective way of improving the quality of the assessments. Tr. at 1170. The Court is aware that there is the potential for inherent intimidation where an assessment team is composed of an employee and that employee's supervisor. However, Plaintiffs presented no evidence that any assessor actually felt coerced into agreeing with his superior's rating or that supervisors, in any instance, tried to bring improper influence to bear. The Court finds the collaborative process represented by the consensus-driven, group assessments, wherein the assessors could not independently affect a person's score without explaining to the group his or her justification for that score, to be an effective, reasonable method to assess the work force. Tr. at 319-21, 1175-76.

Plaintiffs contend that the selection process was not consistently predictive of, or significantly correlated with important elements of work behaviors that are relevant to the jobs for which the individual candidates were being evaluated. However, the mill performed post-assessment reviews. First, the MLT met to review the list of the persons who were rated at the bottom of the job groups as a result of the assessment process. Tr. at 350-51. Using the collective knowledge of all the department heads, the MLT at the draft session looked for people who, although they failed to make the cut in their own job group, might appropriately compete in another job group. Tr. at 227, 267-68, 350-52, 1116. Some of these people were rated a second time in another job group, which resulted in either two rating sheets or a second set of ratings on the original rating sheet. Tr. at 389, 1156.

The MLT members met again in February of 1991 to review the results of the selection process. Tr. at 158-59, 272, 317, 349-51; Exs. 33, 62. After the MLT reviewed the results of the assessment process, each department head was required to present the results of all the assessments in their department to a site review committee to verify the objectivity and integrity of the assessment process. Tr. at 163, 360-61, 388; Ex. 8 (W10006). Over a two-day period, each department head was required to present the assessment results for the job groups in his or her department to a corporate review team, including management personnel from Scott Paper. Tr. at 136-37, 361, 388. The site review committee required that the decisions and scoring be explained by each department head. Tr. at 136-43. Site review member and EEO compliance officer for Scott Paper, William Bubb, asked for clarification on any finding he thought was remarkable. Tr. at 138. As a result of the post-assessment reviews, the MLT determined that several new job groups needed to be created in order to account for relatively unique skills which the mill needed to retain, and that persons having those skills needed to be reassessed for inclusion in these job groups. Tr. at 361-62.

Simply because changes to a few aspects of the selection process might enhance its effectiveness, does not mean that the overall process is ineffective altogether. Unquestionably, a great deal of subjective judgment is present in the system employed here. The Court's painstaking review of the record, particularly of the testimony of those employees of Defendants who devised and implemented the selection plan, persuades the Court that these persons acted throughout their activity with personal decisional integrity. Although the selection process was not flawless, the Court does not find the inconsistencies to be fatal to Defendants' proof.^[26]

The cross-examination of individual Plaintiffs demonstrated persuasively that there were reasons other than their age that led to their comparatively low ratings. For instance, some Plaintiffs did not learn to use the most up-to-date computer system in use at the mill, Tr. at 455-56, 585-86, 666-73; did not volunteer for overtime, Tr. at 813-14; were not able to complete as much work as others, Tr. at 483-86, 490, 1089-90, 703-706; were unwilling to, or did not choose to, attend company schools, Tr. at 525; and some had been unwilling, in the past, to apply for or do other jobs at the mill, Tr. at 454, 672, 814.

By presenting evidence that the selection process itself was well conceived and executed, and that each Plaintiff was less qualified than others occupying comparable positions, Defendants have undercut the importance and persuasiveness of Plaintiffs' statistical proof of disparate impact discrimination. The individual MLT members collectively devised criteria that were job related, and made certain that they understood the criteria and the process. The thrust of Defendants' method of attacking the statistical prima facie case cannot be blunted by reference to the statistics already presented. Plaintiffs have not shown that Defendants' explanations of the development or implementation of the selection process are inherently suspect. Nor have Plaintiffs presented other direct or circumstantial evidence suggesting that Defendants' proffered reasons for the discharge decisions are not true. Although Dr. Mahoney suggested improvements that could be made to the process, Plaintiffs have not identified an alternative method of assessment that would not have resulted in the disparate impact on older workers.

