Terry N. GARRETT, Plaintiff-Appellant, v. HEWLETT-PACKARD COMPANY, a California company authorized to do business in the State of Colorado, Defendant-Appellee, 305 F.3d 1210

SEYMOUR, Circuit Judge. Plaintiff Terry Garrett brought the present action against his former employer, Hewlett-Packard (HP), alleging race and age discrimination in violation of Title VII, 42 U.S.C. §§ 2000e et seq., and the Age Discrimination in Employment Act (ADEA), 29 U.S.C. §§ 621 et seq. Hewlett-Packard filed a motion seeking

summary judgment on all claims. The district court granted the motion except as to Mr. Garrett’s claim that the company failed to raise his pay for discriminatory reasons, which the parties subsequently settled. Mr. Garrett appeals, asserting the district court erred in granting

summary judgment on his claims relating to disparate treatment, retaliation, and constructive discharge. For the reasons set out below, we reverse in part and affirm in part. I In setting forth the facts, we view the evidence in the light most favorable to the non-moving party, as we must when reviewing a grant of

summary judgment. See Sprague v. Thorn Americas, Inc., 129 F.3d 1355, 1359 (10th Cir.1997) (citing Kaul v. Stephan, 83 F.3d 1208, 1212 (10th Cir.1996)). We include in our consideration information set out in the nonmov-ant’s affidavit if that information is “based on personal knowledge and set[s] forth facts that would be admissible in evidence.” Murray v. City of Sapulpa, 45 F.3d 1417, 1422 (10th Cir.1995). We do not consider “eonclusory and self-serving affidavits.” Id. Terry Garrett is an African-American man who was over forty years of age when he left HP. He joined HP in 1972 as a computer engineer.

After working for a few years at the company’s California facilities, Mr. Garrett transferred to HP’s Loveland, Colorado plant. Throughout his tenure at HP, Mr. Garrett’s performance was evaluated regularly by his supervisors. These evaluations were quite consistent in their substantive contents. Supervisors described Mr. Garrett as possessing strengths in the technical arena, a fast learner, and a hard-working, dedicated and dependable employee. Evaluators consistently praised his ease at acquiring new knowledge and adopting new skill sets.

He was described as a team player who balanced technical and social interactions well. He moved between production and development teams easily, as he did between research and customer interactions. By all accounts, Mr. Garrett’s contributions to his projects and the corporation were highly valued. Evaluations were similarly consistent in their critiques of Mr. Garrett’s performance. From the commencement of his tenure, Mr. Garrett displayed a tendency to overanalyze problems. His modus op-erandi appears to have involved checking all possible avenues for a problem prior to settling on a solution and moving on.

Supervisors commented that Mr. Garrett needed to set job priorities for more efficient use of time, and that he overplanned. Supervisors tended to recognize that the qualities critiqued were both a strength and a weakness. Mr. Garrett’s exhaustive approach made his work more reliable, but results were slower in coming. Mr. Garrett’s 1989 evaluation offers an example of this. The evaluator praised Mr. Garrett in the “Quality” category, commenting, “Terry is very thorough and strives for this in everything he does.” Aplt.App., vol. 1 at 118.

In the “Productivity” category, however, the evaluator commented, “Sometimes Terry’s thoroughness and the mis-communication of objectives ... can hurt Terry’s productivity.” Id. at 119. Despite such critiques, Mr. Garrett’s reviews were quite positive. From 1983, when HP’s evaluation forms began including rank, until 1985, Mr. Garrett was ranked as “good.” From 1986 until 1990 his rank was “very good.” Beginning in 1990, however, Mr. Garrett’s evaluations were increasingly negative. At his April 1990 evaluation, Mr. Garrett was ranked as “very good.” Id. at 134.

