Will the kids be alright? Using machine learning to identify predictors and moderators of adolescents’ emotion regulation development


A key developmental challenge in adolescence is acquiring mature emotion regulation skills, which form the foundation for mental health and high-quality relationships throughout life. There is mounting evidence, however, for a "maturity gap" in middle-adolescence: A developmental asymmetry between emotional and regulatory brain circuits, which motivates most youngsters to seek out challenges in life and love, but renders some of them vulnerable to emotion dysregulation. Unfortunately, there is little theory about which risk factors and environmental hazards render adolescents susceptible to emotional difficulties. This Veni-project takes an inductive approach, using machine learning to identify predictors of adolescents’ emotion regulation development, and moderators of parents’ role therein. The resulting insights are consolidated into theory.
First, I will systematically review predictors currently considered relevant in the adolescent emotion regulation literature. Then, I apply SEM-forests, a combination of person-centered structural equation models and the machine learning algorithm “random forests”, to identify which of these factors are most predictive of a drop in adolescents’ emotion regulation abilities. SEM-forests were successfully used to predict a terminal decline in well-being in the elderly, which highlights the feasibility of this approach.
Secondly, my recent research demonstrated that parents remain an important driver of emotion regulation development in adolescence, and that fathers and mothers contribute in distinct ways. I also discovered substantial between-family differences in parental influence. To identify moderators of parental influence, I will apply SEM-forests to person-centered models, which capture within-family predictive effects between parenting and emotion regulation. This will reveal in which families parents play a nurturing, or undermining, role.
A third project synthesizes the existing literature reviewed in Project 1, and data-driven insights from this Veni-project, to develop a new theoretical framework. In doing so, this Veni-project paves the way for a new wave of person-centered research on adolescent emotion regulation.





Dr. C.J. van Lissa

Verbonden aan

Universiteit Utrecht, Faculteit Sociale Wetenschappen, Departement Maatschappijwetenschappen


Dr. C.J. van Lissa


15/01/2020 tot 31/08/2023