Plaintiffs also allege discrimination based on violation of the Maine Human Rights Act, 5 M.R.S.A. §§ 4551-4633. "[T]he employment discrimination provisions of [the MHRA] were intended to be the state counterparts of the Federal Act, complimenting and in certain instances supplementing the federal" statute. Maine Human Rights Commission v. United Paperworkers International Union, 383 A.2d 369, 375 (Me.1978). The Law Court went on to state that "[i]n such circumstances, decisions by federal courts interpreting the federal statutory equivalents of the Maine Act provide significant guidance in the construction of our statute." Id. See also Percy v. Allen, 449 A.2d 337, 342 (Me.1982); Maine Human Rights Commission v. Auburn, 408 A.2d 1253, 1261 (Me.1979); Wells v. Franklin Broadcasting Corp., 403 A.2d 771, 773 n. 4 (Me.1979) (construing Maine's age discrimination statute in light of the federal statute). Because the elements of proving a MHRA age discrimination claim are, with one notable exception^[27], the same as those under the federal statute, the Court finds for Defendants on Plaintiffs' age discrimination claim based on the MHRA for the same reasons stated with regard to the ADEA.

Accordingly, the Court finds for Defendants on the two remaining counts alleging disparate impact age discrimination in violation of the Maine Human Rights Act (Count I) and the Age Discrimination in Employment Act (Count VI). Judgment for the Defendants on each count shall enter forthwith.

So ORDERED.

[1] The Age Discrimination in Employment Act ("ADEA") provides in pertinent part:

It shall be unlawful for an employer

(1) to fail or refuse to hire or to discharge any individual or otherwise discriminate against any individual with respect to his compensation, terms, conditions, or privileges of employment, because of such individual's age;

(2) to limit, segregate, or classify his employees in any way which would deprive or tend to deprive any individual of employment opportunities or otherwise adversely affect his status as an employee, because of such individual's age.

29 U.S.C. § 623(a) (1) & (2).

[2] The Maine Human Rights Act ("MHRA") provides in pertinent part:

1. Unlawful employment. It is unlawful employment discrimination, in violation of this Act, except when based on a bona fide occupational qualification:

A. For any employer to fail or refuse to hire or otherwise discriminate against any applicant for employment because of ... age.... 5 M.R.S.A. § 4572(1) (footnote omitted).

[3] The Court has already decided that the "disparate impact" theory of analysis applies to an ADEA case. Caron v. Scott Paper Co., 834 F. Supp. 33, 38 (D.Me.1993).

[4] Under Wards Cove, after the plaintiff has made out a prima facie case, the only burden that shifts to the defendant is the burden of production, not persuasion. The burden of production means only that unless the party produces some evidence that a fact exists, the judge must find that the fact does not exist. F. James & G. Hazard, Civil Procedure § 7.9 at p. 244 (2d ed. 1977). Once the defendant satisfies the requirement of articulating a nondiscriminatory reason for the disparate impact, the ultimate burden remains with the plaintiff to prove to the trier of fact that the defendant discriminated against plaintiff. A plaintiff can meet his or her ultimate burden of persuasion with respect to a fact if the plaintiff has produced enough evidence to lead the trier of fact "to believe that the existence of [a fact] is more probable than its non-existence." Id.

[5] In section 2 of the CRA, Congress set forth its findings, among which are the following:

(2) the decision of the Supreme Court in Wards Cove Packing Co. v. Atonio, 490 U.S. 642, 109 S. Ct. 2115, 104 L. Ed. 2d 733 (1989) has weakened the scope and effectiveness of Federal civil rights protections; and

(3) legislation is necessary to provide additional protections against unlawful discrimination in employment.

Pub.L.No. 102-166, § 2, 105 stat. 1071, 1071 (1991).