Six months later, his ranking dropped from “very good” to “good” and he was transferred to a different project. Id. at 138. According to the transfer evaluation, the change in rank was a result of the fact that Mr. Garrett had not accomplished as much as expected in the project to which he had been assigned. The evaluation also cited the “increasing talent” of new people in the organization as a reason for his demoted rank. Id. While an employee’s skills might not always fit the needs of a particular project, and changes in the makeup of the engineering staff could make Mr.

Garrett’s ranking in relation to his co-workers drop, the evaluation’s negative tone is new. Specifically, the memorandum suggests that Mr. Garrett “look at other possible career paths where he can take advantage of his skills.” Id. His supervisor offered no explanation as to why managers believed Mr. Garrett should consider leaving a field in which he had performed well for nearly two decades.

On April 9,1990, employees at the Love-land site formed, with HP’s endorsement, a “grassroots” group called Resource Awareness Development and Diversity (RADD), an internal diversity program aimed at promoting the hiring of people of color and fostering relationships with minority firms. Mr. Garrett took “a ‘pioneering’ role” in the group, according to one HP manager. Id. at 164. He was very active in the formation of the group and “contributed greatly to its purpose and direction.” Id. at 165.

He regularly and repeatedly encouraged supervisors to attend RADD meetings, although it appears none of his supervisors ever did so. Mr. Garrett’s active involvement in RADD continued throughout the period at issue in this suit until he resigned from HP in 1994. The period following Mr. Garrett’s engagement in diversity actions at HP coincides with the distinct change in management attitudes towards Mr. Garrett. Mr. Garrett claims that following the creation of RADD, his supervisors became increasingly critical of his work, as his evaluations reflect.

He confronted his supervisor, John Bidwell, about what he considered unequal treatment. When attempts to solve the problem appeared to be fruitless, Mr. Garrett took his complaints to higher-ups. In September 1990, Mr. Garrett met with division managers seeking a transfer away from Mr. Bidwell’s supervision. No such transfer was made at that time. As mentioned above, the following month Mr. Garrett’s ranking was downgraded and he was transferred to a new project within the company. Evaluations from his new supervisor, Chuck Heller, continued to exhibit a negative tone in marked contrast to those in Mr.

Garrett’s preceding two decades with HP. In April 1991, Mr. Heller criticized Mr. Garrett’s failure to acclimate himself to the new project quickly enough, his attitude in facing problems, and his speed at writing code. However, Mr. Heller had apparently done little to introduce Mr. Garrett to, or train him for, the new project. Mr. Garrett described Mr. Heller’s attitude during evaluations as “one of arrogance, one of trying to intimidate.” Id. at 240. Mr. Heller was, according to Mr. Garrett, hyper-critical of Mr. Garrett’s work. Despite the difficult relationship, Mr. Garrett continued to encourage Mr.

Heller to participate in RADD-organized diversity activities. In 1992, Mr. Garrett’s ranking was lowered once more, this time to “action needed” status. Mr. Garrett’s drop in status coincided with his participation in the first ever African-American celebration at the Loveland site, an event he worked very hard to have management attend. Although his rank was downgraded in February, Mr. Garrett was not informed of his drop in ranking until the 1992 review, which was prepared in late April and signed in June. When he requested an explanation of the change in status, Mr.

Heller allegedly offered no more than that Mr. Garrett “fell through the cracks.” Aplt.App., vol. 2 at 299. Following that review, Mr. Garrett sought assistance from Division Manager George Sparks. Mr. Sparks took no action until August 1992, when he transferred Mr. Garrett to yet another project under the supervision of Calvin Erickson. Mr. Garrett’s relationship with Mr. Erickson showed improvement over those with his two previous supervisors. Again, however, Mr. Erickson offered Mr. Garrett little guidance or supervision. More importantly, Mr. Garrett remained in “action needed” status. In May 1993, Mr.

