The moderator variable viewed as heterogeneous regression

Abstract

Considers that while there has been a great variety of recent work dealing with the moderator variable, the meaning of this concept remains unclear. A useful framework for dealing with moderator variables is heterogeneous regression. The heterogeneous model and 2 useful tests of significance are described. A real data example is used as an illustration. It is also shown that this framework is not inconsistent with past work, and some guidelines for employing heterogeneous regression are presented. (PsycInfo Database Record (c) 2025 APA, all rights reserved)