The plaintiff, Mary Sullivan, challenges a Superior Court judge’s order granting summary judgment to the defendant, Liberty Mutual Insurance Company (Liberty), on Sullivan’s sex and age discrimination in employment claims brought pursuant to the antidiscrimination statute, G. L. c. 151B.1 The case arose from Liberty’s implementation in 1999 of a reduction in force in which it permanently discharged eleven employees in its New England region legal department, including Sullivan. Sullivan appealed. We transferred the case here on our own motion to consider the elements for establishing a prima facie case of discrimination under G. L. c. 151B, when the challenged employment action arises in the context of a reduction in force.2 We affirm the order granting summary judgment, but for reasons somewhat different from those of the motion judge.
1. Background. We briefly summarize the basic facts in their light most favorable to Sullivan, the nonmoving party, reserving additional facts for later discussion. In 1986, Sullivan began working for Liberty as an attorney representing the company’s insureds, moving to its Boston office in 1988. In June, 1999, precipitated (according to Liberty) by a decline in its business and a recent merger, Liberty implemented a reduction in force.3 In the months before implementation, Liberty first imposed a *36hiring freeze and then requested its managerial staff to determine whether each of Liberty’s offices was appropriately staffed.
Kenneth A. Latronico, Liberty’s general attorney for its New England region,4 was charged with analyzing what Liberty referred to as “productivity” and “capacity” for all offices within the region.5 After receiving information ffom each office, Latronico determined there was no overstaffing in the Connecticut, New Hampshire, or Canadian offices, but concluded there was overstaffing in the five offices in eastern Massachusetts (Andover, Bedford, Boston, Brockton, and Worcester). He recommended, among other actions, a ten per cent staff reduction in eastern Massachusetts and the eventual closure of the Bedford office by relocating its remaining attorneys to other offices.
After further consultation with senior management, Latronico then solicited recommendations from each office as to which attorneys to lay off. Latronico himself made the layoff recommendations for the Boston office; he recommended the layoff of Steven Hope, David Hartigan, and Sullivan, in that order. On June 15, 1999, Liberty discharged Sullivan and five other attorneys employed in the New England region, including Hope and Hartigan.6 Three of the attorneys who were laid off (includ*37ing Sullivan) worked in the Boston office, two were from the Bedford office, and one was from the Worcester office. Three of the six attorneys terminated were women. Five of the six were over forty years of age; the sixth was thirty-eight years old. At the time of her layoff, Sullivan was forty-nine years old. La-tronico transferred the cases on which Sullivan had been working to six other attorneys, five of whom were men, all of whom were substantially younger than Sullivan.
During her twelve years with Liberty, Sullivan had received various performance evaluations, as had other attorneys employed by Liberty. While Sullivan’s reviews were not the highest when compared to other attorneys in the Boston office, her reviews were generally positive. On no occasion had she received an over-all evaluation falling below “meets expectations,” although beginning as early as 1992, Liberty had noted some concerns with Sullivan’s lack of responsiveness to clients, and, later, her “collegiality” and “human relations skills,” especially concerning her interactions with clerical staff.
On December 15, 1999, Sullivan filed a charge of discrimination with the Massachusetts Commission Against Discrimination, alleging that she was terminated because of her sex and age in violation of G. L. c. 15IB, § 4. On May 4, 2000, Sullivan commenced this action in the Superior Court against Liberty. See G. L. c. 151B, § 5. Over the following twenty months, the parties engaged in discovery, which included twelve depositions and Liberty’s production of over 2,000 pages of documents. On December 20, 2001, Liberty moved for summary judgment on all of Sullivan’s claims, which Sullivan opposed. A judge in the Superior Court allowed Liberty’s motion,7 and a different judge subsequently denied Sullivan’s motion for reconsideration. Sullivan timely appealed from the sum*38mary judgment as well as from an order granting Liberty’s emergency motion to strike portions of Sullivan’s affidavit,8 and an earlier order denying her leave to proceed as “Jane Doe.”9
2. Standard of review. In cases involving claims of employment discrimination, a defendant employer faces a heavy burden if it seeks to obtain summary judgment: summary judgment is disfavored in discrimination cases based on disparate treatment10 because the question of the employer’s state of mind (discriminatory motive) is “elusive and rarely is established by other than circumstantial evidence.” Blare v. Husky Injection Molding Sys. Boston, Inc., 419 Mass. 437, 439 (1995), citing Wheelock College v. Massachusetts Comm’n Against Discrimination, 371 Mass. 130, 137 (1976). This requires “the jury to weigh the credibility of conflicting explanations” of the adverse decision. Blare v. Husky Injection Molding Sys. Boston, Inc., supra at 440. In reviewing an order granting summary judgment in such cases, we of course apply our traditional test and consider the facts in their light most favorable to the nonmov-ing party, drawing all reasonable inferences in her favor. See *39Mass. R. Civ. R 56 (c), 365 Mass. 824 (1974); Blare v. Husky Injection Molding Sys. Boston, Inc., supra at 438. We may also make “an independent compilation of the relevant facts to frame the ultimate legal question whether summary judgment is appropriate.” Matthews v. Ocean Spray Cranberries, Inc., 426 Mass. 122, 123 n.1 (1997). While the standard of review is the same as in all other cases, Liberty, as the moving party, “has the burden of affirmatively demonstrating the absence of a genuine issue of material fact on every relevant issue, even if [Liberty] would not have the burden on an issue if the case were to go to trial.” Id. at 127.
Liberty may satisfy its burden by demonstrating that Sullivan “has no reasonable expectation of proving an essential element of the case at trial.” Id. And although summary judgment is disfavored in employment discrimination cases of disparate treatment, we have upheld summary judgment in favor of an employer where “the plaintiff is unable to offer admissible evidence of the defendant’s discriminatory intent, motive, or state of mind sufficient to carry the plaintiff’s burdens and support a judgment in the plaintiff’s favor.” Id., citing Blare v. Husky Injection Molding Sys. Boston, Inc., supra at 440, and cases cited. We turn now to consider the record in light of these standards, beginning with an overview of the legal requirements imposed on both Sullivan and Liberty in this reduction in force case.
