What Predicts Within-Variance After Correcting for Measurement Error? A Re-Analysis on Podsakoff et al. (2019)

Podsakoff et al. (2019) investigated the average percentage within-variance (PWV) in intraindividual studies in applied psychology and explored study characteristics that predict it. The PWV shows a positive bias under the condition of measurement error, which is often prevalent in intra-individual studies. Therefore, their original results are potentially biased. I complemented their data by the variables’ reliabilities and re-analyzed their data, employing the reliability-adjusted PWV as dependent and refine their state-of-the-art summary. The average reliability-adjusted PWV was .41 across all construct categories and thus substantially lower than the original estimate of .52 (i.e., 21.15% lower). My results shows that daily time referent has a larger negative association with the reliability-adjusted PWV than previously thought. Implications for researchers and reviewers, and limitations are discussed. The reliability-based adjustment of the results of Podsakoff et al. (2019) contributes to previous research by helping reviewers and researchers alike to interpret, assess and compare a study’s results accounting for bias induced by measurement error.