Thursday, November 5, 2009

Probabilities and Symmetrical Access to Information in Harassment Cases

EEOC Image Before I got sidetracked by David Letterman and other current affairs, I was blogging about the application of the Teamsters burden-shifting framework in pattern or practice disparate treatment cases.  Today, I want to revisit the question I posed in my first blog post:  Under what circumstances, and under which antidiscrimination statutes, can the Teamsters method of proof be used by plaintiffs?

As Mike Zimmer noted in a comment to that post, perhaps nomenclature should not be important in determining whether the Teamsters method of proof can be employed in any particular type of case.  Maybe the Teamsters method of proof should just be one of several alternative ways for any plaintiff to establish an inference of disparate treatment discrimination, regardless of the statute or the type of discriminatory conduct at issue.  Perhaps it should not be dispositive that the ADEA does not contain the words “pattern or practice.”  I would tend to agree.  At least, it seems to me, we should undertake a critical examination of the underlying purposes and justifications for the Teamsters burden-shifting proof structure when determining whether or not it should be available to plaintiffs in a particular type of disparate treatment discrimination case. 

When we conduct this sort of critical examination, we can begin to see why systemic harassment claims differ from other types of systemic disparate treatment claims, and why the Teamsters-style “pattern or practice” burden-shifting advocated by the EEOC is inappropriate in systemic harassment cases.  To demonstrate this point, I return to the Court’s reasoning in the Teamsters opinion.  The Teamsters Court held that the McDonnell-Douglas formula was not the only way for a plaintiff to establish a prima facie case of disparate treatment discrimination, and therefore not the only way to create an inference (or a rebuttable presumption) of intentional discrimination.  Int’l B’hood of Teamsters v. U.S., 431 U.S. 324, 358-59 (1977).  (Continued after the jump).

An alternative way to create an inference of discrimination, the Teamsters Court reasoned, was exemplified by the Court’s earlier holding in Franks v. Bowman Transportation Co., 424 U.S. 747 (1976).  In Franks, the plaintiffs alleged “a broad-based policy of employment discrimination.”  If plaintiffs could carry the burden of proving that allegation, then “there were reasonable grounds to infer that individual hiring decisions were made in pursuit of the discriminatory policy and to require the employer to come forth with evidence dispelling that inference.”  Int’l B’hood of Teamsters v. U.S., 431 U.S. 324, 359 (1977).  The Teamsters Court went one step further, and held that statistical evidence could be used to establish a prima facie case and justify an inference of discrimination as to individual claimants.  

Importantly, the Teamsters Court explained the underlying justification for this “pattern or practice” type of burden-shift as follows: 

The holding in Franks that proof of a discriminatory pattern and practice creates a rebuttable presumption in favor of individual relief is consistent with the manner in which presumptions are created generally.  Presumptions shifting the burden of proof are often created to reflect judicial evaluations of probabilities and to conform with a party’s superior access to the proof.  See C. McCormick, Law of Evidence ss 337, 343 (2d ed. 1972); James, Burdens of Proof, 47 Va. L. Rev. 51, 61 (1961).  See also Keyes v. School Dist. No. 1, 413 U.S. 189, 208-209, 93 S.Ct. 2686, 2697, 37 L.Ed.2d 548.  These factors were present in Franks.  Although the prima facie case did not conclusively demonstrate that all of the employer’s decisions were part of the proved discriminatory pattern and practice, it did create a greater likelihood that any single decision was a component of the overall pattern.  Moreover, the finding of a pattern or practice changed the position of the employer to that of a proved wrongdoer.  Finally, the employer was in the best position to show why any individual employee was denied an employment opportunity.  Insofar as the reasons related to available vacancies or the employer’s evaluation of the applicant’s qualifications, the company’s records were the most relevant items of proof.  If the refusal to hire was based on other factors, the employer and its agents knew best what those factors were and the extent to which they influenced the decision-making process.


Teamsters, 431 U.S. at 359, n.45.

 The Court thus recognized two important factors justifying a burden-shift:  (1) judicial evaluations of probabilities, and (2) the parties’ relative access to the proof. 

The first factor – evaluations of probabilities – is a function of the evidence offered at Phase I of a pattern or practice case.  In the typical systemic discrimination case involving hiring, firing, or promotions, statistical evidence of a disparity offered in Phase I can be a good indicator of a relatively high probability that the adverse employment action suffered by any particular individual claimant was the product of discrimination.  Likewise, where an express policy is shown to be facially discriminatory (like the airline's discriminatory “bumping” policy at issue in Trans World Airlines v. Thurston, 469 U.S. 111 (1985)), the probability that the adverse employment action suffered by any particular claimant affected by the policy was the product of unlawful discrimination can be presumed to be high. 