Garrett took his concerns to the Loveland Personnel Manager, Shari Vines, as well as to Larry DesJardin, a supervisor of Mr. Erickson. Mr. Garrett communicated his frustration at being left in “action needed” status, the failure of management to evaluate his performance fairly, and the need for a more supportive environment for minority employees. In June 1993, nearly a year after the transfer, Mr. Erickson evaluated Mr. Garrett. The evaluation abandoned the harsh tone used by supervisors since 1990 and raised Mr. Garrett’s rank back to “band 2” (equivalent to “good”). Id., vol. 1 at 177. All told, Mr.

Garrett spent 15 months in “action needed” status, three months longer than HP policy allows. Id., vol. 2 at 300. Mr. Garrett continued to face obstacles in his work, however. Mr. Garrett’s requests for a computer with greater capacity and other work equipment similar to those provided to other engineers were refused. He often did not have access to materials necessary for completing projects rapidly and his performance suffered as a result. Throughout this period, Mr. Garrett’s activities on behalf of RADD continued. In September 1993, when Mr. Garrett spoke with Mr. Erickson about some RADD matters, Mr.

Erickson replied that he and the other managers who were in the ranking session “were tired of hearing about ‘that diversity stuff.’ ” Id. at 301. In an attempt to solve what he viewed as unequal and unfair treatment, Mr. Garrett began mediation with George Sparks. These sessions failed to bear fruit. In the course of mediation, Mr. Sparks admitted the evaluation and ranking system was wholly subjective. Id. at 302. Mr. Garrett abandoned mediation because he came to believe that Mr. Sparks was more interested in avoiding discrimination litigation than actually solving the problems Mr. Garrett presented. Mr.

Garrett filed his first complaint with the EEOC in February 1994. In April, management informed him that he was being transferred to a new position described as a “Software Integration Specialist,” a position that would not involve a demotion or a cut in pay. Mr. Garrett sought more information about the position, but his queries were ignored. Convinced that he was being set up to fail in this new position, Mr. Garrett resigned from HP on May 2, 1994. He filed a second charge of discrimination with the EEOC in September 1994. After he received a right to sue letter, he filed the complaint in this action.

After many months of discovery, HP filed a motion for

summary judgment. II A. Standard of Review We review the district court’s grant of

summary judgment de novo, applying the same legal standard used by the district court.

Summary judgment is appropriate “if the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue "as to any material fact and that the moving party is entitled to a judgment as a matter of law.” Fed. R.Civ.P. 56(c). When applying this standard, we view the evidence and draw reasonable inferences therefrom in the light most favorable to the nonmov-ing party. English v. Colorado Dep’t of Corr., 248 F.3d 1002, 1007 (10th Cir.2001) (quoting Simms v. Oklahoma ex rel.

Dep’t of Mental Health & Substance Abuse Servs., 165 F.3d 1321, 1326 (10th Cir.1999)). In cases such as this, “ ‘where the nonmoving party will bear the burden of proof at trial on a dispositive issue’ that party must ‘go beyond the pleadings’ and ‘designate specific facts’ so as to ‘make a showing sufficient to establish the existence of an element essential to that party’s case’ in order to survive

summary judgment.” McKnight v. Kimberly Clark Coiy., 149 F.3d 1125, 1128 (10th Cir.1998) (quoting Celotex Corp. v. Catrett, 477 U.S. 317, 322, 106 S.Ct. 2548, 91 L.Ed.2d 265 (1986)). Mr. Garrett does not rely on direct evidence of discrimination in making his case. Rather, he points to circumstantial evidence to demonstrate that HP’s actions were discriminatory. In cases brought under Title VII and the ADEA where circumstantial evidence is the basis for the claim, our analysis at the

summary judgment stage is governed by the burden-shifting framework laid out in McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S.Ct. 1817, 36 L,Ed.2d 668 (1973). See, e.g., Amro v. Boeing Co., 232 F.3d 790, 796 (10th Cir.2000); Beaird v. Seagate Tech., Inc., 145 F.3d 1159, 1165 (10th Cir.1998). The McDonnell Douglas test involves a three-step analysis. First, the plaintiff must prove a prima facie case of discrimination. If the plaintiff satisfies the prima facie requirements, the defendant bears the burden of producing a legitimate, nondiscriminatory reason for its action.

