Christine A. Craft, Appellee/cross-Appellant v. Metromedia, Inc., Appellant/cross-Appellee

Christine Craft was reassigned from coanchor to reporter by KMBC-TV in Kansas City, Missouri, and as a result brought this action against the station’s owner and operator, Metromedia, Inc. Craft alleged that she had been discriminated against on the basis of sex in violation of both Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e et seq. (1982), and the Equal Pay Act of 1963, 29 U.S.C. § 206(d) (1982), and that she had been fraudulently induced into accepting the KMBC employment. The primary focus of the suit was KMBC’s concern with appearance — whether the station’s standards for on-air personnel were stricter and more strictly enforced as to females than as to males and whether the station misrepresented to Craft its intentions as to changing her appearance to persuade her to accept the anchor job. The district court 1 found against Craft on her *1208 Title VII sex discrimination claim and refused a new trial as to the jury verdict against her on her equal pay claim. The court, however, set aside as excessive the jury verdict in Craft’s favor on the fraud claim and ordered a new trial. Craft v. Metromedia, Inc., 572 F.Supp. 868 (W.D.Mo.1983). Craft appeals from these rulings while Metromedia appeals from the judgment on the jury verdict, again in Craft’s favor, in the second fraud trial. We affirm the Title VII and Equal Pay Act judgments against Craft but further conclude that she did not make a submissible case on her fraud claim and thus reverse the judgment against Metromedia on that issue.

Christine Craft commenced her television career in mid-1975 as a weeknight weather reporter at a small station in Salinas, California. After about a year and a half, during which she also worked as an announcer, reporter, and substitute sportscaster, she became the weekend weather anchor at a San Francisco television station. She spent a year there, again handling some additional news and substitute sports assignments.

Craft then was hired by Columbia Broadcasting System to host the “Women In Sports” portion of its network television program “CBS Sports Spectacular.” At the network’s behest she had her hair cut short and bleached blond. She was required to use. black eyebrow pencil and dark red lipstick, and a CBS technician applied heavy makeup before every on-air appearance. “Women in Sports” was discontinued after thirty weeks, and Craft returned to California.

About a year later Craft resumed her television career as a reporter at a station in Santa Barbara, California. She remained there, also serving as coanchor of the late news and occasionally doing sports and weather, until December 1980, when she began work at KMBC.

KMBC, having slipped in its local news ratings, had determined to adopt the coan-chor format used by its two major competitors in Kansas City. Furthermore, it had determined that, because of the perceived “coldness” of its anchor, Scott Feldman, the new position was to be filled by a female to “soften” its news presentation. The station obtained tapes of a number of performers, including Craft, from Media Associates of Dallas. After studying these tapes, KMBC news director Ridge Shannon contacted Craft, among other individuals, to see if she was interested in auditioning for the job. Craft described her unpleasant experience at CBS and made plain that she was not interested if KMBC intended a “makeover” of her appearance. She continued to stress this point while in Kansas City for the audition, and Shannon and R. Kent Replogle, vice president and general manager of KMBC, assured her they planned no changes such as those at CBS. Shannon did mention that KMBC made some use of consultants, and Craft indicated some willingness to work on her appearance and dress.

Craft accepted the coanchor position at KMBC and at the station’s request stopped in Dallas for a meeting with Lynn Wilford of Media Associates before reporting to Kansas City. She made her debut as coan-chor on January 5, 1981, and the testimony is essentially uncontradicted that Shannon and Replogle immediately began having concerns about her clothing and makeup. Wilford came to Kansas City on January 14 to work with Craft on dress as well as on various aspects of her presentation technique. It was during this visit that Wilford for the only time applied Craft’s makeup, and the results on the 6 p.m. news were so unsatisfactory that Craft was allowed to remove the makeup before the 10 p.m. broadcast.

In the following months Shannon continued to make occasional suggestions or criticisms as to certain articles of Craft’s clothing, and Craft was provided with materials, including the book Women’s Dress for Success, on wardrobe and makeup. Then, beginning in April, KMBC arranged for Macy’s Department Store to provide clothing for Craft in exchange for advertising time. Craft was assisted by a consultant *1209 from Macy’s in selecting outfits. She would then return to the KMBC studio, try on the clothing, and appear before camera so tapes could be made to send to Wilford for review.

