Condom use as a dependent variable: a brief commentary about classification of inconsistent users.

Transformation of nonnormally distributed measures of condom use frequency can be problematic for researchers. Distributions are typically dichotomized. We used data collected from 483 university undergraduates in an anonymous, cross-sectional survey to illustrate the value of a screening analysis before dichotomization. Inconsistent condom users were compared to consistent users with respect to 14 measures. Subsequently, inconsistent users were compared to those who never used condoms with regard to the same 14 measures. Findings suggest that a screening analysis is a potentially important aspect of analyzing distributions that assess frequency of condom use.