3. Claims of discrimination. We have construed G. L. c. 151B as containing four elements an employee must prove to prevail on a claim of discrimination in employment: membership in a protected class, harm, discriminatory animus, and causation. See Lipchitz v. Raytheon Co., 434 Mass. 493, 502 (2001). In cases such as this, where the claim is one of discrimination because of sex and age, the first two elements are seldom disputed. Rather, the conflict arises as to the latter two elements. Direct evidence of those elements (discriminatory animus and causation) rarely exists, see Wynn & Wynn, P.C. v. Massachusetts Comm’n Against Discrimination, 431 Mass. 655, 665 (2000), and a plaintiff may therefore establish one or both by indirect or circumstantial evidence using the familiar three-stage, burden-shifting paradigm first set out in McDonnell Dou*40glas Corp. v. Green, 411 U.S. 792, 802-805 (1973) (McDonnell Douglas).11 The three-stage order of proof “does not circumvent the plaintiff’s burden to prove all the essential elements of a discrimination claim, but does permit the jury to infer discriminatory animus and causation from proof that an employer has advanced a false reason for the adverse employment decision, in the absence of direct evidence that the actual motivation was discrimination.” Knight v. Avon Prods., Inc., 438 Mass. 413, 422 (2003). See Abramian v. President & Fellows of Harvard College, 432 Mass. 107, 116 (2000), citing McDonnell Douglas, supra at 802. As we shall explain later, Sullivan has not produced any direct evidence of either age or sex discrimination. See note 24, infra. We confine our discussion to considering whether she has adduced sufficient indirect or circumstantial evidence to survive summary judgment.12
4. The prima facie case. Under the McDonnell Douglas formulation, Sullivan bears the initial burden of establishing by the preponderance of the evidence a prima facie case of discrimination. Her burden is not onerous. See Texas Dep’t of Community Affairs v. Burdine, 450 U.S. 248, 253 (1981). See also Currier v. United Techs. Corp., 393 F.3d 246, 254 (1st Cir. 2004). Sullivan must simply produce sufficient evidence that Liberty’s actions, “if otherwise unexplained, are more likely than not based on the consideration of impermissible factors.” Furnco Constr. Corp. v. Waters, 438 U.S. 567, 577 (1978). As the Supreme Court noted in Texas Dep’t of Community Affairs v. Burdine, supra at 253-254, the prima facie case “serves an important function in the litigation: it eliminates the most common nondiscriminatory reasons for the plaintiff’s rejection.” For this reason, if she establishes a prima facie case, Sullivan is entitled to a “legally mandatory, rebuttable presumption” that Liberty unlawfully terminated her, and she will prevail on her *41claims if Liberty fails to satisfy its burden at the second stage of the framework. Id. at 254 n.7.
Generally, a plaintiff who is terminated from her position establishes a prima facie case of discrimination by producing evidence that she is a member of a class protected by G. L. c. 151B; she performed her job at an acceptable level; she was terminated; and her employer sought to fill her position by hiring another individual with qualifications similar to hers. See Abramian v. President & Fellows of Harvard College, supra. Liberty does not dispute that Sullivan satisfies the first three elements of this formulation of the prima facie case on each of her discrimination claims.13 Indeed it is likely that in a reduction in force case every plaintiff claiming sex or age discrimination can easily satisfy the first three elements of the prima facie case. The first and third elements are straightforward. As for the second element, in a reduction of force case the discharge is not the result of the employer concluding that an individual is not performing well: presumably before layoffs, every plaintiff in such a case is performing “at an acceptable level” or she would have been discharged before a downsizing occurs. Rather, the dispute will almost always concern the fourth element, and therein lies the difficulty.
As generally formulated, the fourth element is nonsensical in a reduction in force case: the plaintiff is not replaced, nor does her employer “seek to fill” the position, for the very purpose of a workforce reorganization is generally to reduce the number of employees. See Lewis v. Boston, 321 F.3d 207, 214 n.6 (1st Cir. 2003) (in typical reduction of force case fourth element is “unworkable because the plaintiff’s position no longer exists”). Yet it cannot be that every individual in the workforce who falls within some protected class (the overwhelming majority of employees) can establish a prima facie case of discrimination simply because she is laid off during a reduction in force. At *42the same time, because an employer seeks to reduce its workforce does not mean that it is free to make its employment decisions on impermissible grounds: “even during a legitimate reorganization or workforce reduction, an employer may not dismiss employees for unlawful discriminatory reasons.” Maresco v. Evans Chemetics, 964 F.2d 106, 111 (2d Cir. 1992), quoting Hagelthorn v. Kennecott Corp., 710 F.2d 76, 81 (2d Cir. 1983). This case presents our first opportunity to consider how the fourth element of the prima facie case must be varied so that a plaintiff who is laid off (or otherwise harmed) during a reduction in force may establish a prima facie case of unlawful discrimination. See Abramian v. President & Fellows of Harvard College, supra (elements of prima facie case “may vary depending on the specific facts of the case”).
We begin by examining how other courts have resolved the issue. Some have concluded that the fourth element “should be ‘relaxed’ when the employee’s layoff occurred in the context of a reduction in force.” Marzano v. Computer Sciences Corp., 91 F.3d 497, 503 (3d Cir. 1996). These courts permit a plaintiff to satisfy that element merely by showing that “other unprotected workers were retained.” Id. at 506, quoting DiBiase v. Smith-Kline Beecham Corp., 48 F.3d 719, 723 n.2 (3d Cir.), cert. denied, 516 U.S. 916 (1995). Accord Coburn v. Pan Am. World Airways, Inc., 711 F.2d 339, 343 (D.C. Cir.), cert. denied, 464 U.S. 994 (1983); Baker v. National State Bank, 312 N.J. Super. 268, 289-290 (App. Div. 1998), aff’d, 161 N.J. 220 (1999). Under that formulation Sullivan would establish a prima facie case by showing, in her sex discrimination claim, that a man was retained during the reduction in force and showing, in her age discrimination claim, that an employee at least five years younger than she was retained. Cf. Knight v. Avon Prods., Inc., supra at 424-425 (age disparity of five years or more between discharged plaintiff and younger replacement satisfies fourth element in age discrimination case). We reject that formulation, as such evidence is insufficient to establish a “legally mandatory, rebuttable presumption” of unlawful discrimination. Texas Dep’t of Community Affairs v. Burdine, supra at 254 n.7.
