the ITF, stated that executive committee members were given no instructions regarding how they should review the Justification Forms prior to signing. Rosen at 151-52. Executive committee member Ronald Kuntzman stated that he "signed with the assumption that [the line managers and higher level managers] were doing their job properly." Kuntzman at 133. In addition, no executive committee member made any changes to or disagreed with any of the Justification Forms presented to them. Rosen at 152. Therefore, it appears that neither the higher level supervisor nor the executive committee member who signed the Justification Form reviewed the reasons why persons were being terminated.
After the executive committee member signed the Justification Forms, the forms were sent to the implementation coordinators. The implementation coordinators' function was to (1) compile lists of who was being terminated, who was being retained, and who was being transferred, and (2) ensure that the Justification Forms, and the other documentation, were complete and bore the requisite signatures. See Rosen at 153. There is a conflict in the record as to whether the implementation coordinators were to review the reasons given for each termination. Rocco Ricciardi, who was an implementation coordinator, stated that he did not believe that it was his job to review the reasons given on the forms for termination, and therefore, he did not closely review the reasons given. Ricciardi at 52. Rosen, however, testified that the implementation coordinators were responsible for ensuring that the terminations reflected on the Justification Forms were consistent with the Guidelines. Rosen at 153. Rosen further stated, however, that no implementation coordinator made any changes in the employees selected for termination. Id. Thus, it appears that the review function performed by the implementation coordinators, if any, was limited.
After the implementation coordinators completed their review, the forms were sent to Swift and Higgins for review. Swift merely reviewed the Justification Forms to ensure that the forms were completed and that the "justification read, . . . , reasonably correctly." Swift at 60. "It was not an intensive review." Id. Higgins checked the justification forms to determine whether the general language on the form was consistent with the Guidelines. Higgins "94 at 99.
After the completion of the Justification Forms, Roche conducted statistical, demographic analysis based on the recommended terminations. This analysis was completed by January 28, 1985. See Miller Aff. Ex. 9 at B010722; Higgins '89 at 58- 60. Roche has claimed the attorney-client privilege with respect to the results of this analysis. Higgins '94 at 110-12. In any event, prior to the completion of the study, Roche had determined that regardless of the outcome, no managers would be asked to revise their selections for termination based in the statistical analysis. Rosen at 186-87, Ex. 6. Such revisions would be requested only if the supporting documentation indicated some "shortcoming." Rosen Ex. 6. After the completion of the study, no decisions were changed. See Rosen at 186-87. Lastly, Roche did not have any contingency plans to delay the RIF if the study revealed any problems. Higgins '94 at 112.
As previously stated, on February 4, 1985 Roche terminated over 1,100 employees at its Nutley, N.J. and Belvidere, N.J. facilities. Of the over 700 Roche employees over the age of 40 who were terminated in Operation Turnabout, 476 have joined this lawsuit as plaintiffs.
2. Expert Reports
In support of their claim that, when terminating these employees, Roche engaged in a pattern or practice of discrimination, plaintiffs submitted, in addition to the evidence just described, two expert reports. First, plaintiffs submitted the report of Dr. Farrell Bloch in which he discusses various statistical analyses that he performed generally addressing the question of whether a disproportionate number of employees over 40 years of age were terminated. Second, plaintiffs submitted the report of Dr. Joel Lefkowitz, an industrial and organizational psychologist, who was asked to opine on the cause of the disparities allegedly demonstrated in Dr. Bloch's report, i.e., whether the age disparities were caused by age discrimination.
i) Report of Dr. Farrell Bloch
The statistical analyses presented in Dr. Bloch's report address the following question: "If underlying termination rates were the same for older and younger workers, then what is the likelihood of observing the actual pattern of termination by age?" See The Relationship between Termination Rates and Age in Operation Turnabout at Hoffmann-La Roche, Inc. at 1 (attached as Exhibit 1 to Miller Affidavit) [hereafter "Bloch Report"].
In answering this question, Dr. Bloch performed a number of different statistical analyses, most of which are reflected in four tables located at the end of the report. Table 1 compares "the termination rates for young and old workers at Hoffmann-La Roche as of the close of business on February 3, 1985 (the day before the reduction in force)." Bloch Report at 2. As reflected in the table, 18.4% of the 4,301 employees at least 40 years of age were terminated, while only 9.3% of the 3,480 employees under the age 40 were terminated. The difference between the 18.4% termination rate and the 9.3% termination rate is the equivalent of 11.45 standard deviations. Standard deviation is "a measure of predicted fluctuations from the expected value of a sample." Hazelwood School District v. United States, 433 U.S. 299, 309 n.14, 53 L. Ed. 2d 768, 97 S. Ct. 2736 (1977). In other words, standard deviation may be used to measure the likelihood that the number of persons, at least 40 years old, actually terminated in Operation Turnabout is the result of something other than age discrimination. The Supreme Court has stated that "as a general rule for such large samples, if the difference between the expected value and the observed value is greater than two or three standard deviations," then the hypothesis that discrimination was not present is suspect. Id. (quoting Castaneda v. Partida, 430 U.S. 482, 496-97 n.17, 51 L. Ed. 2d 498, 97 S. Ct. 1272 (1977)). Table 1 also compares termination rates based on other age dividing lines -- employees at least 45, 50, 55, 65 years old. With these dividing lines, the standard deviation of the disparity in termination rates for younger and older workers range from 7.01 (where 65 years of age is the dividing line) to 14.60 (where 55 years of age is the dividing line).
