Single-Item Versus Multiple-Item Measurement Scales: An Empirical Comparison

Abstract

Common methods variance often is a problem with psychological measures that require respondent self-reports of attitudes, beliefs, perceptions, and the like. The present study examined this problem by comparing multiple-item, Likert-type measures of psychological constructs to single-item, non-Likert-type measures of the same constructs. Using confirmatory factor analysis, the alternative forms were compared on criteria of methods variance and construct validity. Neither method appeared to be empirically better than the other. Unusual situations in which well-developed single-item measures might be appropriate are discussed.