Full opinion text
MEMORANDUM WARRINER, District Judge. I On 23 December 1976 the United States brought suit in this Court against the Commonwealth of Virginia and the Superintendent of the Virginia State Police to enforce the provisions of Title VII of the Civil Rights Act of 1964 as amended, 42 U.S.C. § 2000e et seq. and of § 518(c)(1) of the Omnibus Crime Control and Safe Streets Act of 1968 as amended, 42 U.S.C. § 3766(c)(1), and to protect certain rights guaranteed by the Constitution. By order entered 7 April 1977 plaintiff’s claims under the Constitution and under Title VII were dismissed, and the action proceeded to trial on the merits of plaintiff’s claim under the Omnibus Crime Control and Safe Streets Act of 1968 (hereinafter referred to as the Act). The United States alleges that the Commonwealth of Virginia has discriminated against black applicants for civilian positions and against both black and female applicants for sworn positions in the Virginia State Police. The State Police is denominated a program or activity funded in part with Law Enforcement Assistance Administration (LEAA) funds made available under Chapter 46 of Title 42 of the United States Code, 42 U.S.C. §§ 3701, et seq. The Crime Control Act of 1973 amended Section 518 of the Omnibus Crime Control and Safe Streets Act of 1968, 42 U.S.C. § 3766, to include for the first time a prohibition against, inter alia, racial and sexual discrimination by recipients of funds under the Act. Section 3 of the Crime Control Act of 1973, Pub.L.No. 93-83, provided that the anti-discrimination amendments would take effect on 1 July 1973. Thus, only those acts of discrimination which took place after 1 July 1973 will support a finding of discrimination in violation of § 3766(c). See Hazelwood School District v. United States, 433 U.S. 299, 309, 97 S.Ct. 2736, 53 L.Ed.2d 768 (1977). Evidence of pre-Act discrimination is nonetheless relevant to this case. Hazelwood, supra at 309-10, n. 15, 97 S.Ct. 2736. A showing of pre-Act discrimination would give meaning to post-Act patterns and practices of the defendants. The Court finds that prior to 1 July 1973 the Virginia State Police discriminated against black applicants for civilian positions and against both black and female applicants for sworn positions. The practice of discrimination against female applicants for sworn positions was readily admitted by Captain Meredith S. Urick, who was Personnel and Training Officer for the Virginia State Police from the late 1940’s until his retirement on 1 April 1972. Tr. 5-48. The discrimination against blacks in hiring for sworn and unsworn positions is not admitted by defendants, but is shown by the evidence nonetheless. The statistical evidence of pre-Act discrimination is substantial. The evidence of the hiring of blacks by the Virginia State Police prior to 1969 consists of what the Supreme Court and the Fifth Circuit have termed “the inexorable zero.” Teamsters v. United States, 431 U.S. 324, at 342, n. 23, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977), quoting from United States v. T. I. M. E.-D. C., 517 F.2d 299, at 315 (5th Cir. 1975). In the entire history of the Virginia State Police, only four blacks had ever been hired as troopers before 1973. No blacks had been hired prior to 1969. Plaintiff’s Exhibit 21A. Plaintiff’s Exhibit 8 is a list of 172 black applicants for trooper positions who applied between 1950 and 1972. Although this number is small in proportion to those who applied for trooper positions in this period, it is a substantial number and none were hired. The statistical evidence of pre-Act discrimination alone would support an inference of discrimination. See Roman v. ESB, Inc., 550 F.2d 1343, 1350 (4th Cir. 1976). When coupled with the direct evidence of discrimination, a prima facie case of preAct discrimination is made. The defendant’s evidence is insufficient as a matter of law to rebut the plaintiff’s prima facie case. Teamsters v. United States, 431 U.S. 324, 342-3, n. 24, 97 S.Ct. 1843, 52 L.Ed.2d 396 (1977), citing Alexander v. Louisiana, 405 U.S. 625, 632, 92 S.Ct. 1221, 31 L.Ed.2d 536 (1972). The Court finds that prior to the effective date of the Crime Control Act of 1973, that is, up until 1969, the Virginia State Police maintained an all-white trooper force by purposefully excluding blacks. II The Attorney General of the United States is empowered to bring suit under § 3766(c) whenever he has reason to believe a recipient government agency “has engaged or is engaging in a pattern or practice in violation of the provisions of this section.” 42 U.S.C. § 3766(c)(3). Before proceeding to a detailed analysis of the evidence in this case, the Court must determine what legal standard is appropriate to such “pattern or practice” suits under the Crime Control Act. We deal with a welter of overlapping and interrelated laws, regulations and court decisions which make it difficult to ascertain what burden of proof is imposed on the federal government in such cases. The plaintiff contends that the legal standards applicable to suits under Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e et seq., are intended by Congress to apply to this § 3766 case. This view is bolstered by the clear statement to that effect found in the conference report on Publ.L.No. 94-503, H.R.Rep. 94-1723, 94th Cong., 2d Sess. (1976); U.S.Code Cong. & Admin.News 1976, p. 5374. Plaintiff says that it follows that the legal standard to be applied to this case is the standard enunciated in Griggs v. Duke Power Co., 401 U.S. 424, 91 S.Ct. 849, 28 L.Ed.2d 158 (1971). Under Griggs, proof of discriminatory motive is not necessary in a Title VII case where the plaintiff complains of facially neutral employment practices that exhibit a disparate racial impact. Plaintiff points as well to the regulations promulgated under the Crime Control Act on 18 August 1972, almost a year before § 3766(c) became effective. These regulations impose on recipients a duty to “conduct a continuing program of self-evaluation to ascertain whether any of their . policies . . . directly or indirectly have the effect of denying equal employment opportunities to minority individuals and women.” 28 C.F.R. § 42.306(a). This provision seems to contemplate remedies based on an' adverse impact theory of discrimination. Congress did nothing to require a change in this regulation in 1973 when the anti-discrimination amendments were adopted nor in 1976 when the legislation was re-enacted. Therefore, the plaintiff argues, Congress must have meant for the Griggs adverse impact standard to apply to cases under § 3766(c). In Lau v. Nichols, 414 U.S. 563, 94 S.Ct. 786, 39 L.Ed.2d 1 (1973) the Supreme Court upheld the claims of the plaintiff class brought by non-English-speaking Chinese-American students against the San Francisco public schools under a regulation promulgated by HEW pursuant to Title VI of the Civil Rights Act of 1964 which regulation prohibited practices which had the effect of discriminating even though no discrimination was intended. 414 U.S. at 568, 94 S.Ct. 786. Justice Stewart, joined by Chief Justice Burger and Justice Blackmun, concurred in the result but expressed doubt as to whether Title VI, standing alone, would sustain plaintiff’s position were there no purposeful discrimination. However, the regulations which required affirmative action to overcome any language barriers were, Justice Stewart said, reasonably related to the purposes of the enabling legislation, and therefore binding on the defendant. 