Alan A. Peightal v. Metropolitan Dade County, Metropolitan Fire Department of Dade County, 26 F.3d 1545

Alan A. Peightal v. Metropolitan Dade County, Metropolitan Fire Department of Dade County

26 F.3d 1545

11th Cir.

1994

Check Treatment

ANDERSON, Circuit Judge:

Alan Peightal (“Peightal”) brought an individual reverse discrimination claim against appellee Metropolitan Dade County (“Metro Dade”) because the Dade County Fire Department (“Fire Department”) hired minorities who scored lower than Peightal on the applicant exam, but did not hire Peightal. When Peightal applied for a job as a firefighter, the Fire Department was hiring pursuant to an affirmative action plan (“Plan”) which sought to redress prior discrimination against minorities and women in the Fire Department. The district court upheld the Plan against Peightal’s claims that it violated Title VII and the Equal Protection Clause of the United States Constitution. Our prior decision in Peightal v. Metropolitan Dade County, 940 F.2d 1394 (11th Cir.1991), cert. denied, — U.S. -, 112 S.Ct. 969, 117 L.Ed.2d 134 (1992), affirmed the district court with respect to the Title VII claim but remanded the equal protection claim for consideration in light of City of Richmond v. J.A. Croson Co., 488 U.S. 469, 109 S.Ct. 706, 102 L.Ed.2d 854 (1989). Following remand, the district court conducted a second bench trial and concluded that the Plan satisfied both prongs of the strict scrutiny analysis articulated in Croson. Peightal v. Metropolitan Dade County, 815 F.Supp. 1454 (S.D.Fla.1993). For the reasons set forth below, we affirm the judgment of the district court.

FACTS

On October 18, 1983, Peightal,- a white male, applied for a position as a firefighter with the Fire Department. Peightal took the firefighter examination in October of 1983 along with 3,300 others. Peightal’s score of 98.25 earned him a rank of 28 out of 2,188 persons who passed the test. Upon learning that he had been taken off the “stand by” list of applicants and had not been hired due to the Plan, Peightal filed an Equal Employment Opportunity Commission (“EEOC”) Complaint alleging racial discrimination. The EEOC issued Peightal a “right to sue” letter, resulting in this lawsuit under 42 U.S.C. § 1983 and under Title VII. 1 Following a bench trial in which the district court rendered judgment in favor of Metro Dade, Peightal appealed, resulting in our prior panel opinion. Peightal v. Metropolitan Dade County, 940 F.2d 1394 (11th Cir.1991), cert. denied, — U.S. -, 112 S.Ct. 969, 117 L.Ed.2d 134 (1992). That opinion sets forth in more detail the procedural history and facts surrounding the first appeal. Peightal, 940 F.2d at 1394-98. We affirmed with respect to Peightal’s Title VII claim but remanded Peightal’s equal protection claim for further findings in light of the Supreme Court’s opinion in City of Richmond v. J.A. Croson, 488 U.S. 469, 109 S.Ct. 706, 102 L.Ed.2d 854 (1989). The district court held another bench trial and again entered judgment in favor of Metro Dade. This appeal ensued.

The district court found that the position of firefighter is an entry-level job not requiring specialized skills or training. All applicants must satisfy seven basic requirements to become a firefighter: (1) possess a high school diploma or its equivalent; (2) possess a driver’s license and have the ability to obtain a chauffeur’s license; (3) be at least 18 years old; (4) pass a physical capabilities test; (5) pass a medical examination; (6) pass a personal interview; (7) have corrected vision in both eyes of at least 20/40. Those applicants who satisfy the age and high school graduation requirements are eligible to take the exam.

When Peightal applied in October 1983, the Fire Department was hiring pursuant to a minority preference program that called for the selection of female, black, and Hispanic applicants in accordance with certain goals established for the purpose of increasing the representation of these groups. Be *1549 fore adopting the Plan, Metro Dade 2 conducted an analysis of the Fire Department’s work force which revealed that in 1965 the Department employed 121 firefighters, all but one of whom were white males. By 1975, the Fire Department had increased to 499 firefighters, of whom 89% were white, 8% were black, 3% were Hispanic, and none were female. By 1983, when the Plan at issue was implemented, the number of firefighters in the Department was 921, of whom 74.9% were white, 11.8% were black, 13.8% were Hispanic and 1.3% were female.

After applicants took the firefighter examination, their respective scores were grouped and ranked by the applicants’ particular classification as defined pursuant to the Plan. Six separate categories were used: (i) Black Males, (ii) Black females, (in) White Females, (iv) Hispanic Males, (v) Hispanic Females, and (vi) White Males., Applicants received rankings within these categories on the basis of their test scores, with applicants ranked only against other members of the same category. As found by the district court, the Fire Department then hired in accordance with numerical goals, which were established for each upcoming hiring year and which were based on an analysis of the number of anticipated openings, the probable number of qualified applicants available in each category, and the extent of under-representation in each category.

The Plan called for the usage of a “70% rule” with respect to hiring. According to the Metro Dade “Affirmative Action Policy and Statement: Goals and Timetables:”

[essentially what the 70% rule says is that a significant disparity between minority representation in the service population, in our case, that is Metro Dade’s departments and divisions, may be deemed to exist if the percentage of a particular minority group in the department/agency is not at least 70% of the percentage of that minority in the service population, (emphasis added).

Once the 70% goal is reached, then a significant disparity no longer exists, and the preferential hiring program ends.

Applying the affirmative action plan including its 70% rule, the Fire Department’s affirmative action plan determined that its hiring goals for 1983 should include 15 black males, 29 Hispanic males, 8 black females, 8 Hispanic females, and 7 white females. In 1983, the Fire Department hired 86 firefighters which included 23 white males, 18 black males, 24 Hispanic males, 5 black females, 4 Hispanic females, and 12 white females. 51 of those hired scored lower than Peightal. In March of 1986, Peightal found out that he had been taken off the “stand by” list of applicants.

Prior to adopting the Plan, the Fire Department initiated alternative measures to increase minority representation through race-neutral means. The district court credited the testimony of Jacquelyn Rowe, the former Director of the Fire Department’s Affirmative Action Office, who testified that recruiters travelled to high schools and college campuses to encourage minority students to become firefighters. Additionally, the Fire Department designated a recruitment specialist to make presentations at job fairs and career days at local colleges and high schools. These measures also included outreach programs by minority members of the Fire Department.

In finding that these recruitment efforts achieved limited success, the district court stated that the Fire Department’s evaluation *1550 of test scores for applicants revealed that such efforts would be insufficient to address the problem of under-representation of women and minorities. Seventy percent of applicants taking the examination in 1983 passed and were thus eligible for consideration for employment offers. Hiring new firefighters in strict accordance with the test-ranked eligible list of applicants would have resulted in the hiring of few minority applicants. Of the 100 highest-scoring applicants, only six were Hispanic, one was female, and none were black.

The district court on remand considered statistical evidence comparing the Fire Department’s work force with the population of Dade County between the ages of 18 and 55 (rounded to the nearest whole number). 3 The district court credited the unrebutted testimony of Dr. David Santisteban, the Fire Department’s statistics expert, that statistical evidence regarding visual acuity and high school graduation rates — two additional selection criteria — was unavailable for the general population of Dade County from the United States Census Reports in 1983.

