Daniel F. BRENNAN, Plaintiff, Appellant, v. GTE GOVERNMENT SYSTEMS CORPORATION, Defendant, Appellee

COFFIN, Senior Circuit Judge. Daniel Brennan’s twenty-year-long employment with GTE Government Systems Corporation (“GTE”) ended in March 1993 when he was terminated as part of a reduction in force. Brennan, who was fifty years old at the time of his termination, brought suit claiming that he was discharged as a result of age discrimination in violation of the Age Discrimination in Employment Act, 29 U.S.C. §§ 621-634 (“ADEA”), and Mass. Gen. Laws ch. 151B, § 4 (“eh,151B”). He *24 also alleged a state law misrepresentation claim, asserting that GTE fraudulently and negligently misrepresented its policies and procedures for laying off employees. The district court dismissed the state claim upon summary judgment and granted judgment as a matter of law at the close of evidence. Brennan challenges the rulings, and contests the court’s exclusion of certain evidence. We find error in one of the evidentiary rulings relevant to establishment of a prima facie case. We also conclude that there was sufficient evidence to create jury issues on pretext and age discrimination. We therefore reverse the directed verdicts for GTE on the federal and state age discrimination claims but affirm the dismissal of the misrepresentation claim. I. Factual Background 1 Brennan began working at GTE in 1973 as a senior technician and later held a series of different engineering positions. In 1986, he accepted a position as an integration and test engineer in the Mobile Subscriber Equipment System (“MSE”) division. Created in 1986, the MSE division developed in response to a contract between GTE and the United States Army for a comprehensive communications system. 2 In mid 1992, the division was reorganized and Brennan was assigned to work in the Circuit Switching and Systems Control department, headed by John Van Dolman. His job as an integration and test engineer involved designing and overseeing the MSE testing process. The record contains performance appraisals extending from 1976 to June 1992 consistently describing Brennan as meeting job expectations. In late 1992, GTE assigned Brennan to fill in for Carl Bredberg, a systems engineer whose job was to troubleshoot MSE field problems, after Bredberg suffered a heart attack. In September 1992, MSE engineering department managers Van Dolman and Kevin Flynn each prepared “Work Force Analysis Forms” in anticipation of a reduction in force planned for November 2, 1992. These forms set forth positions to be eliminated, employees performing those functions, and the names of individuals recommended for termination. Van Dolman identified “System Integration and Test Procedure Writer” as the function to be eliminated and Brennan as the only employee for termination, stating that Brennan lacked the skill mix to do the remaining work. The function eliminated by Flynn was “Senior System Engineer.” Two persons, each sixty-three years old, out of nine employees in that position were identified for layoff. GTE, however, did not go forward with the planned reduction in force. Instead, in December 1992, it notified employees of its need to reduce staff as a result of general downsizing in the defense industry, and offered a voluntary separation incentive program and an enhanced early retirement program as “first step[s] in realigning [GTE] staffing levels.” GTE announced that the programs were entirely voluntary, but would be offered only from January 11,1993 until February 5, 1993. Under the early retirement program, available to all employees fifty and older, Brennan could retire with pension and medical benefits that would otherwise not be available to him until January 1,1996. Brennan, however, declined the offered program, believing that his seniority would prevent him from being selected for layoff. On December 7, 1992, Allen Sherwood, manager of the MSE division at that time, developed a “rating and ranking” of employees in the engineering division in anticipation of a layoff. Brennan was listed as the lowest ranking of thirty-three employees at grade levels six and seven. The rating and ranking included three categories of grades (3-5, 6-7, *25 8-11), and then listed the employees who, out of a total of ninety-six, ranked among the bottom thirty percent of all employees. On this list, Brennan ranked ninety-third, or fourth from the lowest. Other terminated employees had similar, although not quite as low, rankings as Brennan. 3 The ages of employees listed in the rating and ranking were admitted as evidence. A high number of fifty and older employees were ranked in the bottom thirty percent. On February 26, 1993, Van Dolman prepared another Work Analysis Form, identifying “System Integration Testing and Test Procedure Writing” as the function to be eliminated. 