414 U.S. at 569-71, 94 S.Ct. 786. In the present case, the regulation was not offered as a basis for liability but is here considered as a guide for the interpretation of the statute which followed it. Defendant, on the other hand, argues the rule of Washington v. Davis, 426 U.S. 229, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976). Washington held that proof of discriminatory motive is necessary to establish a case of race discrimination under the Fifth Amendment (and by implication under the Fourteenth Amendment). Thus, defendant argues, mere disparate impact will not prove a case of discrimination under the Crime Control Act. Defendant arrives at this limitation by arguing that the anti-discrimination section of the Crime Control Act (42 U.S.C. § 3766(c)) was enacted by Congress under the authority of § 5 of the Fourteenth Amendment. Since Washington recognized limits on the Amendment itself, the legislation adopted pursuant to the Amendment cannot exceed those same limits. Thus, if proof of motive is required to support a finding of liability under the Fourteenth Amendment; mere proof of disparate impact must be insufficient to support a finding of liability under § 3766(c), a statute enacted under the authority of that Amendment. This Court considered a similar proposition in the context of Title VII in Friend v. Leidinger, 446 F.Supp. 361, 384-87 (E.D.Va.1977). In Friend this Court decided that proof of discriminatory intent is indeed necessary to establish a case of race discrimination in employment under Title VII where the defendant is either a State or a subdivision of a State. Since Friend was decided, however, two cases have been decided which cast doubt on the analysis of the Court in Friend. In United States v. City of Chicago, 573 F.2d 416, (7th Cir. 1978), the Seventh Circuit held that the legal standard of proving a claim under the Fourteenth Amendment is not incorporated into Title VII by virtue of Washington v. Davis. The court analyzed the analogous cases of Lassiter v. Northampton Election Bd., 360 U.S. 45, 79 S.Ct. 985, 3 L.Ed.2d 1072 (1959) and Katzenbach v. Morgan, 384 U.S. 641, 86 S.Ct. 1717, 16 L.Ed.2d 828 (1966), dealing with English literacy tests for voting. In Lassiter the Supreme Court held that a North Carolina English literacy test did not necessarily violate the Fourteenth Amendment. In Katzenbach v. Morgan the Supreme Court considered the constitutionality of § 4(e) of the Voting Rights Act, which prohibited states from using English literacy tests as a requirement of voting. The Court held that the prior decision in Lassiter was inapposite to the question of the validity of the Voting Rights Act, and the question presented was instead whether the legislation was appropriate to enforce the Equal Protection Clause of the Fourteenth Amendment. 384 U.S. at 649-50, 86 S.Ct. 1717. The Court held in the affirmative. The Seventh Circuit, by analogy, concluded that Title VII, including the impact standard of Griggs, was an appropriate means of enforcing the Fourteenth Amendment in the employment field even as against a State subdivision. 573 F.2d at 422-24. The second case which considered the issue after Friend is United States v. South Carolina, 445 F.Supp. 1094 (D.S.C.1977) (three-judge court), aff'd mem. 434 U.S. 1026, 98 S.Ct. 756, 54 L.Ed.2d 775 (1978). In that case, a unanimous three:judge panel consisting of Haynsworth and Russell, Circuit Judges, and Simons, District Judge, discussed the impact of Washington v. Davis on employment cases under Title VII and concluded that Washington v. Davis did not disturb the application of Griggs and that employment practices which display an adverse impact must be shown to be valid even where the employer is a governmental entity. 445 F.Supp. at 1112. These cases lead the Court to believe that Congress has the power under § 5 of the Fourteenth Amendment to apply the impact standard of employment discrimination to the States in Title VII cases. In any case, the defendant’s argument is based on the assumption that Congress in passing § 3766(c) was acting pursuant to its power under the Fourteenth Amendment. Plaintiff contends, and the Court agrees, that § 3766(c) is more likely based upon the Spending Clause of Article 1, Section 8, of the Constitution. In Lau v. Nichols, supra, the Supreme Court made it clear that Title VI, the model for § 3766(c), was a proper exercise of the Spending Power: “The Federal Government has power to fix the terms on which its money allotments to the States shall be disbursed. Oklahoma v. CSC, 330 U.S. 127, 67 S.Ct. 544, 91 L.Ed. 794 (1947). Whatever may be the limits of that power, . they have not been reached here.” 414 U.S. at 569, 94 S.Ct. at 789. The fact that the anti-discrimination provision is tied directly to a spending program also indicates that it is based upon the Spending Power. Either way, the defendant’s argument fails. In this same connection, however, it must be borne in mind that this is a “pattern or practice” suit brought by the Attorney General against an employer under § 3766(c)(3). Pattern or practice suits in the employment field trace their origins to the DirksenMansfield amendments to the Civil Rights Act of 1964. 110 Cong.Rec. 12818-20 (1964). The Attorney General was authorized to bring a civil action against any employer who he has reasonable cause to believe “is engaged in a pattern or practice of resistance" to, the rights secured by Title VII, “and that the pattern or practice is of such a nature and is intended to deny the full exercise” of Title VII rights. 42 U.S.C. § 2000e-6(a). (Emphasis added). Although this intent language is not found in § 3766(c)(3), which authorized this suit, the connection between Title VII and the anti-discrimination provisions of the Crime Control Act is so close that the intent language of § 2000e-6(a) may be imputed to § 3766(c)(3). The language of § 2000e-6(a) thus plainly contemplates purposeful discrimination as the basis for a suit by the Attorney General. As the Supreme Court has noted, “disparate treatment was the most obvious evil Congress had in mind when it enacted Title VII.” Teamsters, 431 U.S. at 335, n. 15, 97 S.Ct. at 1854. In view of the very language of the statute this Court is of the view that Congress did not intend to subject a governmental employer to liability in a suit brought by the Attorney General unless the employer were engaged in a widespread pattern and deliberate practice of thwarting the goals of Title VII. The courts have, however, applied the Griggs impact standard to pattern and practice eases under Title VII without such analysis. In United States v. Jacksonville Terminal Co., 451 F.2d 418 (5th Cir. 1971), cert. den. 406 U.S. 906, 92 S.Ct. 1607, 31 L.Ed.2d 815 (1972), the Fifth Circuit held that the United States need not prove a specific present intent to discriminate in a pattern or practice suit. The court said: [T]he Act proscribes facially neutral practices perpetuating the effects of past discrimination. If the practices have racially determined effects, proof of subjective intent to discriminate is unnecessary; the Government must show only that the defendant intended to perform the discriminatory act. [451 F.2d at 443.] In United States v. City of Chicago, 573 F.2d 416 (7th Cir. 1978) the Seventh Circuit considered a pattern or practice complaint against the fire department of Chicago, Illinois, brought by the Attorney General under Title VII. Without discussing the question of whether Congress intended that pattern or practice suits be brought against employers who had no discriminatory purpose, the court held that the Griggs standard should apply to the case, and that proof of discriminatory motive was unnecessary. The Eighth Circuit also indicated by implication that discriminatory motive need not be proved in pattern or practice cases in Firefighters Institute for Racial Equality v. City of St. Louis, 549 F.2d 506 (8th Cir. 1977). The court did not distinguish between the claims of the United States and those of the private plaintiffs when it held that discriminatory purpose need not be shown in Title VII cases brought under an adverse impact theory. 