If the defendant does so, the plaintiff must either show that his race, age, gender, or other illegal consideration was a determinative factor in the defendant’s employment decision, or show that the defendant’s explanation for its action was merely pretext. See Kendrick v. Penske Transp. Servs., Inc., 220 F.3d 1220, 1226 (10th Cir.2000). On appeal, Mr. Garrett makes three separate claims of discrimination. He maintains that HP subjected him to discriminatory adverse employment actions on the basis of race and age, retaliated against him for engaging in protected conduct, and constructively discharged him.

We consider each claim separately. B. Adverse employment action In his complaint, Mr. Garrett maintained that HP’s treatment of him with respect to compensation, terms, conditions, and privileges of employment violated 42 U.S.C. § 2000e-2(a)(

l) and 29 U.S.C. § 623(a)(1). Specifically, Mr. Garrett pointed to HP’s failure to raise his pay, its evaluation of his performance, its ranking of his performance, its repeated transferring of him from one project to another, its failure to provide equipment necessary to proper performance of his job, and its failure to provide training necessary to proper performance and advancement in his job. The district court refused to grant

summary judgment on the pay raise portion of his complaint, but otherwise granted

summary judgment on Mr. Garrett’s claims brought under the above-mentioned statutes. As we have mentioned, HP and Mr. Garrett then settled the pay claims. See ApltApp., vol. 2 at 609-21. Mr. Garrett appeals the grant of

summary judgment only as to HP’s ranking of his performance in relation to his fellow engineers. HP does not contest that Mr. Garrett’s claims constitute an adverse employment action. Nor does it contest that Mr. Garrett has satisfied the requirements necessary to make a prima facie case of discrimination. Likewise, Mr. Garrett does not contend that HP failed to offer legitimate, nondiscriminatory reasons for its actions. Rather, he argues that the reasons offered by HP are pretextual. A plaintiff can withstand

summary judgment by presenting evidence sufficient to raise a genuine dispute of material fact regarding whether the defendant’s articulated reason for the adverse employment action is pretextual. See Reeves v. Sanderson Plumbing Prods., Inc., 530 U.S. 133, 147-49, 120 S.Ct. 2097, 147 L.Ed.2d 105 (2000).

A plaintiff can show pretext by revealing “weaknesses, implausibilities, inconsistencies, incoherencies, or contradictions in the employer’s proffered legitimate reasons for its action [such] that a reasonable factfinder could rationally find them unworthy of credence and hence infer that the employer did not act for the asserted non-discriminatory reason.” Morgan v. Hilti, Inc., 108 F.3d 1319, 1323 (10th Cir.1997) (citations and quotations omitted).

More specifically, evidence of pretext may include, but is not limited to, the following: “prior treatment of plaintiff; the employer’s policy and practice regarding minority employment (including statistical data); disturbing procedural irregularities (e.g., falsifying or manipulating ... criteria); and the use of subjective criteria.” Simms, 165 F.3d at 1328 (citing Colon-Sanchez v. Marsh, 733 F.2d 78, 81 (10th Cir.1984); and Beaird, 145 F.3d at 1165).

Our review of the record reflects ample evidence of inconsistent treatment of plaintiff, disturbing procedural irregularities, and the use of subjective criteria. HP offers a number of arguments as to why Mr. Garrett’s pretext argument should fail. First, HP maintained in its brief in support of motion for

summary judgment that employee rankings (which, inter alia, determine salary) are determined by a system that Mr. Garrett did not allege was discriminatory. HP claims rankings are determined at a meeting of all supervisors who supervise employees who are paid on the same pay curve, “for example, all level sixty engineers who perform the same or similar work in a particular division.” ApltApp., vol. 1 at 69. Engineers are ranked from best to worst. HP contended in its brief below that in order to show pretext, “Mr.