4 He listed fourteen persons in this position prior to the reduction in force, out of which five were proposed for separation. Brennan, who was then fifty, along with four others (aged thirty-one, thirty-four, fifty-seven and fifty-nine) were identified as “not possessing] the skills to perform the remaining work available.” The form also set forth the specific skills these workers lacked, namely, “detailed knowledge as required to troubleshoot and resolve System level problems.” Brennan, John Hartgrove (fifty-seven), and John Hurley (thirty-seven), were also described as unable to perform “design related Systems work.” The listed reasons for termination were “wrong skill mix to do work remaining” and “lowest performance rating.” Under “additional comments” was written, “rating and ranking verifies above decisions.” Two days later, Van Dolman completed a Work Force Analysis Form to reduce another section of the MSE division, System Design Activities. Of the four persons he'listed as working in that section, he selected the only over fifty employed, A1 Rieard, for termination. The remaining three employees, all retained, were in their late twenties or early thirties. Van Dolman described Rieard as unable to “perform complex system engineering level design tasks.” Reasons for termination and additional comments referencing the rating and ranking were the same as those listed for Brennan. In late March 1993, Brennan was informed that he would be terminated on April 2,1993. Brennan testified that when he approached Van Dolman about the decision, Van Dolman stated, “I was told it wouldn’t matter to you guys.” Brennan believed “you guys” referred to the fifty and older employees, and “wouldn’t matter” referred to the retirement packages Van Dolman believed to be available to employees aged fifty and older. Van Dolman testified that he could have made the statement, but did not recall doing so. Brennan also testified that during his employment with GTE, he and other older employees were not given new computers or new assignments, which were instead designated for younger employees. The fifty and older employees discharged as a result of the Work Force Analysis Forms, including. Brennan, were the only fifty and older employees remaining in the two MSE departments subject to layoff. Other persons in this age group who had worked in these departments chose voluntary separation or early retirement before the layoffs. II. The Age Discrimination Claims The central aspect of our review involves Brennan’s appeal from directed verdict on his age discrimination claims. The district court found that Brennan had not established the required prima facie case, had not shown that GTE’s reason for discharging him was pretextual, and did not present sufficient evidence to support a finding of discriminatory animus. We review the grant of a motion for judgment as a matter of law de novo, taking the facts most favorable to Brennan. See Russo v. Baxter Healthcare Corp., 140 F.3d 6, 7-8 (1st Cir.1998). Brennan must provide “ ‘more than a scintilla of evidence and may not rely on conjecture or speculation’ to justify the submission of an issue to the jury.” Id. (quoting Katz v. City Metal Co., 87 F.3d 26, 28 (1st *26 Cir.1996)). “The court, however, must ‘not consider the credibility of the witnesses, resolve the conflicts in testimony, or evaluate the weight of the evidence.’” Andrade v. Jamestown Housing Authority, 82 F.3d 1179, 1186 (1st Cir.1996) (citing Wagenmann v. Adams, 829 F.2d 196, 200 (1st Cir.1987)). A verdict may be directed only if, applying these standards, the evidence does not permit a reasonable jury to find in favor of Brennan. See id. As part of his age discrimination ease, Brennan alleged evidentiary errors, which we review for abuse of discretion. See Kelley v. Airborne Freight Corp., 140 F.3d 335, 345-46 (1st Cir.1998). We will not upset a verdict for “any error or defect in the proceeding which does not affect the substantial rights of the parties.” Id. (quoting Fed.R.Civ.P. 61). To review the court’s determinations, we begin by setting forth the burden shifting framework of the controlling discrimination law. We note that until the last step in the framework, discussed infra, Massachusetts law parallels federal law. See Kelley, 140 F.3d at 347-48. A. The burden shifting framework The plaintiffs initial burden to establish a prima facie case of discrimination is “not onerous.” Sanchez v. Puerto Rico Oil Co., 37 F.3d 712, 719 (1st Cir.1994) (quoting Texas Dep’t of Community Affairs v. Burdine, 450 U.S. 248, 253, 101 S.Ct. 1089, 67 L.Ed.2d 207 (1981)). “All that is needed is the production of admissible evidence which, if uncontradicted, would justify a legal conclusion of discrimination.” Id. In a reduction in force ease like this one, a plaintiff must demonstrate: (1) he was at least forty years of age; (2) he met the employer’s legitimate job performance expectations; (3) he experienced adverse employment action; and (4) his employer did not treat age neutrally or younger persons were retained in the same position. See Hidalgo v. Overseas Condado Ins. Agencies, Inc., 120 F.3d 328, 333 (1st Cir.1997) (quoting Goldman v. First Nat’l Bank of Boston, 985 F.2d 1113, 1117 (1st Cir.1993)). Once the plaintiff establishes the prima facie requirements, a rebuttable presumption that the employer engaged in intentional age based discrimination attaches. See Woodman v. Haemonetics Corp., 51 F.3d 1087, 1091 (1st Cir.1995). The burden then shifts to the employer to articulate a legitimate, nondiscriminatory reason for the decision. See Thomas v. Sears, Roebuck & Co., 144 F.3d 31, 32-33 (1st Cir.1998). If the employer meets its burden, the presumption of discrimination vanishes. See Woodman, 51 F.3d at 1091. At this point, although the burden returns to the plaintiff in both instances, state and federal law no longer parallel each other. Under the ADEA, we inquire whether the evidence as a whole would permit a reasonable factfinder to conclude that the proffered reason was pretextual and the true reason was an age-based animus. See Hidalgo, 120 F.3d at 335 (quoting St. Mary’s Honor Ctr. v. Hicks, 509 U.S. 502, 515, 113 S.Ct. 2742, 125 L.Ed.2d 407 (1993)). Under eh. 151B plaintiff’s burden may be limited to demonstrating pretext, without a required showing of discriminatory animus. 