Following a nine-day bench trial, the district court entered judgment for defendants Scott Paper Company and S.D. Warren Company (collectively "Warren") in this age discrimination suit brought by eleven former employees ("plaintiffs"). The plaintiffs alleged that the selection procedures ("selection procedures") used by Warren to reduce by twenty percent the number of salaried employees at its paper mill in Westbrook, Maine, effected an illegal disparate impact on employees over age fifty. In awarding judgment to Warren, the district court found that, though the selection procedures did indeed have a disparate impact on older-age employees, Warren had made a sufficient showing that the procedures were job related and consistent with business necessity. Graffam v. Scott Paper Co., 870 F. Supp. 389, 399- 404 (D. Me. 1994). After a careful review of the record, we affirm.

In late 1990, Warren concluded that it must reduce by twenty percent its salaried work force of approximately 471 employees at its mill in Westbrook, Maine. The conclusion was dictated, at least in part, by Warren's decision to sell the mill and a corresponding requirement that the mill be made more attractive to potential buyers. Consequently, Warren set about creating selection procedures for identifying which employees it would need to discharge in order to meet the desired force reduction goal.

In January 1991, a Mill Leadership Team ("MLT"), consisting of the heads of several departments and the manager of the Westbrook mill, met for a number of days to develop a plan for achieving Warren's reduction goal. As a result, each department identified job functions and job positions that could be eliminated. Department heads divided the jobs in each department-including those to be eliminated-into specific job groups and placed salaried employees subject to the downsizing into the newly-formed groups. Positions with similar functions and responsibilities were grouped together so that employees with comparable skills ultimately would be rated against each other.

At the same time, MLT members collectively defined the selection procedures. As their starting point, the MLT modified procedures previously developed by Warren's corporate offices for use in an employee downsizing. MLT members reviewed drafts of these selection procedures, sought to understand and clarify the assessment criteria, and discussed the applicability of the criteria to the Westbrook mill. After the revisions were included, MLT members tested the new procedures by conducting mock assessments of persons known to the MLT members but who were not subject to discharge.

The final assessment plan allocated a total of 100 points to seven criteria: technical job skills (twenty points); performance (ten points); length of service (ten points); leading-change skills (fifteen points); interpersonal skills (fifteen points); self management (ten points); and versatility (twenty points). The skills included in the technical job skill criterion varied for each individual job group. An individual who received zero technical skill points would not be retained in favor of another employee with a higher technical skill rating regardless of which employee had the higher respective total assessment score. Every employee was awarded the maximum ten performance points, provided that the employee had not received counselling for unacceptable job performance. The common criteria of length of service and leading-change, interpersonal, self-management, and versatility skills ("common criteria") were applied identically to all employees across all job groups.

Once the MLT finished ironing out the revised assessment criteria, teams of at least three individuals were formed to evaluate the salaried employees subject to downsizing. Each team included an MLT member who had participated in developing and discussing the written criteria, and at least one person who possessed substantial first-hand knowledge regarding the skills and past performance of each employee assessed. Every assessor was provided a text explaining the assessment criteria and a list of technical job skills applicable to the particular job group he or she would assess. The teams reached a final consensus rating for each employee through oral discussion and group decision. Following the team assessments, MLT members met to review the process. Additionally, each department head was required to present, explain, and justify the results of the selection process in his or her department to a corporate review team that included management personnel from the mill's corporate headquarters.

The downsizing process resulted in the termination of thirty-nine percent of the salaried employees age fifty years and older but only nine percent of those employees under age fifty. All of the plaintiffs in this case were over age fifty when Warren discharged them as a result of the downsizing effort. Following their discharge, the plaintiffs commenced this action in federal district court, alleging that Warren had discriminated against them on account of their age in violation of the Age Discrimination in Employment Act of 1967 ("ADEA"), 29 U.S.C. Secs. 621-34, and Maine state law. At trial, plaintiffs, pursuing their claim under a theory of disparate impact liability, maintained that the selection procedures, though neutral on their face, effected a substantial disparate impact on older-age employees. Following a nine-day bench trial, the district court entered judgment for Warren, from which the plaintiffs now appeal.

