Assessing Between- and Within-Person Reliabilities of Items and Scale for Daily Procrastination: A Multilevel and Dynamic Approach.

Intensive longitudinal data (ILD) has been collected to capture the dynamic fluctuations of procrastination; however, researchers have typically measured daily procrastination by modifying trait measures (e.g., adding a time reference "today") without adequately testing their reliabilities. The main purpose of this study was to use an advanced approach, dynamic structural equation modeling, to assess the between- and within-person reliabilities of a widely used six-item measure of daily procrastination. A total of 252 participants completed retrospective measures of various types of trait procrastination and daily measures of procrastination over 34 consecutive days. The results showed that the entire scale for daily procrastination and five of its six items had high between- and within-person reliabilities, but one item had much lower reliabilities, suggesting that this item may be inappropriate in everyday contexts. Furthermore, we found moderate to strong associations between the latent trait factor of procrastination and trait measures of procrastination. In addition, we identified substantial between-person variation in person-specific reliabilities and explored its relevant factors. Overall, this study assessed the reliabilities of a daily measure of procrastination, which facilitated future studies to obtain more reliable and consistent results and to better estimate the reliability of ILD.