The paper by David P. MacKinnon and Amanda J. Fairchild provides an overview of the current state of methods for investigating mediating variables in psychological research. Mediating variables are crucial for understanding how antecedent variables influence dependent variables, offering a more detailed understanding of their relationships. The authors discuss various statistical and experimental methods used to assess mediation, including the single-mediator model, which involves equations to measure direct and indirect effects. They highlight the importance of significance testing for the mediated effect, the limitations of the causal-steps approach, and the need for more accurate methods to estimate and test the mediated effect. The paper also covers longitudinal mediation models, which allow for the examination of temporal precedence and stability over time, and models that combine moderation and mediation. Experimental designs, such as blockage and enhancement designs, are discussed as promising methods to combine statistical mediation analysis with experimental manipulation. Future directions include the development of longitudinal models, methods for causal inference, and the refinement of models for different data types and distributions. The authors emphasize the ongoing importance of mediation analysis in psychological research and its potential to provide more accurate insights into the relationships between variables.The paper by David P. MacKinnon and Amanda J. Fairchild provides an overview of the current state of methods for investigating mediating variables in psychological research. Mediating variables are crucial for understanding how antecedent variables influence dependent variables, offering a more detailed understanding of their relationships. The authors discuss various statistical and experimental methods used to assess mediation, including the single-mediator model, which involves equations to measure direct and indirect effects. They highlight the importance of significance testing for the mediated effect, the limitations of the causal-steps approach, and the need for more accurate methods to estimate and test the mediated effect. The paper also covers longitudinal mediation models, which allow for the examination of temporal precedence and stability over time, and models that combine moderation and mediation. Experimental designs, such as blockage and enhancement designs, are discussed as promising methods to combine statistical mediation analysis with experimental manipulation. Future directions include the development of longitudinal models, methods for causal inference, and the refinement of models for different data types and distributions. The authors emphasize the ongoing importance of mediation analysis in psychological research and its potential to provide more accurate insights into the relationships between variables.