Full opinion text
OPINION STAGG, District Judge. Plaintiffs Samuel Carroll and Charles W. Grant filed this Title VII class action on July 25, 1977, asserting an across-the-board attack on defendant Sears, Roebuck and Company’s employment practices, which plaintiffs claim have a disparate impact upon black employees and applicants for employment. Plaintiffs argue that Sears’ discriminatory employment practices disadvantage blacks in the areas of hiring, training opportunities, promotion, compensation, and terminations. In addition, plaintiffs contend that Sears’ testing procedures disparately affect black applicants for employment and promotion. Finally, plaintiffs allege that black employees at Sears have been victims of disparate treatment. This court held a class certification hearing on July 6, 1979 and, in a ruling issued May 14, 1980, certified the following class: All present black employees of defendant Sears, Roebuck and Company’s Shreveport facilities; all former black employees employed at defendant’s Shreveport facilities on or after December 22, 1972; all future black employees of defendant’s Shreveport store; and all unsuccessful black applicants who applied for employment at defendant’s Shreveport facilities and were rejected on or after December 22, 1972. Despite the repeated discovery disputes that hampered the progress of this litigation, the case finally proceeded to trial on January 12, 1981. After considering the testimony and exhibits introduced during the nine days of this trial, and reviewing the parties’ post-trial briefs, the court enters the following findings of fact and conclusions of law. I. STANDARD OF PROOF IN DISPARATE IMPACT CASES This class action is based in part upon a disparate impact theory of discrimination. As the United States Supreme Court has noted, “claims that stress ‘disparate impact’ ... involve employment practices that are facially neutral in their treatment of different groups but that in fact fall more harshly on one group than another and cannot be justified by business necessity.” International Brotherhood of Teamsters v. United States, 431 U.S. 324, 335-36 n. 15, 97 S.Ct. 1843, 1854 n. 15, 52 L.Ed.2d 396 (1977). In a disparate impact case, the court’s inquiry is whether plaintiffs proved a discriminatory effect resulting from defendant’s employment practices. The Supreme Court has observed that “good intent or absence of discriminatory intent does not redeem employment procedures or testing mechanisms that operate as ‘built-in headwinds’ for minority groups.... Congress directed the thrust of the Act to the consequences of employment practices not simply the motivation.” Griggs v. Duke Power Co., 401 U.S. 424, 432, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971) (emphasis in original). Thus, proof of discriminatory intent is not required. Id. International Brotherhood of Teamsters, supra. Initially, plaintiffs have the burden of establishing a prima facie case. They must first introduce evidence of a disparate impact upon a protected class, so as to create an inference that defendant’s employment practices are discriminatory. Griggs, supra; Albemarle Paper Company v. Moody, 422 U.S. 405, 95 S.Ct. 2362, 45 L.Ed.2d 280 (1975). If plaintiffs’ evidence tends to support the existence of a disparate impact, the defendant will be given an opportunity to prove that plaintiffs’ proof is inaccurate or unreliable, and as such fails to establish that a disparate impact, in fact, exists. In Rowe v. General Motors Corp., 457 F.2d 348 (5th Cir. 1972), the Fifth Circuit noted that statistical evidence which shows an apparent disparate impact puts “on the employer the operational burden of demonstrating why, on acceptable reasons, the apparent disparity is not the real one.” Id. at 358. “Thus, both parties are afforded a full opportunity to prove or disprove the existence of a substantial disparate impact indicative of discrimination, with the ultimate burden of proof regarding this prima facie showing on the plaintiff.” Schlei and Grossman, Employment Discrimination Law at 1160 (1976) (hereinafter cited as Schlei and Grossman). If defendant is unable to show that plaintiffs’ initial proof of a disparate impact is unreliable, then plaintiffs will have successfully carried their burden of proving a prima facie case of discrimination. At that point, “the burden shifts to the defendant to prove that the substantial disparate impact is the result of a job-related selection device, a business necessity, a bona fide occupational qualification, or some other explanation which demonstrates that the substantial disparate impact is not the result of unlawful discrimination.” Id. While defendant carries the burden of proving any of these affirmative defenses, the ultimate burden of persuasion is on the plaintiffs. “[A] prima facie case of racial discrimination may be established by nothing more than valid statistical evidence of a significant discriminatory impact of defendant’s practices. ...” Williams v. Tallahassee Motors, Inc., 607 F.2d 689, 691 (5th Cir. 1979); United States v. Hayes International Corp., 456 F.2d 112, 120 (5th Cir. 1972). The courts have often placed heavy reliance upon statistical evidence, stemming in part from the Supreme Court’s aforementioned statement in Griggs that Title VII focuses on “the consequences of employment practices.” The courts have also placed- heavy reliance upon statistical evidence because, as a practical matter, such evidence may provide “the only available avenue of proof”, United States v. Ironworkers, Local 86, 443 F.2d 544, 551 (9th Cir. 1971), cert. denied 404 U.S. 984, 92 S.Ct. 447, 30 L.Ed.2d 367 (1971), and because discrimination “will seldom be admitted by any employer.” Marquez v. Omaha District Sales Office, 440 F.2d 1157, 1162 (8th Cir. 1971). See generally Schlei and Grossman at 1161-93. Though statistical evidence is obviously useful, even essential, to plaintiffs’ presentation, “the usefulness of such statistics varies with the surrounding facts and circumstances, which either support or undermine the inference of discrimination offered by the statistics.” Williams v. Tallahassee Motors, Inc., supra; International Brotherhood of Teamsters v. United States, supra, 431 U.S. at 340, 97 S.Ct. at 1857; United States v. Ironworkers, Local 86, supra, 443 F.2d at 551. Statistics are not an appropriate method of proving a prima facie case unless “such statistics [are] relevant, material and meaningful and not segmented and particularized and fashioned to obtain a desired conclusion.” EEOC v. Datapoint Corp., 570 F.2d 1264, 1269 (5th Cir. 1968); Robinson v. City of Dallas, 514 F.2d 1271 (5th Cir. 1975). In addition, the Fifth Circuit has warned that “undue emphasis” on the use of statistics “may obscure rather than advance the judicial process”, Hester v. Southern Railway Co., 497 F.2d 1374, 1381 (5th Cir. 1974) (footnote omitted), and that a court should avoid becoming involved in a “numbers game”. U. S. v. Hayes International Corp., supra, 456 F.2d at 120. As the Second Circuit Court of Appeals has noted, “after all the technical statistical jargon like ‘one tail’ or ‘two tail’ tests and ‘Chi-Square Test (Yates-corrected)’ as well as the less esoteric numbers and percentages [are] placed before the trial judge, it [is] his job to resolve the issues.” Chance v. Board of Examiners, 458 F.2d 1167, 1173 (2d Cir. 1972). In this case, both parties relied upon statistical evidence