The purpose of this competing continuation grant proposal is to develop, evaluate and apply methodological and statistical procedures to investigate how prevention programs change outcome variables. These mediation analyses assess the link between program effects on the constructs targeted by a prevention program and effects on the outcome. As noted by many researchers and federal agencies, mediation analyses identify the most effective program components and increase understanding of the underlying mechanisms leading to changing outcome variables. Information from mediation analysis can make interventions more powerful, more efficient, and shorter. The P. I. of this grant received a one-year NIDA small grant and four multi-year grants to develop and evaluate mediation analysis in prevention research. This work led to many publications and innovations. The proposed five-year continuation focuses on the further development and refinement of exciting new mediation analysis statistical developments. Four statistical topics represent next steps in this research and include analytical and simulation research as well as applications to etiological and prevention data. The work expands on our development of causal mediation and Bayesian mediation methods that hold great promise for mediation analysis. In Study 1, practical causal mediation and Bayesian mediation analyses for research designs are developed and evaluated. This approach will clarify methods and develop approaches for dealing with violation of testable and untestable assumptions. Study 2 investigates important measurement issues for the investigation of mediation. This work will focus on methods to identify critical facets of mediating variables, approaches to understanding whether mediators and outcomes are redundant, and develop methods for studies with big data. Study 3 continues the development and evaluation of new longitudinal mediation methods for ecological momentary assessment data and other studies with massive data collection. These new methods promise to more accurately model change over time for both individuals and groups of individuals. Study 4 develops methods to uncover subgroups in mediation analysis including causal mediation methods, multilevel models, and new approaches based on residuals for identifying individuals for whom mediating processes differ in effectiveness from other individuals. For each study, we will investigate unique issues with mediation analysis of prevention data including methods for small N and also massive data collection (big data), the RcErLitEicVaANl rCoEle(Soeef imnsetruacstiounrse):ment for mediating mechanisms, and the application of the growing literature on causal methods and Bayesian methods. Study 5 applies new statistical methods to data from several NIH The project further develops a method, statistical mediation analysis, that extracts more information from funded prevention studies providing important feedback about the usefulness of the methods. Study 6 research. Mediation analysis explains how and why prevention and treatments are successful. Mediation disseminates new information about mediation analysis through our website and other media, by analysis improves prevention and treatment so that their effects are greater and even cost less. communication with researchers, and publications from the project.
Showing the most recent 10 out of 42 publications
