Opinion
No. CV-12-615-PHX-GMS
08-08-2012
Equal Opportunity Employment Commission, Plaintiff, v. McLane Company, Inc., Defendant.
ORDER
Pending before the Court is Petitioner EEOC's Motion to Seal (Doc. 29) its in camera submission and to compel Respondent McLane Company, Inc. to provide certain pedigree information regarding its current and former employees. For the reasons stated below, Respondent will not be required to produce further pedigree information.
BACKGROUND
This case involves McLane's use of a physical capacity exam ("PCE") administered by Industrial Physical Capacity Services, Inc. ("IPCS"). The factual background is set out in detail in the Court's previous order detailing the information which Respondent must provide to Petitioner pursuant to Petitioner's investigation of whether, by using the PCE, Respondent is engaging in an employment practice that has a disparate impact on applicants and employees over 40. (Doc. 25). In that order, the Court found that Respondent must provide Petitioner with test results and any adverse employment actions taken against those who failed the test. It found that Respondent need not provide Petitioner with names, contact information, social security numbers, and other pedigree information (other than age) of those who took the test.
Petitioner has now submitted, with the Court's permission, an in camera study analyzing the data regarding those who took the PCE test. (Doc. 30). Petitioner has filed briefing arguing that because the study demonstrates that the test has a disparate impact, the disclosure of names, social security numbers, and contact information of test-takers is necessary to complete their investigation. (Doc. 32).
DISCUSSION
I. Legal Standard
A. Deliberative Privilege
To qualify for the deliberative process privilege, "a document must be both (1) 'predecisional' or antecedent to the adoption of agency policy, and (2) 'deliberative,' meaning 'it must actually be related to the process by which policies are formed.'" Nat'l Wildlife Fed'n v. United States Forest Serv., 861 F.2d 1114, 1117 (9th Cir. 1988) (quoting Jordan v. United States Dep't of Justice, 591 F.2d 753, 774 (D.C. Cir. 1978). "The exercise of judgment in the formulation of a factual statement is not sufficient to lift it to the level of deliberation." Resolution Trust Corp. v. Diamond, 137 F.R.D. 634, 641 (S.D.N.Y. 1991).
B. Disparate Impact
Traditionally, disparate impact claims involve a three-step evaluation. The first step, and the only one relevant here, is establishing whether "an employer uses a 'particular employment practice that causes disparate impact on the basis of race, color, religion, sex, or national origin.'" Ricci v. DeStefano, 129 S. Ct. 2658, 2673 (2009) (quoting 42 U.S.C. § 2000e-2(k)(1)(A)(I)). Although the disparate impact test was developed in the Title VII context, the ADEA permits disparate impact claims based on age. See Smith v. City of Jackson, Miss. , 544 U.S. 228, 240 (2005). Typically, the EEOC relies upon the so-called Uniform Guidelines in the Code of Federal Regulations when determining whether a particular employment practice has a disparate impact. See 29 C.F.R. § 1607.2(C) ("These guidelines apply only to selection procedures which are used as a basis for making employment decisions."). The Uniform Guidelines, as codified in the Code of Federal Regulations, state that "[a] selection rate for any race, sex, or ethnic group which is less than four-fifths (4/5) (or eighty percent) of the rate for the group with the highest rate will generally be regarded by the Federal enforcement agencies as evidence of adverse impact, while a greater than four-fifths rate will generally not be regarded by Federal enforcement agencies as evidence of adverse impact." 29 C.F.R. § 1607.4(D). The Ninth Circuit has typically adopted this so-called "four-fifths" rule when evaluating whether a particular employment action has a disparate impact. See, e.g., Eldredge v. Carpenters 46 Northern California Ctys. Joint Apprenticeship & Training Comm., 833 F.2d 1334, 1339 (9th Cir. 1987) ("EEOC guidelines, as interpretations of Title VII by the enforcing agency, are entitled to great deference.") (internal citations omitted). Its embrace of the rule, however, is not absolute, and it has emphasized that "The Uniform Guidelines are not legally binding." Clady v. Los Angeles Cty., 770 F.2d 1421, 1328 (9th Cir. 1985).
The four-fifths rule is applied to the rates at which employees are selected for a positive employment action. In Bouman v. Block, 940 F.2d 1211, 1225 (9th Cir. 1991), for example, the Ninth Circuit considered an examination that women had passed at a 13% rate and men had passed at a 19% rate to determine that the test violated the 80-percent rule because the women's pass rate was "only 66 percent of the men's pass rate." Had it instead considered failure rates, it would have found that men failed the test 81% of the time, or 93% of women's 87% fail rate, and the test would not have violated the rule.
