Case Information

*1 Before HENRY, HOLLOWAY, and MURPHY, Circuit Judges.

HOLLOWAY, Circuit Judge.

Plaintiffs/appellants Irene Tomsic and Sheila Browning appeal the district court’s grant of summary judgment for defendant/appellee State Farm on their Title VII gender discrimination claims. ^[1] Tomsic v. State Farm Mutual Automobile Ins. Co., 870 F. Supp. 318 (D. Utah 1994). We have jurisdiction under 28 U.S.C. § 1291.

Plaintiffs were hired on April 1, 1991, as trainee agents of defendant. ^[2] Employment as a trainee agent is temporary, lasting at most two years. Employees who successfully complete the training program become agents with the defendant under a contractual relationship but are no longer employees. Plaintiffs were terminated after about nine months in the program. A third trainee who had been hired at the same time as plaintiffs, a male named Toby Gonzales, was retained and successfully completed the program.

Under the terms of the employment contract, defendant was obligated to assign a management employee to provide training and supervision to plaintiffs. The supervisor assigned to train plaintiffs during the first three months of their employment was Dean Olsen. Olsen testified in deposition that plaintiffs were committed to the program, had good *3 attitudes, and showed professionalism in dealing with clients.

After three months, plaintiffs were assigned a new supervisor, Norman Miller, who had been supervising Toby Gonzales from the beginning of his employment. Miller devoted more of his attention to Gonzales than to plaintiffs, according to Browning’s deposition. Miller was supposed to complete weekly progress reports on plaintiffs as well as evaluations every three months. Miller prepared the evaluation for the initial three months of plaintiffs’ employment, even though he had not been their supervisor during that period, and did so without the input of Olsen. Later, Miller sometimes falsified the weekly progress reports by claiming to have spent more time in training plaintiffs than he actually had, the plaintiffs said in their depositions. Miller’s reports on Browning were conflicting; one week she would be rated very highly in all areas, while the next report would say she had a bad attitude and wasn’t a team player. In September 1991, Browning won a contest among both trainees and experienced agents by submitting the highest number of life insurance applications in her district.

Within a month after becoming plaintiffs’ supervisor, Miller told plaintiff Tomsic that he did not think she would succeed; he said her husband made too much money and that she therefore would lack incentive. He asked about Tomsic’s marital relationship and whether her husband had ever had an affair. At about the same time, Miller told plaintiff Browning and Toby Gonzalez that Tomsic would be fired and one of them could have her office. Two weeks later he again told Browning that Tomsic would be fired.

On December 16, 1991, in a tape recorded conversation, Miller told Browning that she had a great future with State Farm but that production was not the most important evaluation factor. He said he was concerned that she devoted so much of her time to her career and that her personal and family life might suffer. He also expressed concern that marital problems might arise in a few years because she would likely be earning more than her husband. Miller later testified that his purpose in this interview, on instructions from his superior, was to try to convince Browning to resign voluntarily.

Three days later the decision was made to fire both plaintiffs, but apparently the decision was not acted on immediately. Plaintiffs were asked to resign in January. Tomsic resigned in late January. Browning refused to resign. On February 26, 1992, a letter was mailed notifying Browning that she was terminated as of February 29. She turned in a resignation on February 28.

The decision to fire plaintiffs was made by Mr. Lee Baumann, after receiving the comments and recommendations of Dave Powell and Norm Miller. Brief of Appellee at 15; Defendants’ Answer, Aplt. App. at 24. Miller did not have the actual authority to terminate a trainee. Gonzales was retained, even though his production was not significantly better than plaintiffs’, according to their deposition testimony, and poor production was ostensibly the primary reason for plaintiffs’ termination.

The District Court’s Memorandum Decision and Order

The district court first analyzed the case according to the four elements of a prima facie case under McDonnell Douglas Corp. v. Green, 411 U.S. 792 (1973). ^[3] The court found that plaintiffs had shown they were members of a protected class and that issues of fact existed with regard to the second and third elements. The court held that plaintiffs’ claims failed because they could not satisfy the fourth element, which the court viewed as requiring plaintiffs to show that State Farm had replaced them with individuals who were not members of the protected class. Because State Farm had neither sought nor hired replacements, plaintiffs could not establish a prima facie case, the judge concluded.