Other courts have advanced different formulations, each seeking to mesh more satisfactorily evidence establishing the prima *43facie case with entitlement by the plaintiff to the presumption of discrimination. The United States Court of Appeals for the First Circuit permits a plaintiff to satisfy the fourth element in an age discrimination case through a narrower, more specific showing that the employer retained unprotected or younger workers “in the same position” as the plaintiff (emphasis added). Currier v. United Techs. Corp., supra at 254. This is consistent with the majority view that permits a plaintiff to satisfy the fourth element by producing some evidence “from which a factfinder might reasonably conclude that the employer intended to discriminate in reaching the decision at issue,” Nichols v. Loral Vought Sys. Corp., 81 F.3d 38, 41 (5th Cir. 1996), quoting Amburgey v. Corhart Refractories Corp., 936 F.2d 805, 812 (5th Cir. 1991), or some similar formulation. Cf., e.g., Currier v. United Techs. Corp., supra (evidence that employer “otherwise did not treat age neutrally” in terminating plaintiff); Fast v. Southern Union Co., 149 F.3d 885, 890 (8th Cir. 1998) (evidence that age was “a factor in the employer’s decision to terminate” plaintiff); Maresco v. Evans Chemetics, supra at 110-111 (evidence that plaintiff was discharged in circumstances “that gave rise to an inference of discrimination”); Barnes v. GenCorp., Inc., 896 F.2d 1457, 1465 (6th Cir.), cert. denied, 498 U.S. 878 (1990) (some evidence “tending to indicate” that employer “singled out the plaintiff for discharge for impermissible reasons”). As we shall explain, a formulation along these lines is more satisfactory.
The purpose of the prima facie case is to identify those circumstances where the employer’s actions, “if left unexplained, are more likely than not based on unlawful discrimination.” Lewis v. Boston, supra at 216. Stated differently, in the typical case where the plaintiff is not hired or is fired, the existing formulation of the prima facie case “ ‘eliminates the most common nondiscriminatory reasons for the plaintiff’s rejection,’ which are lack of competence and lack of job availability, and thereby creates a presumption of discrimination.” Abramian v. President & Fellows of Harvard College, 432 Mass. 107, 116 (2000), quoting Blare v. Husky Injection Molding Sys. Boston, Inc., 419 Mass. 437, 441 (1995). In contrast, in a reduction in force case, “the most obvious *44explanations for the discharge of any one employee are lower proficiency and/or random chance.” Barnes v. GenCorp., Inc., supra at 1466. When these “most common” nondiscriminatory reasons for the plaintiff’s layoff have been eliminated as possible reasons for the employer’s actions, the presumption of discrimination is established, for “it is more likely than not the employer, who we generally assume acts only with some reason, based [its] decision on an impermissible consideration.” Furnco Constr. Corp. v. Waters, 438 U.S. 567, 577 (1978). Because there must be “at least a logical connection between each element of the prima facie case and the illegal discrimination for which it establishes” a presumption, O’Connor v. Consolidated Coin Caterers Corp., 517 U.S. 308, 311-312 (1996), we agree with those courts that recognize that the “mere termination of a competent employee when an employer is making cutbacks due to economic necessity is insufficient to establish a prima facie case.” LaGrant v. Gulf & Western Mfg. Co., 748 F.2d 1087, 1090 (6th Cir. 1984).
The task of formulating a standard to apply in all reduction in force cases is challenging, as the claims may arise in vastly different circumstances. In some reduction in force cases, the fact that an employer retained in the plaintiff’s same position an employee outside the plaintiffs protected class may indeed be “sufficiently probative to allow a factfinder to believe that the employer intentionally discriminated against the plaintiff.” Barnes v. GenCorp., Inc., supra.14 But in others the employer’s retention in the plaintiffs same position of an employee who is outside the plaintiffs protected class, without more, would not necessarily permit a reasonable inference of discrimination.15 In one case, the plaintiff may be the only employee who is laid *45off. In another, several hundred individuals may be affected. Our formulation must encompass all such circumstances.
The formulation of the fourth element of the prima facie case in a reduction in force case therefore cannot be precise. But consistent with the purpose of the prima facie case, and joining the majority of other jurisdictions, we conclude that a plaintiff in a reduction in force case may satisfy the fourth element of her prima facie case by producing some evidence that her layoff occurred in circumstances that would raise a reasonable inference of unlawful discrimination. Cf. Knight v. Avon Prods, Inc., 438 Mass. 413, 425 (2003). See, e.g., Currier v. United Techs. Corp., 393 F.3d 246, 254 (1st Cir. 2004); Nichols v. Loral Vought Sys. Corp., supra; Maresco v. Evans Chemetics, 964 F.2d 106, 111 (2d Cir. 1992); Barnes v. GenCorp., Inc., supra.
We add these observations. First, we repeat that the plaintiff’s initial burden of establishing a prima facie case is not intended to be onerous. See Texas Dep’t of Community Affairs v. Burdine, 450 U.S. 248, 253 (1981). See also Currier v. United Techs. Corp., supra at 254. It is meant to be a “small showing” that is “easily made.” Che v. Massachusetts Bay Transp. Auth., 342 F.3d 31, 38 (1st Cir. 2003), quoting Kosereis v. Rhode Island, 331 F.3d 207, 213 (1st Cir. 2003). Second, in almost all such cases, as in this case, summary judgment follows discovery, during which the employer would have disclosed its reason for selecting a particular plaintiff for “reduction,” i.e., discharge from the work force. As noted by the United States Court of Appeals for the Sixth Circuit, a plaintiff who “was truly singled out for discharge” because of membership in a protected class “should be able to develop enough evidence through the discovery process or otherwise to establish a prima facie case.” Barnes v. GenCorp., Inc., supra at 1465-1466. It is not onerous to require a plaintiff who has had the benefit of discovery to produce some evidence that the decision to lay her off (rather than some other employee) occurred in circumstances that raise an inference of unlawful discrimination. In this case for example, Sullivan established through discovery that some of the attorneys retained by Liberty in its Boston office had *46been assigned a lower rating than she had in their performance evaluations. That a terminated plaintiff has been ranked or evaluated above a younger or male coworker working in the same position who is not laid off is precisely the kind of indirect evidence of discriminatory intent that discovery elicits, which will permit a plaintiff to establish a prima facie case.16
5. Prima facie case of sex and age discrimination. We turn to consider Sullivan’s prima facie showing of sex and age discrimination in light of these requirements.17,18 With regard to *47her sex discrimination claim, to establish her prima facie case, Sullivan has asserted that (1) Liberty “exhibited a bias against women based on hiring, pay, promotion, and firing”;19 (2) Liberty disproportionately terminated women on June 15, *48199920. (3) following her layoff Latronico disproportionately reassigned Sullivan’s cases to male attorneys;21 and (4) Liberty retained male attorneys rated lower than Sullivan.22 Weighing against Sullivan is the uncontested fact that Liberty retained six female attorneys in the Boston office, all of whom were rated higher than she was.