Systemic harassment cases are different.  In a prior post, I discussed the typical lack of meaningful statistical evidence in systemic harassment cases.  If only anecdotal evidence is offered by the EEOC in Phase I of a systemic harassment case, then the EEOC has not shown the same indicator of a high probability of actionable harassment as to the individual claimants that did not testify in Phase I.  In the comments to that post, Jerry Carbo suggested that perhaps the EEOC is offering qualitative research evidence in Phase I of these cases, and not just “anecdotal evidence.”  In my estimation, “qualitative research” is an overly generous characterization of the evidence that the EEOC typically offers in Phase I of a systemic harassment case like Corrections Corp.  Based on my research of systemic harassment cases and my own experience in practice, the EEOC is really just offering the anecdotal evidence of the most sympathetic claimants, without offering a representative sampling of all the results gathered by the EEOC’s investigation, without any expert scientific interpretation of the witnesses’ accounts as qualitative research results, and without employing any recognized validation methods for qualitative research.  In other words, the EEOC is acting purely as an advocate in deciding which witnesses to put on the stand in Phase I, and is not offering scientifically validated qualitative research.  In any event, in the Corrections Corp. case the EEOC itself referred to the evidence it intended to offer in Phase I as “anecdotal.”  It never suggested that it was offering evidence of scientific qualitative research results in order to justify a burden-shift.  Thus, unlike a case involving an expressly discriminatory policy or statistical evidence of racial disparity in hiring, firing, or promotions, the probability factor may not be present (or at least not as strong) in systemic harassment cases.   

As to the second factor – the parties’ relative access to the proof – there is also a difference between ordinary systemic discrimination cases and systemic harassment cases.  In the ordinary disparate treatment discrimination case, the ultimate issue is the employer’s intent in deciding to fire, or to not hire, or to not promote the individual claimant.  Direct evidence of an employer’s actual motive – i.e., what the decision-maker was actually thinking – in making such an adverse employment decision is almost always impossible for a plaintiff to obtain.  As the Teamsters court noted in the passage quoted above, if the reasons for the adverse action were legitimate and non-discriminatory, then the employer and its agents would be in the best position to know those reasons and the extent to which they affected the decision.   

But again, harassment is different.  Evidence of the key issues in harassment cases is generally just as available to individual plaintiffs as it is to employers.  In fact, one element of an individual claim for harassment is a showing that the plaintiff subjectively perceived his or her work environment to be hostile.  The individual plaintiff must have been exposed to the actions, words, or conditions that constitute the hostile work environment in order for them to form the basis for the harassment claim.  The plaintiff knows what he or she was subjected to, and can testify to it.  Further, the plaintiff generally knows who witnessed the incidents – or is at least as likely to know the identities of the witnesses as is the employer.  In short, harassment is necessarily a “public” event.  It happens in the presence of the plaintiff, and not in secret in the supervisor’s office or inside the decision-maker’s mind.  The plaintiff is just as likely (if not more likely) to have access to evidence of the alleged harassment as the employer and its agents.  Harassment cases often boil down to credibility determinations, but that alone does not justify the imposition of a rebuttable presumption that a violation occurred. 

This fundamental difference is reflected in the way that individual harassment claims are resolved, as compared to other types of individual disparate treatment claims.  The McDonnnell-Douglas burden-shifting formula generally will not apply in individual harassment claims - except perhaps in those cases where there is a question about whether plaintiff was harassed because of his or her sex (or race, etc.), and not because of some other reason, like having a generally difficult supervisor that harasses and bullies everyone.  In the ordinary individual harassment case, there is no creation of an inference or a rebuttable presumption that the plaintiff was subjected to a work environment that was so hostile as to constitute a Title VII violation.  Rather, the plaintiff has the burden of establishing that his or her work environment was both objectively and subjectively hostile.  So why should the individual claimants in systemic harassment claims be given the benefit of a rebuttable presumption or an inference that they were the victims of a work environment that was both objectively and subjectively hostile, based on only selective anecdotal evidence that other claimants experienced a hostile work environment?  Perhaps a sufficient level of anecdotal proof in Phase I might be able justify a presumption that any harassment suffered by the other individual claimants was because of sex or race, and not because of some other reason (after all, evidence of this motive question might be only in the harasser's mind), but that is not the inference that the EEOC is advocating.  In systemic harassment cases like Corrections Corp., the EEOC is asking for a presumption, based on the anecdotal evidence offered in Phase I, that all of the individual claimants were subjected to an actionable hostile work environment in violation of Title VII.  As I discussed in my prior post, the district courts remain divided on whether drawing that type of inference is appropriate. 

Putting the “pattern or practice” nomenclature aside, we can and should critically examine the underlying purposes and justifications for burden-shifting to determine when the Teamsters method of proof should be available to plaintiffs in cases alleging systemic discrimination.  By examining (1) the relative probability of a violation as to any individual claimant (in light of statistical, anecdotal, or other evidence offered at Phase I), and (2) the parties’ relative access to evidence, we can see that systemic harassment claims are different from other types of systemic disparate treatment claims.  Unlike Teamsters, the factors justifying a burden-shift are absent, or at least less compelling, in systemic harassment cases.  

-Jason Bent

| Permalink

TrackBack URL for this entry:

Listed below are links to weblogs that reference Probabilities and Symmetrical Access to Information in Harassment Cases:


Post a comment