The district court then considered whether plaintiffs had shown direct evidence of discrimination which would make it unnecessary for them to follow the prima facie case paradigm of McDonnell Douglas . The court found that plaintiffs had not presented direct evidence of discriminatory intent. The statements by Miller to plaintiffs regarding their marital relationships alluded to “arcane” notions of women’s role in society, but it did not appear that Miller himself held such views. Rather, as to the statement regarding Browning and the possibility that she might make more money than her husband, the judge said that Miller’s concern was that such views, if held by others, might create an obstacle to plaintiff Browning’s success with State Farm.

*6 As to the statement regarding Tomsic’s possible lack of motivation because of her husband’s substantial earnings, the judge said that State Farm had a legitimate concern with its trainees’ incentive; it would not be in State Farm’s interest to expend resources training one whose only motivation was to “earn pocket change.” 870 F. Supp. at 324. Further, the judge concluded that plaintiffs had failed to show that Miller’s statements had anything to do with the decision to fire plaintiffs. At most, the judge concluded, plaintiffs could argue that discriminatory intent could be inferred from the statements, but this would not be direct evidence of discrimination. Id. (citing Heim v. State of Utah, 8 F.3d 1541, 1547 (10th Cir. 1993)).

Somewhat less clearly, the district judge also concluded that plaintiffs had failed to offer sufficient evidence that defendant’s proffered reasons for discharging plaintiffs were a mere pretext for discrimination. Id. The judge’s conclusion in this portion of his memorandum follows immediately from his conclusion that Miller’s remarks did not constitute direct evidence of discrimination; it is unclear whether he actually analyzed the issue of whether sufficient evidence of pretext had been submitted.

III

A We review a grant of summary judgment de novo , applying the same standard as the district court under Fed. R. Civ. P. 56(c). Universal Money Centers, Inc. v. A.T.& T, 22 F.3d 1527, 1529 (10th Cir.), cert. denied, 115 S.Ct. 655 (1994). Summary judgment is appropriate *7 if “there is no genuine issue as to any material fact and . . . the moving party is entitled to judgment as a matter of law.” Fed. R. Civ. P. 56(c). We examine the factual record and reasonable inferences therefrom in the light most favorable to the nonmoving party. Id. If there is no genuine issue of material fact in dispute, we must determine whether the district court correctly applied the law. Applied Genetics Internat’l, Inc. v. First Affiliated Securities, Inc., 912 F.2d 1238, 1241 (10th Cir. 1990).

Plaintiffs contend that the district judge’s conclusion that they had failed to satisfy the fourth prong of the McDonnell Douglas prima facie case test was erroneous because that prong as stated in that case is simply inapplicable to the facts and circumstance of this case, which plaintiffs analogize to a failure to promote case. Plaintiffs contend that the particular structure of their employment, in which State Farm does not replace trainee agents who fail, should not insulate State Farm from liability if it unlawfully discriminates in its decisions. Plaintiffs argue that the district court failed to recognize that the McDonnell Douglas structure was never intended to be applied rigidly but instead should be adapted to the varying circumstances presented in particular cases.

We agree that the district court erred in applying the McDonnell Douglas framework without modification to suit the circumstances of this particular controversy. It is not necessary that plaintiffs’ claims fit neatly into a pre-established category such as wrongful termination or failure to promote. We have previously noted the necessity of adapting the *8 framework in the light of the facts presented. In Cole v. Ruidoso Municipal Schools, 43 F.3d 1373, 1380 (10th Cir. 1994), we described the fourth prong of the prima facie case standard as merely requiring a showing that the plaintiff was “treated less favorably than her male counterparts.” ^[4] In McDonnell Douglas itself the Court cautioned that the framework for analysis would require tailoring for the particular facts of individual cases: “The facts necessarily will vary in Title VII cases, and the specification . . . of the prima facie proof required . . . is not necessarily applicable in every respect to differing factual situations.” 411 U.S. at 802 n. 13. This admonition has since been illustrated by the Court’s approach to cases such as Hishon v. King & Spaulding , 467 U.S. 69 (1984), in which the plaintiff brought a claim under Title VII based on the failure of the defendant law firm to recommend her for partnership. Chief Justice Burger, writing for a unanimous Court, there said:

The benefit a plaintiff is denied need not be employment to fall within Title VII’s protection; it need only be a term, condition, or privilege of employment. It is also of no consequence that employment as an associate necessarily ends when an associate becomes a partner. A benefit need not accrue before a person’s employment is completed to be a term, condition, or privilege of that employment relationship. . . . . Accordingly, nothing in the change in status that advancement to partnership might entail means that partnership consideration falls outside the terms of the statute.

