Opinion
Nos. CA-3-76-0834-G, CA-3-76-1593-G.
December 18, 1979.
Kenneth H. Molberg, Dallas, Tex., for plaintiffs.
Lee E. Holt, City Atty., Ronald E. Deutsch, Lois Bacon, Asst. City Attys., Dallas, Tex., for defendants.
ORDER
This order constitutes the court's findings of fact and conclusions of law entered at the end of the liability phase of a trial of class-wide issues asserting claims under Title VII. The suit was filed by Brenda Davis, a black woman, and Cynthia Durbin, a white woman, both former candidates for employment as police officers for the City of Dallas. Together they seek to represent female and black persons who applied for similar positions from August 1, 1973. The issues of class certification have been decided earlier by written order. Because no evidence has been developed that would alter that judgment, class issues will not be further discussed except for the question of whether Durbin may proceed in the face of a failure of the EEOC to defer action on her charge pending referral to the proper state official.
I. The Durbin Charge Referral
It is undisputed that the EEOC erroneously referred the Durbin charge to the Dallas City Attorney's office and not to the District Attorney of Dallas County. Because this also presented a class question, the court received evidence regarding this issue at the trial, there being fact questions concerning whether the EEOC had misled Durbin into believing that a proper referral of her charge had been made such that under the principles of White v. DISD, 581 F.2d 556, 562 (5th Cir. 1978), she ought to be allowed to proceed without compliance with Title VII's deferral requirements.The statutory command, although plain, has from time to time been disobeyed, sometimes in a manner misleading to the charging party. In White v. DISD, supra, the Fifth Circuit en banc found an equitable basis for avoiding the visitation of the EEOC's errors upon an innocent charging party. The escape valve is an affirmative misleading of the charging party by the EEOC. Here Durbin testified that while her charge was pending she was told directly by Mr. Starr, the EEOC investigator, that she need not do anything else and that the correct referral had been made. No letter to that effect was found although Durbin testified that she received one. The court finds that Durbin was misled into believing that the statute had been complied with, and relied to her detriment upon the EEOC. Defendant's motion to dismiss grounded on this contention is DENIED.
II. Defendants' Liability
A. Selecting a Police Officer in Dallas.
The Dallas police force has slowly grown in size from 2,006 in 1973 to approximately 2,050 in 1978. The growth, however, has not kept up with the demand, and at all times vacancies for new police officers existed. With rare exceptions, a police career in Dallas has followed only one path. That path begins with a four-step process leading from citizen to sworn officer in approximately 95 weeks. The four steps are:
1. Initial application; preliminary interview; performance and pencil-and-paper testing; medical and physical ability exam; background investigation; polygraph examination; and applicant interview board.
2. Police Academy (17 weeks).
3. Field Training (13 weeks).
4. Probationary period (26 weeks).
All recruits spend approximately the first two years of their career on patrol duty. The requirement of patrol duty at the outset for all officers has been the rule since 1972, when Title VII was amended to include municipalities within its coverage. Patrol is said to be the "core" work of the force and nearly one-half of all officers are so assigned. After the two-year stint on patrol, some officers move to other departments and some to higher rank. Only one black has ever attained a rank higher than lieutenant. The highest rank ever obtained by a female is lieutenant. Much of this concentration in the lower ranks may be explainable by the effect of pre-act seniority. This court need not explore the question of discrimination in promotions, however. The issues before the court revolve only around claimed discriminatory hiring practices.
B. The Class Contentions.
The pretrial order sets out the claims and general nature of the contentions of the parties. It is sufficient here to explain that the sum of the City's hiring procedures for police officers, including its specific requirement of 45 semester hours of college training, are claimed to be facially neutral but disparate in their effect upon blacks and females.
A requirement of a set number of semester hours has been in effect since October, 1971, when it was 30 hours. In October, 1972, the requirement changed to 45 hours, where it has remained throughout the period relevant to this lawsuit.