which they assert either proves or disproves a discriminatory impact against blacks. Plaintiffs presented the testimony of a political and social scientist, Dr. Marvin Stottlemire. His testimony was accompanied by several graphic exhibits. Defendant responded with the testimony and exhibits of Dr. Arnold Levine, an expert in the fields of applied and theoretical statistics and statistical computation. Plaintiffs contend that Dr. Stottlemire’s presentation establishes that Sears’ employment practices result in a disparate impact on blacks. Defendant argues that Dr. Levine’s statistical presentation affirmatively disproves any allegations of discrimination. In addition, each party contends that the other employed improper methodology in preparing its statistical presentation. II. SEARS’ SHREVEPORT FACILITIES AND AFFIRMATIVE ACTION PROGRAM Before addressing the parties’ statistical evidence, the court sets out the following findings of fact concerning Sears-and its Shreveport facilities. Sears, Roebuck and Company is a national chain of department stores and related facilities, with corporate headquarters in Chicago. During the last decade, the total number of Sears’ employees ranged from 360,000 to 420,000. At the present time, there are about 375,000 Sears employees. In the Sears hierarchy, five administrative areas, or territories, report to the corporate headquarters in Chicago. The territories are divided into zone offices, each of which reports to the appropriate territorial office. Forming the base of the Sears hierarchy are the various retail stores and related facilities. At a given store, certain officials report to the appropriate zone office, while others report to the territorial office. ' During the time period pertinent to this lawsuit, the Shreveport facilities were located in the Southwest Territory, which was headquartered in Dallas; Shreveport is also located within the Dallas zone. The Shreveport facilities consist of the Southern Avenue store, the Retail Distribution Center, the Central Service Unit, and the Southside store. Approximately 400 employees are currently employed at the Southern Avenue store, which also includes an automotive center; 70 to 100 persons are employed at the Retail Distribution Center; approximately 110 employees work at the Central Service Unit; and the Southside store has three or four employees. A major part of Sears’ defense of this lawsuit centered upon the existence of the corporation’s national Affirmative Action Program, instituted in 1968. Various Sears witnesses, most notably Raymond Graham, established the following facts regarding the Affirmative Action Program: 1. In 1968, the top officers at Sears decided to institute a new department, the Equal Rights Department. Sears established this department to aid in compliance with civil rights legislation, including Title VII. 2. Within the new Equal Rights Department, a position known as the Director of Equal Opportunities was established. Sears officials selected Graham to fill this position because of his extensive experience in personnel with Sears, and because of his personal commitment to the civil rights movement. Graham understood that Sears’ management decided to take a leadership role in providing opportunities for minorities. Graham then examined the law and recommended to top officers that a program be fashioned that would produce statistically significant showings of black representation and opportunities in Sears’ work force. The 2,000 Sears facilities at that time were each required to fill out a questionnaire to facilitate the launching of the Affirmative Action Program. 3. Since its issuance in 1965, Executive Order 11246 has required government contractors to comply with a program for equal opportunities for minorities or to be excluded from doing business with the government. At the time Sears’ Affirmative Action Program was being developed, the government had not yet stated any numerical goals for minority representation in an employer’s work force. Later, a government regulation required compliance with Executive Order 11246 through use of numerical goals. While helping to develop the Sears Affirmative Action Program in 1968, Graham anticipated this regulation and built numerical goals into the Sears Affirmative Action Plan. The goals devised by Graham were more stringent than those required by the government. 4. The General Services Administration (GSA) and the Office of Federal Contract Compliance Programs (OFCCP) are charged with overseeing compliance with Executive Order 11246. At the inception of the Affirmative Action Program, Sears voluntarily declared itself to be a governmental contractor, thus subjecting itself to monitoring by the GSA and the OFCCP. Since that time, the agencies have conducted 2,000 “reviews” or “audits” of Sears’ facilities — none of which revealed noncompliance. The GSA performed at least two audits of the Sears Shreveport facilities, and neither indicated any serious problem of noncompliance. 5. After the Affirmative Action Program had been devised, store managers and management at all levels were bombarded with letters and other communications which advised them on the myriad details regarding proper compliance with the program. Management was instructed that they could not expect promotions if their employment records revealed that they had failed to comply adequately with the program. 6. Sears’ Affirmative Action Plan establishes numerical percentage goals for minority representation in the work force at each Sears facility. These goals differ from one facility to another because the goals are based upon minority population in the applicable area. In addition, the plan establishes goals for minority representation in all job positions, not merely for the total percentage of hires. 7. One important aspect of Sears’ Affirmative Action Plan is the “MAG” Program. MAG is an acronym for Mandatory Achievement of Goals, and requires that job vacancies be filled by minorities whenever minorities are underrepresented in that job position. This program is more stringent than that of any Sears’ competitors and, in fact, is more stringent than governmental requirements. If no qualified minority applicant can be found to fill a vacancy in a position in which minorities are underrepresented, the Sears facility must get permission from its territorial office to hire or promote a white person into that vacancy. A facility’s request to hire a white person into such a position is known as a “MAG deviation request”. 8. In addition to the GSA or OFCCP reviews of the progress of the Affirmative Action Plan, and Graham’s overall supervision for compliance, each territorial office has an equal opportunity head who reports to Graham. During much of the period at issue in this case, David Wood was the Southwest Territory equal opportunity head who supervised compliance within the Southwest Territory. Neither Wood nor any of his successors ever reported to Graham that the Shreveport facilities were in any way deficient. 9. Prior to giving his testimony in this case, Graham personally reviewed the overall results of the Affirmative Action Plan at the Shreveport facilities, and found that the facilities are in compliance with governmental requirements, though the stores are still striving to meet the more stringent goals of the Affirmative Action Program. Graham studied the facilities’ progress in the same way he reviews “problem stores” that are called to his attention in Chicago. In Graham’s opinion, the increase in black representation in Sears’ Shreveport work force demonstrates that the Affirmative Action Plan has had “excellent” results. On May 26, 1977, the GSA issued a compliance letter regarding the Shreveport facilities; this letter indicates that a GSA review found that the facilities were in compliance with Executive Order 11246. 