National Longitudinal Study of Adolescent to Adult Health

National Longitudinal Study of Adolescent to Adult Health
Name	National Longitudinal Study of Adolescent to Adult Health
Acronym	Add Health
Country	United States
Sponsor	University of North Carolina at Chapel Hill; National Institute of Child Health and Human Development
Start year	1994
Cohort	Adolescents in grades 7–12 (Wave I)
Sample size	~20,745
Website	(not displayed)

National Longitudinal Study of Adolescent to Adult Health is a large, school-based, longitudinal cohort study that follows a nationally representative sample of adolescents into adulthood. Initiated in the mid-1990s, the study has produced extensive data on health, social, and economic trajectories, informing research across public health, sociology, psychiatry, economics, and demography. Its multidisciplinary design and rich bio-behavioral measures have supported influential work cited in journals, policy reports, and textbooks.

Overview

The project was launched by investigators at University of North Carolina at Chapel Hill with funding from National Institute of Child Health and Human Development, and collaboration with institutions such as Rutgers University, Columbia University, Harvard University, and University of Michigan. The initial fieldwork in 1994–1995 sampled students from schools drawn from a stratified cluster design modeled on procedures used by National Center for Education Statistics and informed by earlier cohort designs like the British Cohort Study and the National Longitudinal Surveys (NLS). The cohort includes oversamples of siblings and twins, enabling analysis comparable to work from Duke University twin registries and longitudinal resources at Watson School of Public Health.

Study Design and Methods

Add Health employed a multistage, stratified, clustered sample with in-school questionnaires, in-home interviews, and biospecimen collection; methods drew on survey best practices from National Health and Nutrition Examination Survey and panel designs used by Panel Study of Income Dynamics. Wave I established baseline measures in 1994–1995; subsequent waves occurred in 1996 (Wave II), 2001–2002 (Wave III), 2008–2009 (Wave IV), and later adult follow-ups analogous to designs at Framingham Heart Study. Field operations were coordinated with survey firms and university centers, and analytic approaches used weighting, imputation, and complex survey variance estimation similar to techniques promoted by American Statistical Association and researchers at University of Chicago and Columbia Business School.

Data Collection and Measures

Data collection combined in-school self-administered questionnaires, in-home interviewer-administered surveys, computer-assisted self-interviews, and biomarker assays. Measures include self-reported health, psychiatric symptom scales comparable to instruments from World Health Organization, substance use histories paralleling items used by Monitoring the Future, sexual behavior inventories aligned with work by Guttmacher Institute, socioeconomic indicators similar to those in Current Population Survey, and geographic linkage to contextual data like census tracts used by United States Census Bureau. Biological measures include anthropometrics, blood pressure, salivary cotinine, dried blood spot assays, and genome-wide genotyping with platforms used in large biobanks at Broad Institute and Wellcome Trust Sanger Institute. Linkage to mortality registries and administrative records has been used in analyses akin to studies from Centers for Disease Control and Prevention and Social Security Administration.

Key Findings and Publications

Analyses from the study have led to high-impact findings on adolescent sexual behavior, mental health trajectories, obesity epidemiology, substance use patterns, peer network influences, and socioeconomic mobility. Influential papers published by investigators affiliated with Princeton University, Yale University, Johns Hopkins University, Columbia University, and Harvard School of Public Health have examined links between adolescent relationship networks and later outcomes, echoing theories from Robert Putnam and empirical approaches used by Nicholas Christakis. Research has connected adolescent depression to adult health outcomes in work cited alongside studies from National Institutes of Health and World Bank. Publications in journals such as those at American Journal of Public Health, The Lancet, JAMA, and Social Science & Medicine have drawn on Add Health data to assess policy-relevant questions addressed also by scholars at Brookings Institution and Urban Institute.

Impact and Uses

The dataset has been used by thousands of investigators across universities and government agencies including Centers for Disease Control and Prevention, National Institutes of Health, and state health departments, shaping interventions and informing educational policy debates similar to contributions from RAND Corporation and Brookings Institution. Findings have influenced clinical guidelines referenced by the American Academy of Pediatrics and contributed to socioeconomic mobility research cited in reports by Organisation for Economic Co-operation and Development. The study’s social network modules have supported replication and extension in projects at Massachusetts Institute of Technology and Stanford University focused on diffusion of behaviors and contagion models associated with work by Duncan Watts and Sinan Aral.

Criticisms and Limitations

Scholars have critiqued aspects of the study, noting potential biases from sample attrition comparable to issues discussed in Panel Study of Income Dynamics literature and challenges in causal inference similar to debates surrounding Observational Study designs. Limitations include reliance on self-report for sensitive behaviors, measurement timing that may miss episodic events, and limited representation for certain immigrant subgroups relative to targeted cohorts like Fragile Families and Child Wellbeing Study. Ethical and privacy concerns about genetic data linkage and geocoded information have been raised in forums including panels at National Academies of Sciences and debates involving Electronic Frontier Foundation. Despite these criticisms, methodological work by researchers at University of California, Los Angeles and Duke University has developed weighting, imputation, and sensitivity analysis techniques to mitigate many limitations.

Category:Cohort studies