Non-response and population representation in studies of adolescent time use

Abstract

Researchers have debated which methods are most valid and reliable for studying time use. One key instrument for measuring time use is the time diary, which has unique analytic properties that, if not adjusted for, can bias estimates. To assess sampling and non-response bias and potential under- or overreports of various activities, we use three different datasets to compare adolescents’ time use. Results of these comparisons are used to show how investigators can statistically adjust time use data to obtain more accurate estimates of time spent in various activities.