In Table 2, Dr. Bloch compares termination rates for employees under 40 years of age and employees at least 40 years of age by division. In Divisions 1, 12 and 14, the disparities in termination rates are the equivalent of 1.0, 0.0, and 1.6 standard deviations. In addition, the disparity in termination rates for Division 8 is equivalent to 2.2 standard deviations. Therefore, for these four divisions, the disparity in termination rates is not particularly, statistically significant. The differences in termination rates in divisions 3, 5, 6, and 7 are more significant. In these divisions, the disparities in termination rates range from 4.3 standard deviations (division 7) to 8.1 standard deviations (division 5).
Table 3 compares the "increased termination probabilities for employees at least forty years old (with other factors held constant)." Bloch Report at 9. By holding constant factors that may influence termination decisions, Dr Bloch sought to isolate the role age played in termination decisions. Dr. Bloch identified four factors to be held constant -- "comporatio", "percent of range", length of service, and division. The comporatio and percent of range relate to an employee's position in their salary range.
Dr. Bloch first held the factors constant individually, and then in various combinations. This resulted in disparities in termination rates ranging from 4.47 standard deviations (both comporatio and length of service held constant) to 11.38 standard deviations (division held constant). See Bloch Report at Table 3.
Dr. Bloch also compared termination rates for employees at least 40 years old while holding merit ratings constant. Arthur Bledsoe, Director of Compensation and Benefits at Roche at the time of Operation Turnabout, stated that "the Merit system compared employees to all others with similar duties and responsibilities." Bledsoe at 49. Employees would then be rated in one of four categories: exceeds expectations, meets all expectations, meets most expectations, and does not meet expectations. Rosen at 113; Miller Aff. Ex. 13. Compensation was then based on these ratings. Bledsoe at 49. Dr. Bloch used the merit ratings as a proxy for performance. Although Dr. Bloch did a number of different tests using the merit ratings, his conclusions with respect to all of these tests can be summarized as follows: "Although the results show that termination rates are higher for those with poorer performance, termination rates are still higher for older workers, holding performance constant." Bloch Report at 5.
Finally, as discussed above, Roche claims that the employees terminated in Operation Turnabout were terminated on the basis of how they were ranked against the other employees in their work group. Roche provided plaintiffs with pools of people who they claim were compared to each other in Operation Turnabout. The pools comprise 1697 of the 7781 Roche employees at Roche employee prior to Operation Turnabout and 385 of the over 1,100 terminated employees. Although the results are not reflected in a table, Dr. Bloch performed statistical analyses on these pools. See Bloch Report at 5-6. After taking into account the pools, the disparity in likelihood of being terminated between those older and younger than 40 years is the equivalent of 5.20 standard deviations. When last merit ratings are held constant, the resulting disparity is equivalent to 3.90 standard deviations for pools 1-115, 2.22 standard deviation for pools 117-143, and 4.62 standard deviations for the combined group.
When average merit ratings over the period 1982-84 are held constant, the respective standard deviations are 4.79, 2.44, and 5.48 standard deviations. See Bloch Report at 5-6.
Based on all of the analyses discussed above, Dr. Bloch concluded that "after analyzing Operation Turnabout in many different ways, holding constant various factors, and using different age dividing lines, I have consistently found that older employees are significantly more likely than younger employees to have been terminated." Bloch Report at 6.
ii) Report of Dr. Joel Lefkowitz
As stated above, plaintiffs also submitted the report of Dr. Joel Lefkowitz, an industrial and organizational psychologist, who was asked to opine on the cause of the disparities allegedly demonstrated in Dr. Bloch's report, i.e., whether the age disparities were caused by age discrimination. The "starting point" of Dr. Lefkowitz's report was the "finding" by Dr. Bloch that Operation Turnabout had a "disparate impact" on older employees. See Critique of the Implementation of the Hoffmann-La Roche, Inc. Cost Reduction Plan "Operation Turnabout" at 2 (attached as Exhibit 4 to Miller Affidavit) [hereinafter "Lefkowitz Report"]. Dr. Lefkowitz "construed [his] task to be an evaluation of whatever evidence may be brought to bear concerning the likely explanation(s) or justification or that disparate impact." Id.
