Adolescents’ Time Use in Context: Institutional, Demographic, and Socioeconomic Factors
Joan Garcia Roman, Centre d''Estudis Demogràfics
Pablo Gracia, Trinity College Dublin
We use five dependent variables that count the number of minutes that the adolescent allocated to each activity on the day of observation, considering only primary activities (we did not include secondary activities in the analyses): (1) daily minutes allocated to educational activities outside schooling activities; (2) daily minutes allocated to watching television in the day of observation; (3) daily minutes allocated to electronic activities; (4) daily minutes allocated to sports; (5) daily minutes on sleeping. Our micro-level independent variables include three categorical measures. Maternal education has three categories based on the maximum level of education obtained by the mother (below high school, high school, college). Family structure differentiates between adolescents living in a two-parent heterosexual family and those living in a single-mother family. Maternal employment establishes a distinction between mothers who were not employed and those who were employed at the moment of the interview. We added two main control variables that are well known to be critical predictors of adolescents’ time use, namely sex (a binary measure of the respondents’ sex) and age (a continuous measure of the respondent’s age) (Wight et al., 2009). Finally, in those analyses where we include all our diaries of study (merging weekends and weekdays) we control for whether the diary referred to a weekday (Monday-Friday) or weekend (Saturday-Sunday).
Our will be based on multivariate statistical regressions. Our empirical strategy will follow three main steps: (1) We will run multivariate analyses for a pooled sample with data on the three countries of study (separately by weekdays and weekends), where we will use weights that correct for the size of each country for both the sample of weekdays and weekends (results remained very similar when we did not balance the presence of countries by using weights); (2) we will employ multivariate analyses for the same pooled sample, but including country dummy variables for our three countries, examining country-level differences after controlling for all our independent variables and control variables in the study; (3) We will run country-level interactions with our three micro variables of interest (maternal education, family structure, maternal employment).
Presented in Session 1102: Families and Households