467 U.S. at 77 (emphasis in original).

We believe that plaintiffs’ position as trainees aspiring to complete their limited *9 period of employment and to be rewarded thereafter with an offer to enter into a continuing contractual relationship with State Farm as independent contractor agents is closely analogous to the position of the plaintiff in Hishon. Hence, we hold that plaintiffs presented sufficient evidence on the fourth element of a prima facie case, to avoid an adverse summary judgment, by proof that they were denied that opportunity while a similarly situated male, Gonzales, was given the opportunity. Thus, the district court erred in holding that plaintiffs had failed to establish a prima facie case.

Defendant contends, inter alia , that plaintiffs did not raise the failure to promote theory in the district court. Plaintiffs did, however, raise this theory at argument on defendant’s motion, Aplt. App. at 261-63, without objection from defendant, and defense counsel addressed the merits of the promotion argument, id. at 267-69. Thus, this issue is properly before us. More fundamentally, plaintiffs assert that their argument is not that this is a failure to promote case per se , but that analogy to failure to promote cases is instructive of the case-sensitive approach to be taken in assessing what a Title VII plaintiff must show to establish a prima facie case. Thus, we perceive plaintiffs’ argument to be that under the particular circumstances here, their terminations must be analyzed with a flexible, rather than rigid, application of the McDonnell Douglas prima facie case standard, which takes into account the fact that their employment was of such a nature that replacements could not expect to be hired in the normal course of defendant’s business. We agree with plaintiffs on this point and hold that the district court erred in concluding that they failed to meet the *10 prima facie case standard.

Plaintiffs also contend that summary judgment was improper, even if they had failed to satisfy the prima facie case element of the McDonnell Douglas analysis, because they presented sufficient direct evidence of discriminatory motive in their terminations. We believe that we should resolve this issue, in spite of our resolution of the previous issue in plaintiffs’ favor, because if plaintiffs prevail on this issue it would affect the course of proceedings on remand. We accept plaintiffs’ legal premise, as we have previously noted that direct evidence of discrimination makes the McDonnell Douglas approach unnecessary. See, e.g., Furr v. AT&T Technologies, Inc., 824 F.2d 1537, 1549 (10th Cir. 1987). We do not agree, however, that plaintiffs’ proof constitutes direct evidence.

A plaintiff in an employment discrimination case proves discrimination by direct evidence when she presents proof of “an existing policy which itself constitutes discrimination . . . .” Ramsey v. City and County of Denver, 907 F.2d 1004, 1008 (10th Cir. 1990) (citing Trans World Airlines, Inc. v. Thurston , 469 U.S. 111, 121 (1985)), cert. denied, 506 U.S. 907(1992). Statements which on their face are expressions of personal opinion, however, can only support an inference of discrimination if the trier of fact finds the inference reasonable, and so constitute only circumstantial or indirect evidence of discrimination against the plaintiff. Id.

In contending that they presented direct evidence of discrimination, plaintiffs rely on *11 the remarks made by the supervisor, Miller, to the effect that Tomsic could be expected to lack motivation because her husband earned a substantial income, while Browning could be expected to encounter problems in her marital relationship, which might adversely affect her performance as an agent, because her husband would probably become uncomfortable with her earning more than he in future years. Initially, we note the fundamental discrepancy in defendant’s position regarding Browning. She purportedly was terminated for poor performance, yet Miller’s remarks to her suggest that he must have believed she eventually could become quite successful. We also note that the district judge appears to have impermissibly drawn inferences from this evidence in favor of State Farm, as he concluded that these remarks did not actually represent Miller’s personal views but only those of others with whom Browning might have to contend. It is fundamental that the jury must be permitted to decide from among the different inferences which might be drawn from the evidence.

