The Equal Employment Opportunity Commission (EEOC) brought this sex discrimination action against The Dial Corporation under Title VII of the Civil Rights Act of 1964 on behalf of a number of women who had applied for work but were not hired. A jury found that Dial had engaged in a pattern or practice of intentional discrimination against women and awarded compensatory damages, and the district court1 concluded that Dial's use of a preemployment strength test had an unlawful disparate impact on female applicants and awarded back pay and benefits. Dial appeals from the denial of its motion for judgment as a matter of law and from the judgment. EEOC cross appeals the denial of back pay to one claimant. We remand one issue but otherwise affirm.

Dial is an international company with a plant located in Fort Madison, Iowa that produces canned meats. Entry level employees at the plant are assigned to the sausage packing area where workers daily lift and carry up to 18,000 pounds of sausage, walking the equivalent of four miles in the process. They are required to carry approximately 35 pounds of sausage at a time and must lift and load the sausage to heights between 30 and 60 inches above the floor. Employees who worked in the sausage packing area experienced a disproportionate number of injuries as compared to the rest of the workers in the plant.

Dial implemented several measures to reduce the injury rate starting in late 1996. These included an ergonomic job rotation, institution of a team approach, lowering the height of machines to decrease lifting pressure for the employees, and conducting periodic safety audits. In 2000 Dial also instituted a strength test used to evaluate potential employees, called the Work Tolerance Screen (WTS). In this test job applicants were asked to carry a 35 pound bar between two frames, approximately 30 and 60 inches off the floor, and to lift and load the bar onto these frames. The applicants were told to work at their "own pace" for seven minutes. An occupational therapist watched the process, documented how many lifts each applicant completed, and recorded her own comments about each candidate's performance. Starting in 2001, the plant nurse, Martha Lutenegger, also watched and documented the process. From the inception of the test, Lutenegger reviewed the test forms and had the ultimate hiring authority.

For many years women and men had worked together in the sausage packing area doing the same job. Forty six percent of the new hires were women in the three years before the WTS was introduced, but the number of women hires dropped to fifteen percent after the test was implemented. During this time period the test was the only change in the company's hiring practices. The percentage of women who passed the test decreased almost each year the test was given, with only eight percent of the women applicants passing in 2002. The overall percentage of women who passed was thirty eight percent while the men's passage rate was ninety seven percent. While overall injuries and strength related injuries among sausage workers declined consistently after 2000 when the test was implemented, the downward trend in injuries had begun in 1998 after the company had instituted measures to reduce injuries.

One of the first applicants to take the WTS was Paula Liles, who applied to Dial in January 2000 and was not hired even though the occupational therapist who administered her test told her she had passed. She filed a discrimination complaint with the Iowa Civil Rights Commission and EEOC in August 2000. On September 24, 2002, EEOC brought this action on behalf of Liles and fifty three other women who had applied to work at Dial and were denied employment after taking the WTS. Twenty four of these applicants had been unable to complete the test.

A jury trial was held in August 2004, and EEOC and Dial offered testimony by competing experts. EEOC presented an expert on industrial organization who testified that the WTS was significantly more difficult than the actual job workers performed at the plant. He explained that although workers did 1.25 lifts per minute on average and rested between lifts, applicants who took the WTS performed 6 lifts per minute on average, usually without any breaks. He also testified that in two of the three years before Dial had implemented the WTS, the women's injury rate had been lower than that of the male workers. EEOC's expert also analyzed the company's written evaluations of the applicants and testified that more men than women were given offers of employment even when they had received similar comments about their performance. EEOC also introduced evidence that the occupational nurse marked some women as failing despite their having completed the full seven minute test.

Dial presented an expert in work physiology, who testified that in his opinion the WTS effectively tested skills which were representative of the actual job, and an industrial and organizational psychologist, who testified that the WTS measured the requirements of the job and that the decrease in injuries could be attributed to the test. Dial also called plant nurse Martha Lutenegger who testified that although she and other Dial managers knew the WTS was screening out more women than men, the decrease in injuries warranted its continued use.

