Employment Litigation · Primary-source case analysis

Hazelwood: Employment Statistics Needed the Right Labor Market and Time Period

By Aneeta Mishra and Timothy Chan

Published August 14, 2026 · Updated October 29, 2026

Decision: Hazelwood Sch. Dist. v. United States, 433 U.S. 299 (1977).

Hazelwood School District v. United States was decided by the Supreme Court on June 27, 1977. It reviewed a federal pattern-or-practice case alleging racial discrimination in teacher hiring after Title VII became applicable to public employers in March 1972.

Qualified teachers were the relevant comparison pool

The proper comparison was between the racial composition of Hazelwood’s teaching staff and the qualified public-school teacher population in the relevant labor market, not the racial composition of teachers and students.

The geographic market could change the inference

The parties disputed whether the labor market included both St. Louis County and the City of St. Louis or only the county. The different percentages could materially strengthen or weaken the statistical inference and required trial-court findings.

Post-Act hiring data mattered

Overall workforce disparities could reflect hiring before Title VII covered the district. Statistics showing whom Hazelwood hired and who applied after March 1972 were relevant to whether the observed pattern resulted from unlawful covered-period decisions.

The case returned for factfinding

The Court vacated and remanded rather than deciding ultimate liability. Courts had to assess the proper market, qualified pool, post-Act hiring, applicant data, and the employer’s rebuttal in the full record.

Justice Stevens dissented because he considered the government’s evidence sufficient on the existing record. Even using the district’s narrower proposed labor market, Hazelwood hired 15 Black teachers where random selection would have suggested about 23. He read that disparity together with the history, subjective hiring process and individual discrimination evidence, and found the district had failed to rebut the government’s case. Justice Brennan joined the majority’s remand but agreed that the statistical difference did not itself rebut the government’s showing; he would allow a closer examination of applicant data. The majority required further trial-court findings. Hazelwood, 433 U.S. at 313–320 (separate opinions).

Key takeaways

• Match the comparison pool to the qualifications and geography of the jobs at issue.
• Separate covered-period decisions from historical workforce composition.
• Use applicant-flow data when reliable and explain missing or distorted data.
• Treat statistics as record evidence requiring contextual and methodological support.