Aaron Starobin (argued), Donald Roy Fraker, Starobin & Associates, Thiensville, Wis., for plaintiffs-appellants.

Fred G. Groiss, Quarles & Brady, Milwaukee, Wis., Patrick W. Ritchey (argued), Reed, Smith, Shaw & McClay, Pittsburgh, Pa., for defendant-appellee.

Before CUMMINGS and CUDAHY, Circuit Judges, and DILLIN, District Judge.1

DILLIN, District Judge.

Plaintiffs John W. Fisher and Richard R. Kirchhoff sued their former employer, Transco Services-Milwaukee, Inc., alleging unlawful discrimination in violation of the Age Discrimination in Employment Act, 29 U.S.C. §§ 621-634 (ADEA). They appeal the District Court's grant of summary judgment in favor of Transco. For the following reasons, we reverse and remand for proceedings consistent with this opinion.

At all relevant times prior to March of 1984, The Great Atlantic & Pacific Tea Company, Inc. (A & P), through its subsidiary Kohl's Food Stores, Inc., owned and operated a retail grocery warehouse where plaintiffs-appellants Fisher and Kirchhoff had worked since 1970 as order selectors (selectors). As selectors, appellants were responsible for processing requests for goods, which included retrieving and loading the goods, and completing some required paperwork.

In March of 1984, A & P contracted with Transco Service Corporation, through its subsidiary, defendant-appellee Transco Services-Milwaukee, Inc. (Transco), to operate the warehouse. Approximately five months after the transfer of management, Transco instituted A & P's "Measured Day Work Program" in order to measure and evaluate the performance of its full-time selectors, including appellants.

At the heart of A & P's Program was a computer which analyzed each incoming order and assigned the number of "leveled minutes" needed for a selector to complete it. In calculating this time, the computer considered the goods requested, particularly their sizes, weights, and locations in the warehouse, and made adjustments for such variables as the distance between the starting area and the loading platform, and the degree of difficulty in loading a particular good.

Every leveled minute of work assigned to a selector earned the selector a proportional number of "rest allowance minutes" for personal needs. The leveled minutes and rest allowance minutes were summed to determine the "standard minutes" needed to complete an order. A selector's performance was calculated by taking the ratio of standard minutes to the actual time expended.

During an eight hour day, which is 480 minutes, a selector was expected under the Program to perform 407 minutes of leveled work and receive 73 minutes of rest allowance. At all times relevant, Transco expected a standard-to-actual ratio of one to one, i.e., 100 percent. Any selector who failed to maintain this ratio on a weekly basis might be subject to discipline.

Specifically, until all selectors achieved a 100 percent performance ratio, those selectors whose performance ratios fell into the lower 20 percent of those working that week were subject to progressive discipline. After the first week of low production, counseling and an oral warning were to be given. If a second week of low production occurred, additional counseling and a written warning were to be given. If a third, another written warning, as well as a one-day suspension; for a fourth, a written warning and a three day suspension. Finally, if the employee's performance level fell in the bottom 20 percent for a fifth time, the employee became subject to discharge. The weeks of poor performance did not need to be consecutive, but if discipline was administered, a one week grace period was to be granted before discipline would again be given.

According to Transco, appellant Fisher failed to rank in the upper 80 percent of selectors for the weeks ending October 13, October 27, and December 15, 1984, and the weeks of January 12 and March 8, 1985. Fisher was terminated on March 8, 1985, at the age of 45. Appellant Kirchhoff failed to rank in the upper 80 percent the first week of the Program's implementation, that is, the week ending September 8, 1984. After an accident rendered him unable to work for nearly two months commencing in November of 1984, Kirchhoff failed to rank in the upper 80 percent for the weeks ending February 2, February 23, March 23 and April 12, 1985. Kirchhoff was terminated on April 12, 1985, at the age of 42. However, according to appellants, they were each erroneously disciplined three times and should not have been terminated. There is evidence to support these contentions.

Transco ended its Program on August 3, 1985, after 48 weeks. During this time 1,182 weekly measurements were taken, of which only 20 were at or above the 100 percent level. As a result of the Program, 11 out of the 52 selectors who worked for at least five weeks were fired, of whom 10 were age 40 or older.