We review the district court's findings of fact only for clear error. See Fed. R. Civ. P. 52(a); Cumpiano v. Banco Santander P.R., 902 F.2d 148, 152 (1st Cir. 1990). Clear error exists when, after reviewing the entire record, we have a "strong, unyielding belief that a mistake has been made." Cumpiano, 902 F.2d at 152; see also Industrial Gen. Corp. v. Sequoia Pac. Sys. Corp., 44 F.3d 40, 43 (1st Cir. 1995). If, however, the district court premised its factual findings on an incorrect view of the law, we are not bound by the clearly erroneous standard. E.g., Brown Daltas & Assocs. v. General Accident Ins. Co., 48 F.3d 30, 36 (1st Cir. 1995). "[T]o the extent that findings of fact can be shown to have been predicated upon, or induced by, errors of law, they will be accorded diminished respect on appeal." Dedham Water Co. v. Cumberland Farms Dairy, Inc., 972 F.2d 453, 457 (1st Cir. 1992).

The plaintiffs contend that Warren's selection procedures, though neutral on their face, resulted in the discharge of a disproportionate number of older-age employees. The theory of disparate impact liability has its roots in the Supreme Court's decision in Griggs v. Duke Power Co., 401 U.S. 424 (1971). In Griggs, the Court held that the use of a facially neutral objective test that tended disproportionately to exclude African-Americans from the employment pool but did not measure skills demonstrably related to job performance violated Title VII's anti-discrimination provisions. In the context of Title VII, Congress codified the disparate impact theory in the Civil Rights Act of 1991. See Pub. L. No. 102- 166, Sec. 3, 105 Stat. 1071 (1991) (listing as one of its purposes "to confirm statutory authority and provide statutory guidelines for the adjudication of disparate impact suits under title VII").

Congress, however, has never explicitly addressed the theory of disparate impact liability in the ADEA context. Moreover, though the Supreme Court has acknowledged that "[t]here are important similarities between [Title VII and the ADEA], ... both in their aims-the elimination of discrimination from the workplace-and in their substantive provisions[,]" Lorillard v. Pons, 434 U.S. 575, 584 (1978), it has nonetheless "never decided whether the disparate impact theory of liability is available under the ADEA." Hazen Paper Co. v. Biggins, 113 S. Ct. 1701, 1706 (1993). Similarly, though we assumed without analysis the applicability of the theory in Holt v. Gamewell Corp., 797 F.2d 36, 37 (1st Cir. 1986), we have never directly addressed the issue. See also Caron v. Scott Paper Co., 834 F. Supp. 33, 35-38 (D. Me. 1993). Again, for purposes of this opinion, we assume arguendo that the district court correctly held that the ADEA supports a claim for age discrimination based on a disparate impact theory of liability.1

As applied in Title VII cases, to prove a claim of disparate impact discrimination, a plaintiff must identify a facially neutral employment practice or policy that causes a statistically discernible disparate impact on a protected employee group. See EEOC v. Steamship Clerks Union, Local 1066, 48 F.3d 594, 601 (1st Cir. 1995). Once the plaintiff has made this initial showing, the defendant must then attempt to debunk the sufficiency of the plaintiff's evidence or, in the alternative, show that the challenged practice is either job related and consistent with business necessity or that it fits within a specific statutory exception.2 Id. at 604. If the employer attempts to justify its actions, the plaintiff may seek to cast doubt on the justification by showing, inter alia, an alternate practice exists that equally protects the employer's putative interest but does not disproportionately burden employees in the protected class. Id.

