UC Berkeley

Although single items can save time and burden in psychology research, concerns about their reliability have made the use of multiple-item measures the default standard practice. Although single items cannot demonstrate internal reliability, their criterion validity can be compared with multiple-item measures. Using ecological momentary assessment data, we evaluated repeated measures correlations and constructed multilevel cross-lagged models to assess concurrent and predictive validity of single- and multiple-item measures. Correlations between the single- and multiple-item measures ranged from .24 to .61. In 27 of 29 unique single-item predictor models, single items demonstrated significant predictive validity, and in one of eight sets of comparisons, a single-item predictor exhibited a larger effect size than its multiple-item counterpart. Although multiple-item measures generally performed better than single items, the added benefit of multiple items was modest in most cases. The present data provide support for the use of single-item measures in intensive longitudinal designs.