Testing moderation effects using non-parametric regressions

Abstract

Listen

Translate article

Testing moderation effects is highly common in the hospitality literature. Most theories in the field depend on variables that alter the nature and direction of the relationship between two variables. While moderation continues to be heavily used, methods for testing moderation effects are not always robust. One common problem that researchers often face is the need to pre-assign a particular functional form. The aim of this note is to address this problem. We describe three different non-parametric models that offer more flexibility in testing moderating effects without a need to pre-impose a specific functional form. We test the three models on an interesting application involving the moderating role of corporate social responsibility (CSR) on the relationship between advertising and firm value. The results revealed interesting moderating effects that go beyond the simple linear moderation.