On May 19 and 20, 1981, Media Associates initiated some research of viewer perceptions of KMBC’s newscasts by conducting four “focus group” discussions. By this technique groups of ten individuals viewed sample video tapes of local news programs and then gave their reactions in sessions moderated by a Media Associates representative. The response to Craft’s appearance was, as summed up by the district court, “overwhelmingly negative.” 572 F.Supp. at 873. When Replogle and Shannon met with Craft the following day to discuss this result, Craft at first wanted to be let out of her contract to return to California. Replogle, however, stated that management was ready to work with her to overcome the problems, and Craft ultimately agreed to cooperate. Thereafter, her wardrobe was more closely supervised, and the “clothing calendar” 2 then mentioned by Replogle was eventually instituted in late July or early August.

As a further follow-up, KMBC and Media Associates in late June conducted a telephone survey of some 400 randomly selected persons in the greater Kansas City viewing area. These persons were asked to respond to a questionnaire specifically drafted to pursue issues raised in the focus group discussions. For example in one segment the survey participants were asked to rank Craft in comparison with the female coanchors at KMBC’s competitors in response to some fourteen statements, four of which dealt with “good looks” or the dress of and image of a “professional anchor woman.” Craft came out trailing in almost every category. Media Associates’ report on the results of this survey, which was conveyed to Shannon and Replogle on August 3, 1981, suggested that Craft was having an extremely adverse impact on KMBC’s acceptance among Kansas City viewers. On August 13 Media Associates recommended that Craft be replaced, and KMBC, after initial resistance, agreed.

The next day Shannon told Craft she was being reassigned to reporter at no loss of pay or contractual benefits. He characterized the results of the research, in the language of the district court, as “devastating and unprecedented in the history of the consultants of Media Associates.” 572 F.Supp. at 874. Craft states that Shannon also told her she was being reassigned because the audience perceived her as too old, too unattractive, and not deferential enough to men. Shannon, however, specifically denies making such a statement, and the district court believed his version of the conversation. Id.

After the weekend Craft sent a telegram to KMBC refusing to accept reassignment, and when further discussions failed to resolve the matter, she returned to Santa Barbara where on September 1, 1981, she commenced work as a coanchor at the television station at which she previously had been employed. This suit followed in four counts: the Title VII sex discrimination and the fraud counts, the allegation of violation of the Equal Pay Act based on the differential between Craft’s and Scott Feld-man’s salaries, and an allegation of prima facie tort based on an intent by KMBC to injure Craft. This last count was abandoned during the first trial in Kansas City.

Following that first trial the district court rejected the findings of the advisory jury 3 on the Title VII sex discrimination claim and entered judgment for Metromedia on that issue. The district court found that KMBC required both male and female on-air personnel to maintain professional, *1210 businesslike appearances “consistent with community standards” and that the station enforced that requirement in an evenhanded, nondiscriminatory manner. 572 F.Supp. at 877-78. Any greater attention to Craft’s appearance, the court concluded, was “tailored to fit her individual needs” and was necessary because of her “below-average aptitude” in matters of clothing and makeup. Id. at 878. The district court also found that Craft had not been constructively discharged because there was insufficient evidence that her working conditions had become so intolerable that she had no choice but to quit or that KMBC had intended to force her resignation. Id. at 879. Finally, the court concluded that the telephone survey conducted by Media Associates had not been discriminatory or designed to effect Craft’s removal as coan-chor and that KMBC had reasonably relied on the survey results as the basis for the personnel change. Id. at 873, 878-79.

As to the Equal Pay Act count, the district court upheld the jury verdict in favor of Metromedia and refused to grant Craft a new trial. It rejected her allegations of error in the jury instructions and in the exclusion of certain testimony and concluded that the weight of the evidence was not contrary to the jury’s finding that Feld-man’s higher salary was based on permissible factors such as his education in broadcasting, his greater experience in major television markets, and his established identity with Kansas City viewers. 4 Id. at 876, 879-80.

The district court, however, granted Me-tromedia a new trial on the fraud count. In a bifurcated procedure the Kansas City jury had found liability and awarded Craft $375,000 in actual damages and then, after additional instructions, had awarded her $125,000 in punitive damages. The court declared itself “firmly convinced” that the verdict was “excessive” and “the result of passion, prejudice, confusion, or mistake on the part of the jury.” Id. at 881. In support it cited the “unprincipled” award of actual damages nearly double that requested by Craft in her complaint (and nearly four times the amount she asked for at closing argument), instructional errors as to wrongful discharge and in the implementation of the bifurcated damages submission, and the “incessant and overwhelming” publicity surrounding the case. Id. at 881-82. The district court ordered that the new trial on the fraud count be held in Joplin, Missouri, with the jury to be sequestered. Id. at 882.