The plaintiffs contend that the district court found that the selection procedures merely identified qualities important for employees to possess in general. They argue that such a finding is equivalent to measuring "the person in the abstract" and not the "person for the job," Griggs, 401 U.S. at 436, and, therefore, is inadequate to establish an affirmative defense to a claim of disparate impact discrimination. The plaintiffs argue further that the district court erred by failing to inquire whether Warren had established that the selection procedures identified specific job behaviors that significantly correlated to successful performance of the specific jobs in question. They maintain that, if the correct standard were to be applied, the evidence is insufficient to support a finding in favor of Warren. We disagree.

Our review of the district court's careful and extensive opinion satisfies us that it did supportably find that Warren's selection procedures measured skills and job behaviors necessary for, and significantly correlated with, successful performance of the jobs in question. The court found that the common criteria used in the selection procedures identified necessary technical and managerial skills "important to all the rated jobs at the mill." To support this finding, the court pointed to Warren's expert, Dr. Richard S. Barrett, a consultant in the field of industrial psychology, who testified at some length that the common criteria described job behaviors required in managerial and technical jobs in a manufacturing facility such as the Westbrook mill. In addition, the court noted that three department heads at the mill, who had participated in designing and implementing the selection procedures, testified that the common criteria fairly represented important skills needed to perform the jobs in their departments. The court also relied on the testimony of Warren's Human Resource Director, Gary Parafinczuk, who explained, inter alia, that, prior to the downsizing, Warren had devoted money and resources to teaching and encouraging the very skills assessed by the common criteria.

The district court also noted that the testimony of the various department heads established that the technical skills portion of the selection procedures (which differed for each job group) identified important skills needed to perform the specific jobs in each job group. Finally, in discrediting the testimony of the plaintiff's expert, Dr. James Mahoney, in favor of Dr. Barrett, the district court expressly rejected the plaintiff's contention that the selection procedures were not consistently predictive of, or significantly correlated with, the necessary skills for the successful performance of the specific jobs in question.

Notwithstanding the district court's careful analysis, the plaintiffs contend that, because Warren applied the same common criteria in equal weights to every job group assessed, the selection procedures could not possibly measure job skills significantly correlated with successful performance of any specific job. We are not convinced: The fact that the skills identified by the common criteria were important in many managerial and technical mill jobs simply does not compel a finding that the skills were unrelated to the particular jobs at issue. Furthermore, the plaintiffs totally disregard the fact that the technical skills and performance criteria directly measured each employee's specific job performance skills and that, although Warren applied the common criteria to all employees, the employees were rated and ranked only within their specific job group by persons familiar with individual employees and group needs. Moreover, we seriously doubt that Warren would have previously committed money and resources to developing the skills measured by the common criteria if they were not directly related to successful job performance. In addition, the employees' own expert admitted on cross- examination that one way to assess the validity of the selection procedures as a tool for measuring skills directly related to job performance would be to examine whether a statistical correlation existed between earlier promotions at the mill and the assessment scores. Subsequently, Dr. James Medoff, an expert in labor statistics retained by Warren, testified that, when he reviewed the data supplied by Warren, he found a strong correlation between those two factors.

In sum, we believe that, in determining whether Warren had sufficiently justified its use of the selection procedures, the district court supportably found that the procedures measured skills necessary for, and correlated with, successful performance of the jobs in question.3

We also believe that the record adequately supports the district court's finding that Warren's implementation of the selection procedures assured that they would fairly identify the employees who most fully possessed the skills and abilities needed for successful job performance. In making this finding the court relied in part on Dr. Barrett's assessment of Warren's efforts. Barrett testified that the team format for rating each employee, calling for open discussion and justification of each assessment and consensus decision-making, enhanced the quality of the rating system. The court also pointed out that all raters who testified stated that they understood the criteria and how to apply them. Furthermore, the court noted that the post-rating review sessions, at which each department head was required to justify his or her decisions, further assured the quality and fairness of the assessments. We have reviewed the record and find that it adequately and convincingly supports the district court's conclusions. Accordingly, the district court's findings are not clearly erroneous.4

For the foregoing reasons, the judgment of the district court is affirmed.