In 1970 the Fire Department work force was 97% white, 2% black, and 1% Hispanic, compared with the population of Dade County between the ages of 18 and 55 which was 49% white, 16% black, and 34% Hispanic. In 1975, 89% white, 8% black, and 3% Hispanic employees comprised the Fire Department, compared with a population which was 49% white, 16% black, and 34% Hispanic. By 1980, the Fire Department consisted of 80% white, 10% black, and 10% Hispanic employees, compared with a relevant population which was 43% white, 17% black, and 34% Hispanic. In 1983, the Fire Department work force was 75% white, 12% black, and 14% Hispanic, compared with the population of Dade County between the ages of 18 and 55 which was 38% white, 18% black, and 42% Hispanic. 4

The district court accepted the uncontro-verted testimony of Dr. Santisteban that the difference between the expected percentage of blacks in the Fire Department, given the percentage in the available labor pool, and the actual percentage was 4.8 standard deviations. The corresponding difference between the expected percentage of Hispanics and the actual percentage was 17.6 standard deviations. Dr. Santisteban testified that 17.6 standard deviations would occur by chance only once in every one billion cases.

Prior to adoption of the Plan, several allegations of racial discrimination were levelled against the Fire Department. Several complaints concerned the entry-level application procedures in the early 1980s, according to the testimony of Jacquelyn Rowe. A lawsuit *1551 filed by an organization of black firefighters, the Progressive .Firefighters Association (“PFA”) was resolved by a settlement agreement, signed in 1982, which required the validation of the Fire Department’s testing procedures and the establishment of a special recruitment team to actively recruit black applicants. The agreement to validate the testing procedures stemmed in part from a recognition of the apparent disparities in performance on the written examination between non-minorities and minorities.- According to the testimony of Edward Donaldson, Chief of the Fire Department from 1978 to 1987, the historically lower level of performance by minorities on the written examination indicated a need to evaluate the examination for bias.

In effectuating this validation process, the Human Services Institute, under contract with Metro Dade, conducted a study 5 in 1983 and concluded in a report written in April 1984 that the Fire Department’s selection procedures were deficient in predicting future job performance. 6 The report stated that a number of the selection procedures had an adverse impact upon minorities. This was particularly true for the written test. The study’s analysis indicated that the relative passing rates were 85% for white males, 56% for black males, 23% for Hispanic males, and 42% for females. 7 Although the report did not recommend that the written test should be discarded, it did suggest circumscribing its effect by de-emphasizing the importance of the score obtained and by placing greater emphasis on other steps in the selection procedure, such as the personal interview.

The application form for entry-level firefighters was at the initial stage self-identifying; thus, the Fire Department did not, as a general practice, screen each applicant’s claim of membership in a specific minority group. If an applicant noted that he was Hispanic, then the Fire Department accepted this self-identification until challenged. Upon a challenge of the asserted minority status of an applicant by another applicant or employee, the affirmative action office would thoroughly investigate the claim by applying the EEOC guidelines to the applicant. The EEOC definition of “Hispanic,” as applied by Metro Dade, includes: “All persons of Mexican, Puerto Rican, Cuban, Central or South American, or other Spanish culture or origin, regardless of race.” Pursuant to the EEOC guidelines, the affirmative action office in its investigation would request that the applicant supply a birth certificate, marriage certificate, or otherwise demonstrate cultural or linguistic ties to Hispanic heritage. The affirmative action office considered family ties, linguistic accent, and Spanish language capabilities as relevant to a claim of Hispanic origin. Pursuant to this investigatory procedure, the affirmative action office would then render its determination concerning the propriety of the applicant’s self-identification.

The district court accepted the testimony of Marcia Saunders, the Director of Affirmative Action, that a number of challenges to applicants’ minority status arose subsequent to initiation of the Plan. The district court credited testimony 8 that several of these challenges involved Hispanic female applicants had married non-Hispanic men, thereby taking non-Hispanic surnames. Additionally, challenges occurred against applicants who had Anglo surnames but who spoke *1552 Spanish or were actually born in Spanish-speaking countries. In all instances, applicants whose minority status was challenged were pronounced qualified by the affirmative action office pursuant to its investigatory procedure.

In upholding the Fire Department’s affirmative action plan, the district court concluded that the Department had demonstrated that the program was justified by the compelling purpose of redressing past discrimination and was narrowly tailored to that goal. This appeal followed, and for the reasons set forth below, we affirm the judgment of the district court.

On this appeal Peightal raises no Title VII challenges. He challenges, as applied to Hispanics, both aspects of the equal protection strict scrutiny analysis undertaken by the district court. 9 A primary contention of Peightal is that the definition of “Hispanic” as adopted by the Fire Department cannot survive Constitutional scrutiny. We address each of Peightal’s arguments in turn below.

DISCUSSION

The Equal Protection Clause of the Fourteenth Amendment provides that “[n]o state shall ... deny to any person within its jurisdiction the equal protection of the laws.” Guaranteed by its terms to the individual, the first section of the Fourteenth Amendment protects from illegitimate uses of race by government officials, Shelley v. Kraemer, 334 U.S. 1, 22, 68 S.Ct. 836, 846, 92 L.Ed. 1161 (1948), constraining state and local actors 10 in their authority to eradicate the effects of private discrimination. City of Richmond v. J.A. Croson Co., 488 U.S. 469, 491-92, 109 S.Ct. 706, 720-21, 102 L.Ed.2d 854 (1989). Accordingly, in assessing the propriety of any classification based upon race in an affirmative action plan, we are required to conduct a strict scrutiny analysis. Cone Corp. v. Hillsborough County, 908 F.2d 908, 913 (11th Cir.), cert. denied, 498 U.S. 983, 111 S.Ct. 516, 112 L.Ed.2d 528 (1990). See Croson, 488 U.S. at 494, 109 S.Ct. at 721 (plurality of four Justices) (“reaffirming] the view expressed by the plurality in Wygant [v. Jackson Bd. of Educ., 476 U.S. 267, 106 S.Ct. 1842, 90 L.Ed.2d 260 (1986) ] that the standard of review under the Equal Protection Clause is not dependent on the race of those burdened or benefitted by a particular classification”); 488 U.S. at 519, 109 S.Ct. at 735 (Kennedy, J., concurring in part and concurring in the judgment) (“accepting] ... rule contained in Justice O’Connor’s opinion”); 488 U.S. at 520, 109 S.Ct. at 735 (Scalia, J., concurring in the judgment) (“agreeing] ... that strict scrutiny must be applied to all governmental classification by race, whether or not its asserted purpose is ‘remedial’ or ‘benign’ ”). Under this analysis, race-conscious relief must be justified by a “compelling state interest,” id. at 505, 109 S.Ct. at 727, with relief “narrowly tailored” to further that interest. Id. at 507-08, 109 S.Ct. at 728-29. A compelling state interest “unquestionably” exists where race-conscious relief is aimed at “past and present discrimination by a state actor.” United States v. Paradise, 480 U.S. 149, 167, 107 S.Ct. 1053, 1064, 94 L.Ed.2d 203 (1987). See Wygant v. Jackson Board of Education, 476 U.S. 267, 286, 106 S.Ct. 1842, 1853, 90 L.Ed.2d 260 (1986) (“The Court is in agreement that ... remedying past or present discrimination ... is a sufficiently weighty state interest to warrant the remedial use of a carefully constructed affirmative action program”) (O’Connor, J., concurring in part and concurring in judgment). Accordingly, we as a reviewing court must ensure that race-conscious relief undertaken *1553 by Metro Dade is narrowly tailored to this remedial purpose. Croson, 488 U.S. at 506-07, 109 S.Ct. at 728-29; Wygant, 476 U.S. at 274, 106 S.Ct. at 1847.

