Abstract:Population-level analyses are inherently complex due to a myriad of latent confounding effects that underlie the interdisciplinary topics of research interest. Despite the mounting demand for generative population models, the limited generalizability to underrepresented groups hinders their widespread adoption in downstream applications. Interpretability and reliability are essential for clinicians and policymakers, while accuracy and precision are prioritized from an engineering standpoint. Thus, in domains such as population neuroscience, the challenge lies in determining a suitable approach to model population data effectively. Notably, the traditional strata-agnostic nature of existing methods in this field reveals a pertinent gap in quantitative techniques that directly capture major sources of population stratification. The emergence of population-scale cohorts, like the Adolescent Brain Cognitive DevelopmentSM (ABCD) Study, provides unparalleled opportunities to explore and characterize neurobehavioral and sociodemographic relationships comprehensively. We propose diversity-aware population modeling, a framework poised to standardize systematic incorporation of diverse attributes, structured with respect to intrinsic population stratification to obtain holistic insights. Here, we leverage Bayesian multilevel regression and poststratification, to elucidate inter-individual differences in the relationships between socioeconomic status (SES) and cognitive development. We constructed 14 varying-intercepts and varying-slopes models to investigate 3 cognitive phenotypes and 5 sociodemographic variables (SDV), across 17 US states and 5 race subgroups. SDVs exhibited systemic socio-spatial effects that served as fundamental drivers of variation in cognitive outcomes. Low SES was disproportionately associated with cognitive development among Black and Hispanic children, while high SES was a robust predictor of cognitive development only among White and Asian children, consistent with the minorities' diminished returns (MDRs) theory. Notably, adversity-susceptible subgroups demonstrated an expressive association with fluid cognition compared to crystallized cognition. Poststratification proved effective in correcting group attribution biases, particularly in Pennsylvania, highlighting sampling discrepancies in US states with the highest percentage of marginalized participants in the ABCD Study©. Our collective analyses underscore the inextricable link between race and geographic location within the US. We emphasize the importance of diversity-aware population models that consider the intersectional composition of society to derive precise and interpretable insights across applicable domains.

Methodological Challenges in Causal Research on Racial and Ethnic Patterns of Cognitive Trajectories: Measurement, Selection, and Bias

On (Cognitive) Bias in School Psychology

Through the lens of causal inference: Decisions and pitfalls of covariate selection

Lifecourse Social Conditions and Racial and Ethnic Patterns of Cognitive Aging

A Causal Framework for Observational Studies of Discrimination

Causal effects in clinical and epidemiological studies via potential outcomes: concepts and analytical approaches.

Causal inference for psychologists who think that causal inference is not for them

Causal Inference Challenges in the Relationship Between Social Determinants and Cardiovascular Outcomes

Diversity-aware Population Models: Quantifying Associations between Socio-Spatial Factors and Cognitive Development in the ABCD Cohort

Cognitive Trajectories and Associated Social and Behavioral Determinants Among Racial/Ethnic Minority Older Adults in the United States

Enhancing causal inference in population-based neuroimaging data in children and adolescents

Navigating complexities of racial disparities in Alzheimer disease biomarkers

The Causal Roadmap and Simulations to Improve the Rigor and Reproducibility of Real-data Applications

Counterfactual Prediction Under Outcome Measurement Error

Bridging the Divide: Tackling Tensions Between Life-Course Epidemiology and Causal Inference

Modeling question difficulty for unbiased cognitive diagnosis: A causal perspective

Introduction to Causal Inference for Psychologists: Testable and Non-Testable Causal and Statistical Assumptions

Reconciling Heterogeneous Effects in Causal Inference

The cultural equivalence of measurement in substance use research.

Associations between Socioeconomic Status, Cognition, and Brain Structure: Evaluating Potential Causal Pathways Through Mechanistic Models of Development

When no answer is better than a wrong answer: a causal perspective on batch effects