Interpreting interrelations across multiple levels in HGLM models: an application in international marketing research

This is the latest version of this item.

Official URL: http://dx.doi.org/10.1108/02651331111107099

Purpose - Although the Use of data from different levels is very common in international marketing research, the practice of employing multi-level analysis techniques is relatively new. The paper aims to provide an application of a specific case of multi-level modelling where the dependent variable is dichotomous, which is often the case in marketing research (e.g. whether a consumer buys the brand or not, whether he/she is aware of the brand or not, etc.) Design/methodology/approach - A hierarchical generalized linear model is employed. Findings - Since this is a technical paper, the authors would like to emphasize the process rather than the empirical findings. In summary, the paper: provides a brief theoretical overview of Hierarchical Linear Modeling and Hierarchical Generalized Linear Modeling; illustrates the application of the method using the domains of consumers within countries and a dichotomous dependent variable; focuses on interpretation of log-odds results; and concludes with practical issues and research implications. Originality/value - The main value of this research is to demonstrate how to employ multi-level models when the dependent variable is dichotomous. Multi-level techniques are quite new in international marketing research, although nested data structures are relatively common in our field. This is a technical paper that guides the researchers as to how to apply and interpret the results when modeling such data with a dichotomous dependent variable.