The plaintiffs have attempted to prove this disparate impact by various statistical analyses of the hiring data. Pointing to numerous calculations by their statistical expert, plaintiffs claim that the disparity between the number of blacks and females that ultimately found their way onto the force and the number said to be "available" is sufficient to rise not only to the level of statistical significance, but to that of "gross disparity." The City relies heavily upon its own calculations in its denial of disparate impact of its college credit requirement or any of the testing and training steps from applicant to sworn officer. The City also contends that its recruiting efforts have been so studied and forceful as to dispel any inference of sex or race discrimination that might flow from any conclusion of statistical significance.
With a test of statistical significance we learn the probability that the observed value could have happened by chance, i.e., the probability that in a random sample of an appropriate test population the variable would exhibit a value as strong as that observed. A test of statistical significance is thus itself based on the results of a hypothetical experiment. We suppose that an infinite number of samples of the same size are drawn from that population. The probability of the observed value occurring by chance is equal to the proportion of the samples in which the value is at least as extreme as the observed value. It has become a convention in social science to accept as statistically significant values which have a probability of occurring by chance 5% of the time or less. See generally, N. Nie, C. Hull, J. Jenkins, K. Steinbrunner, D. Bent, Statistical Package for the Social Sciences 222 (2d ed. 1975).
C. The Statistical Linchpins.
Despite the usual array of complicated statistical exercises, the inescapable conclusion is that liability here turns on one relatively simple question. That question is as follows: In order to measure the discrepancy between actual and potential hires, what figure should properly be compared to the actual number of hires? In other words, the question is how the actual record compares to what it presumably would have been in a sex- or race-neutral hiring process. Plaintiffs urge that the SMSA figure of 39% of work force ought to be used as a benchmark for the adequacy of female hires, but black hiring ought to be measured by a percentage of blacks of total applicants — 22%. The City argues that the measure for black hiring is their percentage of the SMSA or 12% (and not 22%); that female hiring ought to be measured by percentage of net hires or 15.6%. By "availability," the parties are referring to the percentage of the total pool from which hires were drawn accounted for by the group allegedly being discriminated against.
The class disputes the latter figure, arguing that a 17.5% female applicant flow figure should be used. See "Applicant Flow," infra.
The issue of whether plaintiffs established a prima facie case as to black applicants turns on whether general work force or applicant flow figures are used. If the general work force is considered to be the relevant standard, no violation of Title VII is shown; if the applicant flow figures are used, the burden to explain the disparity shifts to the City. With regard to females, use of general work force data, as urged by the plaintiffs, leads to statistically significant deviations from expected hires, while use of applicant flow figures leads to deviations which are not statistically significant.
D. The Hiring of Blacks.
The City's contentions with respect to hiring of blacks are, first, that a prima facie case has not been established, and, alternatively, that the City's vigorous recruiting efforts dispel any inference of racial discrimination. The argument that the plaintiffs did not make out a prima facie case is premised upon the further contention that the court should not compare the hires of the police department with the numbers of blacks that applied for positions, because the number of black applicants was inflated by the recruiting efforts of the City. The argument fails at the outset. The most relevant pool for comparison is surely those who have demonstrated interest and who meet all objective requirements, including the challenged requirement that an applicant have had at least 45 semester hours of college training. The applicant figures ". . . are taken from the City of Dallas Civil Service Registers which include names of applicants certified as eligible for employment. Appointment or rejection (and the reason given) are noted by the department upon return of the applicant's name from the Police Department. By eligible is meant that the applicant has applied, is of the proper age, and has no apparent disqualifications such as physical impairment or educational deficit. The applicant has also been given a Civil Service Examination, but since October, 1971, this examination result has not eliminated any applicant from the remainder of the screening process. . . ." (Defendant's Admission 1. Plaintiff's Exhibit 1).
There is no necessary inconsistency in allowing the use of applicant flow figures by plaintiffs while questioning their use by defendants. The potential vice of using the number of applicants as a benchmark of what the hires ought to have been is that potential applicants may well have been discouraged from applying; failure to recognize the hazards of defensive use of applicant figures might serve, in some cases, to reward successful discriminating practices. See Dothard v. Rawlinson, 433 U.S. 321, 97 S.Ct. 2720, 53 L.Ed.2d 786 (1977).