The Supreme Court has suggested that relying on statistical measurements other than the eighty percent rule may sometimes be appropriate. See, e.g., Castaneda v. Partida, 430 U.S. 482, 496 n.17 (1977) (noting in dicta in a voting discrimination case that "[a]s a general rule for such large samples, if the difference between the expected value and the observed number is greater than two or three standard deviations, then the hypothesis that [the challenged action does not have a disparate impact] would be suspect to a social scientist," but finding discrimination in the case at hand when observed data reflected a difference "of approximately 29 standard deviations."). The Ninth Circuit has not, however, to the Court's knowledge, found disparate impact in any test or employment action when the action did not violate the eighty percent rule. See, e.g., Stout v. Potter, 276 F.3d 1118, 1124 (9th Cir. 2002) ("This means that the rate of selection for women was 81 percent of the rate of interview for men, again demonstrating that no disparate impact was shown."). The Supreme Court has further cautioned that although disparate impact should be evaluated on a case-by-case basis, when relying on statistics to show disparate impact, the "statistical disparities must be sufficiently substantial that they raise such an inference of causation." Watson v. Fort Worth Bank and Trust, 487 U.S. 977, 995 (1988). Modern statistics can demonstrate that, provided a large enough sample size, a discrepancy of less than 4/5 is nevertheless statistically significant. As the Supreme Court has noted in another context, however, "[t]he mathematical conclusion that the disparity between these two figures is 'statistically significant' does not, however, require an a priori finding that these deviations are 'legally significant.'" U.S. v. Test, 550 F.2d 577, 583 (1976).
II. Discussion
The Court has serious doubt as to whether the in camera submission is properly subject to the deliberative process privilege. The Court accepts the EEOC's statement that the analysis represents a preliminary draft, and it is therefore "predecisional." Nat'l Wildlife, 861 F.2d at 1117. Nevertheless, the analysis relates specifically to the facts of the instant case, and involves an investigation as to whether EEOC will proceed; it does not appear on its face to be "related to the process by which polices are formulated" at the EEOC. Id. Nevertheless, for what it's worth, as discussed in detail below, after reviewing the document the Court does not believe that the analysis justifies expanding the scope of the subpoena. Therefore, this Court declines to rule on the motion to compel McLane to provide pedigree information, unless the EEOC provides McLane with the study it has provided to the court in camera. As Petitioner notes, "the only facts contained in the documents are facts provided by Respondent from its data base [sic]." (Doc. 29 at 11). Those underlying facts have already been disclosed.
The Court previously limited the scope of Petitioner's subpoena because Petitioners are investigating whether the PCE has an adverse impact on older test-takers, and the names, addresses, and social security numbers of test-takers is not related to a disparate impact investigation when "[t]he stated definition of the investigation does not suggest that the E.E.O.C. is investigating any acts of discrimination against particular individuals." (Doc. 25). Petitioner argues that because its in camera submission "statistically supports the agency's continued investigation," it now requires the names, social security numbers, and contact information of individual test-takers. (Doc. 29 at 2). It makes three arguments in favor of this conclusion: 1) it argues that it must contact individual test-takers to determine which adverse action, if any, was taken after the test-taker failed the test, 2) social security numbers will allow it to link various data sets and eliminate duplicative data, and 3) recent case law suggests it may be required to contact class members prior to filing a suit. (Doc.29). At least as an initial matter, the Court finds Petitioner's statistical analysis unpersuasive.
Petitioner argues that the "related to" standard of the ADEA is less stringent than the "relevance" standard of Title VII. (Doc. 29). The names, social security numbers, and contact information of individual test-takers are neither related to nor relevant to an investigation as to whether a test, in the aggregate, has a disparate impact on test-takers.
Out of 3,019 testers aged 40 or older when they took the test, 2,319 passed, for a pass rate of 77%. Of the 9,072 individuals under the age of 40 who took the test, 7,726 passed, for a pass rate of 85%. The pass rate for those over forty was 90.5% of the pass rate for those under forty; the test therefore does not violate the four-fifths rule. The Court recognizes that the sample sizes here are large, and that with large sample sizes, statisticians can state with greater reliability whether disparities are statistically significant. It also notes, however, that "[t]he mathematical conclusion that the disparity between these two figures is 'statistically significant' does not, however, require an a priori finding that these deviations are 'legally significant.'" U.S. v. Test, 550 F.2d 577, 583 (1976).
The four-fifths rule has long been generally accepted by the Ninth Circuit. In another context, the circuit has been skeptical of over-reliance on statistical analysis, quoting Mark Twain's observation that in such circumstances, "the remark attributed to Disraeli would often apply with justice and force: 'There are three kinds of lies: lies, damned lies and statistics.'" Fishermen's Finest, Inc. v. Locke, 593 F.3d 886, 901 (9th Cir. 2010) (quoting Mark Twain, Chapters from My Autobiography, 185 North American Review, No. DCXVIII., July 5, 1907, at 465, 471).
There is no evidence that Disraeli ever wrote the quote Twain attributed to him.
--------
The Court is further aware that Petitioner has not filed suit against Respondent, but is merely conducting an investigation. Nevertheless, its stated reasons for requiring additional data rely upon finding that a test that does not violate the four-fifths rule nevertheless has an adverse impact on protected employees. There is no existing authority in the Ninth Circuit to so find. Petitioner can continue its investigation into the impact of the test itself without interviewing individual test-takers, and has not produced evidence to suggest that the test has an adverse impact such that it must conduct such interviews.
IT IS THEREFORE ORDERED:
1. Petitioner may withdraw its motion or inform the Court that it has provided McLane with its in camera submission.
2. The scope of the subpoena will not be broadened.
__________________
G. Murray Snow
United States District Judge