Garrett would have to prove that all supervisors who contributed to the ranking process at any given time ranked him as they did on the basis of his race or age.” Id. (emphasis in original). The district court agreed. HP offered virtually no evidence to support its characterization of its ranking system. It presented no set of objective criteria by which employees are differentiated. In fact, nowhere in the record is it shown how rankings were determined in these meetings.

Contrary to the district court’s conclusion, on this record a jury could reasonably find that in the ranking process, the opinion of an employee’s supervisor was the basis on which rank is established, particularly when the record reflects that rank is so closely linked to the evaluation process. Courts view with skepticism subjective evaluation methods such as the one here. See Simms, 165 F.3d at 1328 (“Evidence of pretext may include .... the use of subjective criteria.”); Bauer v.

Bailar, 647 F.2d 1037, 1046 (10th Cir.1981) (although subjective criteria not wrongful per se, “[ojbvious subjective decision making provides an opportunity for unlawful discrimination.”); see also, e.g., Bergene v. Salt River Project Agric. Improvement & Power Dist., 272 F.3d 1136, 1142 (9th Cir.2001) (subjective criteria evidence of pretext); Lindsey v. Prive Corp., 987 F.2d 324, 327 (5th Cir.1993) (citing with approval Burrus v. United Tele. Co. of Kansas, Inc., 683 F.2d 339, 342 (10th Cir.1982)) (subjective employment criteria may provide opportunities for unlawful discrimination); Bell v.

Bolger, 708 F.2d 1312, 1319-20 (8th Cir.1983) (“subjective promotion procedures are to be closely scrutinized because of their susceptibility to diserimi-natory abuse.”). HP does not contest Mr. Garrett’s allegation that Division Manager George Sparks admitted the ranking and evaluation system was wholly subjective. Mr. Garrett reinforced this assertion with expert opinion that HP’s system is disturbingly lacking in objective standards, means of guidance for poorly evaluated and ranked workers, and oversight by HP executives. Aplt.App., vol. 2 at 441-44.

Absent evidence that HP’s system of ranking and evaluation relies on objective criteria, we hold that Mr. Garrett has satisfied his burden to demonstrate pretext under the third prong of McDonnell Douglas for the purposes of avoiding

summary judgment. Next, HP asserts that Mr. Garrett’s allegedly poor performance is a legitimate, nondiscriminatory reason for differential treatment. While actual performance may constitute a legitimate basis for different treatment, see Furr v. Seagate Tech., Inc., 82 F.3d 980, 988 (10th Cir.1996), on this record there is a fact question regarding whether Mr. Garrett’s performance was the true basis for his poor evaluations and rankings. Mr. Garrett does not deny that in the twenty years he worked at HP, some supervisors noted concerns about efficiency and productivity. However, the mere fact that Mr.

Garrett’s evaluations bear evidence of past criticism of his work habits does not negate the possibility that the justifications given for Mr. Garrett’s drop in rank and negative evaluations between 1990 and 1994 are pretextual. A jury could reasonably infer that Mr. Garrett’s supervisors discriminated against him by inflating and exaggerating longstanding critiques of his performance as a means of exercising racist and ageist animus towards him. The inconsistencies and contradictions noted above between Mr.

Garrett’s 1972 to 1989 evaluations and those from 1990 to 1993 raise disputed issues of fact as to whether Mr. Garrett’s evaluations and ranking were truly the reason for the supervisors’ actions towards their employee. See Morgan, 108 F.3d at 1323. We believe a reasonable factfinder could rationally find HP’s proffered reasons unworthy of credence. Id. HP further contends that Mr. Garrett’s claim relating to his ranking in the “action needed” category must fail because he does not offer comparative data as to others who were placed in that category. In fact, Mr. Garrett did produce such evidence. Mr.