A. Compelling Interest

Determining the parameters of discrimination by a state actor presents an evidentiary inquiry. The exact quantum of evidence required eludes precise definition; the governmental unit must have “ ‘a strong basis in evidence for its conclusion that remedial action was necessary.’ ” Croson, 488 U.S. at 501, 109 S.Ct. at 725 (quoting Wygant, 476 U.S. at 277, 106 S.Ct. at 1849 (plurality opinion)). 11 “Societal discrimination, without more, is too amorphous a basis for imposing a racially classified remedy.” Croson, 488 U.S. at 497, 109 S.Ct. at 724 (quoting Wygant, 476 U.S. at 276, 106 S.Ct. at 1848 (plurality opinion)). This follows because generalized assertions of past discrimination within an industry afford no logical terminus for a remedy. Croson, 488 U.S. at 498, 109 S.Ct. at 724. See Wygant, 476 U.S. at 276, 106 S.Ct. at 1848 (plurality opinion) (“In the absence of particularized findings, a court could uphold remedies that are ageless in their reach into the past, and timeless in their ability to affect the future”). Thus, a state or its subdivision possessing evidence that its. practices are exacerbating a pattern of prior discrimination may undertake race-conscious relief, provided that it identifies the discrimination with some specificity. Croson, 488 U.S. at 504, 109 S.Ct. at 727. In assessing whether the governmental unit has a “strong basis in evidence” for imposing race-conscious relief, the proffered evidence must “approach[] a prima facie case of a constitutional or statutory violation.” Croson, 488 U.S. at 500, 109 S.Ct. at 725. Evidence of “ ‘gross statistical disparities ... may constitute prima facie proof of a pattern or practice of discrimination’ under Title VII” and correspondingly may represent a strong basis in evidence for imposing race-conscious relief. Croson, 488 U.S. at 501, 109 S.Ct. at 725-26 (quoting Hazelwood School Dist. v. United States, 438 U.S. 299, 307-08, 97 S.Ct. 2736, 2741, 53 L.Ed.2d 768 (1977)). Thus, to successfully meet the factual predicate under the compelling interest inquiry, statistical comparisons between the employer’s work force and the composition of the relevant population are probative of a pattern of discrimination. As articulated by the Court in Hazelwood:

[Statistics can be an important source of proof in employment discrimination cases, since ‘absent-explanation, it is ordinarily to be expected that nondiscriminatory hiring practices will in time result in a work force more or less representative of the racial and ethnic composition of the population in the community from which employees are hired.’

Hazelwood, 433 U.S. at 307, 97 S.Ct. at 2736 (quoting International Brotherhood of Teamsters v. United States, 431 U.S. 324, 340 n. 20, 97 S.Ct. 1843, 1956 n. 20, 52 L.Ed.2d 396 (1977)).

i. Characteristics of the job

To adequately assess statistical data, there must be evidence identifying the basic qualifications of an entry-level firefighter and a determination, based upon these *1554 qualifications, of the. relevant statistical pool with which to make the appropriate comparisons. See Croson, 488 U.S. at 501-02, 109 S.Ct. at 725; Long v. City of Saginaw, 911 F.2d 1192, 1199 (6th Cir.1990). In identifying the basic job qualifications, an inquiry must be made to determine whether those qualifications rise to the level of special skills. Croson, 488 U.S. at 501, 109 S.Ct. at 726. This inquiry is necessary because in order to determine discriminatory exclusion, unskilled positions are compared to a different statistical pool than are jobs requiring special skills. See Teamsters v. United States, 431 U.S. 324, 337-38, 97 S.Ct. 1843, 1855-56, 52 L.Ed.2d 396 (1977). This flows from the Court’s recognition that where special skills are implicated, “comparison to the general population (rather than to the smaller group of individuals who possess the necessary qualifications) may have little probative value.” Croson, id. at 501, 109 S.Ct. at 726 (quoting Hazelwood School Dist. v. United States, 433 U.S. 299, 308 n. 13, 97 S.Ct. 2736, 2742 n. 13, 53 L.Ed.2d 768 (1977)). 12 Thus, for positions requiring minimal training or for certain entry level positions, statistical comparison to the racial composition of the relevant population suffices, whereas positions requiring special skills necessitate a determination of the number of minorities qualified to undertake the particular task. Croson, 488 U.S. at 501-02, 109 S.Ct. at 726.

The Supreme Court has offered limited guidance in distinguishing skilled positions from unskilled, illustrating the fact-intensive nature of the inquiry. A position may be deemed unskilled where the skill required “is one that many persons possess or can fairly readily acquire.” Hazelwood, supra, 433 U.S. at 308 n. 13, 97 S.Ct. at 2742 n. 13. Entry level positions require no special prior abilities because they are, in fact, “designed to provide expertise.” Johnson v. Transportation Agency, Santa Clara County, 480 U.S. 616, 632, 107 S.Ct. 1442, 1452, 94 L.Ed.2d 615 (1987). The district court specifically found that the seven prerequisites for the firefighter position at issue here do not rise to the level of special skills. We doubt that even a charitable reading of Peightal’s brief reflects any challenge to this finding of the district court. Moreover, there is no basis for rejecting as clearly erroneous the trial court’s finding that firefighter applicants are “unskilled” for purposes of comparative statistical analysis. Fed.R.Civ.P. 52(a). 13

ii. The relevant population

Upon a determination that the position is entry-level, requiring no special qualifications, the next inquiry is the racial composition of the relevant population for purposes of statistical comparison. We remandéd this case for further refinement of -the relevant population, noting the relative undesirability of general population figures. Peightal, 940 F.2d at 1412 (Tjoflat, J., concurring and dissenting) (“General population figures ... *1555 only serve as a proxy for the qualified applicant pool from which the employer is hiring”). Although the Court in Teamsters upheld the use of general population figures as opposed to a more particularized assessment of the qualified applicant pool, in that case, significant anecdotal evidence of discrimination bolstered the statistics in question. International Bhd. of Teamsters v. United States, 431 U.S. 324, 339, 97 S.Ct. 1843, 1856, 52 L.Ed.2d 396 (1977). 14 The Court in Teamsters cautioned that where the “evidence show[s] that the figures for the general population might not accurately reflect the pool of qualified applicants,” this may diminish the value of the statistical comparison. Teamsters, id. at 339 n. 20, 97 S.Ct. at 1857 n. 20. Following this Court’s suggestion on remand, the district court conducted a statistical analysis which compared the Fire Department’s work force to those members of the general population of Metro Dade who fall within the qualified age group of 18 to 55 year-olds. See Hammon v. Berry, 826 F.2d 73, 77 (D.C.Cir.1987) (denial of rehearing by panel) (“There should be no mistaking the correct benchmark in this case: the relevant labor force consists of persons 20 to 28 years of age in the Washington metropolitan area”) (emphasis omitted). This refinement resulted in a disparity even greater than that suggested by the earlier unrefined data. 15

