Replication Package for The Unequal Effects of Weekday Television on Children’s Academic and Behavioral Skills

This study examines whether children’s weekday television exposure causally affects cognitive and behavioral development, and whether these effects vary across the outcome distribution and socioeconomic status. The central hypothesis is that increased weekday television viewing crowds out cognitively productive activities and parental interaction, leading to weaker academic performance and elevated behavioral problems, particularly at vulnerable points of the outcome distribution. To address endogeneity in television exposure, the analysis uses instrumental variables and IV quantile regression to identify distributional effects rather than average effects alone. The data are drawn from a nationally representative longitudinal child level dataset that links children’s cognitive test scores, behavioral assessments, family characteristics, and time use measures. Outcomes include standardized reading and mathematics assessments and a validated behavioral problems index. Television exposure is measured as weekday viewing time reported by caregivers, and instruments exploit plausibly exogenous variation in family structure and parental constraints. The dataset preserves the panel structure, identifier consistency, and missing value patterns necessary for causal estimation. The empirical results show that increased weekday television exposure has heterogeneous effects across the outcome distribution. Negative effects are generally larger at lower quantiles of cognitive outcomes and for children from lower socioeconomic backgrounds, while behavioral outcomes exhibit stronger adverse responses at higher quantiles of problem behavior. These patterns suggest that television exposure exacerbates existing developmental disparities rather than affecting all children uniformly. Robustness checks using alternative instruments, subgroup analyses, and first stage diagnostics support the stability of the findings. The data and replication code are intended to be used for causal inference on child development outcomes, with careful attention to sample construction, instrument validity, and distributional heterogeneity. Users should interpret coefficients as local quantile treatment effects identified for compliers, rather than average treatment effects, and should account for outcome specific sample restrictions implied by the estimation strategy.