5 See Kelley, 140 F.3d 335 at 347-48; McMillan v. Massachusetts Soc’y for the Prevention of Cruelty to Animals, 140 F.3d 288, 298 & n. 4 (1st Cir.1998); Blare v. Husky Injection Molding Sys. Boston, Inc., 419 Mass. 437, 444-45, 646 N.E.2d 111 (1995). The distinction does not matter in this case, however, as *27 we find that plaintiffs evidence rises to the level required by the ADEA to overcome a directed verdict. B. The prima facie case It is undisputed that, as a fifty-year-old man who was terminated by GTE, Brennan meets the first and third prongs of his prima facie case. As to the second, Brennan has submitted performance reviews dating back to 1976 and running through June 1992. 6 He received acceptable rankings in virtually every appraisal, although he only occasionally received the highest ranking. For example, in the most recent ranking, covering the period of January to June 1992, he was rated “met expectations” for all categories 'of work, except in the area of “control/follow-up” in which he was rated “exceeded expectations.” Out of five possible overall performance ratings, he was given a three, which is described as “Employee meets all individual objectives as established in the performance plan, performs all job responsibilities to meet the requirements of the job and displays a satisfactory degree of competence in the skills required of the job.” And, under “comments,” the review stated, “Dan has the ability to quickly become technically competent in tasks assigned to him and accomplish the task in a timely manner to meet schedules” and “Dan is very supportive to the changing priorities required to support the overall mission.” GTE strives to characterize the “legitimate job expectations” standard as covering what it maintains to be the new, downsized requirements expected of the future MSE division. Such an interpretation would render irrelevant the remaining burden shifting requirements (the employer’s articulation of a legitimate reason would be incorporated, de facto, into the job expectation standard, and the employee’s obligation to demonstrate pretext would have to be fought at the prima facie stage), and would be an “onerous” endeavor, beyond the confines of a prima facie case. We think the gauge therefore should be whether the employee has a record of performing acceptably and meeting the demands of his job. Brennan’s evidence is more than sufficient to meet the second prong of the prima facie ease. It was the fourth prong, lack of age-neutral decisionmaking, where the district court found Brennan’s evidence inadequate. Our analysis demands that we consider whether the district court properly excluded, evidence as to other employees. 7 “[L]ack of neutrality may be manifested either by a facially discriminatory policy or by a policy which, though age-neutral on its face, has the effect of discriminating against older persons, say, by leading inexorably to the retention of younger employees while similarly situated older employees are given their walking papers.” Vega v. Kodak Caribbean, Ltd., 3 F.3d 476, 479 (1st Cir.1993). A judge should not exclude relevant evidence, unless its probative value is substantially outweighed by the danger of unfair prejudice. See McMillan, 140 F.3d at 300 (citing Fed.R.Evid. 403). Evidence that other terminated fifty and older employees had good employment records, skills in certain areas of the job, and superior experience to retained employees is relevant to determining whether a policy has the effect of discriminating against older persons. Among the evidence offered by Brennan but exclud *28 ed by the district court was a comparison between work performed by Tessmer, who was fifty-nine, and a thirty-year-old employee who was retained. Brennan also proffered evidence comparing the work record of Ri-eard, who was fifty-seven, rated at a level of three, and discharged, with that of Craig Poole, who was twenty-eight, given the same rating, trained by Rieard, and retained. This evidence was also excluded. While “proof of a general atmosphere of discrimination is not the equivalent of proof of discrimination against an individual,” it “may add ‘color’ to an employer’s decisionmaking process.” Ruiz v. Posadas de San Juan Assoc., 124 F.3d 243, 249 (1st Cir.1997). Certainly comparative evidence in combination with data showing a disproportionate number of terminated older employees is probative of age discrimination. See Goldman, 985 F.2d at 1119 (excluding numerical evidence because no comparative evidence was introduced). We are not faced here with determining whether this evidence