As to Miller’s remarks to Tomsic, the district judge concluded that Miller was only expressing a legitimate concern of State Farm that it would not be profitable to expend company resources in training her if her only motivation was to “earn pocket change.” No evidence has been identified to us that would suggest that Tomsic had displayed insufficient motivation. Moreover, it is again for the jury to determine the issue of Miller’s intent in making this statement. The district judge may have been correct in his conclusion, or it may be that Miller was displaying stereotyped views of women and of marriage which he actually *12 held and which may have affected his relationship with plaintiffs in other ways. The question is not one capable of resolution on summary judgment.

That said, however, we agree with the district judge that these remarks are insufficient to constitute direct evidence of discriminatory intent in the terminations, even though Miller’s recommendation was considered in State Farm’s decisions. Although “stereotyped remarks can certainly be evidence that gender played a part” in an employment decision, Price Waterhouse v. Hopkins, 490 U.S. 228, 251 (1989) (emphasis in original), statements such as those in issue here which “are on their face expressions of . . . personal opinion, and not an existing policy which itself constitutes discrimination” constitute circumstantial or indirect evidence, not direct evidence, Ramsey v. City and County of Denver, 907 F.2d 1004, 1008 (10th Cir. 1990); accord EEOC v. Wiltel, Inc., 81 F.3d 1508 (10th Cir. 1996); Heim v. State of Utah, 8 F.3d 1541, 1546-47 (10th Cir. 1993). ^[5]

We next consider whether plaintiffs produced sufficient evidence of pretext to *13 withstand defendant’s summary judgment motion. Plaintiffs do not dispute that defendant proffered a legitimate reason for requesting their resignations. Under the McDonnell Douglas paradigm, plaintiffs must counter defendant’s proffer by presenting “enough evidence to support an inference that the employer’s reason was merely pretext, by showing either < that a discriminatory reason more likely motivated the employer or . . . that the employer’s proffered explanation is unworthy of credence.’” Cone v. Longmont United Hospital Ass’n, 14 F.3d 526, 530 (10th Cir. 1994) (quoting Texas Dept. of Community Affairs v. Burdine, 450 U.S. 248, 256 (1981)).

Plaintiffs contend, inter alia , that pretext could be inferred from several circumstances shown by their evidence. First, plaintiffs point to the inconsistency, noted previously, in Mr. Miller’s statements to plaintiff Browning concerning whether her production was the most important factor in her evaluation. Second, plaintiffs contend that the statements by Mr. Miller discussed above could support an inference of discriminatory intent even if, as we have concluded, they do not constitute direct evidence of discrimination. Third, plaintiffs cite to Miller’s statements to others soon after he became plaintiffs’ supervisor that Tomsic would be fired, these conversations having occurred several months before the decision to ask for her resignation was actually made. Plaintiffs do not go into much detail on the significance they attribute to these remarks, but we gather that their argument is that a permissible inference from this evidence is that Miller had prematurely and unfairly prejudged Tomsic solely on the basis of her gender. Fourth, plaintiffs say that the fact that *14 other reasons given by defendant for their discharge, such as abrasiveness and generally poor attitudes, were totally subjective ones can support an inference of pretext. Finally, plaintiffs suggest that the evidence is at least sufficient that a jury could find that State Farm acted on both legitimate and prohibited reasons, making this a “mixed motive” case similar to Price Waterhouse v. Hopkins, 490 U.S. 228 (1988), a contention which we feel we need not consider here. We agree that these evidentiary matters raised factual issues on pretext.