10. One important aspect of the Affirmative Action Program involves the recruiting of minority applicants for employment at Sears. When Esther Colbert Wilkins was Personnel Manager at Sears, she visited area trade schools seeking qualified black applicants, contacted high schools and took part in career days, arranged teas with local black leaders, placed ads in the area newspapers, and recruited black employees from competitors’ stores. J. H. Hutchison, manager of the Southern Avenue store from May 1973 to June 1978, made presentations at Grambling State University and other area black schools with Ms. Wilkins and several department managers. Lyla Render, the current Sears Personnel Manager, has also taken part in such recruiting efforts as career days. There are several relevant factors aside from the Affirmative Action Program which affect applicants for employment at Sears and the makeup of the Sears work force. First, the Sears Personnel Department, which accepts applications for employment, is located within one mile of the State Employment Office. Persons who receive state unemployment benefits must at some point prove that they have been seeking work to continue receiving those benefits. Several witnesses, including Ms. Wilkins, whose testimony was generally unfavorable toward Sears, established that a significant number of blacks proceed from the State Employment Office to Sears’ Personnel Department to apply for work. Thus, the close proximity of the State Employment Office to Sears’ Personnel Department skews Sears’ applicant flow to some extent. Second, the turnover in the full-time work force at Sears’ Shreveport facilities is very low. For example, among Sears’ Shreveport sales force, the average full-time sales person has been with Sears for 14 years. Herb Wolfram, the Southern Avenue store manager after Hutchison’s departure, testified that the Southern Avenue store has one of the lowest turnover rates of any Sears store with which he is familiar. Though Ms. Wilkins’ testimony was highly critical of Sears’ compliance with its Affirmative Action Program, she admitted being aware that it is difficult to meet the Affirmative Action goals in the face of such low turnover. As will be discussed infra, this low turnover plays a major role in the relatively lower representation of minorities in higher-level full-time positions. Third, there is no question that Sears’ hiring and promotion policies and practices are in many respects highly subjective. One must consider Sears’ hiring criteria in conjunction with its hiring practices and procedures. The Personnel Department at Sears is open six days per week, and hires from among the applicant flow to fill specific job vacancies. The department, and in particular the Personnel Manager, determines the need for new hires by keeping track of Sears’ schedule hours, noting those areas where help is needed. The department also becomes aware of job openings when management advises of seasonal needs. If at any given time no positions are available, the department catalogs the applications submitted by “good” applicants and pulls those applications at a later date when positions come open. Ms. Render established that the personnel interviewers immediately begin to “size up” applicants, even if no jobs are open at the time they apply. The hiring criteria or “qualifications” which a personnel interviewer seeks are very subjective for some job positions and less so for others. For example, when evaluating an applicant for a sales position, the applicant’s appearance is of initial concern to the interviewer. As Ms. Render noted, an applicant’s appearance is the first thing an interviewer sees and a sales person’s appearance is the first thing a customer sees. Though satisfactory appearance is an important asset for a sales person, there is no question that “appearance” is a subjective hiring criterion. Most of the other qualifications for sales positions at Sears are similarly subjective. If an applicant’s appearance passes muster, the personnel interviewer will consider such attributes as articulateness, maturity, aggressiveness, and friendliness. Obviously, the ability to speak clearly and correctly is an important quality for a sales person. In regard to the “maturity” factor, as Ms. Render observed, a recent high school graduate would have little credibility with a housewife in giving advice about draperies. Aggressiveness is undoubtedly an important quality for a sales person. Finally, a sales person’s friendliness not only makes customers more willing to deal with that sales person, but also enhances Sears’ overall reputation as a pleasant place to shop. Thus, these hiring criteria for sales persons, like the hiring criterion concerning an applicant’s appearance, are certainly appropriate factors for a retailer to consider in making hiring decisions. At the same time, these criteria are obviously subjective. On the other hand, many of the qualifications for other job positions at Sears are quite objective. For example, experience is a desirable attribute for almost all jobs at Sears. In addition, many of the job positions inherently require objective qualifications. The Sears’ automotive department could not remain in business if it hired unqualified mechanics. The same is true for Sears’ service department, which needs qualified electricians and repairmen. In fact, state law prohibits the hiring of electronic repair technicians not certified by the state. Sears’ promotion criteria are predominantly subjective. In the pretrial order, Sears stipulated that its criteria for promotion include fitness and demonstrated ability, employees’ past work histories and abilities to produce, annual evaluations, and the opinions of supervisors. Sears maintains no written criteria or guidelines for promotion, and does not post notices concerning specific job openings or promotion opportunities. It is undisputed that a majority of the supervisors at Sears are white. Given the importance of supervisory ratings of employees and opinions concerning employees’ work performance, black employees are clearly subjected to a greater risk of conscious or subconscious racial bias than are white employees. The fact that Sears’ hiring and promotion practices are subjective to some degree does not mean, however, that these practices are unlawful. In Hester v. Southern Railway Co., 497 F.2d 1474 (5th Cir. 1974), the Fifth Circuit held that subjective practices “are not violative of Title VII per se. Title VII comes into play only when such practices result in discrimination.” 