The review of a district court's grant of summary judgment is de novo. La Preferida, Inc. v. Cerveceria Modelo, S.A., 914 F.2d 900, 905 (7th Cir.1990). Summary judgment is appropriate where there are no genuine issues of material fact and the moving party is entitled to judgment as a matter of law. Fed.R.Civ.Pro. 56(c). While we view the facts in the light most favorable to the nonmoving party, there is an affirmative burden of production on the nonmoving party to defeat a proper summary judgment motion. Celotex Corp. v. Catrett, 477 U.S. 317, 323-24, 106 S. Ct. 2548, 2552-53, 91 L. Ed. 2d 265 (1986). Specifically, before a court denies summary judgment, it must be determined whether there is sufficient evidence from which a jury could find in favor of the nonmoving party. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 249-50, 106 S. Ct. 2505, 2510-11, 91 L. Ed. 2d 202 (1986)).

To succeed in this action, Fisher and Kirchhoff must prove that they would not have been discharged "but for" their age. LaMontagne v. American Convenience Prods., Inc., 750 F.2d 1405, 1409 (7th Cir.1984). There exist two possible theories under which Fisher and Kirchhoff may assert their claims. First, they may argue that they have suffered disparate treatment because of their age. Second, they may assert that their employer's practice, while not necessarily intended, resulted in a significant disparate impact upon its older workers.

We will first consider the theory of disparate treatment. Under it, appellants may prove their claim in two ways. First they may meet their burden head on by presenting evidence, direct or circumstantial, that age was the determining factor in their discharge. Second, and the more common, they may utilize the indirect, burden-shifting method of proof used in Title VII cases. See McDonnell Douglas Corp. v. Green, 411 U.S. 792, 93 S. Ct. 1817, 36 L. Ed. 2d 668 (1973); Weihaupt v. American Medical Ass'n, 874 F.2d 419, 424 (7th Cir.1989). Appellants have opted for the latter.

Under the indirect method of proving disparate treatment, appellants must set forth a prima facie case to create a presumption of discrimination. In discharge cases such as the present, plaintiffs must show that (1) they were in the protected class (persons over the age of 40), (2) they were doing their jobs well enough to meet their employer's legitimate expectations, (3) they were discharged or demoted, and (4) the employer sought a replacement for them.2 Weihaupt, 874 F.2d at 427.

If the plaintiff is successful, the burden of production shifts to his employer to articulate a legitimate, nondiscriminatory reason for the discharge. If such an articulation is forthcoming, the presumption dissolves, and the burden of production shifts back to the plaintiff to show that the employer's proffered reason is a pretext for discrimination. Id.

It is clear that appellants were members of the class protected by the ADEA, and that they were discharged. In addition, we agree with appellants that until this appeal Transco had not challenged the contention that appellants were replaced and cannot do so now. Thus, remaining at issue is whether the appellants were performing their jobs well enough to meet Transco's legitimate expectations.

Transco's expectations as to the appellants' performances are wholly dependent upon the Program alleged by appellants to be a pretext for discrimination. We find that in this situation, the issue of legitimate performance is best merged with the issue of pretext. See Palucki v. Sears, Roebuck & Co., 879 F.2d 1568, 1571 (7th Cir.1989) (The order-of-proof scheme adopted to guide the litigation of discrimination claims was never meant to be "fetishized"). This approach is logical, for if the conditions placed upon the appellants by the Program are found to be illegitimate, it follows that Transco's expectations were likely pretextual, and vice-versa. Further, to hold otherwise would unnecessarily insulate Transco from the burden of articulating a nondiscriminatory reason for the discharge of the appellants. See Dale v. Chicago Tribune Company, 797 F.2d 458, 463 n. 2 (7th Cir.1986). This would clearly not further the purpose of the ADEA.

For the above reasons, we believe that the appellants have effectively created a rebuttable presumption of discrimination. Thus, the burden of production is shifted to Transco to articulate a legitimate reason for appellants' discharge. This burden is not difficult to satisfy. Transco "need not persuade the court that it was actually motivated by the proffered reasons". Texas Department of Community Affairs v. Burdine, 450 U.S. 248, 254-55, 101 S. Ct. 1089, 1094-95, 67 L. Ed. 2d 207 (1981) (citations omitted). It is sufficient if its proffered reason raises a genuine issue of material fact as to whether it discriminated against the plaintiff. Id.

Transco relies upon appellants' poor performance under the Program as the basis for discharging the appellants. Since the Program was created after extensive studies, we find that Transco has met its burden of production. Therefore, the burden of production now shifts back to Fisher and Kirchhoff to show that the Program was really a pretext for discrimination. In support of a finding of pretext, appellants contend that Transco's Program was unreasonable in its demands and was not implemented in a neutral manner.

