An illustration of modeling moderating variables in cross-national studies

摘要

Purpose - This paper seeks to explore how moderation can and should be modeled in cross-national/cultural contexts. A multi-national study of consumer involvement is utilized to demonstrate proper methods for modeling the different types of moderation. Design/methodology/approach - Using data from a consumer survey regarding wine purchasing preferences conducted in Australia, New Zealand and the USA, the paper demonstrates how to identify moderators of form and of strength. A form moderator is modeled using multiplicative interactions while a strength moderator is modeled using multi-group analyses in structural equation modeling (SEM). Differences in consumers across the three countries are examined from the results. Findings - This study suggests that search behavior is positively influenced by involvement in New Zealand and the USA but not in Australia. It also shows that perceived risk of occasion decreases involvement in all three countries, while partial support for the positive effects of importance of tradition on involvement is found. Furthermore, "perceived risk of occasion," identified as a moderator of form, is found to significantly moderate the relationship between importance of tradition and involvement in the US sample only. Finally, the results demonstrate significant differences across the three samples in relationships among importance of tradition, perceived risk of occasion, involvement, and search behavior, indicating that the country variable has significant moderator effects. Originality/value - Understanding form vs strength moderation is important when evaluating multi-national/cultural differences so that proper methodology can be utilized. This paper provides international marketing researchers with guidelines on how to model interactions and multi-group comparisons using SEM.