Dr. Lefkowitz considered four possible explanations for the disparate impact that Operation Turnabout allegedly had on older workers. First, he considered whether the process by which Roche employees were selected for termination was "valid." Lefkowitz Report at 3. To determine whether this process was valid, Dr. Lefkowitz compared it to a number of "the professionally- and legally-accepted strategies for validating personnel procedures." Id. He concluded that "the employee evaluation ranking system employed by Hoffmann-La Roche as a basis for making the termination decisions fails to meet professional and legal standards of job-relatedness for employee performance appraisals." Id. at 47. According to Dr. Lefkowitz, the system used by Roche did not meet these standards of "job-relatedness" because it was "ad hoc in nature." See Lefkowitz Report at 7-10. It was "ad hoc in nature" because, among other things, "it was not developed in a systematic documented fashion so that its content would reflect job-related dimensions of work performance," id. at 8, "none of the factors [listed in the Guidelines], including the more abstract and vague ones (versatility, flexibility, work habits) are defined in terms of the work behaviors that presumably reflect those attributes--for each applicable job," id., "no instructions are provided concerning the relative 'weighting' of those factors nor how they are to be combined into a composite overall evaluation by which each individual would be ranked," id. at 9, "no training or formal procedures were provided many managers and supervisors who would make the ratings and rankings," id., "no empirical investigation was made of the reliability or consistency of the rankings (i.e. if repeated, would the same rankings result?)," id. at 10, and "no appeals process was provided." Id. Dr. Lefkowitz opined that "[a] professionally developed and legally acceptable performance appraisal system would not have had all of those omissions." Id. Therefore, he concluded that "the disparate impact of the HLR terminations on older employees can not appropriately be 'justified' on the basis of the job-relatedness of the appraisal process." Id. at 23.
Second, Dr. Lefkowitz considered whether age and job performance at Roche were inversely correlated, thereby justifying a positive correlation between age and termination in Operation Turnabout. Dr. Lefkowitz first noted that the Bloch Report found, using the Merit Ratings as a proxy for performance, that although termination rates are higher for those with poorer merit ratings, termination rates are still higher for older workers, holding performance constant. See Lefkowitz Report at 24. Then he discussed "a brief review of the behavioral science literature concerning the relationship between age and job performance." Id. at 24. After reviewing this literature, he concluded that "age is not generally related to performance across a wide variety of jobs," see id. at 32, and therefore, there is no reason to believe that age and job performance at Roche were inversely correlated, thereby, justifying a positive correlation between age and termination in Operation Turnabout.
Third, Dr. Lefkowitz considered whether "the disparate impact reflects, to some degree, unplanned (perhaps unconscious) discrimination in the form of age-stereotyping or age-bias." Id. at 4. In considering this issue, Dr. Lefkowitz reviewed several studies that resulted in findings that differential treatment of older and wronger workers was the result of "unconscious age stereotyping rather than conscious discrimination." See Lefkowitz at 40-41. Based on these studies, he concluded that
(i) age-stereotyping is widespread--perhaps especially so in the business world, (ii) it can be manifested in personnel actions such as terminations, (iii) it is more likely to occur when employees are compared directly with one another as was done by HLR, (iv) it is more likely to occur when not a lot of employee performance data or job description information is provided, as was true for many of the HLR rankings; and (v) researchers recommend that strong assertive steps be taken by management in order to prevent its occurrence--which HLR appears not to have done.
Lefkowitz Report at 43. Therefore, according to Dr. Lefkowitz, "in the absence of support for either of the first two possibilities, the likelihood that the differential impact of the HLR RIF is due to the expression of age bias among those making the termination decisions must be seen as great." Id. at 49.
Lastly, Dr. Lefkowitz noted that the disparate impact that Operation Turnabout had on older employees may be explained by intentional discrimination. However, with respect to the existence of intentional discrimination, Dr. Lefkowitz concluded that he was "not in a position to pass definitive judgment on this possibility " without knowing more about "the structure, functioning, history and culture" of Roche. Lefkowitz at 44 (emphasis added). Dr. Lefkowitz did, however, review deposition testimony of seven Roche employees and concluded that "this testimony is consistent with the possibility of conscious, intended age discrimination." Id. at 49.
Dr. Lefkowitz's conclusions can be summarized as (1) Roche's Guidelines did not meet professional standards for making performance appraisals, (2) according to behavioral science research, performance does not decline with age, (3) age-stereotyping, i.e., unconscious age discrimination, is prevalent in our society and is more likely to be acted on in employment settings when, as in the case of Operation Turnabout, the decisionmakers are not guided by job-related standards of performance, and (4) there is evidence in this case that, while not conclusive, is nonetheless consistent with intentional age discrimination.
