Affiliation:
1. Auburn University
2. University of Baltimore
3. Louisiana State University
Abstract
In moderated multiple regression analysis, several researchers have relied totally or partially on the sign associated with the regression weight of the cross-product term xz to interpret a moderator effect. This study examined a common convention used in such interpretations. First through logical exposition, and then through the use of a Monte Carlo approach to simulate various moderator conditions, it was found that this convention may lead to incorrect inferences regarding interaction effects. Alternative procedures for interpreting moderator effects are recommended.
Subject
Applied Mathematics,Applied Psychology,Developmental and Educational Psychology,Education
Cited by
15 articles.
订阅此论文施引文献
订阅此论文施引文献，注册后可以免费订阅5篇论文的施引文献，订阅后可以查看论文全部施引文献