549 F.2d at 510. Thus, although the Supreme Court has never considered in a State employee case the question of whether discriminatory intent is a necessary element of a prima facie case, the weight of the authority is that proof of impact alone is sufficient. This Court, then, will apply the standard of Title VII to this case with the Griggs impact theory rather than the Washington purposeful discrimination theory. Ill The Court will now proceed to examine the claims of discrimination raised by the complaint. These are three: sex discrimination in hiring for sworn positions with the Virginia State Police, race discrimination in hiring for civilian positions with the State Police, and race discrimination in hiring for sworn positions with the State Police. The plaintiff has sought to prove both disparate treatment and the disparate impact of neutral standards to show that the Virginia State Police have engaged in sex discrimination in hiring trooper applicants. The disparate treatment case rests upon the miniscule number of females hired by the Virginia State Police, the testimony of two female applicants, the failure of the Virginia State Police to recruit females for trooper positions, and the persistence in use of a height and weight requirement for trooper applicants. The plaintiff’s case under a disparate impact theory is based upon the grossly disparate impact of the height and weight requirement on women. The first female trooper was hired by the Virginia State Police in 1976. Until that time there had never been a woman State Trooper in Virginia. Clearly this is an example of “the inexorable zero.” The number of women presently employed and seeking employment is so low that the plaintiff does not argue that the proportion of women in sworn positions in the Virginia State Police should approximate the proportion of women in the civilian labor force in Virginia. Plaintiff does argue, however, that the number of women hired by the Virginia State Police would be substantially greater than it presently is in the absence of discrimination. Thus, the plaintiff’s use of these statistics is in accord with the Supreme Court’s view of the usefulness of such statistics found in Teamsters as proof of disparate treatment. 431 U.S. at 339, n. 20, 97 S.Ct. 1843. Plaintiff’s testimonial evidence of disparate treatment is, on the other hand, very weak. In part it consists of the testimony of two female applicants for trooper positions, Barbara Penn and Wanda Ricciardelli. Barbara Penn, then a secretary in the office of the Assistant United States Attorney, testified that she had an interview with a Lieutenant Abbott of the State Police in the latter part of September 1977. She testified that Lt. Abbott told her that it would take some time to process her application and that he didn’t hold out much hope that they could process her application before her 30th birthday in March of 1978. He invited her to file an application nevertheless. Despite this indication that time was critical Miss Penn did not actually submit her application until several weeks later. At the time of her testimony in February 1978, Miss Penn’s application was being processed by the Virginia State Police. Tr. 1-164-69. Wanda Ricciardelli testified that she applied to the Virginia State Police in January of 1977. In February, she received a letter notifying her that a freeze was in effect and that the State Police were not hiring at that time. In mid-May, Miss Ricciardelli was interviewed by a member of the Virginia State Police. She was interviewed a second time on 8 June 1977. On 22 July 1977 Miss Ricciardelli was notified that she would not be offered employment by the Virginia State Police. She tried repeatedly to ascertain the reason for her rejection, but the State Police told her only that they couldn’t release the reasons for her rejection because they were confidential. She was, however, told that she had passed the age limit of 29 years in June of 1977 and that after that time, she was not eligible for employment. Miss Ricciardelli expressed doubt that this age problem was the only factor in her rejection. Miss Ricciardelli’s testimony indicated that she had been hospitalized for an emotional disturbance in the spring of 1968, and that she had a problem with excessive absences at her secretarial job in Philadelphia. Tr. 3-37-47. The Court does not believe that the evidence shows that either Miss Penn or Miss Ricciardelli was the victim of sex discrimination. When the Court asked counsel for plaintiff what was the relevance of Miss Penn’s testimony, plaintiff’s counsel said: We wanted to show Your Honor that there was an applicant, a female applicant, that we considered to be qualified, who was under a certain degree of time restraint and needed to be considered and was not considered and there’s also, as I think the testimony will show from other witnesses, there is a class approaching that is to be employed sometime in February— THE COURT: Is it the position of the government that because she was a woman and she should have gotten particular treatment? MR. WHITE: If you have a qualified female applicant and you’re in an affirmative action posture to hire applicants, we consider that it might be advisable for the State Police to consider her application, knowing that her 30th birthday was approaching. [Tr. 1-168-69.] This exchange makes it clear that the plaintiff did not contend that Miss Penn was an explicit victim of sex discrimination. The purpose of the testimony of Miss Penn was to show that despite its paucity of female troopers, the State Police were unwilling to bend or waive the application process to recruit additional women. Adherence to such rules is not discrimination. Indeed, had defendant waived the rules because Miss Penn was a woman discrimination might well have been proven. With respect to Miss Ricciardelli, the plaintiff presented no evidence beyond her own suspicion that she was rejected for a reason other than her age, i. e., because of her sex. Defendant argues that even if Miss Ricciardelli’s age were not the reason for her rejection, sufficient other reason has been shown by Miss Rieeiardelli’s own testimony. The Court concurs. There is no evidence that the reasons advanced by the Virginia State Police for their rejection of Miss Ricciardelli are pretexts for sex discrimination. Accordingly, the Court finds that neither Miss Penn nor Miss Ricciardelli was the victim of sex discrimination. In juxtaposition to this testimony the defendant offered the testimony of Trooper Cheryl Nottingham. Trooper Nottingham was the first woman to be hired as a trooper by the Virginia State Police. She