## Facts

- Female full-time, academic-rank faculty in SUNY–New Paltz’s Division of Liberal Arts and Sciences challenged employment practices during academic years 1973–1984.
- Plaintiffs brought a Title VII class action alleging sex discrimination in (1) initial academic rank at hire, (2) promotion to higher rank, and (3) salary.
- Plaintiffs relied primarily on statistical proof, including multiple regression analyses intended to control for legitimate salary determinants and to show a residual pay gap for women, asserted to exceed two standard deviations.
- Defendants argued plaintiffs’ regression models omitted important non-discriminatory variables and did not reflect how rank and salary decisions were made, thereby overstating any sex-based disparity.
- The bench trial lasted roughly nine months and included competing expert statistical evidence and extensive anecdotal testimony about individual employment histories and departmental practices.
- The district court entered judgment for defendants on all Title VII claims (class-wide and individual), while a plaintiff-intervenor prevailed on a separate Equal Pay Act claim; that Equal Pay Act ruling was not appealed.

## Issues

1. Whether the district court clearly erred in finding plaintiffs failed to prove Title VII sex discrimination in initial rank, promotion, and salary.
2. Whether the district court properly evaluated and discounted plaintiffs’ regression analyses and credited defendants’ statistical critiques and other rebuttal evidence.

## Decision

- The Second Circuit affirmed the judgment for defendants on all Title VII claims.
- The court held the district court’s factual findings, including its assessment of competing statistical models, were not clearly erroneous after a lengthy bench trial.
- The unappealed Equal Pay Act judgment in favor of the intervenor remained in effect but did not control the Title VII class claims.

## Legal Principles

- After a bench trial, appellate review of factual findings in Title VII cases is governed by the clearly erroneous standard; an appellate court will not reweigh evidence or retry disputes between experts.
- Statistical proof, including multiple regression analysis, may support an inference of discrimination, but its probative value depends on appropriate model specification and fit with the employer’s decision-making process.
- A factfinder may discount or reject regression results where key explanatory variables are omitted or the methodology does not reflect relevant, non-discriminatory determinants of pay or rank.
- Plaintiffs bear the ultimate burden of persuasion on disparate treatment claims; defendants may rebut statistical inferences by showing methodological flaws and by presenting legitimate explanations supported by the record.
- Anecdotal evidence may corroborate or undermine statistical proof; the absence of consistent corroboration can support a finding of no class-wide discriminatory pattern.

## Conclusion

The Second Circuit upheld the district court’s rejection of class-wide Title VII claims by deferring to the trial court’s credibility and methodology determinations, concluding it was not clear error to find plaintiffs’ regression evidence insufficient and the overall record inadequate to prove systemic sex discrimination in rank, promotion, or salary.


Ottaviani v. State Univ. of N.Y. at New Paltz, 875 F.2d 365 (2d Cir. 1989)