proves age discriminatory animus, but rather are concerned only with whether this evidence indicates lack of neutrality. As we have repeatedly emphasized, this is not a burdensome standard. See Woodman, 51 F.3d at 1091. Combined with the raw data indicating the ages of the persons listed in the rating and ranking and terminated in the reduction in force, the evidence regarding other employees and testimony by Brennan indicating that the company gave preference to younger employees in distributing computers demonstrate “lack of neutrality” sufficient to meet Brennan’s burden of establishing a prima facie ease. C. Establishing pretext and discriminatory animus It is undisputed that GTE has met its burden of production in maintaining that it fired Brennan because MSE testing work was diminishing and Brennan was among the lowest ranked employees performing that type of work. Van Dolman testified that the MSE workload, and specifically the number of plans and procedures written and tests conducted, was diminishing. No longer a new system, the MSE required less testing. David Orlando and David Loomis, both GTE engineering supervisors, confirmed this. And various GTE personnel testified that while Brennan was a good test writer, he lacked the troubleshooting skills and technical knowledge about the system that GTE most needed given the downsizing of new MSE projects. Brennan, in turn, must satisfy his burden of showing pretext, and under federal law, discriminatory animus. He may rely on the same evidence to prove both pretext and discrimination, as long as the evidence rises to the level of meeting both requirements. See Udo v. Tomes, 54 F.3d 9, 13 (1st Cir.1995). He may also rely on the evidence pertaining to his prima facie. See Burdine, 450 U.S. at 255 n. 10, 101 S.Ct. 1089. Brennan offers the following evidence to meet the final hurdle: (1) his experience and performance reviews; (2) GTE’s failure to use standard procedure in creating the Work Force Analysis Forms; (3) absence of a projected work analysis by GTE at the time the Work Force Analysis Forms were prepared; (4) MSE engineering work available after the layoff; (5) disproportionately age-weighted layoffs; (6) favoritism by GTE toward younger engineers; and (7) evidence showing that older qualified employees were discharged and younger, less or equally qualified engineers were retained. We now review these. The performance reviews tell us that Brennan was a competent employee who consistently performed within the expected job range, but did not excel. Again, we note that in June 1992 his review described him as having “the ability to become technically competent in tasks assigned to him.” A reasonable jury could find that this belies GTE’s claim that he lacked the technical competence to troubleshoot problems arising in the MSE division, or, at least, that in June 1992 GTE believed that Brennan was capable of rapidly gaining the technical knowledge required by the company. Support for this view can also be found in GTE’s asking Brennan in late 1992 to fill in for Bredberg, whose work primarily involved troubleshooting system problems. *29 We add to this the evidence presented by Brennan that GTE did not use standard procedure in completing the Work Force Analysis Forms. Deviation from established policy or practice may be evidence of pretext. See Lattimore v. Polaroid Corp., 99 F.3d 456, 466-67 (1st Cir.1996). Robert Hetzel, Director of Human Services at GTE in 1993, testified to the order in which the forms should be completed: identify the function to be eliminated within a particular department and the reason, list the employees performing the function and evaluate them, identify the person or persons who- do not have the skills necessary to do the remaining work, and then identify individuals for layoff, using the criteria, “wrong skill mix,” “poor performance,” “lowest performance rating,” and “lowest seniority.” GTE’s seniority policy requires that, “where ability is comparable, the senior employee will be retained.” Van Dolman’s testimony revealed that, instead of following this procedure, he completed the March 1993 Work Force Analysis Forms by simply filling in names of persons to be laid off that were provided to him by Sherwood from the rating and ranking list created in December 1992. As to the rating and ranking, Van Dolman first provided Sherwood with a list, not in the record, ranking persons in his department. From this list and those put together by other engineering department heads, Sherwood compiled a rating and ranking of all engineering employees. Sherwood explained that he arrived at this list by consulting with department heads. 