by Art Gutman Ph.D., Professor, Florida Institute of Technology
The case is EEOC v. Paramount Staffing, Inc. (210 U.S. Lexis 49042) decided on May 17, 2010 by Judge John Phipps McCalla of the District Court of the Western District of Tennessee. The issue was Paramount’s challenge to admission by the EEOC of a statistical analysis conducted by Dr. David Sharp. The EEOC contends that Paramount, a temporary staffing company, engaged in a pattern or practice of discrimination by favoring Hispanics over Blacks in its referral of applicants for unskilled labor jobs at Technicolor’s Memphis Oaks facility between 2004 and 2006. Judge McCalla ruled that Dr. Sharp’s analysis is admissible.
Paramount used a 3-day application process in which application forms were completed on day 1. Applicants submitted two forms of ID and agreed to drug tests and a background check. Applicants attended an orientation on day 2, and were referred for temporary work on day 3. Paramount contends that the referral was on a first come basis unless the employer requested special skills. The EEOC contends that Paramount’s on-site managers intentionally excluded blacks.
Dr. Sharp’s estimated that 6,309 out of 15,667 placed applicants (40.3%) were Hispanic as compared to 13.4% in the relevant labor pool. Using an exact binomial test focusing on Hispanics alone, Dr. Sharp found the difference between actual and expected placements was 98.73 standard deviations. It was assumed that few Whites (or other groups) applied for these unskilled jobs. Therefore, over-representation of Hispanics was presumed to imply under-representation of blacks. There was a challenge as to how Dr. Sharp defined Hispanic representation (he commissioned a “surname analysis” from another expert), but Judge McCalla accepted the analysis based on a surname analysis conducted in a landmark Supreme Court ruling in Castaneda v. Partida (1977) [430 US 482] (a case in which 44% representation of Mexican-Americans in a grand jury pool was deemed discriminatory because the relevant adult Mexican-American population was over 90%).
Interestingly, both parties agreed that applicant flow data, which was lacking in this case, is the best way to compare expected versus actual employment levels. However, if such were available, the EEOC would have an easier road by claiming adverse impact. As it is, the pattern or practice charge permits the plaintiff to simply articulate its own explanation of the presumed statistical disparity (i.e., first come, first serve), forcing the defense to prove otherwise.
This case is not over. The only issue decided was to admit Dr. Sharp’s statistical analysis. We will keep you abreast of any further developments in this case.