497 F.2d at 1381 (emphasis added). Unless plaintiffs successfully prove that Sears’ subjective practices result in a disparate impact upon blacks, the existence of those practices does not entitle plaintiffs to prevail under Title VII. Finally, the court agrees with Sears’ assessment that this case is distinguishable from the many reported decisions which concern allegations of discrimination against blacks and other minorities at large manufacturing plants. Sears correctly highlights the fact that the general public is not presumptively qualified for most of the job classifications at Sears, unlike the situation facing many manufacturers in which literally anyone in reasonably good .health can perform assembly-line duties or other manual labor. This is true regardless of the fact that many of Sears’ qualifications are measured by subjective criteria. As the Supreme Court noted in Griggs, “Congress did not intend by Title VII ... to guarantee a job to every person regardless of qualifications.” 401 U.S. at 403, 91 S.Ct. at 853. In view of this language, the Fifth Circuit has held that qualifications “are an employer’s prerogative.” Rowe v. General Motors, supra, 457 F.2d at 358. A retail employer such as Sears would be seriously disadvantaged if it were unable to use subjective hiring or promotion practices that are obviously job-related. As mentioned previously, however, the law does not absolutely impose this disadvantage, since subjective practices are not violative of Title VII unless they result in a disparate impact upon a protected class. Accordingly, the court will now consider the parties’ statistical evidence to determine whether Sears’ employment practices have produced a disparate impact upon blacks. III. THE STATISTICAL EVIDENCE (A) Hiring Plaintiffs contend that Sears’ subjective hiring policies result in a disparate impact upon black applicants for hire. In support of this contention, plaintiffs presented Exhibits 33 and 34-40, along with Dr. Stottlemire’s testimony concerning those exhibits. The exhibits reveal that in every year from 1973 through 1979, more blacks than whites applied for employment at Sears, yet a greater percentage of the white applicants was hired in every year except 1975. In every year except 1975 and 1976, the total number of whites hired exceeded the total number of blacks hired, again despite the relatively larger number of black applicants. Dr. Stottlemire analyzed this data in order to determine whether Sears discriminates against black applicants by hiring disproportionately few of them. He concluded that the hiring data contained in plaintiff’s Exhibit 33 warrants that inference. In reaching this conclusion, Dr. Stottlemire relied upon an assumption that qualifications are randomly distributed among black and whites, /. e., that blacks and whites are equally likely to be qualified for each job. If one accepts this assumption, it follows that one would expect equal percentages of successful black applicants and successful white applicants. Exhibit 33 is based upon these principles, and reveals that the actual number of blacks hired in every year except 1975 was less than the “ideal” number of black hires. Dr. Stottlemire’s analysis is fraught with problems. First of all, the court accepts Dr. Levine’s testimony that the black applicant pool is not appearing at random from the experienced labor force. Dr. Levine concluded that there is a significant statistical disparity between the percentage of black applicants and the percentage of experienced black workers in the area labor force. This disparity may be explained in large part by the close proximity of the State Employment Office to Sears’ Personnel Department and by Sears’ efforts pursuant to the Affirmative Action Program to recruit black employees. In addition, Ms. Render established that Sears’ applicant flow includes a large influx of black applicants from the Job Corps, the O.I.C. and from various drug and alcohol rehabilitation programs. The evidence established that “most” of the applicants from the State Employment Office are black; “most” of the Job Corps and handicapped applicants are black; and all of the O.I.C. applicants are black. Obviously, all of the applicants recruited pursuant to the Affirmative Action Program are black. The court concludes that these nonrandom factors result in an artificially high black applicant flow, such that use of that data for any purpose, and in particular for comparison with the percentage of white hires from white applicant flow, is unquestionably suspect. On cross-examination, Dr. Stottlemire admitted that Sears’ recruiting efforts and the close proximity of the State Employment Office, among other factors, could have the effect of inflating black applicant flow. However, given his assumption that qualifications are spread randomly throughout white and black employees, he stated his belief that it does not matter why applicants show up at the Sears Personnel Department since he would expect equal percentages of black and white applicants to be hired, regardless why those applicants actually applied. This brings the court to the second major problem with Dr. Stottlemire’s analysis. Plaintiffs offered no proof which would tend to show that Dr. Stottlemire’s assumption regarding the random distribution of qualifications is accurate. In response to a question posed on cross-examination, Dr. Stottlemire gave no clear answer whether he thinks this assumption is valid for mechanics or other skilled operatives, stating only that he is not a demographer. He simply assumes that if 20 of 100 whites are qualified to be mechanics, then 20 of 100 blacks are qualified for mechanics positions. But, Dr. Stottlemire admitted that he has no basis for this assumption, and that it was merely made for use with his statistical model. He further admitted that the introduction of specific job qualifications for specific job categories “may” have an effect on the assumption. As discussed above, there is no question that most, though certainly not all, of the job categories at Sears do involve specific job qualifications, though these qualifications are in many cases subjective. The Fifth Circuit has decried the use of subjective factors where there is a demonstrated discriminatory impact upon a protected class, but has never indicated that the subjective nature of a hiring criterion renders it any less a “qualification” for the job in question. The United States Supreme Court has implicitly recognized that a particular job may require “special qualifications’ unless “the job skill . . . involved [in that particular job] is one that many persons possess or can fairly readily acquire.” Hazelwood School District v. United States, 433 U.S. 299, 308 n. 13, 97 S.Ct. 2736, 2742 n. 13, 53 L.Ed.2d 768 (1977). It is clear that some of the positions at Sears do not involve special qualifications. However, this court cannot conclude that retail sales positions and clerical positions require job skills which are so commonly available that the general population can be considered presumptively qualified for those positions, regardless of the fact that the qualifications are subjective. Dr. Stottlemire admitted that his model would have to be adjusted to take into account the “qualified applicant pool” if Sears had “compelling data” showing that qualifications are not randomly distributed among blacks and whites. The problem with this reasoning is that Dr. Stottlemire would place upon Sears the burden of disproving the very assumption which is of critical importance to his analysis. Contrary to Dr. Stottlemire’s suggestion, the burden is upon plaintiffs to establish a prima facie case by competent proof; this they cannot do by making a self-serving assumption which underlies their statistical presentation and for which no competent proof was adduced. As mentioned above, Dr. Stottlemire frankly admitted that he had no basis for