ABSTRACT
A mediation relation is composed of two relations, a relation between X and M and a relation between M and Y. Mediation analysis requires some justification for the assumption about the temporal ordering of variables. Mediation analysis also assumes that the correct variables in the micromediational chain were investigated. Mediation studies by different researchers and in different contexts can reduce omitted influence explanations of a mediation model. Information on the reliability and validity of measures in a mediation model is central to accurate interpretation of these models. Three general analyses are outlined: moderator analysis, mediator analysis, and combined moderator and mediator analysis. In an exploratory analysis, these three general analyses could be conducted in any order, although the moderator, mediator, and moderator/mediator sequence is likely to be more consistent with the priorities for the most research studies.