testified that she was not discriminated against in any way on account of her sex either in the application process or in the training process or as a working trooper. Indeed, Trooper Nottingham testified that she was encouraged to apply and become a State Trooper. Miss Nottingham applied for the job of trooper in August of 1976. Her 30th birthday came in October of that year. Plaintiff points to the alleged failure of the Virginia State Police actively to recruit women as evidence of intentional discrimination. The evidence shows that only 38 women applied for trooper positions with the Virginia State Police from 1973 through 1976. Defendant’s Exhibits 110 and 111. This small number of female applicants, plaintiff contends, is the result of the admitted discrimination carried on by the State Police prior to 1973, coupled with the failure of the Virginia State Police to recruit women as a remedial measure. The evidence shows that virtually all of the recruiting efforts expended by the Virginia State Police have been aimed at recruiting black applicants. While it is true that Trooper Green testified that he has sought to recruit women applicants as well as Black applicants, nevertheless, it is clear that the main thrust of his efforts has been at recruiting blacks. Colonel Slane testified that he hoped to add a female trooper to the recruiting team, but wouldn’t be able to do so until a female trooper could graduate and have some experience in the field. Tr. 10-147-48. The Virginia State Police has clearly made a policy decision that the best way to recruit for the kind of people they want is to send a trained and experienced trooper, who is a member of the group they seek to recruit, around to talk to prospective applicants. Thus, the lack of a trained and experienced female trooper in the Virginia State Police has resulted in a lack of recruiting for female applicants. The record shows that the idea of women becoming State troopers is a relatively new one. Colonel Burgess, a police officer of 42 years standing, had never heard of females as troopers prior to 1970. Tr. 7 — 14. Nevertheless, the Court views the failure of the Virginia State Police after the effective date of the Act actively to recruit female applicants as evidence of sex discrimination against women on the part of the Virginia State Police. The most telling evidence of sex discrimination, however, is the existence until 1976 of a height and weight requirement for trooper applicants for the Virginia State Police. Until July, 1976, an applicant for trooper with the Virginia State Police was required to be at least 5'9" tall and weigh 156 pounds. Plaintiff’s evidence showed that these requirements disqualify more than 98% of all women while disqualifying only about half of all men. Plaintiff’s Exhibit No. 16. Defendant’s answer to the plaintiff’s contention that this shows intentional discrimination against women is that the height and weight requirement was instituted and was in effect for many years prior to 1973. During this period the Virginia State Police admittedly would not consider any woman for employment as a State trooper. Her height or weight was no impediment — only her sex. Thus it is specious to argue that the height and weight requirement was intended to exclude women. Defendant also defends its height and weight requirement on grounds that the Federal Bureau of Investigation used a height and weight requirement for employing special agents until as late as June of 1975. Defendant does not argue that the FBI is incapable of discriminating against women. Instead, its argument is that the Virginia State Police was mindful that the FBI is an arm of the Attorney General of the United States and it reasonably assumed that the U. S. Department of Justice would not maintain the policy if it were prohibited by law. Defendant maintains that its reliance on the standards set by the FBI is evidence of good faith and the nondiscriminatory purpose of the Virginia State Police. Recognizing that in years past no woman had indicated an interest in becoming a State trooper (a woman police officer was “unheard of”) and that even in recent years it is rare for a woman to be so motivated, the Court cannot lay at the feet of the Virginia State Police all the blame for the lack of female troopers on its force. However, after the Congress acted in 1973 to bar sex discrimination to recipients of LEAA funds the State Police had a legal duty to remove barriers imposed by it which had the effect of discriminating against female applicants. The defendant had a duty to review its employment practices when it became subject to the Act, and either to modify those practices which resulted in extreme adverse impact to a protected group, such as women, or to seek to justify by empirical evidence the necessity for such standards. The defendant did not remove the height and weight barrier until July 1976 and in this suit it made no effort to justify the retention of the requirement on empirical grounds. The only evidence of the job-relatedness of the height and weight requirement found in the record is the opinions of five members of the Virginia State Police that height and weight are useful in police work. This evidence was not offered to prove job-relatedness but to prove a good faith belief of job-relatedness. There is no evidence that justified a 5'9" requirement or a 159 pound weight requirement. Such evidence would be necessary to establish the job-relatedness of the standard used by the Virginia State Police. In fact, the only empirical evidence in the record as to the validity of the height and weight requirement is Dr. Edgerton’s finding that neither height nor weight of the officers correlated substantially with any criterion of job performance in his 1973 validity study. Defendant’s Exhibit 12, p. 96. It was not, of course, the implementation of the height and weight requirement in the pre-Act period when no women were considered, nor even its continued use after the Act when it arguably could be considered a bona fide job-related facially neutral requirement that was illegal. But the persistence of the practice well after the enactment without evidence of its enhancement of job performance can only mean that the State Police intended to keep the number of women on its force low. Thus, the Court concludes that from and after the effective date of the Act the defendant purposefully discriminated against women in employment as State troopers. The foregoing discussion also affirms plaintiff’s contention that the height and weight requirement is an illegal employment practice under the disparate impact theory of Griggs. IV The plaintiff alleges that the Virginia State Police discriminates on the basis of race in hiring for civilian positions. Plaintiff’s case is based almost entirely upon the statistical evidence on the numbers of blacks hired by the defendant compared with the numbers of blacks that would have been hired by the defendant if the defendant had hired blacks in proportion to their representation in the civilian labor force in Virginia. Plaintiff seeks to prove its case by a statistical analysis involving the use of the standard deviation of a population sample. The argument is that if there were no discrimination practiced by the Virginia State Police the proportion of blacks employed in civilian positions by the Virginia State Police would be roughly equal to the proportion of blacks in the civilian