Finally, we conducted a series of sensitivity analysis to evaluate the robustness of our results. First, we reran all path analyses with a subsample of individuals aged 25 (typical age for finishing schooling in the U.S; U.S. Census Bureau, 2018) or older. Second, we conducted exploratory analyses by including education polygenic score, educational attainment, and income from both fathers and mothers in the same model for analyses presented in Figures 1–2. For Figure 3, due to the small sample size and limited power, we ran a father and mother combined path model using the mean scores of parental education polygenic score, parental educational attainment, parental income, and parental bonding. In this model, for individuals who only had data available for one parent, the available parental data was used instead of mean scores. Finally, we conducted multi-group path analyses to examine whether the patterns of associations differed across offspring sex. This study was not preregistered.