At first glance there is merit to the City's argument that there is a "Catch-22" to the use of figures generated by an employer's strong recruiting effort. Use of applicant figures would mean that an effort to achieve a balanced work force would tend to increase an employer's exposure to liability for Title VII violations. If such a threat actually exists, use of applicant figures in courts would arguably not further the congressional purpose behind the statute. But the City's argument depends on an unproven assumption. If increased recruiting is accompanied, as it ought to be, by increased hiring, use of applicant flow figures will probably not prejudice an employer. For the City's argument to be valid, it would have to show that the strong recruitment efforts were directed at an applicant pool less well qualified than the applicant pool it would have had without the strong recruitment.
The statute in disparate impact cases strikes at employment practices with proscribed effects; that the practice was not born of a racist or sexist animus is no defense. As the Supreme Court has stated, ". . . Proof of discriminatory motive . . . is not required under a disparate-impact theory." International Brotherhood of Teamsters v. United States, 431 U.S. 324, 335 n. 15, 97 S.Ct. 1843, 1855 n. 15, 52 L.Ed.2d 396 (1977). It follows, then, that the fact that the City was acting in good faith is no reason not to use the data which is most probative in measuring the evenness of its selection system.
Applying the 22% availability figure yields a statistically significant disparity between the expected number of blacks hired and the actual number hired. From January, 1973, through December, 1978, the police force hired 945 persons, of whom 130 were black. The difference between these 130 actual hires and the 211 expected hires (945 multiplied by 22%) is 80.7, which is equal to 6.31 standard deviations (based on a binominal distribution). The probability that this variance was the result of random, race-neutral factors is less than .0001, certainly statistically significant given the commonly accepted statistical significance level of .05.
Under the law a plaintiff presents a prima facie case when a case of gross disparity is proffered. It then passes to the City to show that the disparity is the result of some force other than unlawful discrimination. The City offered no such explanations except generalized statements about blacks' reluctance to engage in police work because of their attitude "toward the man." Apart from its typecasting bias, that argument blinks at the reality of the numbers of blacks that did apply. The question of why blacks have been disparately failed has not been answered. The answer may lie in the testing and schooling system. Tests may be permissible and job requirements may be legal even if they produce a disparate impact upon minorities, but they must be shown to be legally job related. The City has not validated its processes and is left here with no legal answer. Class-wide liability is thus established with regard to the hiring of blacks.
E. The Hiring of Females.
As with blacks, the crucial issue with respect to hiring of females is whether there is an unacceptable disparity between the rate of women hired and the percentage of women "available" for employment. The City has not suggested a nondiscriminatory reason for any difference, and, instead, has contended that the plaintiffs failed to prove a gross disparity. By contrast to the availability figures for blacks, the percentage of women in the general work force is far greater than the percentage of women in the pool of actual applicants. Not surprisingly, the plaintiffs urge that the general work force figure — 39% — is the relevant figure here, while the defendants point to applicant flow as the relevant standard. Before examining the parties' arguments in this regard, the court turns to a brief history of the City's hiring practices.
Before Congress extended the reach of Title VII to municipalities, there were no female police officers in Dallas, although there were a few "policewomen," whose total number never exceeded ten. Job qualifications and assignments differed. In November, 1971, for example, the respective qualifications were as follows:
Age Height Weight Vision Coll. Ed.
M 19 1/2-38 Min. 5'6" Min. 130 Min. 20/70C Min. 30 hrs.
F 22-35 5'2"-5'10" 110-166 Min. 20/30C Min. 30 hrs.
The education requirement had differed until May, 1971. Until then a female was required to have a college degree, while a male was required to have a high school diploma or its equivalent. As previously noted, in 1974 all officers were assigned to two years of patrol duty immediately following their field training. By August 31, 1973, 71 females were listed in the City's EEO-4 report. By June, 1979, females in nonsupervisory uniformed positions numbered 118 of the 1,609 persons in that category. The 1979 figure represents a doubling of the number of females (59) in June, 1974.