Some of the evidence on which Sullivan has relied is clearly inadmissible. See notes 8, 19, supra. We nevertheless conclude that Sullivan has established a prima facie case of sex discrimination. Specifically, there is undisputed, admissible evidence that Liberty retained in its Boston office male attorneys with the same job classification as Sullivan, who had been rated lower than she had in the last performance evaluation conducted before the reduction in force. That Liberty retained lower-rated, similarly situated male employees eliminates one of the “most obvious explanations for the discharge” — her “lower proficiency.” Barnes v. GenCorp., Inc., 896 F.2d 1457, 1466 (6th Cir.), cert. denied, 498 U.S. 878 (1990). This evidence alone, and certainly in combination with other admissible evidence, would permit a jury reasonably to conclude that Liberty’s termination of Sullivan, “if otherwise unexplained, [is] more likely than not based on the consideration *49of impermissible factors.” Furnco Constr. Corp. v. Waters, 438 U.S. 567, 577 (1978).
Turning to Sullivan’s age discrimination claim, Sullivan relies on the following evidence to establish her prima facie case: (1) five of the six discharged attorneys were over the age of forty, including all three attorneys from the Boston office23; (2) what Sullivan terms “ageist remarks” attributed to Latronico concerning another attorney in the Boston office24; (3) the average age of attorneys in the Boston office decreased as a result of the layoffs25; (4) Latronico reassigned all of Sullivan’s cases to younger attorneys26; (5) Liberty retained “lower-rated” attorneys, who were younger than Sullivan27; and (6) other *50evidence that, Sullivan asserts, is probative of Liberty’s discriminatory animus against older attorneys.28 Weighing against her is the uncontested fact that Liberty retained three attorneys over forty years of age in the Boston office with the same job classification as Sullivan, one of whom, at fifty-eight years of age, was older than she was.
We conclude that Sullivan has established a prima facie case of age discrimination because the layoff occurred in circumstances that raise a reasonable inference of unlawful age discrimination. Liberty’s retention of lower-rated, younger attorneys in its Boston office with the same job classification as Sullivan, particularly when combined with the fact that five of the six attorneys Liberty discharged were over forty years of age, would permit a reasonable jury to conclude that Liberty’s termination of Sullivan, “if otherwise unexplained, [is] more likely than not based on the consideration of impermissible factors.” Furnco Constr. Corp. v. Waters, supra. In particular, Liberty’s retention of lower-rated, similarly situated employees who are substantially younger than Sullivan permits a rebuttable presumption of age discrimination by eliminating one of the most obvious explanations for the discharge — her “lower proficiency.” Barnes v. GenCorp., Inc., supra.
6. Employer’s burden to rebut prima facie case. Sullivan having established a prima facie case of both sex and age discrimination, the burden shifts to Liberty in the second stage of the McDonnell Douglas framework. Liberty may satisfy its burden “by articulating ‘a lawful reason or reasons for its employment decision [and] producing] credible evidence to show that the reason or reasons advanced were the real reasons.’ ” Abramian v. President & Fellows of Harvard College, 432 Mass. 107, 116 (2000), quoting Blare v. Husky Injection Molding Sys. Boston, Inc., 419 Mass. 437, 442 (1995). *51Liberty’s burden at this stage is one of production and not persuasion; Liberty “need not prove that the reasons were nondiscriminatory.” Abramian v. President & Fellows of Harvard College, supra at 117.
As to Sullivan’s age discrimination claim, “[t]he question is why, given [Liberty’s] need to reduce [its] workforce, [it] chose to discharge the older rather than the younger employee,” Thornbrough v. Columbus & Greenville R.R., 760 F.2d 633, 645 (5th Cir. 1985), abrogated on other grounds by St. Mary’s Honor Ctr. v. Hicks, 509 U.S. 502 (1993). As to Sullivan’s sex discrimination claim, Liberty similarly must explain why it chose to discharge Sullivan, a woman, rather than one of its male attorneys in her same position. Although Liberty’s burden at this stage is one of production only, Liberty “nevertheless retains an incentive to persuade the trier of fact that the employment decision was lawful. Thus, the defendant normally will attempt to prove the factual basis for its explanation.” Texas Dep’t of Community Affairs v. Burdine, 450 U.S. 248, 258 (1981). We conclude that Liberty has met its burden by explaining the circumstances that led to the reduction in force, and by producing evidence of nondiscriminatory reasons to support Sullivan’s selection for layoff. In other words, Liberty proffered admissible evidence to explain that in selecting Sullivan as one of six attorneys to lay off, it had business reasons for doing so unrelated to her age and her sex.
As in all reductions in force, Liberty was forced to choose whom to terminate from among several presumably qualified attorneys. Although Liberty had evaluated its several attorneys over the preceding years, it was under no legal requirement to lay off the attorneys with the lowest performance ratings. While Liberty could (and did) take into account the performance evaluations of each attorney, Liberty could also take into consideration other business needs that a substantial reduction in its professional staff would surely occasion. It could consider the particular expertise of an attorney, or whether two candidates for layoff had overlapping expertise. It could in short determine which of the attorneys would best meet its ongoing business needs, provided it did not make the selection for impermissible reasons. The record establishes that Liberty undertook just such *52a review and reached a conclusion that Sullivan was less suited for the needs of the new environment for reasons related to neither her age nor her sex. We summarize that evidence.
Much revolved around Liberty’s determination that Sullivan had failed to respond appropriately to new opportunities to offset declining business in the Boston office. Latronico explained that attorneys working in Liberty’s field offices generally provided litigation representation to Liberty’s clients. In some cases, however, Liberty retained outside counsel for that purpose, a decision that was, in each case, made jointly by Liberty’s regional attorney and Liberty’s claims department staff. Environmental litigation was handled differently. Liberty had established an environmental unit with its own claims staff who had “complete discretion” to refer cases to outside counsel. In fact, the staff generally referred more environmental cases to outside counsel than to Liberty’s field offices. In other words, the attorneys employed in Liberty’s field offices were competing directly with outside counsel for environmental work. As legal work in Liberty’s Boston office declined, Latronico was hopeful that the quality of its services would encourage the environmental claims staff to increase their referrals to his office: environmental cases referred to the Boston office therefore had to be handled “in an appropriate and efficient manner.”
For whatever reasons, Sullivan did not meet those expectations. By way of example, in one environmental case, Sullivan delayed for over six months a.response to a plaintiff’s motion for summary judgment despite repeated inquiries and demands from the environmental claims department. The situation was sufficiently perilous that counsel for the environmental unit, together with the adjuster for the case, met with Latronico to discuss their “concern” with Sullivan’s performance.29 Latronico reassigned the case to another attorney.