On retrial the jury again found for Craft, awarding $225,000 actual and $100,000 punitive damages. After the district court denied motions for judgment notwithstanding the verdict and for another new trial, Metromedia filed this appeal and Craft cross-appealed from the adverse determinations on her Title VII and Equal Pay Act claims.

I.

Craft’s attacks on the district court’s disposition of her Title VII sex discrimination claim are essentially factual, rather than legal as she contends. Her central position is that KMBC’s appearance standards were based on stereotyped characterizations of the sexes and were applied to women more constantly and vigorously than they were applied to men. Her further contentions that the district court erred in finding certain grooming cases controlling and in allowing Metromedia to rebut her prima facie case with stereotypical customer preferences follow only if we accept her central factual position.

A.

As an initial matter Craft argues that, since her case was one involving di *1211 rect evidence of discrimination, the district court erred in failing to shift the burden to Metromedia to prove by a preponderance of the evidence that it would have taken the challenged actions even in the absence of gender bias. E.g., Bell v. Birmingham Linen Service, 715 F.2d 1552, 1557 (11th Cir.1983), cert. denied, — U.S. - 104 S.Ct. 2385, 81 L.Ed.2d 344 (1984); Lee v. Russell County Board of Education, 684 F.2d 769, 774 (11th Cir.1982). The cases relied on by Craft, however, involved direct testimony of discrimination accepted by the trier of fact. Bell, 715 F.2d at 1557; Lee, 684 F.2d at 774-75. It is apparent that the district court did not accept any such testimony in this case. Thus, the burden-shifting rule for which Craft argues could have no application here, and we need not give further consideration to whether it should be adopted by this circuit. See Clay v. Hyatt Regency Hotel, 724 F.2d 721, 724 (8th Cir.1984) (no direct evidence that employer acted with a discriminatory motive); see also Norcross v. Sneed, 755 F.2d 113, 118 (8th Cir.1985) (to same effect in handicapped discrimination suit under 29 U.S.C. § 794 (1982)).

In the alternative, Craft argues that she also established a Title VII violation through use of circumstantial evidence and inferences under the McDonnell Douglas pattern of proof. 5 The district court plainly considered whether KMBC’s position was pretextual, which indicates that it reached its decision through this mode of analysis. On review, however, we need not discuss the record in terms of “prima facie case,” “articulation of a legitimate, nondiscriminatory reason,” and “pretext” but instead may look directly to the ultimate factual issue: whether the defendant intentionally discriminated against the plaintiff. United States Postal Service Board of Governors v. Aikens, 460 U.S. 711, 715, 103 S.Ct. 1478, 1482, 75 L.Ed.2d 403 (1983); see Easley v. Empire Inc., 757 F.2d 923, 929-30 (8th Cir.1985). As the district court recognized, the burden of proof on this issue always remains with the plaintiff. Craft, 572 F.Supp. at 876-77 (quoting Texas Department of Community Affairs v. Burdine, 450 U.S. 248, 253, 101 S.Ct. 1089, 1093, 67 L.Ed.2d 207 (1981)). A finding of intentional discrimination is a factual finding, Pullman-Standard v. Swint, 456 U.S. 273, 102 S.Ct. 1781, 72 L.Ed.2d 66 (1982), cited in Anderson v. City of Bessemer City, — U.S. — 105 S.Ct. 1504, 1508, 84 L.Ed.2d 518 (1985), and the court found against Craft.

Finally, Craft argues that the district court erred in finding her claim of discrimination in employment conditions controlled by. grooming cases such as Knott v. Missouri Pacific Railroad, 527 F.2d 1249 (8th Cir.1975). In Knott we held that appearance regulations making distinctions on the basis of sex will not support allegations of discrimination when the standards are reasonable and are enforced as to both sexes in an evenhanded manner. Id. at 1252. Again what Craft really contests is the district court’s factual determination that KMBC’s actions concerning the appearance of its on-air personnel fell within the Knott guidelines.