labor force. A slight deviation would not be strong evidence of discrimination. A major deviation would be significant. Therefore, the plaintiff has used the standard deviation analysis to show that the numbers of blacks hired by the Virginia State Police in civilian positions is significantly smaller than the numbers of blacks that would have been hired if blacks had been hired in the same proportion as their representation in the civilian labor force. This analysis has resulted in Table I: TABLE I Statistical Analysis of Black Employees in Civilian Positions with the.Virginia State Police, 1973-76. (Source: Plaintiff's Exhibit 23D) Difference Blacks Between Total Actually Expected Actual & Standard Employed Employed Employed Expected Deviation Number of Standard Deviations Difference 12/31/73 396 21 68 47 7.51 6.3 12/31/74 412 18 70 52 7.66 6.8 .12/31/75 438 22 75 43 7.90 5.4 12/31/76 451 28 77 49 8.01 6.1 Table I is derived by first multiplying the total number of persons employed in civilian positions for each of the years 1973 through 1976 by the number .172 which is the proportion of blacks in the civilian labor force in Virginia as of the 1970 census. This calculation results in the “expected employed” statistic — the number of blacks who would have been in the employ of the Virginia State Police had the Virginia State Police employed blacks in proportion to their membership in the civilian labor force. The plaintiff then subtracts the number of blacks actually employed by the Virginia State Police from this figure for expected employees and divides the figure thus obtained by the standard deviation for the population. The number thus obtained is the number of standard deviations between the number of blacks who would have been employed had the hiring of the Virginia State Police been random, and the number of blacks actually hired. If that number were no more than two or three, then the inference that hiring in the Virginia State Police was unbiased with respect to race would be supported. When, as here, the number is five or more, the inference that hiring in the Virginia State Police for civilian positions is unbiased with respect to race is not supported. This analysis necessarily makes several assumptions. First, it is assumed that the civilian labor force for the entire Commonwealth of Virginia is the labor pool from which the Virginia State Police draws its civilian employees. The evidence shows, however, that the bulk of its civilian employees are located at the State Police Headquarters in Richmond and thus are drawn from the Richmond area. Accordingly, the civilian labor force statistics for the State as a whole are not applicable. Plaintiff’s analysis does not take into account the present effects of pre-Act hiring. The analysis further fails to take into account other population characteristics which are not shown to be randomly distributed with regard to race. One such nonrandom population characteristic may be the distribution of various skills required by the civilian positions offered by the Virginia State Police. Another such characteristic, shown by the evidence to be concentrated among blacks, is an antipathy toward law enforcement as a career. Tr. 9-104. The entire history of black deprivation which has made affirmative action necessary refutes an assumption of even or random distribution of job skills and motivations. As the Supreme Court has said, “[wjhen special qualifications are required to fill particular jobs, comparison to the general population (rather than to the smaller group of individuals who possess the necessary qualifications) may have little probative value.” Hazelwood, 433 U.S. at 308, n. 13, 97 S.Ct. at 2742. A review of Plaintiff’s Exhibit 22, however, shows that the civilian employees of the Virginia State Police include persons in such positions as accountants, building construction inspectors, clerk stenographers, clerk typists, computer programmers, computer systems analysts, highway equipment mechanics, radio engineers, radio technicians, photographers, and other skilled positions. In the absence, of any evidence from plaintiff as to the percentage of blacks holding the indicated skills, or, alternatively, evidence that the skills are commonly held and easily learned, cf. Hazelwood, 433 U.S. at 308, n. 13, 97 S.Ct. 2736, the Court simply has no basis for finding that blacks are so underrepresented in the skilled civilian positions that racial discrimination may be inferred. The Court has attempted to separate out from the civilian job titles found in Plaintiff’s Exhibit No. 22B positions that presumably many persons can fill or which can be filled with minimal training. The distribution of black and white employees in these positions is set out in Table II-A. The remainder of the positions are set out in Table II-B. TABLE XI-A Civilian Employees in Unskilled Positions as of 31 December 1976 . (Source: Plaintiff's Exhibit 22B) White Black Clerk D 6 Clerk C 19 Clerk B. 47 Clerk-Mess. B 2 Comp. Oper. A 2 Comp. Oper. B 2 Cook A Corr. Off. Cust. Worker Fing. Tech. A. Fing. Tech. B Food Serv. Aid A Off. -Dup. Mac. Op. A Painter PhotoCpy. Eqp. Oper. B Rec. Info. Dispatcher Staff Photo. Stat. A Storekeeper Asst. TOTAL (Percent) 102(84.3) 19(15.7) TABLE II-B Civilian Employees in Skilled Positions as of 31 December 1976 (Source: Plaintiff's Exhibit 22B) White Black Acct. B 1 Acct. A 1 ADP Mgr. A 1 ADP PC Supv. 1 ADP P. Tech A. 1 Const. Insp. 1 Bldg. & Gds. Supv. A 1 Bldg. Crim. Rec. Asst. Dir. 1 Cent. Crim. Rec. Exc. Dir. 1 Cent. Crim. Rec. Supv. 1 Cent. Clerk-Steno D 15 Clerk-Steno C. 53 Clerk-Steno B 10 Clerk-Typist C 22 Clerk-Typist B 48 Comp. Oper. Supv. 1 Prog. B 2 Comp. Prog. A 3 Comp. Syst. An. A 2 Comp. Syst. Dev. Supv. 1 Comp. Sec. 1 Conf. Data Entry Op. A 20 Data Entry Supv. 1 Diet. Unit Supv. A Equip. Supv. A 1 Hwy. Equip. Rep. For. A 1 Hwy. Equip. Mech. A 5 Hwy. Equip. Mech. B 1 Hwy. Supv. A Hsk. Dir. A 1 Info. Police Dispatcher 84 Pur. & Stores Supv. B 1 Radio Asst. Eng. 1 These tables, the Court recognizes, have very little probative value. There is, for instance, no evidence in the record to support the Court’s determination that the jobs found in Table I are those which require skills that many persons possess or may easily acquire. The Court’s choice was based entirely on the name of the position as set forth in Exhibit 22B. Nevertheless, it is apparent that the proportion of blacks in those jobs which the Court speculates are those in which comparison with general population statistics is appropriate, is very close to the proportion of blacks in the State civilian labor force according to census data. These figures do not suffice to prove that the Virginia State Police have not discriminated against blacks in hiring for civilian positions. However, it is not the burden of the Virginia State Police to prove that they have not discriminated. It remains the burden of the United States to prove that they have discriminated. Roman v. ESB, Inc., 550 F.2d 1343, 1350 (4th Cir. 1976). The plaintiff introduced the testimony of only one witness to show that there had been racial discrimination in hiring for civilian positions in the Virginia State Police. Miss Evelyn C. Faison testified that she had taken the civil service test for a clerical position with the State of Virginia and passed it in 1972. She testified that in 1974 she applied for a clerical position with the Virginia State Police. Her application, Defendant’s