Despite the fact that Transco's standard under the Program, i.e., a 100 percent performance level, was seemingly unreasonable in that this level was achieved very rarely by any selector, and in most weeks was not achieved at all, we agree with the District Court that this does not, standing alone, show intentional discrimination. Palucki, 879 F.2d at 1571. However, we believe that this standard, when coupled with the alleged manner of the Program's implementation, raises a genuine issue of material fact as to the Program being a pretext for age discrimination.

Appellants argue that erroneous disciplines occurred under the Program twenty-two times, of which 18 were applied against older workers and four against younger ones, and that the rate at which Transco disciplined workers decreased substantially after it had terminated several of its older workers. Because of this, appellants contend that eight younger workers who should have been fired were not, and that six older workers, including appellants, were fired when they should not have been.

To support its argument, appellants rely upon exhibit 6 to their expert's deposition. This exhibit sets forth the names and ages of Transco's warehouse workers, the performance level achieved for each week under the Program, whether the performance rendered an employee in the bottom 20% of those receiving scores, whether discipline had been given, and the form of the discipline.

An examination of exhibit 6 reveals that some errors were indeed committed. For example, for the week ending 10/13/84, 35 selectors were measured and none achieved a 100 percent level. Thus, seven of these constituted the lower 20 percent. However, appellant Fisher, who had received a performance level of 70.0% was disciplined despite his being one person above the 20 percent cut-off, while a younger selector, who had a level of 69.7%, was not.

Transco argues that this error and others "are nothing more than rounding problems in ... [appellants' expert's] data base." However, this is merely an assertion and is not documented. In sum, we believe that whether the alleged errors present in Transco's implementation of the Program, coupled with the very high standards under the Program, render the Program a pretext for age discrimination presents a genuine issue of material fact.

Next to consider is the least developed aspect of this case, the appellants' claim of disparate impact.3 To establish unlawful disparate impact under the ADEA, as under Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e et seq., an employee must first show that an employer utilizes a particular employment practice that results in a disparate impact upon a protected class of workers. If the employee makes such a showing, the employer must then set forth how the challenged practice serves, in a significant way, its legitimate employment goals, i.e., the business necessity of the practice. However, no proof of discriminatory intent is necessary.4

The appellants maintain that Transco's Program, while having facially neutral criteria, fell more harshly upon Transco's older employees. In granting Transco's motion for summary judgment, the District Court held, as the appellees now urge, that no conclusions could be gleaned from the statistical samples, as they were too small and too poorly presented. In the alternative, the District Court found that appellants had not rebutted Transco's claim that its Program was consistent with a business necessity.

Eleven terminations took place as a result of the Program, and 52 selectors were involved, of which 27 were age 40 or above. Of the 11 terminated, 10 were of a protected age. It is true that we, like other courts, are reluctant to render conclusions in a disparate impact claim based upon small statistical samples. See Soria v. Ozinga Bros., Inc., 704 F.2d 990, 995 (7th Cir.1983); Williams v. Tallahassee Motors, 607 F.2d 689, 691-92 (5th Cir.1979), cert. denied, 449 U.S. 858, 101 S. Ct. 159, 66 L. Ed. 2d 74 (1980). However, this is not a hard and fast rule and generally some substantial evidence of the unreliability of the statistics is asserted before it is applied. See e.g., Soria, 704 F.2d at 995.

In this case, where 10 of 27 older employees were terminated, as against 1 of 25 younger ones, it does not require expertise in differential equations to observe that an adverse ratio of approximately 10 to 1 is disproportionate. Thus, while the statistical analysis will have to be fully developed and presented at trial, we believe that the granting of summary judgment in favor of Transco on this issue by the District Court was premature.

The District Court found in the alternative that even if the appellants could show a disparate impact, the appellants had not rebutted Transco's claim that the Program was justified by business necessity. We disagree.

The standard required under the Program was so high that out of 1,182 attempts to meet it, the workers were successful only 20 times. Moreover, the Program was abandoned after 48 weeks, after several older employees had been eliminated, and was replete with errors in its administration. Under these circumstances, we believe that the District Court's grant of summary judgment was inappropriate. See Allen v. Seidman, 881 F.2d 375, 380 (7th Cir.1989). Simply put, whether the Program implemented at Transco was consistent with business necessity likewise presents a genuine issue of material fact.

For each of the above reasons, the District Court's order granting Transco's motion for summary judgment is reversed and the case is remanded for further proceedings consistent with this opinion. Circuit Rule 36 shall apply.