Because many of Craft’s arguments boldly seek reevaluation of the evidence and rejection of the district court’s factual findings, we make clear our limited standard of review in such cases. Factual findings may be set aside only when clearly erroneous. See Fed.R.Civ.P. 52(a). Our “foremost principle” is that “ ‘a finding is “clearly erroneous” when, although there is evidence to support it, the reviewing court on the entire evidence is left with the defi *1212 nite and firm conviction that a mistake has been committed.’ ” Anderson v. City of Bessemer City, — U.S. -, 105 S.Ct. 1504, 1511, 84 L.Ed.2d 518 (1985) (quoting United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 542, 92 L.Ed. 746 (1948)); see Hoefelman v. Conservation Commission of Missouri, 718 F.2d 281, 285 (8th Cir.1983). We may not duplicate the function of the district court by making our own determination of the facts and reversing if we believe we would have decided the case differently: “Where there are two permissible views of the evidence, the factfinder’s choice between them cannot be clearly erroneous.” Anderson, 105 S.Ct. at 1512; Inwood Laboratories v. Ives Laboratories, 456 U.S. 844, 857-58, 102 S.Ct. 2182, 2190-91, 72 L.Ed.2d 606 (1982).

Furthermore, we have stated that the burden is on the objecting party to clearly demonstrate error in the factual findings. Aetna Casualty & Surety Co. v. General Electric Co., 758 F.2d 319, 323 (8th Cir.1985); Hunt v. Pan American Energy, 540 F.2d 894, 901 (8th Cir.1976), cert. denied, 429 U.S. 1062, 97 S.Ct. 786, 50 L.Ed.2d 778 (1977); Reilly v. United States, 513 F.2d 147,150 n. 2 (8th Cir.1975). The evidence is to be construed in the light most favorable to the party prevailing below. Tautfest v. City of Lincoln, 742 F.2d 477, 481 (8th Cir.1984); United Barge Co. v. Notre Dame Fleeting & Towing Service, 568 F.2d 599, 602 (8th Cir.1978).

B.

We turn first to the most prominent factual element of Craft’s case — her testimony that Shannon told her she was being reassigned because she was too old, too unattractive, and not deferential enough to men or because the audience perceived her as such. Shannon, however, specifically denied making that statement. The only additional evidence 6 was that of Feldman, who testified on the stand at trial that he did not recall Shannon making such a comment but who had said in his deposition that Shannon had done so — again diametrically opposite testimony, this time from one witness. Credibility thus was central to the district court’s finding on this point.

Respect for the role of the district court is especially crucial as to determinations which rest on “variations in demeanor and tone of voice” of the witnesses and “the listener’s understanding of and belief in what is said.” Anderson, 105 S.Ct. at 1512 (emphasis added). The Supreme Court has cautioned:

[WJhen a trial judge’s finding is based on his decision to credit the testimony of one of two or more witnesses, each of whom has told a coherent and facially plausible story that is not contradicted by extrinsic evidence, that finding, if not internally inconsistent, can virtually never be clear error.

Id. at 1513 (emphasis added). The district court was in the best position to determine whether to believe Shannon or Craft, and there are no circumstances suggesting any basis for finding clear error in the district court’s choice in favor of Shannon.

Craft next argues, as do the amici curiae, that the district court clearly erred in failing to find that KMBC enforced appearance standards more strictly as to female *1213 than as to male on-air personnel. 7 In support of her position she points to the evidence most favorable to her, which shows that only females were subject to daily scrutiny of their appearance or were ever required to change clothes at the station before going on the air and that no male was ever directed to take time from his journalistic duties to select clothing, with the help of a consultant, from Macy’s and to test that clothing on camera for the approval of another consultant. The district court, however, concluded that such facts, in light of other evidence, showed only that KMBC was concerned with the appearance of all its on-air personnel and that it took measures appropriate to individual situations, characteristics, and shortcomings. Craft, 572 F.Supp. at 875, 876, 878. The court found that KMBC had only gradually increased its attentions to Craft’s appearance, with the institution of the clothing calendar representing the culmination of efforts made necessary when lesser suggestions, such as those complied with by other personnel, proved ineffective. Id. at 878.

The evidence shows that one female, Brenda Williams, had never been the subject of criticism for her appearance while numerous males on occasion had been given specific directions as to their individual shortcomings. Mike Placke, for example, had been told to lose weight, to get better-fitting clothes, to refrain from wearing sweaters under jackets, and to tie his necktie in a certain manner. Similarly, Michael Mahoney had been told to lose weight and to pay more attention to his wardrobe and hairstyle, while Bob Werley had been told to try wearing contact lenses and to get a hair piece and had been given a makeup chart on a form similar to that used with Craft. Shannon had discussed hair or moustache problems with Tim Richardson, Dave Dusik, Stan Carmack, and Corrice Collins, and Craig Sager had been advised to improve his wardrobe. Even Feldman on at least two occasions had received specific directives regarding his choice of shirts, and the Macy’s clothing consultant who worked with Craft testified that she had also met with Feldman and Dusik. Media Associates eventually prepared wardrobe “dos” and “don’ts” for male as well as female newscasters. 8 The district court found that the male personnel generally complied with grooming suggestions.

