On measuring segregation in samples with small units

William J. Carrington, Kenneth R. Troske

Research output: Contribution to journalArticlepeer-review

101 Scopus citations


Standard indexes of segregation measure a sample’s distance from evenness, which occurs when each sample unit (e.g., an occupation) has the population share of both the minority and majority groups. We show that random allocation of individuals to units generates substantial unevenness among small units and hence that standard segregation indexes reflect random allocation as well as systematic group segregation. We then modify two popular indexes so that they measure deviations from random allocation rather than deviations from evenness. An empirical example suggests that these modified indexes provide improved measures of the systematic component of group segregation.

Original languageEnglish
Pages (from-to)402-409
Number of pages8
JournalJournal of Business and Economic Statistics
Issue number4
StatePublished - Oct 1997


  • Random allocation
  • Segregation
  • Segregation index

ASJC Scopus subject areas

  • Statistics and Probability
  • Social Sciences (miscellaneous)
  • Economics and Econometrics
  • Statistics, Probability and Uncertainty


Dive into the research topics of 'On measuring segregation in samples with small units'. Together they form a unique fingerprint.

Cite this