Psychometric Considerations for Conducting Observations Using Time-Sampling Procedures

Abstract

Time-sampling techniques are popular among school practitioners and researchers alike when conducting observations of student behavior. Despite their popularity, the psychometric properties of such observation data have not received much attention, likely given that both procedures and targets vary widely across individual coding systems. A systematic literature review was therefore conducted to provide insight into how various procedural aspects may influence the psychometric properties of time-sampling data. Specifically, the influences of the number of target behaviors, time-sampling method, interval length, observation length, and number of observations on the psychometric properties of time-sampling data are explored. Both implications for practice and future research are discussed.