Peightal argues that the methodology employed by the district court failed to properly take into account minimum requirements for entry-level firefighters. However, except for a conclusory reference to age restrictions, high school education, and language proficiency, Peightal argues without reference to evidence or logic. We uphold the district court’s finding that the relevant qualified age group consisted of those persons 18 to 55 years of age. With respect to the other minimum requirements, the district court specifically found that further refinement was not feasible because data was not available, and appellant failed to adduce any evidence to the contrary. For example, the level of education of minorities, although available in the 1990 United States Census, was unavailable through the census in the 1970s and 1980s, according to testimony credited by the district court; appellant offered no contrary testimony on this issue. The thorough inquiry conducted by the district court with respect to the availability of data concerning each of the minimum requirements coupled with the appellant’s failure to adduce more refined statistical data fulfills this Court’s prior directive, given “the difficulty, if not impossibility in obtaining labor market figures that precisely identify the racial composition of the qualified applicant pool.” Peightal, 940 F.2d at 1412 (Tjoflat, J., concurring and dissenting).

iii. Proper geographic scope

A further statistical consideration articulated by the Court in Croson is the determination of the proper geographic scope for the relevant labor pool. Croson, 488 U.S. at 504, 109 S.Ct. at 727. In the instant case, the use of the Dade County geographical boundaries has gone unchallenged. Accordingly, we decline to disturb the district court’s findings in this regard.

iv. Evidence of statistical disparity

Having ascertained the nature of the job, the relevant population, and proper geographic scope, we turn to the evidence of statistical disparity presented to the district court. Evidence that the statistical imbalance between minorities and non-minorities in the relevant work force and available labor pool constitutes a gross disparity, and thus a prima facie case of a constitutional or statutory violation, may justify a public employer’s adoption of racial or gender preferences. Cone Corp. v. Hillsborough County, 908 F.2d 908, 915-16 (11th Cir.), cert. denied, 498 U.S. 983, 111 S.Ct. 516, 112 L.Ed.2d 528 (1990); Croson, 488 U.S. at 500, 109 S.Ct. at 724-25; see also Wygant, 476 U.S. at 274-75, 106 S.Ct. at 1847 (plurality opinion), and id. at 293, 106 S.Ct. at 1857 (O’Connor, J., concur *1556 ring). Statistical analysis of- market data should not be accorded talismanic significance. See Watson v. Fort Worth Bank & Trust, 487 U.S. 977, 995-96 n. 3, 108 S.Ct. 2777, 2789-90 n. 3, 101 L.Ed.2d 827 (1988) (“We have not suggested that any particular number of standard deviations can determine whether a plaintiff has made out a prima facie case in the complex area of employment discrimination”); cf. Hazelwood, 433 U.S. at 318, 97 S.Ct. at 2747 (Stevens, J., dissenting) (“absolute precision in the analysis of market data is too much to expect”). Rather, disparate impact analysis should be considered “a numerical comparison that will help identify a possibly unfair, discriminatory hurdle interposed between the eligible minority applicant and success,” Stuart v. Roache, 951 F.2d 446, 451 (1st Cir.1991), cert. denied, — U.S. -, 112 S.Ct. 1948, 118 L.Ed.2d 553 (1992), viewed flexibly to accommodate the complexities inherent in reverse discrimination cases. Accordingly, as a general proposition, where “ ‘the difference between, the expected value and the observed number is greater than two or three standard deviations,’. then the hypothesis that [employees] were hired without regard to race would be suspect.” Hazelwood, 433 U.S. at 308 n. 13, 97 S.Ct. 2742 n. 13 (quoting Castaneda v. Partida, 430 U.S. 482, 497 n. 17, 97 S.Ct. 1272, 1281 n. 17, 51 L.Ed.2d 498 (1977)). The district court credited the evidence presented that the difference between the expected and actual percentage of Hispanics in the Fire Department, given the percentage in the age-eligible population, was 17.6 standard deviations. 16 According to the unrebutted testimony of Dr. Santisteban, the statistics expert for the Fire Department, 17.6 standard deviations would occur by chance only once in every one billion cases. Accordingly, we uphold the district court’s determination that evidence of a statistical disparity and thus constituted the requisite “strong basis in evidence” mandated by Croson.

As a final effort to undermine the statistics presented by the Fire Department, Peightal argues that the proffered statistical analysis fails to factor in the effect of the influx of Latin American immigrants into Dade County in the late 1970s and 1980s. Peightal urges that the statistics could not properly distinguish underrepresentation in the Fire Department based upon past discrimination from that caused by sudden demographic changes. 17 The district court found that the gross statistical disparity between the actual and expected percentage of Hispanics in the Fire Department predated the large migrations from Cuba, Haiti, and Nicaragua. Specifically, in 1975 Hispanics accounted for 34% of Metro Dade’s general population between the ages of 18 and 55, whereas a mere 3% of the Department’s firefighters were Hispanic. The district court properly concluded that Peightal had failed to adduce any evidence to support his speculations or undermine the validity of the statistical evidence which was credited by the district court. 18 We uphold the district *1557 court’s findings that Peightal failed to rebut the Fire Department’s statistical analysis either by demonstrating that it was flawed in its methodology or by introducing contrasting data which indicated that significant disparities were not present.

Accordingly, from these findings, coupled with Peightal’s failure to present contrasting statistical data, we readily conclude that Metro Dade has demonstrated a compelling government interest in remedying the effects of prior discrimination. We turn now to the second prong of the strict scrutiny analysis outlined in Croson to determine whether the Plan is narrowly tailored to satisfy the compelling interest Metro Dade has demonstrated.

B. Narrowly Tailored

.To invoke race-conscious relief, a state or local government must utilize means which are specifically and narrowly tailored to accomplish its remedial purpose. Croson, 488 U.S. at 507-08, 109 S.Ct. at 729-30; Wygant, 476 U.S. at 280, 106 S.Ct. at 1850 (plurality opinion). “Racial classifications are simply too pernicious to permit any but the most exact connection between justification and classification.” Wygant, Id. (quoting Fullilove, 448 U.S. at 537, 100 S.Ct. at 2805 (Stevens, J., dissenting)). In determining the propriety of race-conscious relief, a court must address several factors which include:

the necessity for the relief and the efficacy of alternative remedies; the flexibility and duration of the relief, including the availability of waiver provisions, the relationship of the numerical goals to the relevant labor market; and the impact of the relief on the rights of third parties.