The environmental unit’s dissatisfaction with Sullivan’s performance occurred at the very moment the Boston office’s *53legal business was declining.30 By January, 1999, Liberty had imposed a hiring freeze in the field offices. As its financial circumstances deteriorated, by late 1998 or early 1999, Liberty had embarked on its “productivity and capacity” analysis of its workforce, including all of its legal staff. We have earlier summarized the details of that analysis. See note 5, supra. Here we note only that the assessment of individual attorneys by La-tronico and others31 during that exercise included a wide range of factors that are typical for a reduction in force evaluation.32
Liberty has also produced evidence explaining why it retained three male attorneys in the Boston office who had lower performance ratings than did Sullivan. One had expertise that Sullivan did not have.33 The second, also rated lower than Sullivan, was highly regarded by the claims department, in part *54because he provided excellent customer service.34 This was considered helpful for future business in the Boston office. Liberty retained a third male attorney in the Boston office rated lower than Sullivan, who was forty-four years of age at the time of the reduction in force. This attorney’s most recent evaluation noted that his “affable, straightforward nature” helped him “get along well with internal and external clients.”
After recommending that Sullivan be laid off, Latronico was required to explain his recommendation to Liberty’s human resource department. See note 31, supra. We have reviewed Latronico’s reasons given at that time,35 long before Latronico had any reason to anticipate that Sullivan would commence litigation against Liberty, in light of all of the evidence subsequently adduced during discovery. The evidence is consistent with the reasons for Sullivan’s layoff first articulated by Latronico. We agree with the motion judge who concluded that Liberty articulated a credible nondiscriminatory reason for its choice to terminate Sullivan “by demonstrating that it was engaged in a legitimate reduction in force and chose Sullivan to be terminated on the basis of a history of sub-par performance evaluations combined with recent serious complaints from clients regarding her handling of cases.”
7. Evidence of discrimination. Because Liberty has succeeded “in carrying its burden of production, the McDonnell Douglas framework — with its presumptions and burdens — is no longer relevant.” St. Mary’s Honor Ctr. v. Hicks, 509 U.S. 502, 510 (1993). The presumption of discrimination disappears. The burden returns to Sullivan to establish that the basis of Liberty’s *55decision was unlawful discrimination “by adducing evidence that the reasons given by [Liberty] for its.actions were mere pretexts to hide such discrimination.” Lewis v. Boston, 321 F.3d 207, 214 (1st Cir. 2003). “This may be accomplished by showing that the reasons advanced by [Liberty] for making the adverse decision are not true.” Abramian v. President & Fellows of Harvard College, 432 Mass. 107, 117 (2000). Once again we agree with the motion judge who concluded that Sullivan has not established with credible evidence that Liberty’s proffered reasons for her layoff were merely a pretext for a true intent of sex discrimination. We also conclude that Sullivan has failed to make a sufficient showing with respect to her age discrimination claims.
Sullivan first argues that the disproportionate impact the reduction in force had on female and older attorneys raises inferences of discrimination.36 Sullivan points to evidence, noted above, that the percentage of female attorneys and the percentage of attorneys over the age of forty declined following the reduction in force, and that after her dismissal Latronico reassigned her cases disproportionately to younger, male attorneys.37 Although this evidence helped Sullivan establish her prima facie cases, it has limited probative value at this stage. The evidence fails to rebut Liberty’s proffered reasons for laying off Sullivan and does not, by itself, create reasonable inferences of discriminatory animus and causation. See, e.g., Rummery v. Illinois Bell Tel. Co., 250 F.3d 553, 559 (7th Cir. 2001) (statistics that do not account for employer’s legitimate nondiscriminatory explanations do not establish pretext); Coleman v. Quaker Oats Co., 232 F.3d 1271, 1283 (9th Cir. 2000), cert. denied sub nom. Gentile v. Quaker Oats Co., 533 U.S. 950 (2001) (statistics that do not account for factors pertinent to selection of plaintiff for *56discharge do not alone establish pretext). Sullivan’s evidence similarly fails to eliminate other explanations for the disproportionate statistics, such as random chance (given the small discrepancies and small sample size involved here) or the actual distribution of aptitudes or expertise among attorneys of differing ages and genders both before and after the reduction in force. See, e.g., Smith v. Xerox Corp., 196 F.3d 358, 371 (2d Cir. 1999) (statistics that do not account for other possible causes of disparity do not establish pretext); Fallis v. Kerr-McGee Corp., 944 F.2d 743, 747 (10th Cir. 1991) (statistics must eliminate nondiscriminatory reasons for disparity).
Next, Sullivan attacks the process that Liberty used to select her for layoff. She argues that Liberty improperly relied on subjective and unreliable assessments, compared her to employees it should not have, and failed to compare her to employees it should have.38 Even were we to conclude that Sullivan is correct, our task is not to evaluate the soundness of Liberty’s decision making, but to ensure it does not mask discriminatory animus. See Mesnick v. General Elec. Co., 950 F.2d 816, 825 (1st Cir. 1991), cert. denied, 504 U.S. 985 (1992) (“Courts may not sit as super personnel departments, assessing the merits — or even the rationality — of employers’ nondiscriminatory business decisions”). Sullivan’s criticisms of Liberty’s selection process, which Liberty has explained in abundant detail, fail to reveal any hidden animus.
Last, Sullivan argues that Liberty’s opinions concerning her handling of the environmental case and her human relations skills were incorrect.39 Again, even were we to conclude that Sullivan neither mishandled the environmental case nor was *57responsible for the difficulties that arose with her support staff, these facts do not establish that Liberty’s stated reasons for terminating her were pretextual. There is ample, uncontroverted evidence that the negative impression Latronico, staff in the environmental unit, and others had formed of Sulhvan’s abilities was a primary reason she was selected for layoff. There is no evidence that would allow a fact finder reasonably to conclude that the environmental unit had not complained to Latronico about Sullivan’s lack of substantive legal knowledge and responsiveness. Sullivan does not challenge whether Latronico truly believed that her mishandling of the environmental case jeopardized the Boston office’s ability to retain environmental cases during a period of declining caseloads. She also fails to point to any evidence rebutting Liberty’s nondiscriminatory explanations for retaining younger, male attorneys with her job classification who had received lower performance ratings than she had. Nor does she point to any evidence suggesting that any of the attorneys to whom she believes Liberty should have compared her had performance failings comparable to those Liberty believed she had. In summary, Sulhvan’s evidence does not permit a reasonable inference that Liberty selected her for layoff for any reason other than her own performance.