As to the statements by Miller on which plaintiffs rely, defendant contends that these statements were, at most, “stray remarks” of the kind found insufficient to establish a jury question on pretext in Cone v. Longmont United Hospital Ass’n, 14 F.3d 526 (10th Cir. 1994), and Rea v. Martin Marietta Corp., 29 F.3d 1450 (10th Cir. 1994). We disagree. In Cone , an Age Discrimination in Employment Act case, we began our discussion on this point by noting that remarks exhibiting bias which refer directly to the plaintiff may support an inference of discrimination. 14 F.3d at 531. We said that a plaintiff must show a nexus between the allegedly discriminatory statements and the employer’s decision. Id. We found in that case that there was no such nexus and that an inference of bias could not be supported from the remarks there at issue for two reasons. First, we found that one of the speakers had not been shown to have had any role in the employment decision. Id. Second, as to a remark which was made by one involved in the employment decision, we found that his very general remark that “long-term employees have a diminishing return,” had not been shown to have had any connection with a discriminatory motive in terminating the plaintiff. Id. We find *15 significant differences in the circumstances presented in the instant case, however.

The remarks at issue in this case were directed to the plaintiffs individually, unlike the very general statements in Cone. Moreover, as we have noted, Miller’s recommendation was clearly before the decision maker, Baumann. We believe that this is a sufficient nexus to the decision to ask for plaintiffs’ resignation; a jury could infer from Miller’s remarks in the total context presented that unlawful bias was a motivating factor in defendant’s decision. ^[6] In the context of the case, we also think it is significant that Miller at one critical point offered an inconsistent explanation of his evaluation of the performance of plaintiff Browning. As noted, on December 16, 1991, three days before State Farm decided to ask for plaintiffs’ resignations, Miller discussed with Browning her future with State Farm. He told her in that interview that production was not the most important criterion, a statement inconsistent with the justification State Farm later gave for its actions. Such inconsistencies can support an inference that the proffered reasons for plaintiffs’ discharge were pretextual. See Cole v. Ruidoso Municipal Schools, 43 F.3d at 1380. Miller was also inconsistent in his evaluations of plaintiffs, as we have noted.

In short, we conclude that the district court was in error in holding that plaintiffs had failed to produce sufficient evidence of pretext to withstand the summary judgment motion. A jury may infer discriminatory animus from “the simple showing of pretext,” and it may *16 find illegal discrimination upon a prima facie case and pretext. Randle v. City of Aurora, 69 F.3d 441, 451 (10th Cir. 1995). In sum, we hold that plaintiffs produced sufficient evidence to require submission to a jury. ^[7]

Accordingly, the summary judgment is REVERSED and the case is REMANDED for further proceedings in accord with this opinion.

[1] At the hearing on defendant’s motion for summary judgment, the plaintiffs stipulated to dismissal of their Title VII claim against Norman Miller and to dismissal of all claims against State Farm except the Title VII claim and a claim for intentional infliction of emotional distress. On appeal, plaintiffs challenge only the adverse judgment on the Title VII claim.

[2] Because we are reviewing a summary judgment in favor of defendant, we state the facts in the light most favorable to plaintiffs.

[3] The judge stated the four elements as: (1) plaintiffs belong to a protected group; (2) they were qualified for their jobs; (3) they were terminated despite their qualifications; and (4) after their termination the employer hired someone or sought applicants for plaintiffs’ vacated positions whose qualifications were no better than those of plaintiffs. 870 F. Supp at 323.

[4] The district court did not have the benefit of Cole to guide its decision. In Cole, however, we simply restated the requirement of the fourth prong as set out in Rea v. Martin Marietta Corp., 29 F.3d 1450, 1454 (10th Cir. 1994), a reduction in force case which also was not stating new law.

[5] Plaintiffs also argue that summary judgment should have been denied defendant, even if they failed to establish a prima facie case, on the basis that plaintiffs had carried the burden of showing sufficient indirect evidence to establish a probability that, but for the plaintiffs’ status, the challenged employment decision would have favored the plaintiffs. They rely on Notari v. Denver Water Dept., 971 F.2d 585 (10th Cir. 1992), as establishing this method of proof as an alternative to the more familiar McDonnell Douglas framework. Defendant in response contends that Notari was a case of reverse discrimination and that its holding should not be applied in other contexts. Our holding that the district court erred in concluding that plaintiffs had failed to show a prima facie case makes it unnecessary for us to reach this alternative argument.

[6] Our analysis in Rea followed closely the analysis in Cone. Therefore, defendant’s reliance on that case is similarly misplaced.

[7] On remand, the district court will have the benefit of Randle and Cole, which he did not have before his ruling in this case. They should be applied because of their pertinence here.