3. Anecdotal Evidence of Intentional Age Discrimination15
In support of their claim that Roche engaged in a pattern or practice of discrimination, plaintiffs also rely on evidence in the record regarding, in their words, "age discriminatory attitudes" by Roche executive committee members and other high level managers. See Plaintiffs' Supplementation Brief at 30. This evidence is comprised of deposition testimony that several Roche managers, at various points in time prior to Operation Turnabout, made age-related comments and testimony that Roche compiled various reports on which employees' ages were listed.
First, plaintiffs point to evidence in the record which, according to plaintiffs, demonstrates that "President Lerner himself expressed concerns regarding the company's average age." Plaintiffs' Supplementation Brief at 30. Peter DePaolo, who was manager of analytical services at the time of Operation Turnabout, testified that he was told by Dr. Ann Goetz that, at a managers' meeting, in about 1983 or 1984, attended by Dr. Goetz, Lerner stated that the average age of Roche was 44 years old and that was too old. See DePaolo at 49-52.
George Reilly, the Director of Finance and Administration for Roche at the time of Operation Turnabout, testified that in approximately early 1981, after either an executive committee meeting or a planning meeting, Anthony Maris, an executive committee member, went to Reilly's office to discuss a topic that was brought up at the meeting. In the course of the discussion, "Maris indicated that he had just heard a comment from Mr. Lerner stating that the organization is loaded with gray-haired burnouts." Reilly at 15-16. Reilly also testified in his deposition that "[he] always felt that Lerner had a history of terminating senior employees." Id. at 134. Although at his deposition Reilly was unable to specify any senior employees who Lerner terminated, in response to a subsequent interrogatory Reilly provided a list of seventeen Roche employees who were terminated by Lerner or one of his direct subordinates. See Miller Aff. Ex. 14. However, Reilly also stated in response to an interrogatory that he had "no personal, direct knowledge of whether Mr. Lerner's decisions to terminate any particular individual was influenced by age." See Miller Aff. Ex. 14.
Plaintiff Gabriel Saucy, who at the time of Operation Turnabout was the Director of Chemical Process Development in the Research and Development Division, testified that, after a tour in early 1983 of the two laboratories which Dr. Saucy directed, Lerner conveyed to Saucy his observation that a lot of the people in the two laboratories were "middle-aged." See Saucy at 93-94. Thereafter, Dr. Saucy sent a memorandum to Lerner showing the age of the personnel in the labs. Id. at 98. The memo also states that "the attached [the list of average ages] confirms that we have a rather 'middle-aged' department!" See Saucy Ex. 7.
Second, plaintiffs point to evidence in the record that, according to plaintiffs, demonstrates that other senior Roche executives were biased against older employees. For example, according to the deposition of Dr. Saucy, sometime between December of 1982 and March of 1983, at a luncheon meeting with upper level managers in the Research and Development division, Dr. Ronald Kuntzman proposed replacing senior level Ph.D.'s who were past 40 years old with younger Ph.D.'s. See Saucy at 69. Kuntzman gave the following reasons: First, he claimed that scientists over 40 were no longer were creative. Second, people over 40 in research were no longer knowledgeable in terms of state-of-the-art knowledge about new science education. Third, younger Ph.D.'s by contrast would be at the cutting edge of science education in terms of knowledge. Lastly, younger persons would be less expensive to employ. See Saucy at 69-70. In his deposition, Dr. Kuntzman stated that he sought to hire new Ph.D.'s, in part, because
in any scientific organization you have to bring in people continuously that have just completed their training. They have new technology, they have the most modern thinking of the research in their field of anyone and so you constantly want to bring in new Ph.D.s into an organization. It is just important for the research effort to be able to do that.
Kuntzman at 103-04. He also testified that while new Ph.D.'s (i.e., Ph.D.'s who have recently completed their training) are generally younger than Ph.D.'s who completed their training some years ago, they are not necessarily younger. See id. at 104-05. Kuntzman was appointed by Lerner to the positions of Vice President of Research and Development and executive committee member in October of 1984. Id. at 16.
Dr. Herbert Weissbach, the head Roche's Institute of Molecular Biology, was asked the following question at his deposition: "Was there a desire by you or by Dr. Kuntzman to focus your hiring on scientists who were young and talented?" Weissbach at 46. Weissbach answered:
I think it's true in any organization that you want to always have a constant stream of young, dynamic individuals coming in, building up their career, leaving so that a place can stay vibrant. That's true whether it be a university department, a pharmaceutical company, the RIMB [Roche Institute of Molecular Biology].