United States v. Paradise, 480 U.S. 149, 171, 107 S.Ct. 1053, 1066, 94 L.Ed.2d 203 (1987) (plurality opinion). See Howard v. McLucas, 871 F.2d 1000, 1008 (11th Cir.), cert. denied, 493 U.S. 1002, 110 S.Ct. 560, 107 L.Ed.2d 555 (1989). Put another way, three distinct inquiries fuel the narrow tailoring analysis: (1) the prior consideration of race-neutral alternative measures; (2) the relationship between the degree of preference and the effects of prior discrimination; and (3) the severity of the burden placed upon third parties. 19 See Croson, 488 U.S. at 489, 504-510, 109 S.Ct. at 719, 727-30; Fullilove, 448 U.S. at 463-67, 484, 486-88, 100 S.Ct. at 2767-69, 2777, 2779-80.

i. Alternative remedies

To determine whether a race-conscious remedy is narrowly tailored to serve a compelling state interest, an initial inquiry should be undertaken to determine whether the local government considered “the use of race-neutral means to increase minority” participation. Croson, 488 U.S. at 507, 109 S.Ct. at 729. In Croson, the Court observed that the record contained no evidence that the Richmond City Council considered any alternatives to the racial quota adopted. Id. The Court contrasted this absence of evidence to the program upheld in Fullilove, in which Congress “carefully examined and rejected race-neutral alternatives before enacting the MBE set-aside.” Id. In assessing the propriety of race-neutral measures, a balance must be struck between their desirability and efficacy. Because “preferential programs may only reinforce common stereotypes holding that certain groups are unable to achieve success without special protection,” Regents of University of California v. Bakke, 438 U.S. 265, 298, 98 S.Ct. 2733, 2752, 57 L.Ed.2d 750 (1978) (opinion of Powell, J.), race-neutral means are to be favored, subject to the caveat that a government entity need not “exhaust every alternative, however irrational, costly, unreasonable, and unlikely to succeed_” Coral Const. Co. v. King County, 941 F.2d 910, 923 (9th Cir.1991), cert. denied, — U.S. -, 112 S.Ct. 875, 116 L.Ed.2d 780 (1992) (emphasis in original).

In this case, the district court found that the Fire Department did initially implement race-neutral measures. Specifically, the Fire Department initiated high school and college *1558 recruiting programs to provide information and to solicit applications from young minorities and women for firefighting positions. The Fire Department additionally held outreach programs which minority firefighters spearheaded. Finally, the Fire Department designated a recruitment specialist to organize and implement a special recruitment program. The duties of the recruitment,specialist included presentations at job fairs and career days at local colleges designed to apprise minorities and women of fire service career opportunities. The district court found that these extensive recruitment and outreach programs achieved only limited success, with the Fire Department’s rank-ordered list overwhelmingly dominated by white males subsequent to the programs’ implementation. Had the Fire Department strictly adhered to the rank ordering in 1988, the resulting class of firefighters would have consisted almost entirely of white males, with no blacks, no women, and only a handful of Hispanics. 20 Accordingly, we uphold the district court’s determination as not clearly erroneous that the Fire Department satisfactorily considered and implemented race-neutral means before deeming them ineffective and undertaking race-conscious measures.

ii. Relationship of relief to prior discrimination

In assessing whether a race-conscious measure is narrowly tailored, the second inquiry undertaken ensures that the degree of remedial preference is tied to the effects of past disadvantages or discrimination. Croson, 488 U.S. at 507-508, 109 S.Ct. at 729-30; Fullilove, 448 U.S. at 463-67, 100 S.Ct. at 2767-69; 100 S.Ct. at 2791-93 (Powell, J., concurring). This inquiry analyzes the flexibility and duration of the remedy as well as the relationship of the numerical goals to the relevant labor market. United States v. Paradise, 480 U.S. at 171, 107 S.Ct. at 1066.

In considering whether a race-conscious remedy meets this requirement, courts should emphasize that the government entity should adopt a program for a limited duration to facilitate its periodic reevaluation. See, e.g., Fullilove, 448 U.S. at 513, 100 S.Ct. at 2792 (Powell, J., concurring) (“[t]he temporary nature of this remedy ensures that a race-conscious program will not last longer than the discriminatory effects it is designed to eliminate”). Limiting the duration of relief also ensures that courts will not “uphold remedies that are ageless in their reach into the past, and timeless in their ability to affect the future.” See Wygant, 476 U.S. at 276, 106 S.Ct. at 1848 (plurality opinion). Race-conscious relief must be particularly circumscribed to avoid reinforcement of “a manner of thinking by race that was the source of the injustice and that will, if it endures within our society, be the source of more injustice still.” Croson, 488 U.S. at 527-28, 109 S.Ct. at 740 (Scalia, J., concurring in the judgment). Flexibility forbids mechanistic application of fixed quotas. Unyielding quotas operate under the unrealistic premise that “minorities will choose a particular trade in lockstep proportion to their representation in the local population.” Croson, 488 U.S. at 507, 109 S.Ct. at 729. By contrast, case-by-case utilization goals, including percentage preferences, eschew making the color of an applicant’s skin the sole relevant consideration and thus are less problematic from an equal protection standpoint. Coral Construction Co. v. King County, 941 F.2d 910, *1559 924 (9th Cir.1991), cert. denied, — U.S. -, 112 S.Ct. 875, 116 L.Ed.2d 780 (1992).

Turning first to the duration of the Fire Department’s program, the district court found that the Plan was of limited duration because it contains a provision ensuring that the program will end when it achieves its 70% hiring goal. According to the testimony of Jacquelyn Rowe, former director of the Affirmative Action Office, “[t]he plan is reviewed annually ... [T]he plan is remedial in nature, and it addresses problems of under-utilization. If they don’t exist, then there is no need to set a goal.” Significantly, the goals in the instant case were reviewed before each hiring year, and a new goal established, based upon the extent of underrepresentation remaining, the probable number of qualified minority applicants, and the anticipated number of openings. These findings of the district court support the court’s conclusion that the Plan is narrowly tailored.

Peightal next argues that the Plan created a de facto quota instead of a goal. Peightal urges that because the Fire Department grouped applicants into six categories of racial and ethnic classifications, with applicants compared only against members of their respective group, this effectively isolated each group’s members from overall competition, thus creating a quota. We decline appellant’s invitation to revisit this issue, given our adoption of the district court’s fact-finding prior to remand that the Plan utilized a goal as opposed to a quota. Peightal, 940 F.2d at 1396-97 and n. 10, 1410. 21

Peightal contends that the Plan is nevertheless constitutionally flawed because of its definition of “Hispanic.” 22 Peightal contends that the amorphousness of the definition, coupled with the self-identification procedure utilized on the application form, creates an improper classification for remedial purposes. The crux of Peightal’s argument is that the Plan is both over-and-under-inclusive in its preferential treatment of His-panies. Peightal argues that the Plan is over-inclusive because it encompasses the same variety of ethnic and racial characteristics as are found in the generalized classification of all English speaking nations; for example, it includes persons who can trace their ancestry to Spain, irrespective of language or culture.

The EEOC definition of “Hispanic,” as applied by Metro Dade, includes: “All persons of Mexican, Puerto Rican, Cuban, Central or South American, or other Spanish culture or origin, regardless of race.” The EEOC guidelines also require that a person’s claim of identification with a certain racial or ethnic group “should accompany strong visible indication that the person culturally and linguistically identifies with the group he or she claims.” (emphasis in original). The district court found that Metro Dade, in applying the definition of “Hispanic” articulated by the EEOC, also applied this limiting guideline. The district court, thus, specifically rejected Peightal’s effort to east the classification procedure utilized by the Department solely as self-identifying. The investigatory procedures utilized by the affirmative action office operated as a check to the self-identification procedure and also operated to narrow the definition of the protected group to those with strong visible indications of cultural and *1560 linguistic identification with the group, thus ensuring that non-minorities would not erroneously receive race-conscious relief.