Sulhvan’s proof is insufficient, as a matter of law, to show that, when she was terminated, Liberty had a discriminatory intent, motive, or state of mind based on her age or sex and that any such animus was “a material and important ingredient in the discharge.” Knight v. Avon Prods., Inc., 438 Mass. 413, 426-427 (2003). See Matthews v. Ocean Spray Cranberries, Inc., 426 Mass. 122, 127 (1997).
Judgment affirmed.
The motion judge also granted summary judgment to Liberty on a breach of contract claim asserted by Sullivan. Although Sullivan appealed from the summary judgment order in its entirety, she makes no argument concerning her breach of contract claim. We therefore consider it waived. See Mass. R. A. R 16 (a) (4), as amended, 367 Mass. 921 (1975).
We acknowledge the amicus brief filed jointly by the New England Legal Foundation and the Associated Industries of Massachusetts.
Sullivan refers to the resulting layoffs as a “so-called” or “alleged” reductian in force. “In some [reduction in force] employment discrimination cases, the courts focus their inquiry on whether a [reduction in force] actually occurred or whether the employer has claimed falsely a [reduction in force] in order to conceal age discrimination. See, e.g., Hardin v. Hussmann Corp., 45 F.3d 262, 265 (8th Cir. 1995) (concluding that a [reduction in force] did in *36fact occur); Oxman v. WLS-TV, 846 F.2d 448, 456 (7th Cir. 1988) (treating the employer’s structural reorganization as a legitimate, nondiscriminatory reason for the plaintiff’s termination).” Note, The Prima Facie Case of Age Discrimination in Reduction-in-Force Cases, 94 Mich. L. Rev. 832, 833 n.12 (1995) (Note). Here, because Sullivan fails to develop any separate argument that Liberty’s actions pertaining to her were not part of a reduction in force and because the record contains ample evidence that Liberty carried out an actual reduction in force, we will assume that Sullivan was laid off as part of a reduction in force.
The New England region included offices in Massachusetts, Connecticut, New Hampshire, and Canada. Latronico in fact held two positions: regional general attorney for Liberty’s New England region, and resident attorney for Liberty’s Boston office. The resident attorneys from Liberty’s other New England region offices reported to Latronico.
The analysis examined several factors including the volume of cases handled by each office, cases pending, cases closed, cost per case, and the cost of utilizing inside staff or outside counsel. According to Liberty, its purpose was to determine whether each office was staffed appropriately for emerging business trends.
On that day Sullivan was hand delivered a letter from Latronico stating in *37part: “As a result of a capacity study, Liberty Mutual has examined the positions and functions in the New England Region Legal Department. We regret to inform you that, as a result of this process, your employment is being terminated.”
The judge in the Superior Court concluded that Sullivan had failed to establish a prima facie case of age discrimination because she could not show “that she was replaced by a ‘substantially younger person’ or otherwise show *38that age was a determinative factor in her lay off.” On her sex discrimination claim, Liberty conceded that Sullivan had established a prima facie case, see note 13, infra, but the judge concluded Sullivan had not established “with credible evidence” that Liberty’s “proffered reasons” for her layoff were “merely a pretext meant to cover for a true intent of sex discrimination.”
In November, 2002, the judge in the Superior Court allowed Liberty’s emergency motion to strike six paragraphs of Sullivan’s affidavit filed in opposition to Liberty’s motion for summary judgment. Because we conclude that Sullivan cannot prevail on her claims even if the paragraphs are not struck, we do not reach Sullivan’s appeal from the allowance of the motion to strike. See note 19, infra.
Because Sullivan has not argued this point on appeal, her claim is waived. See Mass. R. A. P. 16 (a) (4).
A plaintiff may prove a charge of employment discrimination under one of two theories — disparate treatment or disparate impact. Sullivan’s claims are based on a theory of disparate treatment: she argues that Liberty terminated her because of her gender or her age. “A plaintiff who brings her claim under the disparate treatment theory must demonstrate that her employer intentionally treated her less favorably than others because of her race, gender, religion, national origin, or age. . . . The disparate impact theory addresses employment practices that are facially neutral but allegedly affect members of a protected class more harshly than those outside the protected class. . . . The disparate impact theory does not require proof of discriminatory motive.” Note, supra at 835 n.17, citing International Bhd. of Teamsters v. United States, 431 U.S. 324, 335 n.15 (1977).
Although not binding on this court, it “is our practice to apply Federal case law construing the Federal antidiscrimination statutes in interpreting G. L. c. 151B.” Wheatley v. American Tel. & Tel. Co., 418 Mass. 394, 397 (1994).
For an explication of the evidentiary burdens in cases where there is direct evidence of discrimination, see generally Wynn & Wynn, P.C. v. Massachusetts Comm’n Against Discrimination, 431 Mass. 655, 664-671 (2000).
In its summary judgment motion, Liberty conceded that Sullivan had established a prima facie case of sex discrimination. The concession was apparently made in light of Finney v. Madico, Inc., 42 Mass. App. Ct. 46, 50 (1997), which could be read to suggest that, in a sex discrimination case concerning a reduction in force, the plaintiff does not have to establish the fourth element of the prima facie case. Our ruling today should dispel any such misapprehension.
Consider a hypothetical case where a hospital employed five nursing supervisors, four of whom are women. In implementing a reduction in force, the hospital retained the four female nurse supervisors, but discharged the only male supervisor. The fact that the hospital retained all women nursing supervisors and discharged the only man is sufficient, by itself, to raise a reasonable inference that the hospital selected the man for discharge because of his sex. See note 16, infra.
Consider a hypothetical case where a computer company discharged one hundred of its information technology employees. Many of the terminated employees are over the age of forty. The employer retains employees ranging from ages nineteen to sixty-four years. The company’s retention of one, or *45even a few, employees younger than the terminated employee over the age of forty is, without more, insufficient to raise a reasonable inference of age discrimination.
Discovery may also give rise to probative statistical evidence. For example, evidence that a reduction in force has a disproportionate impact on members of a protected class sometimes may help establish a prima facie case of discrimination, even in a disparate treatment case. See, e.g., Lipchitz v. Raytheon Co., 434 Mass. 493, 508-09 (2001) (statistical evidence may support inference that particular decision was made because of discriminatory animus); Smith College v. Massachusetts Comm’n Against Discrimination, 376 Mass. 221, 228 n.9 (1978) (“In a proper case, gross statistical disparities alone may constitute prima facie proof of a practice of discrimination”).