The jury was asked to decide whether Dial had engaged in a pattern or practice of intentional discrimination against female job applicants, the date on which any such discrimination began, and a question relating to damages. The jury returned its verdict on August 23, 2004. It found Dial had engaged in a pattern or practice of intentional discrimination beginning in April 2001. The jury awarded a total of $30,003 in compensatory damages to the nine claimants who testified at trial and declined to assess punitive damages. Dial moved for judgment as a matter of law, alleging there was insufficient evidence for a reasonable jury to have found intentional discrimination. The motion was denied on February 3, 2005, but the district court eliminated nominal damages awarded to two applicants who had been rejected before April 2001 (the date when Dial's intentional discrimination began according to the jury verdict).

Following the jury trial the parties submitted additional evidence and briefs relating to the disparate impact allegations. The district court ruled on these issues in sixteen pages of detailed findings of fact and conclusions of law issued on February 3, 2005. It found that the WTS had had a discriminatory effect, that Dial had not demonstrated that the WTS was a business necessity or shown either content or criterion validity, and that Dial had not effectively controlled for other variables which may have caused the decline in injuries, including other safety measures that Dial had implemented starting in 1996.

After the court issued its findings and conclusions, Dial offered employment to all of the claimants in the spring of 2005. Further discovery and submissions followed, as well as additional briefing. The district court found that the claimants who had been unable to complete the full seven minutes of the WTS were also entitled to relief and determined the amount of back pay and interest to which each applicant was entitled. Back pay was calculated from the date the district court found the applicants should have been hired up to the date of Dial's offer of employment, less any wages earned elsewhere during that period. Health care benefits were awarded in the amount Dial would have paid for premiums, minus any benefits the women had received in the meantime. The range of the back pay awarded to the individual applicants varied from a high of $120,236 to a low of $920, and the individual health benefits ranged from $30,385 to $882.

An additional issue was raised in respect to one of the women who had accepted Dial's reinstatement offer, Heather Wright-Bradley. She had a criminal record predating her initial application, which included a number of convictions and at least one felony. Dial dismissed her after learning about her criminal history in a background check done after her reinstatement. The district court held a telephone conference with the parties to address whether Wright-Bradley should receive back pay under the circumstances. Dial's general counsel stated that the company had a policy on background checks during the period when the WTS was given which would have uncovered her criminal record. The district court concluded that Dial would have terminated her on account of her criminal record had she been hired in 2000 and that she was therefore not entitled to back pay.

On appeal Dial challenges the district court's denial of its motion for judgment as a matter of law, arguing there was insufficient evidence for a jury to find intentional discrimination. Dial also attacks the district court's findings of disparate impact and claims it proved that the WTS was a business necessity because it drastically decreased the number of injuries in the sausage production area of the plant. It contends the district court should not have awarded any back pay to applicants who were not strong enough to complete the WTS, it should have used the company's tenure data to calculate back pay, and that no applicant should have been awarded health care benefits without proof of any actual costs incurred. EEOC disagrees and cross appeals the denial of back pay to Wright-Bradley. It argues that Dial did not prove that at the time she was hired it had a policy in place to terminate new employees with similar criminal backgrounds.

Dial first argues that EEOC did not establish a pattern or practice of intentional sex discrimination and that its motion for judgment as a matter of law should therefore have been granted. EEOC responds that the jury had sufficient evidence on which to base its decision. We review the district court's denial of Dial's motion for judgment as a matter of law de novo, using the same standard as the district court. Ollie v. Titan Tire Corp., 336 F.3d 680, 685 (8th Cir.2003). The reviewing court must decide whether there is sufficient evidence to support the jury's verdict when examined in the light most favorable to the verdict. Id. Judgment as a matter of law is only appropriate when there is no reasonable inference to be made from the evidence which can sustain the verdict. Id.