[6] Section 703(k) (1) provides in part:

(A) An unlawful employment practice based on disparate impact is established under this title only if

(i) a complaining party demonstrates that a respondent uses a particular employment practice that causes a disparate impact on the basis of race, color, religion, sex, or national origin and the respondent fails to demonstrate that the challenged practice is job related for the position in question and consistent with business necessity; or (ii) the complaining party makes the demonstration described in subparagraph (C) with respect to an alternative employment practice and the respondent refuses to adopt such alternative employment practice.

42 U.S.C. § 2000e-2(k) (1).

[7] Defendants rely on Letourneau v. Casa Mia, Inc., 804 F. Supp. 389 (D.Me.1992), to support their argument that the 1991 amendments to the CRA are inapplicable to this case. Letourneau is distinguishable from the instant case. Namely, the Court in Letourneau held that the compensatory and punitive damage provisions added by the CRA of 1991 were not available for conduct occurring before the effective date of November 21, 1991. See also LaPlante v. United Parcel Service, Inc., 810 F. Supp. 19 (D.Me.1993). In this case, however, the Court is faced with deciding the retroactive application of another provision of the CRA. This requires the Court to analyze anew the distinct statutory provisions here in question.

[8] The dichotomy between substantive and procedural law arises in a number of contexts. For example, in a conflicts case, the First Circuit held that

It is apparent ... that burden of proof does not fall within either category of "substance" or "procedure" by virtue of any intrinsic compulsion, but the matter has been made to turn upon the purpose at hand to be served by the classification.

Sampson v. Channell, 110 F.2d 754, 756 (1st Cir.1940). The Sampson court went on to find the particular rule before it to be substantive. At least one court has found the burden of proof to be procedural. See United Securities Corp. v. Bruton, 213 A.2d 892, 893-94 (D.C.1965) (upholding the application of the U.C.C.'s burden of proof standard in a trial involving a pre-Code transaction, and stressing that the matter involved was not substantive but procedural).

In a law review article discussing this dichotomy, the author argues that rules controlling the burden of proof fall within a third category which he labels "substantively-based procedural rules." D. Risinger, "Substance" and "Procedure" Revisited, 30 UCLA L.Rev. 189 (1982). Other commentators think that the whole distinction is meaningless. Alies, Substance and Procedure in the Conflict of Laws, 39 Mich.L.Rev. 392 (1941) ("The distinction between substantive and procedural law is an illusion.").

[9] The only other provision discussing retroactivity is section 109(c), which deals with extraterritorial employment and provides, "The amendments made by this section shall not apply with respect to conduct occurring before the date of enactment of this Act." It would seem that this section also would be unnecessary if the act were prospective only.

[10] Stender v. Lucky Stores, 780 F. Supp. 1302, 1304 (N.D.Cal.1992) ("only interpretation of section 402 that does not make clause (b) meaningless is that the Act is applicable to cases which were pending at the time of its enactment, with the exception of Wards Cove"); Graham v. Bodine Elec. Co., 782 F. Supp. 74, 76 (N.D.Ill.1992); Andrade v. Crawford & Co., 786 F. Supp. 1302, 1304 (N.D.Ohio 1992); Robinson v. Davis Memorial Goodwill Indus., 790 F. Supp. 325, 327-28 (D.D.C.1992). But see Mozee v. American Commercial Marine Serv. Co., 963 F.2d 929, 932-33 (7th Cir.1992).

[11] Although not directly applicable, the Court finds support for this position in the 1991 amendments to the CRA. Section 105 provides in part:

With respect to demonstrating that a particular employment practice causes a disparate impact as described in subparagraph (A) (i), the complaining party shall demonstrate that each particular challenged employment practice causes a disparate impact, except that if the complaining party can demonstrate to the court that the elements of a respondent's decisionmaking process are not capable of separation for analysis, the decisionmaking process may be analyzed as one employment practice.

42 U.S.C. 2000e-2(k) (1) (B) (i).

[12] The Supreme Court has recognized that for disparate impact mode of proof, standard deviation analysis is an appropriate and meaningful method for establishing a prima facie case of discrimination. Castaneda v. Partida, 430 U.S. 482, 496-97 n. 17, 97 S. Ct. 1272, 1281-82 n. 17, 51 L. Ed. 2d 498 (1977); Hazelwood School District v. United States, 433 U.S. 299, 308-09 n. 14, 97 S. Ct. 2736, 2742-43 n. 14, 53 L. Ed. 2d 768 (1977).