8 He. described the “factors [that] came up in the discussions leading to the merging of the three grades into the [lowest thirty percent] column” as “principally value, in terms of doing the residual MSE, and a lot of work that was coming.” It is clear that seniority was not considered in this ranking. Although thin, this evidence that standard procedure was not followed is directly relevant to Brennan’s burden of demonstrating pretext. The record also raises questions about whether the layoffs were needed; Brennan points to the two Work Force Analysis Forms completed in September 1992 by Van Dolman and Flynn, in which he was the only person slated for layoff in his group and two persons were slated for layoff in Flynn’s group. Of the nine senior system engineers in Flynn’s group in 1992, only four remained in 1993. This reduction, in combination with the number of fifty or older employees who took early retirement, significantly reduced GTE’s work force — by a greater number,.in fact, than targeted in the September 1992 Work Force Analysis Forms. As to Brennan’s contention that the company lacked knowledge of the projected work load, Sherwood testified that at the time he created the rating and ranking he did not know how many test writing positions remained for the future MSE work, and Flynn testified that he did not remember any discussion during the development of the rating and ranking concerning the specific work remaining in the MSE. Brennan suggests from this that the company’s subsequent contention, that it needed to terminate test writers, was not true, and simply a cover for its intention to lay him and other older employees off through a constructive device of a rating and ranking. While certainly limited, we cannot say that, in combination with the remaining evidence, this testimony could not have helped to lead a reasonable jury to suspect that the reasons articulated by GTE for downsizing its workforce were not true. Brennan also introduced evidence that GTE hired persons for engineering positions *30 and that MSE work was available after his layoff. Such evidence, in undercutting the employer’s proffered reason for reducing the workforce, is probative of pretext. See Martin v. Envelope Division of Westvaco Corp., 850 F.Supp. 83, 91 (D.Mass.1994). While GTE presented plentiful testimony that MSE testing work was on the decline, we think that a factfinder could reasonably conclude, looking at the advertisements for new engineers and the number of persons hired, together with other evidence, that the reductions were not in fact needed. Finally, Brennan points to evidence of discriminatory animus based on age: younger engineers received preference in the distribution of new computers and new assignments, a high proportion of older persons were discharged and ranked low in the rating and ranking, and Van Dolman’s statement, “I was told it wouldn’t matter to you guys.” After examining all the evidence, the jury might have concluded, as did Hetzel, that the disproportionately high number of older persons ranked low and discharged was not indicative of discrimination. But this determination is rightly within the jury’s province, not the court’s. And although Van Dolman testified that he could not recall making the statement, supra, neither did he deny making it. In context, his testimony could leave a reasonable jury to conclude that Van Dol-man expected Brennan to receive full benefits 9 and went along with the layoffs with that assurance in mind. We find therefore that Brennan has demonstrated sufficient evidence of pretext and discriminatory age animus to withstand a directed verdict. His age discrimination claims should proceed before a jury. III. The Misrepresentation Claim Our review is de novo to determine whether the jury, drawing all inferences in Brennan’s favor, could reasonably have rendered a verdict for him on this issue. See Coastal Fuels of Puerto Rico, Inc. v. Caribbean Petroleum Corp., 79 F.3d 182, 196 (1st Cir.1996); Fed.R.Civ.P. 56(c). Brennan argues that he was misled by fraudulent and negligent misrepresentation—specifically, GTE’s seniority policy—to decline participation in the Early Retirement Program. He alleges that GTE had a duty to disclose to him that it would use a ranking system rather than performance reviews to determine who would be terminated, and that he suffered economic loss because he relied on the performance reviews to gauge the likelihood that he would be laid off from his job. We can dispose of this issue quickly. Under Massachusetts law, claims of fraudulent and negligent misrepresentation require a false representation of material fact, knowledge of falsity or carelessness on the part of the defendant, and a reasonable reliance by the plaintiff. See Knapp Shoes, Inc. v. Sylvania Shoe Mfg. Corp., 72 F.3d 190, 199-200 (1st Cir.1995); see also Fox v. F & J Gattozzi Corp., 41 Mass.App.Ct. 581, 587, 672 N.E.2d 547, 551 (1996) (requiring “justifiable” reliance by the plaintiff). Brennan offers no evidence that GTE represented to him that it would only use the performance evaluations in ranking employees and determining whether they were “comparable” for the purpose of invoking the company’s seniority policy. Even if a jury could conclude that GTE made such a representation, Brennan’s reliance on that fact as security that he would not be laid off is not reasonable or justifiable. Brennan was fairly consistently ranked a three out of five. He presents no evidence that he knew the number of persons the company would terminate (he offers the opposite in support of his pretext argument), the levels at which other persons were ranked, or the seniority other employees had in comparison to him. It may be that GTE’s layoff would require that it terminate all *31 level three employees and some level two employees. In that