A pattern or practice of intentional sex discrimination must be shown by proving "regular and purposeful" discrimination by a preponderance of the evidence, Int'l Brotherhood of Teamsters v. U.S., 431 U.S. 324, 339, 360, 97 S. Ct. 1843, 52 L. Ed. 2d 396 (1977). EEOC must show that more than an isolated act of discrimination occurred and that "discrimination was the company's standard operating procedure," id., but statistics combined with anecdotal examples of discrimination may establish a pattern or practice of regular, purposeful discrimination. Morgan v. United Parcel Service of America, Inc., 380 F.3d 459, 463-64 (8th Cir.2004). Moreover, discriminatory intent can be inferred from the mere fact of differences in treatment, Teamsters, 431 U.S. at 335 n. 15, 97 S. Ct. 1843.

Statistical disparities are significant if the difference between the expected number and the observed number is greater than two or three standard deviations. Hazelwood Sch. Dist. v. U.S., 433 U.S. 299, 308 n. 14, 97 S. Ct. 2736, 53 L. Ed. 2d 768 (1977). Here, the disparity between hiring of men and women showed nearly ten standard deviations. The percentage of women who passed the WTS declined with each implementation of the test. Despite knowing about the statistical difference, Dial continued to use the WTS. Dial argues that EEOC's statistics are inapplicable because men and women are not similarly situated and have profound physiological differences. There was evidence, however, that women and men worked the same job together for many years before the WTS was instituted. There was also evidence of women and men receiving similar comments on their test forms, but only the males receiving offers of employment.

Dial attacks the jury's finding that intentional discrimination began in April 2001, a month when the WTS was not in use and no particular identifiable discriminatory action was alleged. EEOC responds that the jury appeared to have found April 2001 to be the month when Dial must have known of the discriminatory effect of the WTS, but nonetheless continued to use it for future hiring periods. A reasonable jury could discredit Lutenegger's testimony that the decrease in injuries was the company's motivation for continuing to use the WTS. A reasonable jury could also have found that the differing treatment of males and females supported an inference of intentional discrimination. We conclude that the evidence was sufficient for a reasonable jury to find that there was a pattern and practice of intentional discrimination against women and that the district court did not err by denying Dial's motion for judgment as a matter of law.

Dial objects to the district court's findings of disparate impact and its conclusion that the company failed to prove the WTS was necessary to establish effective and safe job performance. We review the district court's factual findings regarding disparate impact for clear error and its legal findings de novo. Fed. R.Civ.P. 52(a). In a disparate impact case, once the plaintiff establishes a prima facie case the employer must show the practice at issue is "related to safe and efficient job performance and is consistent with business necessity." Firefighters Inst. for Racial Equality v. City of St. Louis, 220 F.3d 898, 904 (8th Cir.2000). An employer using the business necessity defense must prove that the practice was related to the specific job and the required skills and physical requirements of the position. Belk v. Southwestern Bell Telephone Co., 194 F.3d 946, 951 (8th Cir.1999). Although a validity study of an employment test can be sufficient to prove business necessity, it is not necessary if the employer demonstrates the procedure is sufficiently related to safe and efficient job performance. Hawkins v. Anheuser-Busch, Inc., 697 F.2d 810, 815-16 (8th Cir.1983). If the employer demonstrates business necessity, the plaintiff can still prevail by showing there is a less discriminatory alternative. Firefighters, 220 F.3d at 904.

Dial contends the WTS was shown by its experts to have both content and criterion validity. Under EEOC guidelines, "A content validity study should consist of data showing that the content of the selection procedure is representative of important aspects of performance on the job for which the candidates are to be evaluated." 29 C.F.R. § 1607.5(B). Dial's physiology expert testified that the WTS was highly representative of the actions required by the job, and Dial claims that his testimony was not rebutted by EEOC which had no physiology witness. The district court was persuaded by EEOC's expert in industrial organization and his testimony "that a crucial aspect of the WTS is more difficult than the sausage making jobs themselves" and that the average applicant had to perform four times as many lifts as current employees and had no rest breaks. There was also evidence that in a testing environment where hiring is contingent upon test performance, applicants tend to work as fast as possible during the test in order to outperform the competition.