Beginning in October, 1972, applicants, male and female, were required to have 45 hours of college credit. Plaintiff has not shown that this requirement, a requirement not validated by defendants, had much of a disparate impact on females. In the absence of a precise index being offered at trial, this court, relying on some peripheral indicators, finds it not unreasonable to assume that this requirement had, in a very rough sense, similar impact on potential female applicants as potential male applicants. On a national basis, there were actually more females than males enrolled as first-time freshmen in institutions of higher education in the fall of 1976 and fall of 1977. Marquis Academic Media, Standard Education Almanac 1979-80 275 (12th ed. 1979). And in the years 1970-75, females accounted for 45-48% of first-time degree-credit enrollment in institutions of higher education. W. Grant C. Lind, Digest of Education Statistics 1979 93 (1979). Females accounted for between 43 and 45% of bachelor's degrees earned in the years 1970-76. M. Golladay J. Noell, The Condition of Education 138 (1978).
The initial question that must be answered with regard to the allegedly discriminatory hiring of females is whether the relevant figure against which the rate of actual hiring is to be measured is the availability of women in the general work force, or the percentage of women among those who actually applied. The plaintiffs have shown that the percentage of women in the general work force in the Dallas-Fort Worth SMSA is 39%, and argue that the representation of females on the police force should approximate that percentage. If this general work force figure is accepted as the benchmark, liability follows as a matter of course; juxtaposed to actual hires, a gross disparity is presented, thereby shifting to the City the burden of offering an explanation for the difference.
There is no inflexible rule governing the statistical figures to be used in a Title VII case. In Hazelwood School District v. United States, 433 U.S. 299, 97 S.Ct. 2736, 53 L.Ed.2d 768 (1977), involving teacher hiring practices, the Court indicated that the determination of what figures are relevant for statistical purposes is a rather flexible one to be made in view of the circumstances presented in each case. There, the Court stated:
In determining which of the two [labor pool] figures — or, very possibly what intermediate figure — provides the most accurate basis for comparison to the hiring figures at Hazelwood, it will be necessary to evaluate [various] considerations [listing five such considerations] . . . It is thus clear that a determination of the appropriate comparative figures in this case will depend upon further evaluation by the trial court. As this court admonished in Teamsters: "[s]tatistics . . . come in infinite variety . . . Their usefulness depends upon all of the surrounding facts and circumstances." 431 U.S., at 340 [ 97 S.Ct., at 1856-1857]. Only the trial court is in a position to make the appropriate determination. . . .433 U.S. at 311-12, 97 S.Ct. at 2743-44. See also Mayor v. Educational Equality League, 415 U.S. 605, 94 S.Ct. 1323, 39 L.Ed.2d 630 (1974). In Hazelwood, the Supreme Court considered that the relevant comparative figures were between the racial composition of the staff with the racial composition of the qualified public school teacher population in the relevant labor market. Hazelwood, supra, at 308, 97 S.Ct. 2736. The school district objected to the heavy reliance on these data on the ground that applicant flow data, showing the actual percentage of white and Negro applicants for teaching positions at Hazelwood, would be firmer proof. Id. at n. 13. The Court noted that here there was no clear evidence on exactly how many of the job applicants in each of the school years were Negroes. Significantly, the Court stated that were competent proof of this applicant flow to be adduced on remand, "it would, of course, be very relevant." Id.
In Dothard v. Rawlinson, supra, decided on the same day as Hazelwood, the Supreme Court reviewed a three-judge district court's finding of sex discrimination that was based on general work force figures. The Court rejected the appellant's argument that a showing of disproportionate impact on women based on those work force statistics should not suffice to establish a prima facie case. But the Court did not hold that use of general work force data was the only course open to the district court. The Court merely held that "[t]here is no requirement . . . that a statistical showing of disproportionate impact must always be based on analysis of the characteristics of actual applicants." (Emphasis added). Id. 433 U.S. at 330, 97 S.Ct. at 2727. Indeed, in Hazelwood, the Court, after referring to the "very relevant" nature of comparative applicant flow statistics, noted "Cf. Dothard v. Rawlinson, post, 433 U.S. at 330 [, 97 S.Ct. 2720]." By way of further support of this interpretation of Dothard, the Supreme Court in New York City Transit Authority v. Beazer, 440 U.S. 568, 99 S.Ct. 1355, 59 L.Ed.2d 587 (1979), noted:
Although "a statistical showing of disproportionate impact [need not] always be based on an analysis of the characteristics of actual applicants," Dothard v. Rawlinson, 433 U.S. 321, 330 [ 97 S.Ct. 2720, 2727, 53 L.Ed.2d 786], "evidence showing that the figures for the general population might not accurately reflect the pool of qualified job applicants" undermines the significance of such figures. Teamsters v. United States, supra [431 U.S.], at 340 n. 20 [ 97 S.Ct. at 1857 n. 20.Id. at 586 n. 29, 99 S.Ct. at 1366 n. 29.