[13] Prof. Roberts agreed that where a group is subdivided before selections are made from it, a regression analysis would determine more specifically than standard deviation analysis whether the actual result of a selection process deviates significantly from the expected result. Tr. 30-31.

[14] Presumably, the inference Defendants want the Court to make is that the jobs to be cut were in the job groups with the highest numbers of older workers. Defendants presented no definitive evidence to support their theory that the disparate impact millwide resulted because job groups with the higher numbers of older workers sustained the majority of the job loss. In fact, the Court has been frustrated in its attempt to reconstruct this explanation or find any definitive support for this theory in the record. The Court has been unable to determine even the total number of job groups. The number seems to vary from 40-odd to 60-odd.

[15] The job grouping and job shedding were part of the universe of the assessment process. Because this is a reduction-in-force case, there is no dispute about the reason why Plaintiffs are no longer working at the mill. Unlike a case where the reason for dismissal or failure to hire is in question, it is stipulated here that all Plaintiffs lost their jobs as a result of the corporate downsizing. In this way, reduction-in-force cases are unique in that the inference of cause results, in part, from the decisional process itself.

[16] The Supreme Court has found that where the result obtained by application of the challenged selection process differs from the result expected in a random selection process by at least two or three standard deviations, a presumption of discrimination is justified. Castaneda, 430 U.S. at 496-97 n. 17, 97 S. Ct. at 1281-82 n. 17; Hazelwood School District, 433 U.S. at 308-09 n. 14, 97 S. Ct. at 2742-43 n. 14. In this case, the statistics show a variation of seven standard deviations.

[17] The Court is satisfied this is correct because it may be assumed, absent some reason to believe the opposite is true, that job skills, as well as other criteria in the assessment process, will be equally distributed across all age groups. In a work-force-reduction case, unlike a hiring case, we assume that all of the employees are qualified to perform their jobs since they were actually performing them at the time of their discharge.

[18] In Griggs, the Court identified the "touchstone [as] business necessity," Griggs, 401 U.S. at 431, 91 S. Ct. at 853, while in the very next sentence stating, "If an employment practice which operates to exclude [blacks] cannot be shown to be related to job performance, the practice is prohibited." Id. In many subsequent decisions, courts have applied this language. Washington v. Davis, 426 U.S. 229, 96 S. Ct. 2040, 48 L. Ed. 2d 597 (1976); Albemarle Paper, 422 U.S. 405, 95 S. Ct. 2362; Boston Chapter, NAACP, Inc. v. Beecher, 504 F.2d 1017 (1st Cir.1974). The Griggs Court also used the term "job related." Griggs, 401 U.S. at 436, 91 S. Ct. at 856. Ultimately, the Griggs Court concluded that the challenged job requirement at issue did not bear "a demonstrable relationship to successful performance of the jobs for which ... [it was applied]." Id. at 431, 91 S. Ct. at 853.

[19] In Dothard v. Rawlinson, 433 U.S. 321, 97 S. Ct. 2720, 53 L. Ed. 2d 786 (1977), the Supreme Court formulated a more onerous burden for defendant stating that "a discriminatory employment practice [challenged on disparate impact grounds] must be shown to be necessary to the safe and efficient job performance to survive a Title VII challenge." Id. at 332 n. 14, 97 S. Ct. at 2728 n. 14. A few years later, the Court preferred a lesser burden of business justification in New York City Transit Authority v. Beazer, 440 U.S. 568, 99 S. Ct. 1355, 59 L. Ed. 2d 587 (1979). The Court observed that the district court had noted that the Authority's use of safety and efficiency were "significantly served ... [by the] rule as it applies to all methadone users including those who are seeking employment in nonsafety-sensitive positions." Id. at 587 n. 31, 99 S. Ct. at 1366 n. 31. The Court went on to conclude that the record demonstrated that the "rule bears a `manifest relationship to the employment in question.'" Id. (quoting Griggs, 401 U.S. at 432, 91 S.Ct. at 854). With this decision, the Court seemed to equate the necessary or "manifest relationship" requirement with a standard that required that the policy or practice having a disparate impact "significantly serve" the employer's goals.