Dial argues the WTS was criterion valid because both overall injuries and strength related injuries decreased dramatically following the implementation of the WTS. The EEOC guidelines establish that criterion validity can be shown by "empirical data demonstrating that the selection procedure is predictive of or significantly correlated with important elements of job performance." 29 C.F.R. § 1607.5(B). Although Dial claims that the decrease in injuries shows that the WTS enabled it to predict which applicants could safely handle the strenuous nature of the work, the sausage plant injuries started decreasing before the WTS was implemented. Moreover, the injury rate for women employees was lower than that for men in two of the three years before Dial implemented the WTS. The evidence did not require the district court to find that the decrease in injuries resulted from the implementation of the WTS instead of the other safety mechanisms Dial started to put in place in 1996.

Dial contends finally that the district court improperly gave it the burden to establish that there was no less discriminatory alternative to the WTS. Dial claims the burden should have been allocated to EEOC as part of the burden shifting framework in disparate impact cases, Firefighters, 220 F.3d at 904. Since Dial failed to demonstrate that the WTS was a business necessity, however, EEOC never was required to show the absence of a nondiscriminatory alternative. Part of the employer's burden to establish business necessity is to demonstrate the need for the challenged procedure, Kirby v. Colony Furniture Co., 613 F.2d 696, 705 n. 6 (8th Cir.1980), and the court found that Dial had not shown that its other safety measures "could not produce the same results." We conclude that the district court findings in its disparate impact analysis were not clearly erroneous, and we see no legal error in its conclusions on liability.

Dial claims the district court committed error by awarding back pay and benefits to all but one of the claimants even though twenty four women had been unable to complete the WTS. But once an employer is found liable for a Title VII violation, the district court is obligated to grant "the most complete relief possible." King v. Staley, 849 F.2d 1143, 1144 (8th Cir.1988). There is a strong presumption that an employee who has suffered discrimination should receive back pay. E.E.O.C. v. Rath Packing Co., 787 F.2d 318, 329 (8th Cir.1986). This presumption can be overcome only if back pay would "frustrate the central statutory purposes of eradicating discrimination throughout the economy and making persons whole for injuries suffered through past discrimination." Albemarle Paper Co. v. Moody, 422 U.S. 405, 421, 95 S. Ct. 2362, 45 L. Ed. 2d 280 (1975). The trial court "has broad equitable discretion to fashion back pay awards in order to make the Title VII victim whole." E.E.O.C. v. Delight Wholesale Co., 973 F.2d 664, 669-70 (8th Cir. 1992). The district court's finding of discrimination was based on Dial's use of the WTS and the evidence that the test was more difficult than the actual job. Women who were not hired because they were unable to complete the WTS suffered losses as a consequence, and Dial did not overcome the presumption in favor of awarding back pay in respect to these claimants.

Dial also claims the district court erred by refusing to use its employee tenure data in calculating the amount of back pay because the plant's high turnover rate suggests the claimants would not have been employed for the entire back pay period. The district court applied the well established rule for calculating back pay—the difference between the amount the claimant would have earned absent the discrimination and the amount of wages actually earned during the relevant period. See Hartley v. Dillard's, Inc., 310 F.3d 1054, 1062 (8th Cir.2002). This was consistent with Title VII's dual purposes of compensating victims and deterring future discrimination, as well as the district court's obligation to grant "the most complete relief possible." King, 849 F.2d at 1144.

Dial also challenges the award of lost medical premiums, arguing the claimants should have been required to prove they incurred medical expenses. Our court has not decided whether out of pocket expenses are required before health care benefits can be awarded, see Tolan v. Levi Strauss & Co., 867 F.2d 467, 470 (8th Cir.1989), and other circuits are divided on the issue. In the view of the Fourth Circuit, Congress intended fringe benefits to be part of the monetary award compensating claimants for the discrimination they suffered. See Fariss v. Lynchburg Foundry, 769 F.2d 958, 965-66 (4th Cir. 1985) (awarding medical benefits to widow of age discrimination victim without requiring proof of out of pocket medical insurance costs); see also Blackwell v. Sun Elec. Corp., 696 F.2d 1176, 1185-86 (6th Cir.1983) (granting the amount of health care premiums to claimant as part of recovery); but see Galindo v. Stoody Co., 793 F.2d 1502, 1517 (9th Cir.1986) (reimbursing only out of pocket expenses incurred to obtain health care).

