Topic 5 - Moderator and other Non-Linear Effects

Pick-a-Point Technique

selects levels of a continuous predictor and moderator based on high, average, and low scores on the moderator and the mean and SD of the IV

Covariate

variables that are important to the expanation of the DV but are not the focus of the analysis

Homogeneity of Regression in moderated regression

An assumption that a covariate has the same effect at each level of the moderator

Variable Centering

A statistical transformation in which each score in a continuous distribution is subtracted from the mean score, producing a mean of 0

Moderators

In some way change the association between a predictor and outcome variable

Mediators

are influenced by a predictor and in turn effect the outcome

Johnson-Neyman Test

An alternative to simple slopes analysis when the moderator is continuous, testing multiple levels of the continuous moderator simultaneously