Garrett contends another engineer in his classification, referred to in the record as JD-13, a younger, white worker hired in 1988, was similarly situated but treated preferentially. HP counters by pointing out that Mr. Garrett “admits” evaluations of JD-13 demonstrate that he, too, had performance problems and that Mr. Heller “went out of his way to criticize” JD-13’s work. Aple. Br. at 31. The record demonstrates that this is true. In fact, the discrete criticisms of JD-13 read in ways quite similar to those of Mr. Garrett. For example, JD-13 “lets details dominate objectives.” Aplt.App., vol. 2 at 392.

JD-13 “needs to be sure that his work is of bullet proof quality.” Id. at 409. When read as a whole, however, the two sets of evaluations are a study in contrasts. Mr. Heller, who evaluated JD-13 between 1988 and 1993, expressed great hope for his employee despite his faults. In his evaluations of Mr. Garrett, similar faults became grounds for suggesting that the employee leave the company and seek another career. Even more telling are the rankings. Mr.

Garrett’s rank sank from “very good,” to “good,” to “action needed” between 1989 and 1993, while JD-13’s rose from “good” to “very good.” The difference in their evaluations and ranks supports Mr. Garrett’s assertion of discrimination. Based on these differences, we hold Mr. Garrett has raised a genuine dispute of material fact that HP’s legitimate nondiscriminatory reasons are pretextual. Mr. Garrett’s pretext argument regarding the “action needed” ranking focuses on procedural irregularities in the ranking process. In Mohammed v.

Callaway, 698 F.2d 395, 401 (10th Cir.1983), we held that “disturbing procedural irregularities” can satisfy the requirements of a pretext claim. Here, the HP ranking committee placed Mr. Garrett in the “action needed” category, the lowest possible category allowed without being terminated, in February 1992. Mr. Garrett was not informed of his lowered rank until two months later, giving him neither notice nor the opportunity to attend to the factors that allegedly caused his placement in that category. Mr.

Garrett then stayed in “action needed” status until May 1993, notwithstanding an HP policy that only allows an employee to be ranked in the “action needed” category for twelve months. These deviations from normal company procedure, especially when seen in light of the other circumstances described above, provide support for Mr. Garrett’s pretext argument. We conclude Mr. Garrett has met the requirements necessary to a showing of pretext. When viewed in the aggregate, his proffered evidence is “sufficient to raise a genuine doubt about Defendant’s motivation.... ” EEOC v.

Horizon/CMS Healthcare Corp., 220 F.3d 1184, 1200 (10th Cir.2000). C. Retaliation Mr. Garrett also appeals the district court’s grant of

summary judgment to HP on his claim of retaliation. “A plaintiff establishes a prima facie ease of retaliation by showing: (1) he or she is engaged in protected opposition to discrimination; (2) he or she was subject to adverse employment action; and (3) a causal connection exists between the protected activity and the adverse action.” Kendrick, 220 F.3d at 1234; see also Amro, 232 F.3d at 799. HP contends Mr. Garrett has failed to establish the second and third elements necessary to a prima facie retaliation claim. HP asserts that Mr.

Garrett failed to satisfy the third element of the prima facie claim: a causal connection between his protected activities and the adverse action. HP contends Mr. Garrett has made “no showing of any relationship” between his participation in RADD and the adverse employment actions. We disagree. Mr. Garrett’s active involvement with RADD began in March 1990. In September 1990, Mr. Garrett met with division managers and accused his supervisor, Mr. Bidwell, of not evaluating his work fairly. Mr. Garrett asserted that he stated at the meeting that he believed his race and diversity activities were factors in Mr.