The Supreme Court in Croson cautioned that a plan’s gross over-inclusiveness of racial groups that have never suffered from prior discrimination may indicate that race-conscious relief is not narrowly tailored. Croson, 488 U.S. at 505, 109 S.Ct. at 728 (quoting Wygant, 476 U.S. at 284 n. 13, 106 S.Ct. at 1852 n. 13 (“haphazard inclusion of racial groups ‘further illustrates the undifferentiated nature of the plan’ ”)). Whereas the program invalidated in Croson included groups for which there was no evidence of prior discrimination, Metro Dade adduced ample evidence of prior discrimination for the “group” of Hispanics identified by its Plan. Contained within Peightal’s assertion concerning the impropriety of this classification is the implication that certain “subgroups” should not have been included within the “group” of Hispanics. Nevertheless, the district court implicitly found that application of the definition in conjunction with the EEOC guidelines operated effectively to narrow the defined class to exclude those persons having no cultural or linguistic identification with the Hispanic group. The affirmative action office interviewed applicants or employees whose minority status was challenged, applying the EEOC guidelines to the person’s self-identification. Significantly, Peightal failed to adduce one instance in which application of the definition and guides lines produced an over-inclusive result, whereby a non-minority received race-conscious relief.

Although a number of affirmative action programs have tried to develop a narrower substitute for the term “Hispanic,” Peightal has conceded that no case has invalidated the EEOC definition applied by Metro Dade. See, e.g., Mackin v. City of Boston, 969 F.2d at 1274 n. 2, cert. denied, — U.S. -, 113 S.Ct. 1043, 122 L.Ed.2d 352 (1993) (“Spanish-surnamed”); Fullilove, 448 U.S. at 464, 100 S.Ct. at 2767 (“Spanish-Americans”); and Croson, 488 U.S. at 478, 109 S.Ct. at 713 (“Spanish speaking”). Further, these alternative definitions may similarly suffer from the infirmity of over-inclusiveness under the analysis employed by Peightal. For example, Spanish-surnamed individuals could include those individuals who have married someone of Spanish ancestry but who themselves lack Spanish heritage. Similarly, the group of Spanish-speaking individuals could include persons of Caucasian ancestry who have simply chosen to learn a second language. Peightal has offered no alternative definition to that created by the EEOC and has offered no case containing a definition that could not produce the over-inclusiveness of which he complains. 23 From the foregoing, we conclude that the Plan is not over-inclusive.

*1561 Peightal also attacks the Plan as under-inclusive, arguing that it benefits those of Spanish descent, while excluding, and therefore discriminating against, persons of other European national origin. Peightal claims that Brazilians of Portuguese descent, Greeks, Jews, Italians, and Iranians are culturally and linguistically diseemable, and may have suffered similar discrimination, but have been denied benign preference in hiring by Metro Dade’s Plan. 24 Peightal also asserts that it is logically inconsistent to deny Brazilians of Portuguese origin hiring preference while according such a preference to Brazilians of Spanish origin. Peightal has not offered evidence or pointed to evidence submitted by Metro Dade to support an assertion that prior discrimination exists against the groups he has identified. Absent such evidence, implementation of race-conscious relief in favor of such groups would have violated the teachings of Croson, 488 U.S. at 506, 109 S.Ct. at 728. Indeed, Peigh-tal has not proffered any evidence in support of his illustrations of underinclusiveness. 25 Accordingly, we hold that the Plan is not under-inclusive.

iii. Effect on third parties

Finally, the third inquiry undertaken to determine the propriety of race-conscious relief under the second prong of the strict scrutiny analysis is to ascertain whether the impact of the relief upon third parties is unduly burdensome. Wygant, 476 U.S. at 281-83, 106 S.Ct. at 1850-52 (opinion of Powell, J.); Fullilove, 448 U.S. at 484, 100 S.Ct. at 2777-79 (plurality opinion). No constitutional defect necessarily arises from the disappointment of non-minorities asked to share the burden in curing the effects of prior discrimination. Fullilove, 448 U.S. at 484, 100 S.Ct. at 2778. For example, in Fullilove, the statute at issue mandated a 10% MBE set-aside. The Court upheld this requirement in part because the “actual ‘burden’ shouldered by non-minority firms [was] relatively light.” Id. Critical to the analysis of the appropriate burden to be shouldered by non-minorities is the extent to which the relief disrupts settled “rights and expectations.” Wygant, 476 U.S. at 283, 106 S.Ct. at 1851. As elaborated by the Court in Wy-gant:

In cases involving valid hiring goals, the burden to be borne by innocent individuals is diffused to a considerable, extent among society generally. Though hiring goals may burden some innocent individuals, théy simply do not impose the same kind of injury that layoffs impose. Denial of a future employment opportunity as not as intrusive as loss of an existing job.

*1562 476 U.S. at 282-83, 106 S.Ct. at 1851 (opinion of Powell, J.) (emphasis in original). 26 See Jansen v. City of Cincinnati, 977 F.2d 238, 242 (6th Cir.1992), cert. denied, — U.S. -, 113 S.Ct. 2344, 124 L.Ed.2d 254 (1993) (“ ‘although initial employment opportunities coupled •with hiring goals may burden some innocent individuals, they do not impose the same type of intrusive injuries that layoffs, which result in loss of job expectancy, security, and seniority, involve’ ”) (quoting Long v. City of Saginaw, 911 F.2d 1192, 1196-97 (6th Cir.1990)). The district court found that the Plan did not require layoffs and did not pose an absolute bar to the hiring of non-minorities. Thus, disappointed non-minority applicants under this Plan shoulder a type of diffuse burden similar to that approved in Fullilove and Wygant.

Further, the study conducted by the Human Services Institute on behalf of Metro Dade confirmed prior suspicions that the written firefighter test had an adverse impact upon minorities and that the test was not a valid predictor of subsequent job performance. Specifically, 85% of white males taking the examination passed, as compared with 23% of Hispanic males. 27 Given the test’s unreliability and its adverse impact on minorities, we cannot say that Peightal should be entitled to relief based upon the rigid rank-ordering produced from such a flawed examination or that Metro Dade’s departure from these rankings placed an undue onus upon Peightal. From the foregoing, we hold that the Plan did not unduly burden non-minority applicants.

Having analyzed Metro Dade’s consideration and implementation of alternative remedies, the relationship between the remedy and prior discrimination, and the impact of the Plan upon third parties, we conclude that the Plan is narrowly tailored.

CONCLUSION

From the foregoing, we hold that Metro Dade’s Plan survives both prongs of the strict scrutiny analysis articulated in Croson. Metro Dade’s evidence of gross statistical disparities between the population of Hispanics in Metro Dade and the relevant qualified labor pool, coupled with Peightal’s failure to adduce contrary statistical evidence, satisfies Croson’s compelling interest requirement. Evidence that Metro Dade adequately considered and implemented alternative remedies to race-conscious relief, taken in tandem with the flexibility and limited duration of the Plan, our determination that the EEOC definition as applied by Metro Dade was not over-or-under-inclusive, and our conclusion that the Plan did not unnecessarily burden third parties, leads us to hold that the Plan satisfies the narrow tailoring requirement mandated by Croson. Accordingly, we AFFIRM the judgment of the district court.

AFFIRMED.

Notes

. Peightal alleged the violation of both Title VII and the equal protection clause as the predicate for his claim for recovery under § 1983.

. Metropolitan Dade County is a sovereign entity comprehending specific powers as to the 26 municipalities and the unincorporated area within the geographical boundaries of the county. The Florida Constitution provides:

The electors of Dade County, Florida, are grante.d to power to adopt, revise, and amend from time to time a home rule charter of government for Dade County, Florida, under which the Board of County Commissioners of Dade County shall be the governing body.

