Exploring Father-Adolescent Closeness: A Random Forest Approach

Background and Purpose: Adolescents who feel close to their family members are less likely to experience negative outcomes in adulthood, such as substance abuse and intimate partner violence (Steiner et al., 2019). Having close family relationships helps adolescents develop positive coping mechanisms to withstand a variety of challenges (Gervais & Jose, 2020). Previous research has demonstrated the importance of father-adolescent relationships for adolescent development (Fagan, 2022; O'Gara et al, 2019). Yet, partly due to data and methodological limitations, less is known about the relative importance of factors promoting positive father-adolescent relationships (Trahan et al., 2021). The purpose of this study is to advance the understanding of the relative importance of factors that may be related to adolescents’ perception of their relationship with their fathers.

Methods: The data used in this study are from the Future of Families and Child Wellbeing Study—a cohort study of children born to mostly unmarried parents in 20 large U.S. cities. Interviews were conducted with mothers and fathers around the time of the child’s birth, then at ages 1, 3, 5, 9, and 15. Children were interviewed at ages 9 and 15. The data for this study (n = 2,927) primarily draw upon the father and child surveys. The dependent variable is the child’s answer to the question “How close do you feel to your dad?” measured with a 4-point Likert scale ranging from “not very close” to “extremely close.” The independent variables include a comprehensive set of developmentally appropriate father engagement indicators measured across all waves of the study (e.g., singing songs at age 3, talking about things that matter at age 9) and a set of sociodemographic variables (e.g., father’s residential status, father’s number of children). Random forest modeling (Breiman, 2001), a non-parametric machine learning method, is used to assess the relative “importance” of over 100 independent variables for father-adolescent closeness.

Results: Overall, recent father engagement was most important for father-adolescent closeness. The top variables in order of importance were how well the child feels they talk about things that matter with their father (age 15), the father living with the child at ages 5 and 9, the frequency the child feels the father engages in play and chore activities with them (age 15), the frequency they talk about current events (age 15), the frequency they talk about the child’s day (age 15), the father living with the child at age 3, the father helping with homework (age 15), the mother’s household income-to-poverty ratio (age 9), and the mother and father being married (age 5). All these variables were positively associated with father-adolescent closeness.

Conclusions and Implications: While father engagement early in a child’s life is important, this study suggests that what matters most to adolescents is how they currently interact with their fathers. Policies and interventions that enable fathers to spend time with their children will likely help adolescents navigate challenges they face, such as resisting and challenging racist and oppressive environments.