Opinion
Nos. 117, 264, Dockets 88-7400, 88-7438.
June 1, 1989.
Appeal from the United States District Court for the Southern District of New York.
Before MESKILL, PIERCE and WINTER, Circuit Judges.
Counsel for Ms. Greenberg and amicus curiae the NAACP Legal Defense and Educational Fund, Inc. raise two arguments on this petition for rehearing. One is a legal argument, the other a factual one.
The legal argument claims that the panel erroneously held that counsel may not preliminarily review statistical data obtained in discovery to determine whether the data is sufficiently probative of discrimination to justify hiring an expert to perform sophisticated computer or other statistical analysis. The panel decision neither held nor implied that such a proposition was the law. Our decision held only that counsel, in making broad and costly discovery requests for data that is to be subjected to statistical analysis, should inform the district court of the precise use to which the data is to be put. Absent such information, a decision regarding the appropriate scope of discovery is not possible. If, for example, a preliminary review by counsel is intended before a decision to hire an expert is made, it may be possible for the adversary to reduce its costs considerably by providing the limited data necessary to that preliminary review with more elaborate discovery to follow only upon a determination that retaining an expert is appropriate. Our decision thus requires no more than candor and precision on the part of counsel seeking discovery.
We do not mean to suggest that such a discovery request should not include all the data that would be needed for a comprehensive review by an expert. Indeed, it should, because there is naturally much overlap between the data needed for a preliminary evaluation and that needed for a comprehensive review. The adversary might well find it more economical to compile the more complete data initially than to risk having to reexamine all of the same files upon a finding that a comprehensive review was appropriate. We hold only that the party seeking such data must make its intentions with regard to professional analysis clear so the adversary can select the most economical means of compliance.
The factual argument concerns whether counsel based their discovery requests upon representations as to subsequent analysis by professional statisticians and whether the analysis utilized by counsel constituted such professional analysis. We adhere to our ruling that the papers filed by plaintiff's counsel indicated that professional analysis would be made of the data requested. See 870 F.2d 926, 937-38. Indeed, the letter to the Magistrate from one of plaintiff's counsel requesting a settlement conference explicitly stated that she understood that professional analysis by an expert would be made of the data if a settlement was not reached.
We turn now to the claim that the analysis performed by one of plaintiff's counsel on the data constituted professional analysis as contemplated by the papers filed on plaintiff's behalf. Although counsel never made such a claim either in this court or in the district court — the pertinent submission stated only that a "rough analysis" by hand examination had been made without any indication that that analysis was of a professional nature or was available for examination by the court — she is now prepared to offer evidence from a statistician who has personally reviewed the hand examination performed by counsel. That evidence consists of his opinion that the hand examination constituted appropriate professional statistical, albeit preliminary, analysis indicating that it was highly unlikely that further analysis would lead to probative evidence.
One of the affidavits submitted in connection with the petition for rehearing states:
FRED THOMPSON, being of full age and duly sworn, upon his oath deposes and says that:
1. I am the Grace and Elmer Goudy Professor of Public Management and Policy at the Geo. H. Atkinson Graduate School of Management, Willamette University, Salem OR, where I teach courses in a number of subjects, including applied statistics. I hold a PhD in economics and political science from. Claremont Graduate School. I am the author or coauthor of over sixty scholarly publications, of which some fifteen are primarily quantitative in nature. I have testified as an expert witness before committees of the California State Senate and the Canadian Parliament, as well as before the Defense Secretary's Commission on Base Realignment and Closure. I have also performed statistical analysis on behalf of the National Institute of Drug Abuse and the National Highway Traffic Safety Administration.
2. I have examined the statistical analysis performed by Susan Meredith, counsel for the plaintiff, of the data produced by defendent [sic], Hilton International, in response to the plaintiff's second motion to compel discovery. While her analysis is unsophisticated, it is adequate to estimate the likelihood that further analysis would produce results that would support her client's case.
3. Ms. Meredith dichotomized the population into male and female groups and computed the mean rate of promotion and the mean annual rate of salary increase for both groups. Based upon the results of this analysis, it is highly unlikely that more refined statistical analysis (including OLS regression, GLS regression, or analysis of variance) of this data set would produce results that would show a pattern and practice of discrimination. Ms. Meredith also compared initial placement patterns for the two groups, taking account of prior experience and education. Again it is highly unlikely that more refined statistical analysis of this data set would produce results that would show a pattern and practice of discrimination.
4. It is always proper to conduct a preliminary analysis of a data set (such as the simple cross tabulations of the data performed by Ms. Meredith), both as a means of estimating the payoff to further analysis (in many cases qualitative results are sufficient) and as a basis for subsequent weighting of findings via Bayes theorem, prior to the conduct of more costly analysis. That is, of course, precisely what Ms. Meredith did. No expert in statistical decision theory would fault her approach.
