R2 effect-size measures for mediation analysis.

R(2) effect-size measures are presented to assess variance accounted for in mediation models. The measures offer a means to evaluate both component paths and the overall mediated effect in mediation models. Statistical simulation results indicate acceptable bias across varying parameter and sample-size combinations. The measures are applied to a real-world example using data from a team-based health promotion program to improve the nutrition and exercise habits of firefighters. SAS and SPSS computer code are also provided for researchers to compute the measures in their own data.

Mediation designs for tobacco prevention research.

This paper describes research designs and statistical analyses to investigate how tobacco prevention programs achieve their effects on tobacco use. A theoretical approach to program development and evaluation useful for any prevention program guides the analysis. The theoretical approach focuses on action theory for how the program affects mediating variables and on conceptual theory for how mediating variables are related to tobacco use. Information on the mediating mechanisms by which tobacco prevention programs achieve effects is useful for the development of efficient programs and provides a test of the theoretical basis of prevention efforts. Examples of these potential mediating mechanisms are described including mediated effects through attitudes, social norms, beliefs about positive consequences, and accessibility to tobacco. Prior research provides evidence that changes in social norms are a critical mediating mechanism for successful tobacco prevention. Analysis of mediating variables in single group designs with multiple mediators are described as well as multiple group randomized designs which are the most likely to accurately uncover important mediating mechanisms. More complicated dismantling and constructive designs are described and illustrated based on current findings from tobacco research. Mediation analysis for categorical outcomes and more complicated statistical methods are outlined.