See Constitution, State of Florida, Home Rule Amendment Relating to Dade County, Art. VII, Sec. 11 (adopted Nov. 6, 1956). The Home Charter for Metropolitan Dade County, (Official Records Book 182, p. 667, Public Records of Dade County, Florida) (adopted May 21, 1957), provides that the "Board of County Commissioners shall be the legislative and the governing body of the county and shall have the power to carry on a central metropolitan government." Id., Ar. I., Sec. 1.01.

. At the first trial, only general population data were considered. On remand, the district court considered refined data which included age restrictions for the entry-level applicants so that data would more closely approximate the qualified applicant pool. Peightal, 940 F.2d at 1412 (Tjoflat, J., concurring and dissenting). The district court determined that the Fire Department did not receive applications from persons over age 55, and applicants under age 18 were ineligible for the exam.

. The relevant statistical evidence is summarized as follows:

[[Image here]]

. The Institute issued the study, entitled, “Job Analysis and Selection Procedures Study for the Positions of Firefighter, Administrative Officer I, and Social Worker I,” in April, 1984. Dr. David Santisteban, the Project Director, submitted the Final Report to Metro Dade.

. The study compared test scores with subsequent performance in fire college and first year on-the-job performance evaluations.

. The report also noted disparate passing rates for the physical capability test with the rates being 94% for Hispanic males, 90% for white males, 76% for black males, and 19% for fe-. males. Dr. Santisteban testified that the physical capability test had an adverse impact on women, with the swimming component of the test having an adverse impact on blacks. The study concluded that the physical capability test bore no relationship to the physical exertions required of a firefighter and accordingly suggested substitution of a more practical test to better assess and correspond with on-the-job requirements.

.The district court noted that no written documentation of these challenges was presented at trial.

. Peightal's brief raises no Title VII challenge, and no equal protection challenge as to blacks or females. His challenge focuses only on Hispanics, the proof of prior discrimination against Hispanics, and the corresponding development of goals with respect to Hispanics.

. The Supreme Court has distinguished programs mandated by Congress and those undertaken by states and other local governmental entities. As articulated by Justice Brennan in Metro Broadcasting, Inc. v. F.C.C., 497 U.S. 547, 110 S.Ct. 2997, 111 L.Ed.2d 445 (1990):

[M]uch of the language and reasoning in Cro-son reaffirmed the lesson of Fullilove [v. Klutznick, 448 U.S. 448, 100 S.Ct. 2758, 65 L.Ed.2d 902 (1980)] that race-conscious classifications adopted by Congress to address racial and ethnic discrimination are subject to a different standard than such classifications prescribed by state and local governments.

497 U.S. at 565, 110 S.Ct. at 3009.

. Justice Powell suggested a stronger evidentia-ry standard when he wrote that "judicial, legislative, or administrative findings of constitutional or statutory violations” may be necessary to trigger a governmental interest in remedying past discrimination. Regents of Univ. of Cal. v. Bakke, 438 U.S. 265, 307, 98 S.Ct. 2733, 2757, 57 L.Ed.2d 750 (1978). Justice Powell, however, later wrote in support of a less strict standard, authorizing race conscious relief where the public entity had a "strong basis in evidence.” Wygant, 476 U.S. at 277, 106 S.Ct. at 1940 (plurality). See also id. at 289, 106 S.Ct. at 1855 (O’Connor, J., concurring in part and concurring in the judgment) ("agreeing] with the plurality that a contemporaneous or antecedent finding of past discrimination by a court or other competent body is not a constitutional prerequisite”). Other circuits have also focused and relied on the "strong basis in evidence” language to determine if sufficient evidence of past discrimination existed to justify a racial preference. See United Black Firefighters Assoc. v. City of Akron, 976 F.2d 999, 1009 (6th Cir.1992); Stuart v. Roache, 951 F.2d 446, 449 (1st Cir.1991), cert. denied, — U.S. -, 112 S.Ct, 1948, 118 L.Ed.2d 553 (1992); Donaghy v. City of Omaha, 933 F.2d 1448, 1458 (8th Cir.1991); Davis v. City and County of San Francisco, 890 F.2d 1438, 1447 (9th Cir.1989), cert. denied, 498 U.S. 897, 111 S.Ct. 248, 112 L.Ed.2d 206 (1990).

. For positions requiring special skills, the relevant statistical population must be refined to screen for the effects of mere societal discrimination in access to training and skills. See United States v. City and County of San Francisco, et al., 696 F.Supp. 1287, 1303 (N.D.Cal.1988), aff'd in part and modified in part on other grounds sub nom. Davis v. City of San Francisco, 890 F.2d 1438 (9th Cir.1989), cert. denied, 498 U.S. 897, 111 S.Ct. 248, 112 L.Ed.2d 206 (1990).

. The driver’s license requirement at issue here is similar to that in Teamsters, supra, in which the Court determined that the ability to drive a truck did not constitute a special skill. Teamsters, 431 U.S. 324, 329 n. 3, 97 S.Ct. 1843, 1852 n. 3, 52 L.Ed.2d 396 (1977) (statistical comparison between minority truck drivers and relevant population probative of discriminatory exclusion). The educational requirement of a high school diploma or its equivalent is markedly less than pre-employment training or certification or college education requirements associated with public school teachers in Hazelwood, supra. Hazelwood, 433 U.S. at 308 and n. 13, 97 S.Ct. at 2742 and n. 13 (relevant statistical pool is the racial composition of the qualified public school teacher in the relevant labor market). The Ninth Circuit has also upheld the finding that the position of firefighter is entry-level, requiring no special skills. Davis v. City and County of San Francisco, 890 F.2d 1438, 1447 (9th Cir.1989), cert. denied, 498 U.S. 897, 111 S.Ct. 248, 112 L.Ed.2d 206 (1990). See also Bennett v. Arrington, 806 F.Supp. 926 (N.D.Ala.1992), aff'd in part and rev’d in part, In re Birmingham Reverse Discrimination Employment Litigation, 20 F.3d 1525 (1994); Mackin v. City of Boston, No. 89-2025-S, 1991 WL 349619 (D.Mass.1991), aff'd, 969 F.2d 1273 (1st Cir.1992), cert. denied, - U.S. -, 113 S.Ct. 1043, 122 L.Ed.2d 352 (1993).

. Additionally, further statistical refinements could not obscure the “inexorable zero" of minority drivers in that case. Teamsters, 431 U.S. at 342 n. 23, 97 S.Ct. at 1858 n. 23.

. Earlier statistical disparity data indicated a standard deviation, of 14.5, Peightal, 940 F.2d at 1406 n. 33, whereas the current data indicates a standard deviation of 17.6.

.The Second Circuit recently explored the contours of statistical analysis in this regard:

Standard deviation analysis measures the probability that a result is a random deviation from a predicted result — the more standard deviations the lower the probability the result is a random one. Social scientists consider a finding of two standard deviations significant, meaning there is about one chance in 20 that the explanation for the deviation could be random and the deviation must be accounted for by some factor other than chance. A finding of two or three standard deviations (one in 384 chance the result is random) is generally considered highly probative of discriminatory treatment.