Normally we would be inclined to reject this proffer as untimely. Counsel is correct, however, in noting that the documents produced in response to the discovery requests at issue and referred to at page 939 of the opinion were not filed in the district court and were not before it when it ruled on the Rule 11 motion. These documents were obtained by this court after argument of the appeal without objection by plaintiff's counsel. Although these documents may not have been essential to our holding, examination of them certainly reinforced our conclusion that the discovery requests were not made in good faith in light of the failure to retain an expert. Notwithstanding the failure to object to our request for, and examination of, these documents, we are disinclined to impose sanctions without further proceedings in circumstances in which materials not before the district court may have played a role in our decision and where counsel, although belatedly, also offers to amplify the record with other relevant matter. We therefore remand to the district court for a hearing.
On remand, the district court should determine whether the "rough analysis" performed by plaintiff's counsel was adequate by accepted professional standards to indicate that retention of an expert was not justified. Upon remand, plaintiff's counsel should promptly provide the district court with the actual analysis performed and later examined by the statistical expert (but not submitted to us with his affidavit), see note 2 supra, as well as the affidavits submitted in connection with the petition for rehearing. We will return the discovery materials obtained after argument to defendant who will provide them to the district court, if they are needed, under conditions assuring their confidentiality.
The results of the hand analysis are relevant because those results are said by counsel to be the sole cause of not retaining an expert. See op. at 933, Petition for Rehearing at 9 ("Had counsel's preliminary analysis been favorable, the data would have been sent to the experts for further analysis").
In the future, requests for costly statistical compilation useful only for professional analysis should be accompanied by reasonably precise representations as to counsel's intentions with regard to preliminary analysis and to retention of an expert. Those resisting such discovery can then be given the option of producing only that data necessary to preliminary analysis with more elaborate production to follow if the preliminary analysis indicates that more sophisticated examination would be useful. See note 1 supra.
Unmentioned in the petition for rehearing (erroneously filed in Ms. Greenberg's name) is the fact that vacating the award of Rule 11 sanctions against Ms. Greenberg's counsel reopens the issue of whether an award of attorney's fees should be made against Ms. Greenberg under 42 U.S.C. § 2000e-5(k) (1982). See op. at 940 (Rule 11 sanctions against counsel obviate need to remand for a determination whether to award fees under Section 2000e-5(k)). Without resolving the issue, we note that the case for awarding fees is non-frivolous. The demands for costly discovery were made after our decision in Martin v. Citibank, N.A., 762 F.2d 212 (2d Cir. 1985), rendered plaintiff's constructive discharge claim at best wafer thin, at worst frivolous. Without the constructive discharge claim, of course, Ms. Greenberg's damage claim was concededly only $1200. Moreover, discovery as to the thirty-one employees comparable to Ms. Greenberg had not resulted in evidence of gender discrimination. Nevertheless, there is no indication that Ms. Greenberg affirmatively participated in the aggressive pursuit of the costly discovery, compare Brown v. Fairleigh Dickinson University, 560 F.Supp. 391, 407 n. 7 (D.N.J. 1983), and Ms. Greenberg's straitened financial circumstances would make such an award inequitable and probably superfluous. Prate v. Freedman, 583 F.2d 42, 48 (2d Cir. 1978). We therefore decline to remand the issue of an award of fees under Section 2000e-5(k).
We add one note. Amicus curiae asserts that the wide-spread use of Rule 11 in civil rights cases has discouraged private counsel from taking such cases. We wish to point out that the panel opinion in the instant matter was at pains to fashion standards to protect counsel from sanctions in asserting non-frivolous Title VII claims. We trust that this opinion has made clear that the basis for our prior award of sanctions, now vacated and remanded for a factual hearing, was not the assertion of a Title VII claim. Rather, it was arguable misconduct in discovery that was in no way unique to civil rights cases, see Apex Oil Co. v. Belcher Company of New York, 855 F.2d 1009 (2d Cir. 1988) (award of sanctions under 28 U.S.C. § 1927 for failing to confer in good faith to settle discovery disputes in litigation over sale of goods), and followed our recent reaffirmation in unmistakable terms that we will respond to abusive discovery practices with appropriate sanctions. Update Art, Inc. v. Modiin Pub. Ltd., 843 F.2d 67 (2d Cir. 1988) (affirming entry of $475, 406 judgment against party refusing discovery in copyright action).
The petition for rehearing is granted. The case is remanded for proceedings consistent with this opinion.