Workweek Estimate-Diary Differences and Regression to the Mean
Published: Jan 2013

This paper confronts the assertion that differences between questionnaire-based “stylized” and diary-based estimates of paid work time can be understood as a regression to the mean. We use a 3-year panel study with repeated 7-day time diaries to investigate the plausibility of the hypothesized regression explanation, along with a simple weekly paid work “grid” (now a standard element in the Eurostat Harmonised Time Use Study design). These are deployed to demonstrate that an alternative mechanism related to the variability of the workweek, and hence to the survey respondents’ knowledge (or ignorance) of their own work hours—can explain various features of the gap between the two sorts of estimates.

No current keywords


Download Paper