his assumption. The court rejects this assumption for lack of supportive proof, and rejects plaintiffs’ methodology insofar as it relies on applicant flow data that has not been adjusted in any way to take qualifications into account. In Fisher v. Proctor and Gamble Mfg. Co., 613 F.2d 527 (5th Cir. 1980), the Fifth Circuit indicated that “a prima facie case may be shown without evidence of qualifications where the inference of discrimination is supported by a compelling level of racial underrepresentation in a sizable work force.” Id. at 544 (emphasis added). Under this principle, it is arguable that plaintiffs in an appropriate case need not take qualifications into account in proving a prima facie case of discrimination in hiring if they present “compelling” data that a defendant hires too few blacks. In this case, however, plaintiffs’ own data reflects the relatively large percentage of black applicants hired by Sears during the relevant period. Plaintiffs’ Exhibit 33 shows that 1,953 applicants were hired from 1973 to 1979. In a community in which the black population has ranged from 30 to 34 per cent during the relevant period, 885 of those hired, or 45.3 per cent, were black. It is also true that 8,779 of the 15,544 applicants during the period, or 56.5 per cent, were black, but the court has already found that the black applicant tally is artificially inflated by certain nonrandom factors, so a comparison between the percentage of black applicants and the percentage of successful black applicants is misleading. Plaintiffs’ failure to present compelling evidence that Sears hires too few blacks renders the foregoing language from Fisher inapplicable. More importantly, plaintiffs’ raw applicant flow data, unadjusted for qualifications, is of questionable probative value. The court agrees with Sears’ assessment of Dr. Stottlemire’s assumption that qualifications are randomly distributed among blacks and whites: [The assumption] flies in the face of common knowledge and experience. The very reason for training programs for minorities, and for affirmative action programs, is the recognition that for historical reasons blacks are not as qualified for all jobs as whites. Sears’ Posttrial Brief at 10-11. Several Supreme Court opinions support this position. In United Steelworkers v. Weber, 1443 U.S. 193, 99 S.Ct. 2721, 61 L.Ed.2d 480 (1979), the Court discussed the history and purpose of Title VII: Congress’ primary concern in enacting the prohibition against racial discrimination in Title VII . . . was with “the plight of the Negro in our economy.” Before 1964, blacks were largely relegated to “unskilled and semi-skilled jobs”. Because of automation the number of such jobs was rapidly decreasing. As a consequence “the relative position of the Negro worker [was] steadily worsening. In 1947 the non-white unemployment rate was only 64 per cent higher than the white rate; in 1962 it was 124 per cent higher.” Congress considered this a serious social problem.... Congress feared that the goals of the Civil Rights Act— the integration of blacks into the mainstream of American society — could not be achieved unless this trend were reversed. And Congress recognized that would not be possible unless blacks were able to secure jobs “which have a future”. 443 S.Ct. at 202 99 S.Ct. at 2727 (citations omitted). The Supreme Court earlier recognized that blacks are disadvantaged in competing for jobs because they “have long received inferior education in segregated schools”. Griggs v. Duke Power Co., supra, 401 U.S. at 430, 91 S.Ct. at 853. Thus, Congress enacted Title VII with an eye toward blacks’ economic “plight” caused by historically inferior education and the historical relegation to unskilled or semi-skilled jobs. Sears has attempted to face this problem by recruiting large numbers of black applicants, which increases Sears’ chances of finding qualified black employees so that the affirmative action goals may be met. Dr. Stottlemire’s assumption of random distribution of qualifications is clearly unrealistic, aside from the fact that it was not proven. In summary, the court finds that plaintiffs’ statistical presentation contained in plaintiffs’ Exhibits 33 and 34-40, fails to demonstrate the existence of a disparate impact upon black applicants for hire at Sears, and thus fails to establish a prima facie case of discrimination in hiring, for the following reasons: (1) Plaintiffs’ reliance upon raw .applicant flow data is misplaced in the context of a work force wherein many job categories require “special qualifications”. Generally, an attempt to utilize applicant flow data to prove discrimination in hiring is a sound approach: the Fifth Circuit has ruled that “the most direct route to proof of racial discrimination in hiring is proof of disparity between the percentage of blacks among those applying for a particular position and the percentage of blacks among those hired... . ” Hester v. Southern Railway Co., supra, 497 F.2d at 1379. More recently, the Fifth Circuit observed that “actual” applicant flow data is “often the best measurement”. Phillips v. Joint Legislative Comm., at p. 1025 (5th Cir. Feb. 23, 1981). However, this data must be adjusted in appropriate circumstances to take special qualifications into account; as one source observes, Hester “clearly supports the view that the plaintiff’s statistics must focus on the qualified labor market.” Schlei and Grossman, supra, at 1174 (emphasis added). Plaintiffs’ failure to adjust the applicant flow data for qualifications renders that data unreliable as a basis for comparison, and the Fifth Circuit recognized in Robinson v. Union Carbide Corp., 538 F.2d 652 (5th Cir. 1976) that a district court may reject statistical analyses which are based upon unreliable applicant flow data. (2) Apparently, plaintiffs’ only reason for failing to adjust the applicant flow data for qualifications was their trust in Dr. Stottlemire’s assumption that qualifications are randomly distributed among blacks and whites. But plaintiffs offered no proof to establish the validity of this assumption, and Dr. Stottlemire admitted he had no basis for the assumption. Moreover, the assumption is unrealistic in that it fails to consider the historical disadvantages which past discrimination has visited upon blacks in our society. The effects of past discrimination certainly constitute a nonrandom factor that would have a profound influence upon the “random distribution” of qualifications among blacks and whites. (3) The data contained in plaintiffs’ Exhibit 33 reveals that 45.3 per cent of all persons hired from 1973 to 1979 were black, a figure which compares very favorably with the general population figures of 30 to 34 per cent during the pertinent period. This fact precludes any finding by the court that plaintiffs have demonstrated such a compelling underrepresentation of black hires that a prima facie case may be demonstrated without the necessity of any evidence of or adjustment for qualifications. (4) Finally, the use of applicant flow data is particularly suspect because of the inflated numbers of black applicants. Dr. Stottlemire was untroubled by this factor only because of his assumption regarding random distribution of qualifications. Having rejected this assumption, the court questions the reliability of statistics based upon the skewed black applicant flow data. This factor, in combination with plaintiffs’ failure to adjust the' applicant