The first stage of the McDonnell Douglas paradigm, the prima facie case, however, should not be the occasion for battling statisticians. Especially when, as here, the sample sizes are small, the inquiry at this threshold stage is whether a jury, guided by their collective day-to-day experiences with probability, could reasonably conclude from statistical and other indirect evidence that discrimination was more likely than not the reason for the plaintiff’s layoff. The third stage is the more appropriate stage for the employer to establish that the plaintiff’s statistical evidence is unreliable or not probative of discrimination because the statistics do not account for factors pertinent to the employer’s selection process. See discussion infra.
Liberty has produced evidence it normally would not be required to produce until after Sullivan had established her prima facie case. Some courts have suggested that in these circumstances it is not necessary to analyze a plaintiff’s prima facie showing, but to proceed directly to the third stage of the McDonnell Douglas formulation. See, e.g., Johnson v. Allyn & Bacon, Inc., 731 F.2d 64, 70 (1st Cir.), cert. denied, 469 U.S. 1018 (1984), quoting Sweeney v. Research Found. of State Univ. of N.Y., 711 F.2d 1179, 1184 (2d Cir. 1983) (“a reviewing court ‘need not linger long over the question of whether [the plaintiff] in fact established a prima facie case if the defendant has met its burden of articulating a legitimate non-discriminatory reason for its actions”). See also Rummery v. Illinois Bell Tel. Co., 250 F.3d 553, 556-557 (7th Cir. 2001) (assuming, without deciding, that plaintiff established prima facie case where employer has proffered explanation for plaintiff’s discharge). We proceed with the McDonnell Douglas analysis because it remains “a sensible, orderly way to evaluate the evidence in light of common experience as it bears on the critical question of discrimination.” Furnco Constr. Corp. v. Waters, 438 U.S. 567, 577 (1978).
Sullivan’s statement of material facts in dispute, filed in opposition to *47Liberty’s motion for summary judgment, did not comply with the requirements of Rule 9A(b)(5) of the Rules of the Superior Court (1998). The rule “is an ‘anti-ferreting’ rule designed to assist a trial judge in the all too typical situation in which the parties throw a foot-high mass of undifferentiated material at the judge,” who must then determine whether the record contains any material facts in dispute. Dziamba v. Warner & Stackpole LLP, 56 Mass. App. Ct. 397, 399 (2002). See Alsina-Ortiz v. LaBoy, 400 F.3d 77, 80-81 (1st Cir. 2005) (discussing similar Federal District Court local rule). Rule 9A(b)(5) provides a strong incentive for a party opposing summary judgment to file a compliant opposition: “For purposes of the motion for summary judgment, facts contained in [the moving party’s statement of facts] shall be deemed to have been admitted unless controverted in the manner set forth [by this rule]” (emphasis added).
Sullivan consistently failed to cite to any record evidence to support her denials, and improperly began almost all of her responses to Liberty’s 166 numbered paragraphs with identical, and often irrelevant, boilerplate language. Although she included an eleven-page, separately numbered set of thirty-nine paragraphs in which she recited her view of the facts, mostly with citations to the record, she did not cite to Liberty’s corresponding paragraphs. Sullivan’s opposition does not permit a judge to avoid “ferreting” through the record to determine whether any material facts are in dispute — precisely what rule 9A(b)(5) was promulgated to prevent. The parties disagree whether the motion judge deemed Liberty’s facts as admitted pursuant to this rule. Because we conclude that summary judgment must enter for Liberty even if we take all of Sullivan’s “facts” as disputed, we need not resolve the matter.
In an affidavit opposing Liberty’s motion for summary judgment, Sullivan asserted: (a) Latronico never asked that she mentor younger attorneys, in contrast to male attorneys, whose performance Latronico had reviewed “in terms of mentoring of younger attorneys”; (b) men were promoted to positions heading legal offices, while women were not; (c) women received lower evaluation scores than men with respect to the 1996 and 1997 evaluations, and “while males comprise only [sixty-nine per cent] of the total, they are receiving 87.5 [per cent] of the high upper third continuum ratings” in the 1999 evaluations; (d) the average length of service for women is shorter than that for men; (e) a younger male attorney hired approximately four years after Sullivan, who had the same job classification (Grade 94), was paid more than she was ($81,300 versus $79,100).
Sullivan also cites deposition testimony from a former Liberty employee, the resident attorney of the Brockton office, who opined that “women in the department should have risen to a higher level than I ever saw anyone acquire, anyone acquire that was a woman.” As noted above, portions of Sullivan’s affidavit were struck by the motion judge. See note 8, supra. Because none of the “facts” to which Sullivan points alters our decision, we do not address the question of their admissibility. See Mass. R. Civ. P. 56 (e), 365 Mass. 824 (1974).
Eighteen (or thirty-one per cent) of the fifty-eight attorneys in the offices affected by the reduction in force were women, yet fifty per cent of the six discharged attorneys were women. Taking a broader view, sixty-five per cent of the workforce, including support staff, in the affected offices were women, yet eight (or seventy-three per cent) of the eleven employees laid off were women. As a result of the layoffs, the percentage of attorneys who were female dropped from thirty-one per cent to twenty-nine per cent.
Latronico assigned eighty-nine per cent of Sullivan’s cases to five men, assigning the remaining eleven per cent to just one woman, even though twenty-nine per cent of the attorneys retained were women.
In Liberty’s final annual evaluation shortly before Sullivan’s layoff, covering the period from January 1 to December 31, 1998, Latronico gave Sullivan a performance rating of “92,” on a scale on which New England region employees had received ratings ranging from “76” to “123.” The evaluations were completed in March, 1999. Liberty retained three male attorneys in the Boston office with the same job classification (Grade 94) as Sullivan. The three men had ratings of “88,” “90,” and “91,” i.e., they were rated lower than Sullivan. Additionally, Liberty retained at least three other lower-rated male attorneys in its Andover and Brockton offices, although those attorneys were not classified as “Grade 94” attorneys.
Eighty-three per cent of the attorneys terminated in all affected offices were age forty or above, whereas fifty-six per cent of the total number of attorneys in those offices were age forty or above.