ease, Brennan’s seniority would not have impacted his retention. We therefore affirm the district court’s grant of summary judgment on this claim. IV. Conclusion On the record before us, we find that Brennan has introduced sufficient'evidence to submit his age discrimination claims to the jury, but insufficient evidence on his misrepresentation claim. We emphasize again that we make no credibility judgments and do not weigh the evidence; our analysis has been restricted to determining whether Brennan presented enough evidence to proceed before the jury. We remand the age discrimination claims to the district court. Affirmed in part and reversed in part. Costs to appellant. Footnotes [1] As this case comes to us from summary judgment and directed verdict, we review the facts in the light most favorable to Brennan. See Hodgens v. General Dynamics Corp., 144 F.3d 151, 155-56 (1st Cir.1998); Russo v. Baxter Healthcare Corp., 140 F.3d 6, 7-8 (1st Cir.1998). [2] Prior to the MSE, the United States Army contracted communication services such as telephone, radio and security on a piecemeal basis. The MSE division was generated to provide a system that covered all areas of communications. Because the MSE was a new system, it required extensive testing. [3] The record indicates that the three employees ranked lower than Brennan (all over fifty) took advantage of the voluntary separation incentive program in February 1993. [4] Although the title is slightly different from that identified by Van Dolman in the 1992 Work Force Analysis Form, we understand the position to be the same in both instances. [5] We recognize that there is some question as to whether Massachusetts is indeed a "pretext-only" state as defined in Blare v. Husky Injection Molding Sys. Boston, Inc., 419 Mass. 437, 443, 646 N.E.2d 111, 116 (1995). A more recent case describing the third step in a ch. 15IB analysis used language that parallels federal law, which requires a higher standard. See Matthews v. Ocean Spray Cranberries, Inc., 426 Mass. 122, 128, 686 N.E.2d 1303, 1309 (1997). In McMillan v. Massachusetts Society for the Prevention of Cruelty to Animals, 140 F.3d 288, 298 n. 4 (1st Cir.1998), we declined to find that this language, which was taken directly from Blare, signified a change in the "pretext only” standard. But three judges from our court indicated that "the viability of Blare, or at least the most common and expansive reading of it, has been cast in doubt by Matthews." McMillan v. Massachusetts Society for the Prevention of Cruelty to Animals, 140 F.3d at 295-96 (1st Cir.1998). [6] The performance appraisals cover time spans ranging from twenty-six to six months. [7] Brennan also appeals the exclusion of statistical evidence on the proportion of fifty and older employees who were terminated and who received low rankings in the rating and ranking. While we think the focus of actuarial work is such as to sufficiently qualify Brennan’s proposed expert witness for the limited statistical analyses offered in this case, we cannot say that the exclusion was an abuse of the broad discretion granted to the district court. See Moulton v. Rival Co., 116 F.3d 22, 27 (1st Cir.1997) ("A trial court has wide discretion in determining the admissibility of expert testimony, and we will reverse its decision only when there has been a clear abuse of discretion.”). We find this particularly true where, as here, other evidence admitted served much the same purpose as the excluded evidence. The district court admitted raw data indicating the ages of the persons listed in the rating and ranking and terminated in the reduction in force. The data, which was not so overwhelming or large to be inaccessible to the jury, permitted the factfinder to discern any existing age disparities. Even if error, there would be no prejudice. [8] Sherwood testified that, after each department manager submitted a list ranking members of their departments, we convened a fairly lengthy meeting; and for each of these three groups ... we would start at the top and talk about the top persons on the three lists, the strengths and weaknesses, and do our best to compare the people after some discussion. I acted as the arbitrator, and I ordered one to three. And we probably did that for probably the top third, and then we went and we would talk about the bottom person on each of the department managers’ lists. And sometimes they lead the discussions, and then I would order them and work our way up to the middle parts of it. We didn't argue about a lot, other than to kind of make sure the order was right. Sherwood estimated that this meeting took about half a day, and that this meeting was the third or fourth in a series of such meetings. [9] Van Dolman testified: As part of that discussion, Dan ... felt that he was going to be given a package as a result of the layoff that was less than what he ... would have received if he had taken the [voluntary early retirement] package.... I’m not quite sure what I said, you know. I probably—I was surprised, because I—he mentioned some specifics which I don't really recall, but I remember the specifics had to do with his not meeting certain things relative to the requirements of the [involuntary severance] package. (Emphasis added.)