Exhibit 148, shows that she did not list the fact that she had taken and passed a State civil service examination for a clerical position. The work experience listed in her application shows work as a presser in a cleaning establishment, as a utility aide in a paper company, as an elevator operator at the Medical College of Virginia and in the cleaning department of a cleaning establishment. On the basis of her application Miss Faison was offered a food service job at the Virginia State Police Headquarters. The plaintiff argues that this incident shows the method by which blacks were channeled into lower paying clerical and food service positions in the Virginia State Police. In fact, it shows nothing of the kind. The job Miss Faison was offered was commensurate with her abilities as they appeared from her application. Thus, the testimony of Miss Faison is of no value in showing discrimination by the defendant against blacks in hiring for civilian positions. Plaintiff also argues that the concentration of black civilian employees in food service and custodial positions shows racial discrimination on the part of the Virginia State Police. The evidence shows that of 15 civilian employees in the jobs of cook, food service aide, and custodial worker, six are white and nine are black. However, there is no evidence in the record to show whether this disproportional assignment of black workers to custodial and food service positions is the result of post-Act hiring, or if this pattern is merely a holdover from the time before the effective date of the Act. In the latter case, no liability could be found on the basis of these statistics. Hazelwood, 433 U.S. at 309, 97 S.Ct. 2736. The evidence that the Court does have, however, shows that the proportion of Macks in these positions has declined from 77% in 1974 to 60% in 1976. (See Table III). Finally, there is no evidence in the record that food service or custodial jobs are lower paid or less desirable jobs in the Virginia State Police than such jobs as clerks, clerk typists or office duplicating machine operators. TABLE III Civilian Employees in Food Service and Custodial Positions by Race as of 31 December 1974 and as of 31 December 1976 (Source: Plaintiff’s Exhibit 22B) 1974 1976 White Black White Black Cook A 2 2 Cust. Worker 3 3 5 2 Food Svc. Aid A 5 1 5 TOTAL 3 10 5 ' 9 Percent 23.1% 76.9% 40.0% 60.0% The Court treats separately the civilian position of police dispatcher. Applicants for the job of police dispatcher with the Virginia State Police were required to take the same written examination taken by applicants for trooper positions. This written examination is discussed in Part IV of this memorandum. The analysis of adverse impact found there applies equally to the test when used to test applicants for dispatcher position. Thus, the test, as will be seen in Part IV, is shown to have an adverse impact upon blacks, and may be used as a selection device only if it has also been shown to be a valid predictor of job performance. The defendant has not even attempted to show that the written test is in any way related to any criterion of job performance in the job of police dispatcher. Unlike the job of trooper, successful completion of a training program is not a requirement for the job of dispatcher. The evidence shows, it is true, that some dispatchers are sent to the State Police training school, if space permits, to familiarize them more fully with the tasks of the troopers they support. However, it is clear from the evidence that the training school is not a requirement of the job of dispatcher. Thus, the Court finds that it was a violation of § 3766(c) for the Virginia State Police to use the written test as a selection device for applicants for- the dispatcher position. Plaintiff’s Exhibit 15 shows that only five black applicants for dispatcher positions failed the State Police written test from 10 May 1973 to 31 December 1975. The record does not show how many of these were failed prior to the effective date of the Act, 1 July 1973. It is clear in any event that the unlawful employment practice found by the Court had a very small impact. Nevertheless, relief must be fashioned to make whole those persons who are the victims of 'that unlawful employment practice. V The Court will now consider the main thrust of plaintiff’s case, that the Virginia State Police discriminated against blacks in hiring for trooper positions. Plaintiff seeks to prove its case through five basic modes of proof. First, the plaintiff has offered statistical evidence on the hiring of blacks by the Virginia State Police compared with the statistics on the proportion of blacks in what is alleged to be the relevant labor market. Second, the plaintiff contends that the use of a written test from the effective date of the Act to December 1975 had an adverse impact upon black applicants and that the test has not been shown to be validated as job-related. Third, the plaintiff contends that the background investigation of trooper applicants was conducted in such a way as to discriminate against black applicants. Fourth, the plaintiff has presented the testimony of several black applicants and conditionally appointed troopers who claim to have been discriminated against in the employment process by the Virginia State Police. Fifth, the plaintiff contends that the minority recruiting program of the Virginia State Police is inadequate, and that the failure of the Virginia State Police to run an adequate minority recruiting program is evidence of discrimination. A. The race of Virginia State Police sworn personnel on the last day of December 1973 to the last day of December 1976 is shown in Table IV. TABLE XV Virginia State Troopers by Race 1973-1976 Date Total Employed White (7„) Black(%) 31 December 1973 1055 1047 (99.2%) 8 (0.8%) 31 December 1974 1096 1087 (99.2%) 9 (0.8%) 31 December 1975 1104 1093 (99.2%) 11(1.0%) 31 December 1976 1132 1101 (97.3%) 31(2.7%) These figures, of course, reflect the results of hiring practices of the Virginia State Police over a period of decades. This Court must necessarily concern itself with the hiring record of the Virginia State Police since the July 1973 effective date of the Act. The statistics on hiring in the Virginia State Police since 1973 are set out in Table V. TABLE V Troopers Hired by the Virginia State Police by Race 1973-1977 Year Total Hired White (%) Black (%) 1973 91 85 (93-, 4%) 6 (6.6%) 1974 96 93 (96.9%) 3 (3.1%) 1975 64 62 (96.9%) 2 (3.1%) 1976 86 63 (73.3%) 23(26.7%) 1977 (As of 63 52 (82.5%) 11(17.5%) 12 August 1977) TOTAL "if The Supreme Court’s opinion in Hazel-wood indicates that there are two comparisons that may be made using the data found in Table V. One is to compare the proportion of blacks hired by the Virginia State Police with the proportion of blacks in the relevant labor market. The other comparison that may be made is to compare the proportion of blacks hired with the proportion of blacks among those who sought employment with the Virginia State Police. a. Before the first comparison can be drawn, however, it is necessary to determine what constitutes the relevant labor market for the position of trooper in the Virginia State Police. Plaintiff contends that the best statistics available for a relevant labor market are the figures from the 1970 census for the civilian labor force aged 21 to 29 in Virginia. Defendant argues, however, that these undifferentiated work force