Though not necessary to this opinion, it should be pointed out that much of the Court's criticism of the use of the flow figures in Dothard do not apply in this case. In Dothard there was evidence that the challenged employment practices — the use of height and weight requirements for employment as prison guards — had operated to discourage women who did not meet these easily-measured qualifications from applying for work as prison guards; thus the applicant flow data were suspect, and the district court apparently had good reason to use general work force data as the standard against which to measure actual hiring data. In this case, the City of Dallas did not impose easily-measured height and weight requirements; thus it is less likely that a potential applicant could determine his or her physical qualifications before actually making application and undergoing tests. Indeed, the fact that so many persons, male and female, failed the physical tests after applying indicates that the existence of the tests did not serve to discourage applications by those unqualified to pass them. This circumstance makes less plausible the argument that applicant flow data are reflective of a kind of self-effectuating preselection process.
An argument that the fact that since a large percentage of women pass the tests indicates that the existence of the tests serves to discourage women applicants in relative terms, and hence has deleterious effects on women is only superficially attractive. Equally probable speculations exist: (1) the particular form or weighting of the results of the various tests may favor females; (2) for reasons for which the City of Dallas cannot be held responsible, potential female applicants for police jobs may be more realistic than potential male applicants in assessing their own physical capabilities, and so relatively fewer genuinely unfit females may be applying.
The applicant flow here is not "tainted" by the education requirements either. See n. 5, supra.
We see then that under Dothard and Hazelwood, use of comparative labor force as well as comparative applicant flow data can be used, the choice depending on the surrounding circumstances. Under the circumstances of this case, it is far more likely that the general work force data are less reliable and, hence, less relevant than applicant flow data. The City's recruiter testified that she had found a marked job resistance to police patrol work among females. The City's ability tests actually operated in favor of women, and the City actively recruited women with the help of a full-time employee. Yet despite these factors enhancing job opportunities for women in the police department, only 15% to 18% of the department's applicant pool were women, compared to 39% women in the general labor pool. The most plausible explanation for the disparity, an explanation borne out by the evidence discussed, is that women were not during the relevant period interested in police work in the same proportion as were men. Moreover, the city recruiter, a black female, testified persuasively that substantial numbers of females decided against police work upon learning that assignment to patrol duty was mandatory. Thus the flow of female applicants is on these facts a far more meaningful standard against which to compare actual hiring of women than is the percentage of women in the general work force.
These job preferences of females may be born of attitudes conditioned by societal sexist values. But frustration with the realization that equality of opportunity untouched by gender remains a social goal and not an achieved reality, must not be visited on this employer in the form of liability.
There were 945 persons hired by the police department from January, 1973, through December, 1978, of which 147 were women. The expected number of women hires (17.5% multiplied by 945) is 165.375, leaving a disparity of 18.375 between actual and expected hires, or -1.57 standard deviations (using the binomial distribution). The probability that this disparity resulted from sex-neutral hiring procedures is .0582, or greater than one in 20, which does not reach the commonly accepted .05 level of statistical significance. Thus applying these figures, there is no statistically significant disparity between actual and expected hires, and no prima facie case of employment discrimination has been shown. The City is therefore not liable to this class of female applicants.
The plaintiffs essentially performed two sets of calculations, each based on the 17.5% availability figure. One set of calculations was based on the full number of persons hired (gross hires) by the police department during various time periods, while the other set was based on the number of persons hired reduced by the number of persons terminated (net hires). To use the latter set of figures would be to concern ourselves with the resultant of two practices: hiring, which is the subject of this suit, and termination, which is not. The "net hires" figure would be useful in analysis of hiring practices only if the employer were seeking to avoid a discrimination charge by "cycling" minorities — that is, by hiring a given number of minorities each year and terminating a like number during the year. In that situation the "net hires" figure would give the number of true hires. But there is no evidence that such "cycling" was going on here.