Bidwell’s treatment. Three weeks later, Mr. Bidwell transferred Mr. Garrett to a different project, and suggested he pursue other career paths. Mr. Garrett was informed at this meeting that his rank had been lowered from “very good” to “good.” Mr. Garrett continued his work with RADD, approaching his new supervisor, Charles Heller, numerous times about participating in the group’s activities, specifically an African-American history celebration in February 1992. That same month, Mr. Garrett’s was placed in “action needed” status. He was not informed of his demoted ranking until two months later. Mr.

Heller offered Mr. Garrett no assistance in finding ways to improve his performance or ranking. In June, Mr. Garrett was transferred away from Mr. Heller’s supervision. A causal connection is established where the plaintiff presents “ ‘evidence of circumstances that justify an inference of retaliatory motive, such as protected conduct closely followed by adverse action.’ ” Bullington v. United Air Lines, Inc., 186 F.3d 1301, 1320 (10th Cir.1999) (quoting Burns v. United Tel. Co. of Kansas, Inc., 683 F.2d 339, 343 (10th Cir.1982)). Relying on our decision in Murray, HP contends Mr.

Garrett has presented insufficient evidence to raise an inference of a causal connection. Murray is distinguishable for a number of reasons. First, that decision dealt with the pretext prong of the McDonnell Douglas formula, Murray, 45 F.3d at 1422, while HP is asserting Mr. Garrett failed to meet prima facie standard for retahation. The standard for proving a prima facie case, however, is low.

The critical inquiry is “whether the plaintiff has demonstrated that the adverse employment action occurred ‘under circumstances which give rise to an inference of unlawful discrimination.’ ” Kendrick, 220 F.3d at 1227 (quoting Texas Dep’t of Community Affairs v. Burdine, 450 U.S. 248, 253-54, 101 S.Ct. 1089, 67 L.Ed.2d 207 (1981)). Thus, at this stage Mr. Garrett need not overcome HP’s proffered legitimate, nondiscriminatory reasons. Second, in Murray, the plaintiff had only one possible “connection”: the fact that he filed complaints with the EEOC and was subsequently fired. Here, Mr.

Garrett has presented not only evidence of temporal connections, but he has also documented evidence of a marked shift in the attitudes and treatment of him by his supervisors. The evidence presented by Mr. Garrett satisfies the requirements necessary to make a prima facie case of retaliation. Because the district court grounded its decision in its belief that Mr. Garrett failed to raise a genuine issue of material fact as to whether HP’s legitimate, nondiscriminatory explanations for its actions were pre-textual, we reach this matter as well. As with his disparate treatment claim, Mr.

Garrett points to procedural irregularities and the use of subjective criteria in placing him in “action needed” status. For the reasons stated supra relating to the disparate treatment claim, we believe there is a genuine issue of material fact as to Mr. Garrett’s retaliation claim. D. Constructive Discharge Finally, Mr. Garrett appeals the district court’s grant of

summary judgment to defendants on his constructive discharge claim. In support of this claim, Mr. Garrett points to an extended period of intimidating behavior by supervisors that caused him to suffer lower performance evaluations as well as repeated denials of his requests to transfer. Constructive discharge occurs “when ‘the employer by its illegal discriminatory acts has made working conditions so difficult that a reasonable person in the employee’s position would feel compelled to resign.’ ” Sanchez v. Denver Public Schools, 164 F.3d 527, 534 (10th Cir.1998) (quoting Derr v.

Gulf Oil Corp., 796 F.2d 340, 344 (10th Cir.1986)); see also Bolden v. PRC, Inc., 43 F.3d 545, 552 (10th Cir.1994). The bar is quite high in such cases: a plaintiff must show he had no other choice but to quit. Sanchez, 164 F.3d at 534. While we do not adopt the district court’s reasoning here for granting

summary judgment on behalf of HP, we do agree with the result. We are not convinced the facts alleged by Mr. Garrett raise an inference that he was left with no choice but to resign when he did. This is especially so because Mr. Garrett resigned before he had complete details as to the position into which HP was in the process of transferring him. Ill At the