An interpretation of this record as showing that KMBC enforced its appearance standards equally as to males and females in response to individual problems is neither “illogical” nor “implausible” and has “support in inferences that may be drawn from the facts.” Anderson, 105 S.Ct. at 1513. We cannot conclude that the district court erred in this respect.

Our determination is not altered by the testimony, cited by Craft, of the consultant Wilford or of other female KMBC employees to the effect that the station put more emphasis on the appearance of females. This testimony represents only the opinions and impressions of the witnesses as to KMBC’s policy, and the district court by implication rejected such opinions, making its own finding to the contrary on the record. 9 The court also apparently chose *1214 not to believe testimony by Pam Whiting that Shannon had said it was more important for she and Craft than for Feldman to look good on the air; Whiting herself had been subject to a great deal of attention as to her appearance and had eventually left the station, and on the stand she expressed the opinion that appearance was overemphasized in television news and said she hoped Craft prevailed. Finally, the district court attributed the use in the survey of questions on appearance and good looks only as to female anchors as prompted not by sexual bias but by the concerns raised as to Craft by the focus groups. Craft, 572 F.Supp. at 873. The viewers were asked to rank the male anchors on more points concerning personality and “coldness,” Feldman’s perceived weaknesses. The inclusion of a makeup criterion and a requirement that Craft not deviate from her clothing calendar in the “standards of performance” by which her work was to be evaluated, in contrast to the absence of any appearance objectives in Feldman’s “standards of performance,” similarly just reflected management’s efforts to pursue with personnel their individual weaknesses. 10

Craft further argues, however, again joined by the amici curiae, that even if KMBC was evenhanded in applying its appearance standards, the district court erred in failing to recognize that the standards themselves were discriminatory. 11 She contends that she was forced to conform to a stereotypical image of how a woman anchor should appear. The evidence included a communication from the consultant, Wilford, to Shannon suggesting that Craft purchase more blouses with “feminine touches,” such as bows and ruffles, because many of her clothes were “too masculine.” The general wardrobe hints for females developed by Media Associates warned that women with “soft” hairstyles and looks should wear blazers to establish their authority and credibility while women with short “masculine” hairstyles shouldn’t wear “masculine” clothing in dark colors and with strong lines because they would appear too “aggressive.” The district court found that Craft had been hired to “soften” KMBC’s news presentation. Id. at 871.

The “dos” and “don’ts” for anchors suggested that while males should remember “professional image,” female anchors were to remember “professional elegance” (emphasis added). The outfits on the clothing calendar given to Craft were dominated by recognized fashion labels. Wilford testified that in April she had told Craft not to wear the same outfit more than once every three to four weeks because people would start calling in about it; males, however, Wilford said, could wear an outfit every week and a suit even twice within the same week if combined with a different tie. Wilford further testified that viewers — particularly other women — criticize women more severely than men for their appearance on camera and that women’s dress is more complex and demanding because “society has made it that way.” Craft’s argument is that these differing standards as to females reflect customer preferences, which a number of cases have held cannot justify discriminatory practices. E.g., Diaz v. Pan American World Airways, 442 F.2d 385, 389 (5th Cir.), cert. denied, 404 U.S. 950, 92 S.Ct. 275, 30 L.Ed.2d 267 (1971); see also Gerdom v. Continental Airlines, 692 F.2d 602, 609 (9th Cir.1982), cert. dismissed, 460 *1215 U.S. 1074, 103 S.Ct. 1534, 75 L.Ed.2d 954 (1983); Fernandez v. Wynn Oil Co., 653 F.2d 1273, 1276-77 (9th Cir.1981).