flow data for qualifications, renders that data unreliable. Sears pointed out these problems with Dr. Stottlemire’s methodology and also submitted Dr. Levine’s statistical presentation, which allegedly disproves any inference that Sears’ employment practices result in a disparate impact against blacks. In his presentation, Dr. Levine has attempted to avoid Dr. Stottlemire’s mistake of failing to take qualifications into account by comparing the composition of Sears’ labor force with the labor force in the general area. The 1970 United States Census data, in conjunction with the 1975 updated material, contains information regarding racial representation in various EEO job categories. Table 1A shows the percentages of blacks in each EEO job category in the area labor force. The information in this table was extracted from the 1970 U.S. Census data. The “Totals” column adds Bossier Parish, Caddo Parish and Shreveport SMSA data to represent a weighted average of these areas so that no relevant area is excluded. Table 1B compares the percentage of blacks in the area labor force, from Table 1A, with the percentage of blacks in the Sears labor force. The last column on Table 1B, the “1975 census update (Shreveport)” shows that the 1975 update figures are very comparable to the 1970 census data so that the passage of time evidently has not affected the validity of the 1970 data. Table IB demonstrates the gains which blacks made in the Sears work force from 1973 to 1977. Black representation in the more “desirable” job categories increased significantly while black representation in the operatives, laborers and service categories, to which blacks have traditionally been relegated, actually decreased. As Dr. Levine noted, Table 172 of the 1970 Census data shows that, at that time, the experienced labor force in this area was 28 per cent black. He concluded that a statistician would expect Sears’ work force to be 28 per cent black “if everything is random”, i. e., if no discrimination is taking place. Plaintiffs’ Exhibit 49 shows that Sears’ work force has been in excess of 28 per cent black in every year except 1974, when blacks represented 26.8 per cent of the work force. Exhibit 49 also reveals the impressive increases in the percentage of black representation at Sears, undoubtedly due in part to the affirmative action program. Tables IF and 1G, and Figure 1G, reflect the principle that, in determining the number of “expected” black hires, it is not valid to compare the percentages of blacks at Sears to the percentages of blacks in the SMSA. Rather, one must assign appropriate weight to the occurrence of each job category in the SMSA in proportion to its occurrence in the Sears workforce. Dr. Levine assigned weights, or “Q values” to the SMSA figures, since the job structure at Sears is not the sanie as the job structure of the SMSA area in general. This procedure adjusts the labor force data to reflect the availability of jobs in the Sears labor force. According to Dr. Levine, this procedure yields the “hypothetical qualified people” in the area. Once the “Q calculation” is performed by year and by job category, it appears that blacks are overrepresented when one considers the Sears work force in comparison with the weighted SMSA work force. In several exhibits which are relative not only to the issue of hiring, but also to the issue of job assignment which will be discussed infra, Dr. Levine has attempted to show that the actual number of black hires exceeds the expected number of black hires for all positions except laborers. Table 2D shows that in 1977 more whites were hired into the laborer category than were expected. Defendant’s Exhibit 45 shows a similar result in 1978 in the laborers’ category, and in 1979 in the laborers and operatives categories. Based upon these exhibits, Dr. Levine concluded as follows: (1) When one removes the laborers category from this data, the number of black hires exceeds the expected number of black hires in every year; (2) The overrepresentation of white hires in the laborers’ category can be explained by the strategy of the affirmative action plan to place more whites into traditionally black jobs. On rebuttal, plaintiffs offered several objections to Dr. Levine’s methodology. First, both Dr. Stottlemire and Dr. Kenneth Hintz criticized Dr. Levine’s usage of the 1970 census data because such data reflects the effect of institutional racism and past discrimination against blacks. On this point, Dr. Levine admitted that society-wide discrimination would undermine the usefulness of the 1970 census data; but he noted that there is much data indicating that the census information is, in fact, useful. More importantly, Dr. Levine observed that if one “fudge factors” the 1970 percentages upward to account for any effect of past discrimination, Sears would still meet those higher percentages since Sears’ percentages are higher than those in the SMSA. Second, Dr. Stottlemire questioned Dr. Levine’s assumption that the distribution of qualifications among applicants for employment at Sears will reflect SMSA percentages. However, in Robinson v. Union Carbide Corp., supra, the Fifth Circuit upheld a district court’s utilization of area labor force data, noting that “on many occasions, federal courts have compared the composition of the company’s work force to the composition of the labor force in the surrounding area.” 538 F.2d at 658 (citations omitted). Dr. Hintz raised a more serious objection to the use of the 1970 census data. In many of Dr. Levine’s comparisons, such as those in Table IB and Table 1C, the 1970 census figures are frozen while the Sears data is permitted to advance over time. Dr. Hintz observed that the advances one sees in the Sears labor, force may have been occurring in the area labor force as well. Though this point is theoretically well-taken, Dr. Hintz admitted on cross-examination that he had no evidence that the black representation in the area labor force has increased from 1970 to 1980 as it has in the Sears labor force. In fact, Dr. Hintz acknowledged that the 1975 manpower update data contained in Table 2B appears to show that in all but one EEO category black representation was decreasing from 1970 to 1975. Finally, Dr. Hintz criticized Dr. Levine’s hypothesis that the Sears applicants will mirror the SMSA data. According to Dr. Hintz, a demographer would prefer to utilize applicant flow data. However, as discussed previously, Dr. Hintz acknowledged that applicant flow data must be adjusted for any known deficiencies in order to be useful, yet plaintiffs failed to adjust the applicant flow data to take qualifications into account. Given the choice between unadjusted applicant flow data and Dr. Levine’s assumptions concerning the use of area labor force data, the court chooses to accept the latter. The court has already noted the problems inherent in plaintiffs’ statistical presentation, and has found that the presentation fails to establish a prima facie case of discrimination in hiring. Moreover, the court finds that Dr. Levine’s statistical analysis reveals the impressive success of the affirmative action program in increasing black representation in the Sears work force. Not only did plaintiffs fail to prove their allegations of discrimination in hiring, but the totality of the evidence shows that, in fact, Sears’ employment practices do not result in a disparate impact upon black applicants for