Ethnic diversity as a multilevel construct: the combined effects of dissimilarity, group diversity, and societal status on learning performance in work groups

Felix C. Brodbeck, Yves R.F. Guillaume, Nick J. Lee

Research output: Contribution to journalArticlepeer-review

Abstract

The authors present a model of the multilevel effects of diversity on individual learning performance in work groups. For ethnically diverse work groups, the model predicts that group diversity elicits either positive or negative effects on individual learning performance, depending on whether a focal individual’s ethnic dissimilarity from other group members is high or low. By further considering the societal status of an individual’s ethnic origin within society (Anglo versus non-Anglo for our U.K. context), the authors hypothesize that the model’s predictions hold more strongly for non-Anglo group members than for Anglo group members. We test this model with data from 412 individuals working on a 24-week business simulation in 87 four- to seven-person groups with varying degrees of ethnic diversity. Two of the three hypotheses derived from the model received full support and one hypothesis received partial support. Implications for theory development, methods, and practice in applied group diversity research are discussed.
Original languageEnglish
Pages (from-to)1198-1218
Number of pages21
JournalJournal of Cross-Cultural Psychology
Volume42
Issue number7
Early online date30 Sept 2010
DOIs
Publication statusPublished - Oct 2011

Keywords

  • ethnic dissimilarity
  • group diversity
  • relational demographics
  • societal status
  • information-elaboration
  • small groups
  • multilevel models
  • individual learning
  • performance
  • minority/majority influence
  • business game simulation

Fingerprint

Dive into the research topics of 'Ethnic diversity as a multilevel construct: the combined effects of dissimilarity, group diversity, and societal status on learning performance in work groups'. Together they form a unique fingerprint.

Cite this