Comparisons of self-report measures of sleep duration: an actigraphic validation experiment

Abstract

This study has explored in-depth the importance of the variables sleep quality and quantity, their effects on human functioning, and how we currently measure them. Evidence from past research and pilot data suggests that there are flaws in current self-report sleep duration measures, such as used in the Pittsburgh Sleep Quality Index (PSQI), that may be addressed easily to improve the predictive validity of self-reported measures with actigraphy-recorded sleep duration. The current study compared three formats of the PSQI (i.e., standard, rounding instructions, drop-down menu)to determine if format change would lead to increased predictive validity.. Secondarily, outputs from solely the PSQI surveys were analyzed to determine if the presence of heaping (i.e., rounding to the whole hour) decreased with alternative formats. It was hypothesized that simple changes to the format of self-reported sleep duration could lead to an increase in correlation between self-report and objective measures of sleep quantity and a decrease in heaping. Results supported the hypothesis that heaping can be significantly reduced by modifications to the PSQI. Although the results indicate some improved predictive validity, the analyses were underpowered, and statistically significant results were not obtained.