statistics include many persons who could not meet the minimum requirements for the job of State trooper. The defendant offers statistics which purport to include those individuals who at the time of the 1970 census were 21 through 29 years of age, had completed twelve years of education or had a General Education Development (GED) certificate, were not disabled for six months or more, were not in jail or in a long term hospital, and were in the “labor force” as defined by the Bureau of Census. The plaintiff’s method of calculating the relevant labor market results in a relevant labor market that is 16.3% black. The defendant’s method gives a relevant labor market that is 10.67% black. The method chosen by the defendant appears to be by far the more relevant. However, the plaintiff raises two objections to defendant’s approach. First, the plaintiff argues that the adjustments made by the defendant are made not in an attempt to more closely define the proper labor market but in an attempt to reduce the proportion of blacks in the relevant labor market. Second, the plaintiff argues that the defendant’s figures are not up-to-date with respect to the proportion of blacks in the relevant age group and with respect to the proportion of blacks who have achieved the required educational level. The Court believes that the plaintiff’s first objection is not well taken. Despite the view of plaintiff’s industrial psychologist that aggregate data should be used even where more precise data is available because there is no agreement among experts as to what adjustments are appropriate, the Supreme Court in Hazelwood made it clear that the district court should define the relevant labor market with as much precision as possible. 433 U.S. at 312, 97 S.Ct. 2736. The entire purpose of a comparison between hiring statistics and labor force statistics is to show that the hiring of the defendant is or is not random with respect to race. Unless other non-random factors are first accounted for, this purpose will necessarily be defeated, and the entire analysis will be utterly lacking in probative value. It would be simply a cover for statistically justifying a pre-coneeived conclusion. The plaintiff’s second objection is far more valid. Plaintiff presented the testimony of Dr. Larry Suter, Chief of the Education Statistics Branch of the Population Branch of the Bureau of the Census, to show that certain population trends since 1970 have resulted in an increase in the proportion of blacks in the defendant’s defined relevant labor force. Dr. Suter testified that the proportion of blacks in the 21-29 year old age group has increased in Virginia since the' latest decennial census. He also testified that the proportion of blacks meeting the minimum educational standards of the Virginia State Police has increased in Virginia since 1970. Dr. Suter suggested increasing the proportion of blacks in the eligible labor force as defined by the defendants from 10.67% to either 12.47% or 15.47% in 1977. Tr. p. 11-13. Dr. Suter testified that his analysis of the increase in black educational attainment since 1970 would indicate that the eligible labor force for the Virginia State Police as defined in Defendant’s Exhibit 119 should be increased by perhaps 1.8% or even by as much as 4.8% because of the increase in the proportion of blacks in the 20-29 age group and the increase in the proportion of blacks who are high school graduates. The 1.8% figure comes from Dr. Suter’s analysis of the current population survey for the nation as a whole. Tr. p. 11 — 11. The 4.8% figure is derived from the Virginia segment of the national current population survey, but Dr. Suter was reluctant to put great confidence in the Virginia figure because the current population survey was based on an intended national sample, not a sample reflective of the population of Virginia. The Virginia sample is so small that the standard error of the sample is relatively high. In addition, Dr. Suter’s analysis is limited to persons not in the armed forces, a fertile recruiting ground for the State Police. Also, Dr. Suter’s competence in the field of determining relevant labor markets "was not conceded by defendant nor, indeed, was it claimed by Dr. Suter. Tr. 11-5. Finally, Dr. Suter’s testimony was not claimed to be and was not presented by the plaintiff in an effort to prove the proportion of blacks in the eligible labor force. It was instead presented on rebuttal to show that the figures in Defendant’s Exhibit 119 should not be accepted by the Court as an accurate portrayal of the eligible labor force because the figures are now eight years old and change has taken place. Dr. Suter’s testimony clearly supports this objection. The plaintiff argues that defendant’s calculations based on the 1970 census are incorrect because they are out-of-date. Unfortunately for the Court, the same argument applies with equal force to plaintiff’s statistics. The purpose of establishing the relevant labor market of trooper applicants to the Virginia State Police is to raise an inference, by comparison of that figure with the number of blacks actually hired, that the Virginia State Police discriminated against blacks in hiring. Thus, the burden of proving the relevant labor market rests squarely upon the plaintiff in this case. The foregoing discussion makes it clear that the plaintiff has not carried its burden. Neither party has offered the testimony of an expert in demography or of a witness qualified by his experience or training to testify as to the racial breakdown of the relevant labor market for a trooper in the Virginia State Police for any period covered by this law suit. Plaintiff protests that other courts in similar cases have used the aggregate, undifferentiated census figures it offered in this case. If this be true it still is not justification in this Court’s view for imposing liability upon a defendant on the basis of unreliable or missing evidence. This case has been pending since December 1976. The United States does not lack the resources to develop the proof necessary to show the relevant labor market. The reliance upon unanalyzed data copied from the 1970 census, even coupled with Dr. Suter’s projections, cannot replace reliable proof in an important case such as this. In Hazelwood, the district court failed to make a finding of fact on the relevant labor market for teachers in the St. Louis area. The court ruled no discrimination upon a .finding that the proportion of black teachers was roughly equal to the proportion of black pupils in the school district. The Supreme Court disapproved of this comparison, but also disapproved of the action of the court of appeals in accepting the plaintiff’s argument that the relevant labor market should be determined by including teachers from the entire St. Louis metropolitan area. The Supreme Court remanded the case to the district court for a precise finding on the actual relevant labor market of qualified teachers in the area from which applicants were drawn. In the present case, unlike Hazelwood, the plaintiff has failed to introduce evidence sufficient for a finding of the relevant labor market. Under these circumstances, no comparisons can validly be drawn and the standard deviation data offered by plaintiff in reliance upon Castaneda v. Partida, 430 U.S. 482, 496-97, n. 17, 97 S.Ct. 1272, 51 L.Ed.2d 498 (1977), is meaningless. b. Even if the relevant labor market were, as plaintiff