summary judgment stage, “the judge’s function is not himself to weigh the evidence and determine the truth of the matter but to determine whether there is a genuine issue for trial.” Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 249, 106 S.Ct. 2505, 91 L.Ed.2d 202 (1986). Mr. Garrett has demonstrated that there are genuine issues of material fact warranting a trial in this case. For the reasons set out above, we REVERSE the district court’s grant of

summary judgment to HP on Mr. Garrett’s claims of disparate treatment under Title VII and the ADEA and retaliation under Title VII. We AFFIRM the district court’s grant of

summary judgment to HP on Mr. Garrett’s claim of constructive discharge. . At the time he was hired by HP in 1972, Mr. Garrett possessed a Bachelor of Science degree in electrical engineering, and a Master of Science degree in material science, both from the University of California at Berkeley. . HP’s ranking system changed a number of times over the course of the period relevant to this case. From 1989 to 1991, the evaluation form used a six-point scale wherein 1 was exceptional, 2 excellent, 3 very good, 4 good, 5 acceptable, and 6 unacceptable.

In 1992, the evaluation form employed a five-point scale wherein 4 was the highest (no adjective attached), 1 the lowest, and below 1 was "action needed.” In 1993, the rating system changed again. 5 was now the highest, 1 was the "action needed” category. Below 1 was "unacceptable.” "New to Job" was another possible rank, apparently outside the categories above it. See Aplt.App., vol. 1 at 122, 134, 145, 157, 177. . At the hearing on defendant's motion for

summary judgment, the magistrate judge specifically requested further explanation of the system from HP's counsel. See Aplt.App., vol. 3 at 3. . Taking a similar tack to its argument relating to the ranking system, HP maintains that to make his case, Mr. Garrett "must show that all of the evaluating supervisors who criticized his productivity between 1972 and 1994 did so on the basis of race or age.” Aple. Br. at 18. As with its ranking argument, HP offers no legal basis for this expansive claim. .In its brief, HP repeatedly suggests that Mr.

Garrett’s pretext claim fails because he does not offer any "direct” evidence of discrimination (i.e. racist comments, ageist drawings, or the like). Such evidence is not necessary to prove a discrimination claim under Title VII or the ADEA. See, e.g., Kendrick v. Penske Transp.

Servs., Inc., 220 F.3d 1220, 1225 (10th Cir.2000) ("A plaintiff alleging discrimination ... may prove intentional discrimination through either direct evidence of discrimination (e.g., oral or written statements on the part of a defendant showing a discriminatory motivation) or indirect (i.e., circumstantial) evidence of discrimination.”); see also Perry v. Woodward, 199 F.3d 1126, 1135 (10th Cir.1999). Indeed, the burden-shifting structure, including prong three, pretext, is precisely the means by which a plaintiff may prove discrimination with circumstantial rather than direct evidence. See id. .

A notable difference between the two evaluations is that neither seems to account for the amount of time the employee had spent on the project for which he was evaluated. During the relevant time period, JD-13 worked in only one project while Mr. Garrett worked in four different projects. Mr. Garrett's expert witness noted his concern with the lack of attention to such differences in experience and familiarity with the assigned project. Aplt.App., vol. 2, at 444. . As it did in its argument on the disparate treatment claims, HP contends that because ranking is determined by all supervisors, Mr.

Garrett must show that all supervisors had retaliatory motives in ranking him as they did. We reject this argument for the reasons set out above. . HP finds Mr. Garrett’s allegation that his involvement in RADD led his supervisors to retaliate against him "most surprising" because, they assert, "Hewlett-Packard created, supported, promoted, and maintained the RADD program....” Aple. Br. at 28 (emphasis in original). Even if this is so, the mere fact that some in the upper echelons of a company support a program does not necessarily mean everyone in middle management does as well. .

Our holding here relates only to Mr. Garrett’s retaliation claims under Title VII. He has presented, and the record contains, no evidence as to protected activities relating to age discrimination.