The district court, however, found that KMBC’s appearance standards were based instead on permissible factors. 572 F.Supp. at 877. While there may have been some emphasis on the feminine stereotype of “softness” and bows and ruffles and on the fashionableness of female anchors, the evidence suggests such concerns were incidental to a true focus on consistency of appearance, proper coordination of colors and textures, the effects of studio lighting on clothing and makeup, and the greater degree of conservatism thought necessary in the Kansas City market. The “dos” and “don’ts” for female anchors addressed the need to avoid, for example, tight sweaters or overly “sexy” clothing and extreme “high fashion” or “sporty” outfits while the male “dos” and “don’ts” similarly cautioned against “frivolous” colors and “extreme” textures and styles as damaging to the “authority” of newscasters. These criteria do not implicate the primary thrust of Title VII, which is to prompt employers to “discard outmoded sex stereotypes posing distinct employment disadvantages for one sex.” Knott v. Missouri Pacific Railroad, 527 F.2d 1249, 1251 (8th Cir.1975); see generally Note, Title VII Limits on Discrimination Against Television Anchorwomen on the Basis of Age-Related Appearance, 85 Colum.L.Rev. 190, 201 (1985). 12

Courts have recognized that the appearance of a company’s employees may contribute greatly to the company’s image and success with the public and thus that a reasonable dress or grooming code is a proper management prerogative. E.g., Fa-gan v. National Cash Register Co., 481 F.2d 1115, 1124-25 (D.C.Cir.1973); La Von Lanigan v. Bartlett & Co. Grain, 466 F.Supp. 1388, 1392 (W.D.Mo.1979). Evidence showed a particular concern with appearance in television; the district court stated that reasonable appearance requirements were “obviously critical” to KMBC’s economic well-being; and even Craft admitted she recognized that television was a visual medium and that on-air personnel would need to wear appropriate clothes and makeup. Craft, 572 F.Supp. at 877; see Note, supra, at 201 (“As television is a visual medium, television networks and local stations clearly have a right to require both male and female anchors to maintain a professional appearance while on camera.”). While we believe the record shows an overemphasis by KMBC on appearance, we are not the proper forum in which to debate the relationship between newsgath-ering and dissemination and considerations of appearance and presentation — i.e., questions of substance versus image — in television journalism. The record does not leave us with the “definite and firm conviction” that the district court erred or adopted an impermissible view of the evidence when it concluded that KMBC’s appearance standards were shaped only by neutral professional and technical consider *1216 ations and not by any stereotypical notions of female roles and images.

We further reject the argument of Craft that the perception, discussed above, that she would add warmth and “eomfortability” to the newscast defined a stereotypical “female” role, secondary to that of the male anchor, into which she was forced in terms of story assignments and division of duties. First, we observe that the research survey asked participants to rank both male and female anchors in response to the statement “comfortable to watch.” In addition, the district court rejected Craft’s factual contention that Feldman was assigned the lead story more frequently and was given more live stories while she was assigned the human interest, personal, and humorous stories. The court, for example, while acknowledging the testimony of news producer Sandra Woodward that she initially was instructed to give Feldman about two-thirds of the lead stories, also accepted testimony of Shannon that that policy had been designed to ease the transition to coanchors for Craft and the viewers and was discontinued after a few weeks, after which time the lead stories were equitably divided. Craft, 572 F.Supp. at 872, 875. The court further found that the on-site reporting and hard news stories were equitably distributed. Id. at 875. Finally, Shannon testified that Craft herself suggested her Thursday arts and entertainment segment to take advantage of her cultural background and her contacts from Santa Barbara with various Hollywood personalities. Craft cites no facts sufficient for us to find the district court’s interpretation of the record in this regard clearly erroneous.

We also cannot conclude that the district court erred when it held that Craft was not impermissibly removed as anchor on the basis of gender. While we disposed earlier of the alleged “too old, too unattractive” statement and have also already discussed the legitimate reasons why the survey questions focused on appearance only as to the female and not as to the male anchors, Craft further argues that her reassignment was based on a discriminatory interpretation of the survey results. She contends that discrimination may be inferred in that KMBC attributed all the negative viewer reaction to her despite Feldman’s earlier problems in the ratings and the failure of most respondents to name him as a positive reason for watching the station. She further alleges that she was unfairly burdened with having to offset Feldman’s weaknesses. The inferences on which she relies, however, are insufficient to justify setting aside the conclusions of the district court.

Craft essentially is asking us to reject the district court’s findings accepting Media Associate’s interpretation of the survey data and to substitute our own interpretation thereof. The evaluation of expert testimony, however, is primarily the function of the district court. Hoefelman v. Conservation Commission of Missouri, 718 F.2d 281, 284-85 (8th Cir.1983). The court here specifically found that the survey had been “conducted in accordance with generally accepted principles of survey research” and that its results were “trustworthy.” Craft, 572 F.Supp. at 873. Craft scored low on many neutral points such as knowledge of Kansas City, journalism ability, and apparent enjoyment of her job, and the court found “simply no evidence” that the survey had been designed to effect her removal because of her gender.