contends, 16.3% black, the Castaneda analysis would not support a finding of racial discrimination. First, the Court considers it inappropriate to apply the Castaneda analysis to the statistics for the total sworn work force of the Virginia State Police. Because only such discrimination as took place after the effective date of the Act may be considered in making a finding of liability in this case, only the hiring behavior of the defendant since 1 July 1973 is relevant to the question of liability. Thus, the proper statistics to which the Castaneda analysis may be applied are the figures for troopers hired from 1973 to 1977. More precisely the proper statistics to use in this analysis would be the statistics showing the racial makeup of persons offered employment with the Virginia State Police in sworn positions after 1 July 1973. The Virginia State Police has no power to conscript persons into its work force. If 16.3% of those persons offered employment in sworn positions by the Virginia State Police were black, the plaintiff’s argument would fail even if no blacks accepted the offer from the State Police (assuming nothing more were shown). Thus, the fact that the plaintiff uses only hiring statistics necessarily requires the Court to assume, in order to find liability on the basis of those statistics, that the applicants from both races accepted employment when offered at the same rate. Nothing in the record will support this assumption, and this fact makes the Court uneasy about using plaintiff’s Castaneda analysis to impose liability upon the defendant. In any case, the plaintiff’s Castaneda analysis is set out in Table VI. TABLE VI Statistical Analysis of Black Employees in Trooper Positions with the Virginia State Police 1973-77 (Source: Plaintiff's Exhibit 21B) Difference Number of Blacks Blacks Between Standard Total Actually Expected Actual & Standard Deviations Employed Hired Hired_Expected_Deviation Difference 1973 91 6 14.8 -8.8 3.52 -2.5 1974 96 3 15.6 -12.6 3.62 -3.5 1975 64 2 10.4 -8.4 2.95 -2.8 1976 86 23 14.0 +9.0 3.42 +2.6 1977 63 11 10.3 +0.7 2.93 +0.2 Table VI shows that the difference between the number of black troopers actually hired and the number of black troopers that would have been hired if the defendant had hired blacks in proportion to their numbers in the civilian labor force was significant in 1973, 1974 and 1975. In each case, the difference between the number actually hired and the number expected to be hired was more than two standard deviations from the mean. However, the numbers involved are very small. For example, if in 1973 the Virginia State Police had hired two more blacks than they actually hired, 8 instead of 6, the difference between the actual number hired and the expected hired would have been only 6.8, less than 2 standard deviations difference. Furthermore, in 1976 the defendant hired many more blacks than the' plaintiff’s formula would have expected it to hire. In 1977, the defendant hired almost exactly the number that plaintiff’s formula expected it to hire. The Court does not believe that it is a coincidence that defendant’s statistics on hiring of blacks should change so dramatically at the same time that the defendant abolished its written test, that is, the end of 1975. Thus, the Court believes that the low numbers of blacks hired prior to 1976, far from being the telltale sign of intentional discrimination, is in fact, the unintentional (and apparently undesired) result of defendant’s use of a written test as a selection device. It should be noted that if the defendant’s 10.67% figure be used then the hiring practices during the overall post-Act period are proportionate to random hiring in the designated population and no inference of discrimination arises. As will be shown hereafter, the Virginia State Police force is not composed of persons randomly hired. The hiring of State troopers is a highly selective process. c. The other comparison which may be made is a comparison between the hiring of blacks and the applicant flow of blacks. If the proportion of blacks offered employment by the defendant is significantly smaller than the proportion of blacks among applicants, then the inference that the defendant is discriminating against blacks with respect to hiring for trooper positions is supported. If, on the other hand, the proportions are roughly equal, the position of the plaintiff that the defendant discriminates against blacks in hiring is not supported. In making this analysis, the Court is faced immediately with two serious problems in the evidence presented by the plaintiff. First, the applicant statistics introduced by the plaintiff in Plaintiff’s Exhibit 34 are, for the years 1974, 1975, and 1976, not broken down between applications received for civilian positions and applications received for trooper positions. Tr. 2-21— AM. The second problem is that in any given year the persons listed in Plaintiff’s Exhibit 34B are not necessarily the same persons who applied during that particular year, but may include persons who applied during an earlier year. Thus, year-by-year comparisons are impossible. Taken as a whole, however, the applicant-flow data available shows that from 1973 through 1976 10.8% of all applicants for both civilian and sworn positions were black. Plaintiff’s Exhibit 34A. Table VI shows that over the same period 34 of 337 or 10.1% of all troopers hired were black. Thus, the analysis of such applicant-flow data as is available to the Court fails to reveal any evidence of a pattern or practice of discrimination. The Court notes that in 1973, 1974 and 1975 the proportion of blacks hired as troopers was much lower than the proportion of blacks hired in 1976 and 1977. Although, as has been shown, year-by-year comparisons of applicant flow data and hiring data are difficult or impossible to draw and the proper comparison in any case is between the number of applicants and the number of persons offered employment with the Virginia State Police, the Court nevertheless recognizes that the relatively low number of blacks hired may be a telltale sign of discrimination in the early years immediately following the effective date of the Act. The plaintiff has attempted to attribute the increased hiring of blacks in 1976 and 1977 to a combination of the pressure put on the defendant by the Justice Department in the negotiations that preceded the filing of this lawsuit and the desire of Dr. Phillip Ash for a sample of at least 20 black troopers for his differential validity study. Defendant did not display evidence of bending to threats or pressure from the Justice Department. It was because defendant refused to bend that the suit was instituted. The Court believes that the abolition in 1976 of the written test used by the Virginia State Police was the factor which led to the sharp increase in hiring of blacks. This test had a substantial adverse impact on blacks and its abolition coincided with a sharp increase in the hiring of blacks. Further, by the middle of the 1970’s the minority recruiting program commenced in 1973 had begun to have favorable effects. B. From the effective date of the Act until December 1975 the Virginia State Police administered a written test to all applicants for trooper positions consisting of a battery of three tests: the Army Alpha Test, the Otis Test of Intelligence, and the O’Rouke Test of Mechanical Aptitude. Plaintiff’s Exhibit 11. The defendant does not dispute pla