The employee, David Hartigan, like Sullivan, was forty-nine years of age when he was laid off. He testified at his deposition that, in a meeting that took place in 1998, a team leader told him that Latronico thought that Hartigan “was not adapting to the new ways of doing business at Liberty” and that he “was stuck in the old ways.” Hartigan also testified that in March, 1999, Latronico told him that he was “not adapting to the new Liberty Mutual way of doing things,” and that on April 19, 1999, Latronico told him: “[Y]ou are part of the old guard. You have never adapted to the new system at Liberty Mutual. You simply do not fit in around here anymore.”
Sullivan argues that these remarks are direct evidence that Liberty terminated her because of her age. We disagree. The remarks are facially ambiguous and may reflect managerial concerns regarding an employee who declines to adapt to changed business practices, rather than Latronico’s preference for more youthful workers. (Latronico, himself, was older than both Sullivan and Hartigan.) Even were we to infer that the remarks reflected Latroni-co’s age bias against Hartigan, any such bias cannot be extended to Sullivan. Finally, were we to accept the comments as direct evidence of discriminatory animus against Sullivan, we cannot conclude that the remarks are direct evidence of causation, i.e., that Latronico terminated Sullivan because of her age. The connection between these remarks and Sullivan’s layoff is too attenuated to constitute direct evidence that Sullivan’s layoff was tainted by unlawful age discrimination.
Following the layoffs the average age of attorneys in the Boston office dropped by one year, from 40.5 to 39.5 years.
Sullivan was forty-nine years old when she was terminated. Latronico assigned Sullivan’s cases to six other attorneys who ranged in age from thirty-one to forty-four years.
The younger attorneys were also male. See note 22, supra. Two of the attorneys had the same job classification as Sullivan and were in the Boston office; both were forty-four years of age.
The evidence includes deposition testimony from a former resident attorney of Liberty’s Brockton office in which he discussed four persons over forty years of age who were demoted or replaced by younger persons, which he said “seemed to he somewhat of a pattern.” He also stated his belief that “the age of the attorney involved” was one of the factors Latronico considered in making decisions and that Latronico had “a preference for younger attorneys.” Sullivan also asserted that Liberty paid more money to a younger employee who held Sullivan’s same job classification. See note 19, supra.
In his affidavit, environmental counsel states that he was “dissatisfied” with Sullivan’s continued delays “despite the reminders and requests,” and that he had concerns about her “substantive knowledge” in this area of the law based on her written submissions. He also expressed “dissatisfaction” with Sullivan’s “demeanor” when responding to the concerns he expressed.
Latronico explained at his deposition that the “primary” reason for selecting Sullivan as one attorney among several to be laid off “was a loss of very valuable work coming into the office at a time when the office was seeing a decline in work.” He explained that the “environmental unit was unwilling to send more environmental cases to our Boston office because they were given the impression that the lawyers who were handling our environmental work, namely Ms. Sullivan and Mr. Hartigan, were too busy to give the kind of time and attention and status reporting that they demanded and expected from the outside firms that handled the environmental work.”
Latronico collaborated with Kristin Ciotti, the legal office administrator for the New England region, on the productivity and capacity analysis and the resulting recommendations. They forwarded their recommendations to Liberty’s human resources department, which asked Latronico and Ciotti to complete an employee summary sheet for each employee they recommended for layoff. The form required Latronico and Ciotti to explain why they recommended the employee for layoff rather than up to six other comparable employees (comparators) selected for retention. Latronico and Ciotti chose the comparators for each employee, and completed the forms before implementing the reduction in force.
Among the factors were “job responsibilities” as represented by an employee’s “grade,” or job classification, “experience level,” “human relations abilities” including relationships with clients and Liberty’s claims department, “legal abilities,” “support and achievement of corporate goals,” and “corporate needs in terms of [trends] in the cases and the work coming into the office and the types of lawyers [Liberty] needed to run the business.”
The attorney, who was forty-four years old, was one of only two attorneys in Liberty’s New England region who had specific expertise in and handled longshore worker claims. In addition, Liberty asserts that his lower rating was due to personal organization issues, which Latronico judged were “less critical” than Sullivan’s interpersonal issues.
Sullivan’s employee summary sheet, see note 31, supra, indicated that Sullivan’s performance issues were “more problematic” than this attorney’s “because they relate directly to the representation of our clients, and to the atmosphere in the office.”
In the employee summary sheet he completed for Sullivan, Latronico characterized Sullivan’s legal ability as “sporadic,” noting that, despite “some good results,” she is “generally disorganized and tends to neglect her cases,” including a recent environmental case. He specifically noted that she “was not providing status reports” to the claims department, and that the environmental unit had “expressed a lack of confidence in her substantive legal abilities.” He also cited Sullivan’s “poor human relations skills, especially in regards to support staff,” and explained that Sullivan’s “treatment of her last secretary was an important factor” in that secretary’s recent resignation.
As noted earlier, see note 16, supra, statistical evidence may be used to prove intentional discrimination in a disparate treatment case such as this. See Smith v. Xerox Corp., 196 F.3d 358, 370 (2d Cir. 1999) (“In contrast to a disparate impact claim where the focus is on how a facially neutral employment practice affects a protected group, a disparate treatment claim looks at how an individual was treated compared to her similarly situated coworkers. Thus, statistical analyses that compare coworkers who competed directly against each other to receive a benefit, here selection for retention, are appropriate”).
See notes 20-21, 23, 25-26, supra.
Sullivan challenges several aspects of Liberty’s selection process: (1) Liberty relied in part on a 1998 field survey in comparing Sullivan to other attorneys, but the survey had a low response rate and Sullivan’s comparatively low rating was from a single respondent; (2) Liberty compared Sullivan to attorneys not similarly situated to her, and failed to compare her to lower-rated, similarly situated attorneys, including attorneys from other Massachusetts offices; and (3) Liberty’s selection process used subjective factors that could be influenced by discriminatory animus.
Sullivan argues that she did not mishandle the environmental case and that Liberty based its decision in part on Sullivan’s “difficulty working with support staff,” a difficulty she attributes to the support staff rather than herself. Record evidence documents problems between Sullivan and her secretaries dating back to 1994, a recurrent theme in her annual evaluations.
*57Sullivan cites deposition testimony given by several of her former colleagues that contradicts Liberty’s view of her abilities.
Sullivan also points out that Liberty’s March, 1999, evaluation of her placed her within the range of effective performance, and made no mention of the environmental unit’s concerns. Sullivan asserts that Latronico knew of the concerns before issuing her evaluation. The record does not support this assertion. The environmental unit communicated their concerns about Sullivan to Latronico at a meeting in April, 1999. The last formal evaluations preceding the reduction in force were issued on March 19, 1999, and covered performance during calendar year 1998.