Choosing Between Parametric and Nonparametric Tests

A fundamental analysis decision confronting researchers in psychology and education is the choice between parametric and nonparametric tests. Despite the statistical and substantive implications of this important decision, many researchers unerringly employ parametric tests and thus ignore the advantages of their nonparametric counterparts. One justification for this behavior has been the absence of guidelines for choosing between these procedures. A second has been the lack of a comprehensive nonparametric test that is computationally manageable. In this article, the author discusses several statistical and substantive criteria that can be used to choose between parametric and nonparametric tests. A non‐parametric test capable of testing a number of statistical hypotheses using existing computer packages is also presented. Recommendations are made encouraging researchers to routinely use nonparametric tests in their data analytic work.