hire. (B) Job Assignment Sears allegedly discriminates against blacks in job assignment in two ways: first, by assigning a disproportionate number of black hires into part-time positions; second, by placing a disproportionate number of black hires into lower-paying, less prestigious job categories. In other words, plaintiffs contend that those blacks who manage to get hired are placed into less lucrative and/or less desirable jobs. These contentions will be considered separately. (1) Part-time placement There are two categories of part-time employees at Sears: “part-time” employees and “part-time regular” employees. “Part-time” employees generally work less than 30 hours per week, except in rush seasons when they sometimes are required to work 40 hours per week. “Part-time regular” employees are part-time employees who have been on the Sears payroll for one year or longer. - Part-time and part-time regular employees who work 16 40-hour weeks are automatically promoted to full-time. Obviously, “full-time” employees are those who work 40 hours per week. Raymond Graham, J. H. Hutchison, Herb Wolfram, Lyla Render and other Sears witnesses established the importance of part-time employees to a business which is primarily engaged in retail merchandising. The seasonal nature of retail sales dictates that more employees are needed in peak seasons, such as Christmas. Raymond Graham established that in some Sears facilities 50 to 60 per cent of the total annual work hours ours are performed by part-time employees. In most cases, employees who perform the same job are paid the same wage. In addition to receiving less total pay because they work fewer hours, part-time and part-time regular employees receive fewer benefits than full-time employees receive. Plaintiffs contend that Sears discriminates against blacks by placing a disproportionate number into part-time and part-time regular positions where they receive less total pay and fewer benefits. In support of this contention, plaintiffs presented Exhibits 41, 42A-F, 43-48, 50-55 and 56 along with the testimony of Dr. Stottlemire. Before discussing these exhibits, the court must set out the limitations of this inquiry. First, J. H. Hutchison, the Southern Avenue store manager from May, 1973 through June, 1978 established that there were few full-time hires during his tenure at the store because of low turnover and the weak economy. Hutchison stated that “most” full-time employees were promoted to full-time from part-time positions. Herb Wolfram, the Southern Avenue store manager following Hutchison, was more precise about the extent to which full-time positions are filled through promotion from part-time: he estimated that 90 per cent of all full-time employees moved up from the part-time ranks. With this factor in mind, it appears that very few newly-hired employees of either race, perhaps as low as 10 per cent of all full-time employees, are initially assigned into full-time positions. Thus, if blacks are underrepresented in full-time positions, that circumstance is more likely indicative of discriminatory promotion practices rather than discriminatory job assignment practices. Plaintiff’s allegations of discrimination in promotions will be discussed infra. Second, much of the statistical data presented by both parties involved part-time regular employees, as well as part-time employees. As noted above, part-time regular employees are merely part-time employees who have been on Sears’ payroll in excess of one year. It is clear that no one, black or white, is assigned to a part-time regular position. Therefore, the part-time regular classification is irrelevant in regard to the issue of discrimination in job assignment. Third, Sears’ witnesses established conclusively that many, if not most, of the part-time employees, black and white, were initially assigned into part-time positions because they applied for those positions. Raymond Graham established that many applicants for employment at Sears seek part-time work. Some of the typical reasons why applicants desire part-time positions include the following: (1) some applicants have other jobs; (2) some applicants are housewives who want to supplement their family income; and (3) some applicants are college students. For one notable example, Ms. Render stated that all of the part-time porters presently employed at Sears sought part-time employment because they are holding down other jobs. In summary, plaintiffs’ persistent attempts to characterize part-time employees as unfortunate underlings in Sears’ lowest caste are unrealistic in the sense that they fail to consider (1) the importance of part-time employees to a retail merchandising establishment, (2) the fact that the great majority of full-time employees were promoted to those positions from part-time positions, or (3) the fact that many of the part-time employees are initially assigned into those positions because they request them. Plaintiffs’ Exhibit 41 analyzes full-time and part-time hiring in light of the applicants’ requests either for full-time or part-time positions. This data covers only April, 1978 through March, 1980 because the race of an applicant could not be determined from information contained on the application prior to this period. Plaintiffs assumed for the purpose of this exhibit that persons who were hired part-time requested part-time positions, and that persons hired full-time requested full-time positions. This assumption is erroneous. On a Sears application form, an applicant may express a preference for either part-time or full-time work by checking an appropriately marked box. Ms. Render established that many applicants checked both the part-time and full-time boxes, while some applicants check neither. In fact, some of the class members who testified at this trial checked both boxes to indicate that they would consider accepting any available work. Even assuming for the sake of argument that this assumption does not affect the reliability of the data contained in Exhibit 41, the court finds that this exhibit does not demonstrate that black hires are discriminatorily assigned to part-time rather than full-time positions. To begin with, this exhibit offers strong evidence of the effect of this country’s weak economy during the two-year period covered by the data. During this two-year period, 4,031 out of 4,384 applicants for employment at Sears, fully 92 per cent of all applicants, were not hired. Although 70.6 per cent of all applicants requested full-time positions, only 2 per cent received full-time employment. More importantly, 79.3 per cent of all white applicants who were hired were assigned to part-time positions, and 85.3 per cent of the black hires received part-time assignments. On cross-examination, Dr. Stottlemire admitted that this is a small difference. In addition, there is very little difference between the percentage of black and white applicants who sought part-time work: 28.3 per cent of all black applicants requested part-time work, while 31.1 per cent of all white applicants requested part-time. Finally, Exhibit 41 shows that 57.5 per cent of all applicants hired were white; 42.5 per cent of the hires were black. Again, it should be noted that the 42.5 per cent black hires during this period well exceeds the 30-34 per cent black population in the general area. Ms. Render and her staff at Sears made a study of all applications during the period covered by Exhibit 41 to determine