[20] In Wards Cove, the Court employed Beazer-type language as the standard: "[T]he dispositive issue is whether a challenged practice serves, in a significant way, the legitimate employment goals of the employer.... The touchstone of this inquiry is a reasoned review of the employer's justification for his use of the challenged practice." Wards Cove, 490 U.S. at 659, 109 S. Ct. at 2126. The Court went on to caution that "a mere insubstantial justification ... will not suffice, because such a low standard of review would permit discrimination to be practiced through the use of spurious, seemingly neutral employment practices." Id. However, the Court warned that there is "no requirement that the challenged practice be `essential' or `indispensable' to the employer's business for it to pass muster." Id.

[21] The elimination of one's job function directly affected one's chances of being retained since the employee had to then compete in a foreign job group. Tr. at 1064. Two of the eleven Plaintiffs, Neal Cobisi and Norman Wedge, held jobs that were eliminated. Tr. at 1016-17.

[22] In general, Dr. Mahoney was highly critical of Defendants' selection process. He testified that an employer needs to identify the kind of observable behavior for each job. Tr. at 843. Most important, to ensure reliability of the process, Dr. Mahoney thought raters should have been required to undergo two to four days of training in order to make accurate assessments, Tr. at 843-44, 859, and as part of the training run a "pilot." Tr. at 844. The raters, therefore, received less training than recommended by Dr. Mahoney. Tr. at 858-59, 1138-39, 428, 431-50, 1087-88. The Court finds that the record supports Warren having incorporated each of these points into its selection process and, therefore, understands Dr. Mahoney's criticisms to be aimed at the sufficiency of Warren's efforts.

[23] Plaintiffs correctly point out that the versatility criterion required the raters to assess the future potential and performance of employees. However, under most circumstanceswhether during the initial hiring, promotion, or the reduction in forceit will be necessary for employers to project, to a certain extent, what an employee's performance will be in the future. The problem with the versatility criterion is that it was not shown that employees had the same opportunity to acquire multiple job skills or job functions.

[24] The Court has found five employees under age 50 who had lower assessment scores and were retained over five employees over age 50. It was explained that the exceptions arose in the review process when it was discovered that either an individual's job skill was absolutely necessary and could not be found elsewhere at the mill, Christain Depo. at 109, or that an individual was initially placed in the wrong job group. Tr. at 82, 124, 387.

[25] For example, Dr. Mahoney testified that essential terms used in the instrument were so ambiguous and ill-defined that, even if the rating categories were applied narrowly to a set of job skills, raters could not understand what they were supposed to do. Tr. at 858-59.

[26] The flaws in the process have been the subject of much intense analysis by the Court to determine what effect in larger evidentiary terms should properly be found to result. The Court's concerns regarding the process are as follows. MLT members had three distinct roles in the process: team leader, coach, and possibly the final decision maker if no consensus could be achieved. The Court has questioned whether the MLT member's tripartite role could be completed without conflict of interest to the other roles. The selection process was not consistent for all employee assessments. For example: some employees' scores were changed after reconsideration; some employees were assessed more than once; some employees were drafted, after elimination, to other job groups; and job groups were created after the assessments were complete. Although it was not the uniform for all employees, the Court is persuaded as a matter of fact that this selection process was sufficiently consistent to be job related.

[27] This distinction in the federal and the Maine age discrimination statutes is that the Maine statute does not limit its protection to a particular age group. All persons are covered by the MHRA, while only individuals over age 40 are given federal protection. Compare 5 M.R.S.A. § 4574(2) & (3) with 29 U.S.C. § 631(a). See also Graffam v. Scott Paper Co., 848 F. Supp. 1 (D.Me. 1994) (discussing the propriety of age sub-grouping).