Waisome v. Port Authority, 948 F.2d 1370, 1376 (2d Cir.1991) (citations omitted). The statistical significance of standard deviation analysis depends, of course, upon the size of the sample utilized.

. Further, Peightal argues that the undifferentiated population statistics fail to account for the large number of Hispanics who would not qualify to take the exam because of their failure-to meet the high school diploma requirements. This argument was addressed by the district ■ court when it found that no differentiated population statistics exist based upon the level of education of minorities. Because Peightal failed to present his own statistical data, the district court properly found Peightal's speculation regarding the impact of educational statistics to be unconvincing.

. Had Peightal introduced evidence of more precise population statistics, that would have been significant. Consonant with the teaching of Freeman v. Pitts, - U.S. -, -, -, 112 S.Ct. 1430, 1437-38, 1447-48, 118 L.Ed.2d 108 (1992), knowledge of demographic shifts is important in differentiating between patterns of mi *1557 nority representation in state organizations which represent state action, and are thus constitutionally cognizable, from private choices, which generally do not raise such concerns.

. The precise formulation of these factors chosen for consideration under the second prong of the strict scrutiny analysis will vary with the particular circumstances of the case.

. This discrepancy between minorities and non-minorities flowed from flaws in the testing procedure. In 1982, the Fire Department agreed, pursuant to a settlement agreement, to conduct a validation study of the firefighter test. A study conducted from 1983 to 1984 by the Human Services Institute on behalf of Metro Dade confirmed a number of deficiencies in the test taken by Peightal and other applicants in 1983, deficiencies of which the Fire Department was already on notice in 1983. First, the study concluded that the firefighter test was not a valid predictor of subsequent job performance. High scores simply did not translate into high performance evaluations. Second, the study revealed that the written test had an adverse impact upon minorities. Specifically, 85% of white males taking the examination passed, as compared with 23% of Hispanic males. The passing rate for black males was 56%, and the female passing rate was 42%. Given the test's unreliability and its adverse impact on minorities, we cannot say that Metro Dade lacked justification in departing from the rigid rank-ordering produced by the test results.

. Peightal specifically argued to the district court in the first proceeding that the imposition of a quota violated Title VII. Plaintiff's Trial Memorandum at 16. The district court specifically made a contrary finding of fact, determining that the Plan imposed goals and not quotas. The hiring goals in 1983 were not rigidly applied, as is evident from the presence of numerous minority applicants who did not receive employment offers, despite the Department’s failure to meet its goals. The Fire Department had anticipated hiring 29 Hispanic males, eight black females, and eight Hispanic females. The actual class of recruits consisted of 24 Hispanic males, five black females, and four Hispanic females. Further, in significant respects, the hiring goals utilized here are similar to those approved in Cone Corp. v. Hillsborough County, 908 F.2d 908, 916-17 (11th Cir.), cert. denied, 498 U.S. 983, 111 S.Ct. 516, 112 L.Ed.2d 528 (1990). In both cases, the hiring goals utilized possessed ample flexibility, in stark contrast to the mandatory 30% set-aside invalidated in Croson. For example, in the instant case, goals were established for each hiring year. Similarly, in Cone, hiring goals were established for each project. Id. at 917.

. Peightal does not challenge the definition of black or of female utilized by the Department.

. Peightal relies upon two district court cases to assert that national origin did not suffice to qualify an individual as Hispanic for purposes of challenging discriminatory employment practices. See Nieves v. Metropolitan Dade County, 598 F.Supp. 955 (S.D.Fla.1984); Budinsky v. Corning Glass Works, 425 F.Supp. 786 (W.D.Pa.1977). Peightal argues that in Nieves the court examined racial and physical distinctions in assessing Nieves' claim of Hispanic origin. In Nieves, the court questioned whether the plaintiff had satisfied his prima facie burden of demonstrating that he was a member of the protected class of Hispanics because Nieves failed to list himself as Hispanic on his employment and failed to notify any personnel in County management of his status. 598 F.Supp. at 962. Thus, the district court turned to outward indicia of Nieves' Hispanic origin to determine whether the supervisor had other information from which he may have concluded that Nieves was Hispanic, despite Nieves' failure to claim such status. Id. In any event, the district court assumed that Nieves did qualify as Hispanic, despite the absence of any reference to his Hispanic origin and absent Spanish language abilities. Id. at 960 and 962. The court then concluded that Nieves did not receive the promotion because he was not the best qualified candidate. Id. Thus, the opinion accepted Nieves' self-identification as a Hispanic at trial, without any corroboration from racial or physical distinctions, and despite the contradiction with Nieves' earlier self-identification on the employment form.

Peightal urges that Budinsky supports the proposition that more than national origin must attach before "Hispanics" would be accorded the classification of minority. 425 F.Supp. at 788. Budinsky held that the Civil Rights Act of 1870, 42 U.S.C. § 1981, did not encompass claims of employment discrimination raised by an individual solely on the basis of Slavic national origin. Id. at 787. In dicta, the court noted that section 1981’s coverage had been extended to Hispanic persons in Miranda v. Clothing Workers, Local 208, 10 FEP Cases 557, 1974 WL 221 (D.N.J.1971) because of racial misperceptions:

Hispanic persons and Indians, like blacks, have been traditional victims of group discrimination, and however inaccurately or stupidly, *1561 are frequently and even commonly subject to a racial’ identification as 'non-whites.' There is accordingly both a practical need and a logical reason to extend § 1981's proscription against exclusively ‘racial’ employment discrimination to these groups of potential discriminatees.

425 F.Supp. at 788. Title VII embraces a broader scope than that of § 1981. That both have been construed to protect Hispanics bolsters the conclusion that prejudice against a group identified as Hispanics has existed which transcends the bounds of both race and national origin. To suggest that the definition adopted by the EEOC is flawed because it may accommodate one set of prejudices but not another illustrates the faulty reasoning attendant to the development of the prejudice, not to its definition. In any event, the EEOC definition’s supplementation by guidelines administered by affirmative action officers ensured that a definition created to protect those victimized by prior discrimination was not over-inclusively applied, as found by the district court.

. Peightal has not alleged that he is a member of one of the culturally and linguistically identifiable groups which he names.

. The most troubling facet of defining a group who have been the victims of prior discrimination is the derivation of the definition itself. As noted by the Tenth Circuit: “Prejudice is as irrational as is the selection of groups against Whom it is directed. It is thus a matter of practice or attitude in the community, it is usage or image based on all the mistaken concepts of race.' ” Manzanares v. Safeway Stores, Inc., 593 F.2d 968, 971 (10th Cir.1979). Because of the irrationality of the definitional process underlying social stereotypes, it is equally vexatious to develop a definition, or to criticize one that has been developed, without resort to the same stereotypes which application of the definition seeks to eradicate. Despite the difficulties, we conclude that the EEOC definition adopted by the district court, narrowed as it is by the requirement of strong visible indications of cultural and linguistic identification with the group, is consistent with the case law and satisfies the narrow tailoring prong.

. In Steelworkers v. Weber, 443 U.S. 193, 208, 99 S.Ct. 2721, 2729, 61 L.Ed.2d 480 (1979), the Court approved the challenged hiring program in part because the program did not "unnecessarily trammel the interest of the white employees.” Because Weber did not involve an equal protection challenge, it is not directly applicable to this case.

. The passing rate for black males was 56%, and the female passing rate was 42%.