Integrative Data Analysis: A Novel Methodological Framework for Data Harmonization and Pooling
Integrative data analysis (IDA) is the modeling of raw data that have been pooled across two or more independent samples. Advantages of IDA include economy (i.e., reuse of extant data), power (i.e., large combined sample sizes), the potential to address new questions not answerable by a single contributing study (e.g., combining longitudinal studies to cover a broader swath of the lifespan), and the opportunity to build a more cumulative science (i.e., examining the similarity of effects across studies and potential reasons for dissimilarities). There are also methodological challenges associated with IDA, including the need to account for sampling heterogeneity across studies, to develop commensurate measures across studies, and to account for multiple sources of study differences as they impact hypothesis testing. IDA is particularly promising when applied to the study of substance use and abuse within and across time. In this talk we provide a general overview of IDA, explore the advantages and disadvantages of this approach, and describe future avenues for developing IDA as a general methodological framework within the psychological sciences.Patrick Curran is a professor in the Department of Psychology at UNC-Chapel Hill