Health care benefits are an important element of an employee's overall employment package, and Dial does not contest that it would have awarded claimants health care benefits had they been hired. The district court only required Dial to compensate the claimants for the amount of health care premiums that would have been part of their employment package had they not suffered discrimination. No reimbursement for health care costs incurred by uninsured claimants was awarded. The court's limited award was reasonable, for "[t]his insurance coverage, not the proceeds, is the benefit for which the employer must be held liable." Fariss, 769 F.2d at 965.

EEOC cross appeals the denial of back pay to Wright-Bradley. EEOC argues that Dial did not overcome the presumption in favor of awarding back pay to her as a victim of Title VII violations. See Rath Packing, 787 F.2d at 329. Dial responds that it should not have to contribute back pay to Wright-Bradley because she was convicted of a felony before she applied in 2000, a background check would have revealed her criminal record, and she would have been terminated. EEOC disputes Dial's factual assertions and argues that McKennon v. Nashville Banner Publishing Co., 513 U.S. 352, 115 S. Ct. 879, 130 L. Ed. 2d 852 (1995), supports an award of back pay to Wright-Bradley.

In McKennon, the Supreme Court concluded that an employer's belated discovery of wrongdoing by a dismissed employee should not completely bar an award of back pay because of the congressional "objective of forcing employers to consider and examine their motivations, and of penalizing them for employment decisions that spring from [] discrimination." Id.; see also Sellers v. Mineta, 358 F.3d 1058, 1061-62 (8th Cir.2004) (applying the after acquired evidence rule to Title VII cases). The Supreme Court decided that back pay should be awarded, but only "from the date of the unlawful discharge to the date the new information was discovered" absent findings of "extraordinary equitable circumstances." McKennon, 513 U.S. at 362, 115 S. Ct. 879. The district court distinguished McKennon on the basis that the misconduct there had occurred during the plaintiff's employment and concluded that back pay would result in a windfall to Wright-Bradley. Dial argues that this is a mixed motives case so McKennon does not apply, but Wright-Bradley's criminal record could not have been a motive for Dial's not hiring her since it was unaware of it at the time.

Dial's general counsel represented in the court's telephone conference that its offers of employment in 2000 were contingent on passing a background check. Counsel also stated that Dial had not previously dismissed an employee due to such a check since it had not discovered felony convictions before but that Dial had terminated five employees with criminal records. EEOC objected that "we've got a factual problem here," that Dial had offered no evidence to establish that the described policy existed, and that Dial's job application stated only that a hiring offer was contingent on a drug test and a check of previous employment. EEOC added that it had asked Dial during discovery for copies of background checks completed on employees hired near the time Wright-Bradley took the WTS, but Dial responded that it did not have such evidence.

After examining the record we conclude that there are disputed factual issues here on the question of whether Wright-Bradley should be awarded back pay. The statements of Dial's counsel on which the district court relied were not made under oath, and no other evidence of Dial's policy has been proffered. Like the employer in McKennon, Dial learned about Wright-Bradley's wrongdoing only after its discriminatory hiring decision was made. Under the after acquired evidence framework, Dial has the burden of proving that Dial would have terminated Wright-Bradley in 2000 because of her criminal background. See McKennon, 513 U.S. at 362-63, 115 S. Ct. 879. If Dial produces such proof, the district court should consider whether Wright-Bradley is entitled to any amount of back pay. See id.

In sum, we affirm the district court's denial of judgment as a matter of law, its findings of disparate impact, and its award of back pay and benefits to all claimants except Wright-Bradley. Her claim for back pay is remanded for further proceedings consistent with this opinion.