Strengthening causal inference in understanding the relationship between childhood adversity and adolescent psychopathology

Thesis

Adverse childhood experiences (ACEs) are a global public health concern, with robust evidence linking ACEs to lasting mental, physical, and psychosocial difficulties across development. Yet despite the proliferation of ACE research over the past two decades, (1) no universal paradigm for operationalising ACEs exists, (2) the extent to which these associations are confounded by preexisting vulnerabilities remains uncertain, and (3) addressing genetic confounding represents an ongoing challenge. This thesis aims to strengthen causal inference in understanding the relationship between childhood adversity and adolescent psychopathology by applying data-driven, quasi-experimental, and genetically informed methods on contemporary longitudinal cohorts from the UK and US.

Chapter 1 reviews the current literature on ACEs and psychopathology, highlighting conceptual and methodological limitations in the field. Chapter 2 applies an exploratory data-driven approach to identify underlying dimensions of adversity across UK and US cohorts – parental threat, deprivation, and victimisation – and finds that victimisation emerges as the strongest predictor of adolescent psychopathology. Chapter 3 uses propensity score matching to strengthen causal inference about the effects of individual ACEs, revealing victimisation and emotional adversities as robust predictors of psychopathology even after accounting for early life vulnerabilities. In contrast, associations with household-level adversities were substantially attenuated, suggesting they were largely explained by pre-existing confounding. Chapter 4 introduces the novel approach of genetically adjusted propensity score (GAPS) matching to address measured genetic confounding. It reveals gene–environment correlations between polygenic scores and ACE exposures, and further demonstrates that victimisation and emotional adversities predict psychopathology independently of both genetic and environmental confounding. While associations with household-level adversities largely diminished once confounding was addressed, parental psychopathology was no longer significantly associated with adolescent psychopathology after GAPS adjustment, suggesting genetic confounding. Finally, Chapter 5 synthesises these findings, discusses theoretical and practical implications, and outlines directions for future research.

 Overall, this thesis advances the study of childhood adversity by: (1) clarifying how ACEs can be meaningfully operationalised through data-driven dimensions, (2) strengthening causal inference using quasi-experimental methods, and (3) addressing genetic confounding with a novel genetically informed design. Together, these contributions provide a more rigorous foundation for understanding how childhood adversity shapes adolescent psychopathology, over and above measured confounding from genetic and environmental influences.

Authors: Chow, ARW

• DOI: 10.5287/ora-1avvvz155
• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English
• Keywords: adverse childhood experiences; childhood adversity; adolescent psychopathology; causal inference; adolescent mental health; aces
• Subjects: Psychology; Psychiatric epidemiology; Psychiatry