the effect, if any, of further breaking down the applicant data to reflect those applicants who requested “both” part-time and full-time, and those applicants who requested neither. Ms. Render testified that 39.4 per cent of all black applicants sought part-time work, as opposed to 43 per cent of the white applicants. Her tabulation also showed that 79 per cent of all black hires were assigned part-time positions, while 76.8 per cent of all white hires were assigned part-time. Dr. Levine testified that there is no significant difference between Ms. Render’s figures and those contained in plaintiffs’ Exhibit 41. He concluded that, regardless which figures are used, the data reveals no significant disparity between part-time hires by race or applicants for part-time by race. The court accepts this conclusion. Plaintiffs’ other exhibits which deal with part-time placement are Exhibits 49, 50-55 and 56. Exhibit 49 demonstrates the numbers and percentages of whites and blacks employed in full-time, part-time and part-time regular positions for each year from 1974 through 1980. Dr. Stottlemire admitted that Exhibit 49 is “not very useful” because blacks are not compared to whites but to the entire work force as a whole. However, Dr. Stottlemire attempted to make the relevant comparisons, for example, black full-time to white full-time, in Exhibits 50-55. The court agrees with Dr. Stottlemire’s conclusion that Exhibit 49 is not useful as proof of discrimination in job assignment, but Exhibit 49 does offer powerful evidence of the commendable success of Sears’ Affirmative Action Program. The exhibit reveals the following changes in the numbers and percentages of black and white employees from 1974 to 1980: Dr. Levine established that the 13 per cent decrease in Sears’ total work force should trigger a concomitant decrease for each racial group if the process were racially neutral. Yet, the figures reveal a substantial decrease in white employees and a substantial increase in black employees. The court concurs in Dr. Levine’s conclusion that Sears’ Affirmative Action Program is the nonrandom factor which most likely explains the significant increase in black employees while the work force as a whole declines. Plaintiffs’ Exhibits 50-55 contain Dr. Stottlemire’s comparisons between the numbers and percentages of black and white full-time, part-time regular and part-time employees for each year from 1974 through 1979 for selected job categories. However, these exhibits are of marginal utility in establishing plaintiffs’ allegations of discrimination in assignments to part-time positions. First, hiring is but one element which affects these job compositions. For example, the court has already noted that the average employment period for full-time sales people at Sears is 14 years. These exhibits unquestionably take into account employees who were hired prior to the effective date of Title VII. In addition, as discussed previously, Sears’ traditional low turnover in full-time positions has a definite effect upon the numbers and percentages of both white and black full-time employees over the period covered by these exhibits. Thus, Dr. Stottlemire admitted that the entire history of the Sears work force is “rolled into” the various compositions of job classifications contained in these exhibits. Though Exhibits 50-55 offer inconclusive evidence regarding placement of blacks into part-time positions, they, like Exhibit 49, reveal the striking gains made by black employees in certain categories. For example, the data concerning sales persons shows black representation increasing at a time when the total sales force was decreasing and the white sales force was decreasing dramatically. It is particularly noteworthy that black full-time sales persons increased by 40 per cent during the period, while white full-time sales persons decreased by 27.5 per cent. Over the entire period, the total number of black sales persons increased very slightly (by two employees), while the total number of sales persons decreased by 25.4 per cent. Finally, plaintiffs’ Exhibit 56 reveals the percentages of full-time black employees in selected job categories. Unfortunately, this exhibit is subject to many of the same shortcomings as the previously-discussed exhibits. Like Exhibits 50-55, this exhibit contains data relating to the overall composition of several job classifications, which inherently involves the entire history of each job classification rather than simply the hiring element. Accordingly, the data contained in this exhibit does not conclusively indicate anything about initial assignment of blacks into full-time positions. A more important problem with this exhibit is that Exhibit 56, in many job categories, deals with very small numbers; a change of one or two employees could radically affect the percentages. For example, Exhibit 56 shows that, in 1977,14.3 per cent of all full-time clerical employees were black. However, Ms. Render established that this percentage indicates only that one out of seven full-time clerical employees was black. Similar circumstances exist for several of the job categories contained in this exhibit. In this regard, the Fifth Circuit has recognized that statistical information which involves an extremely small universe is of questionable probative value in proving discrimination. See, e. g., Ochoa v. Monsanto Company, 473 F.2d 318 (5th Cir. 1973); Rogillio v. Diamond Shamrock Chemical, 446 F.Supp. 423, 428 (S.D.Tex.1977). For these reasons, the court finds that plaintiffs’ Exhibits 41, 49, 50-55 and 56 are inconclusive and fail to establish that blacks are disproportionately assigned into part-time positions. (2) Placement into “lower status” job categories In support of their contention that Sears discriminates against blacks by placing a disproportionate number of black hires into lower-paying, less-prestigious job categories, plaintiffs introduced Exhibits 42A-F. These exhibits purport to show the numbers of whites and blacks hired into specific job classifications from 1974 through 1979. Based upon these exhibits, Dr. Stottlemire concluded that there is an association between the type of job a person receives and his race. Plaintiffs contended in their pretrial memorandum that the exhibits illustrate “the disproportionate assignment of black hires to lower status jobs.” Plaintiffs’ Pretrial Memorandum at 17. The court finds these exhibits fatally defective in several respects: (a) Dr. Stottlemire admitted that the figures contained in these exhibits do not guarantee that the people shown in each position were hired into that position. Rather, the exhibits reflect the positions held at the end of a given year by employees who were hired during that year. For example, if a person was hired in early 1974 as a sales person, but was later transferred to a clerical position and remained in that position until the end of 1974, Exhibit 42A would carry that person as a clerical hire despite the fact that he was initially assigned to a sales position. In addition, Dr. Stottlemire testified that Exhibit 42 reflects only those employees who were hired during a given year and remained on the payroll at the end of that year. Thus, if someone was hired in January of 1974 and was fired or quit in February of the same year, he would not appear on Exhibit 42A within the category to which he was assigned. Furthermore, a comparison of Exhibits 42A-F with Exhibit 33 shows