ASAC 2005 Najam Saqib (student) Toronto, Ontario Joe Zhang (student) Ed Bruning I.H. Asper School of Business University of Manitoba ACCOUNTING FOR INTERACTION EFFECTS AND CONSUMER HETEROGENEITY IN AN ORDERED PROBIT ANALYSIS OF INTERNATIONAL AIR TRAVELLER PREFERENCES In this paper, the authors analyze product attributes and consumer characteristics in an international service selection decision. Previous work has focused on main effects without exploring the significance of interaction or individual heterogeneity effects. Using conjoint and ordered probit analyses, we find that including interaction and respondent specific dummy variables alters the impacts of main effect variables. Furthermore, accounting for interaction and consumer heterogeneity increases the predictive ability of the consumer preference model. Introduction From an international marketer’s point of view, an important reason to understand how consumer preferences are formed is so that consumers can be segmented for brand positioning, targeting, and promotion. Moreover, increased dependence on international trade makes it even more worthwhile for marketers to understand factors, such as country-of-origin, that influence consumers’ preferences for products and services. Research has shown that one's impression of a country shapes attitudes and impacts purchase behaviours with regard to that country's offerings. In studying factors that affect preferences for international air carriers, our research objectives are twofold: 1) to investigate interaction effects of specific consumer characteristics and product attributes upon air carrier preferences, and 2) to ascertain the impact upon both interaction effects as well as main variable effects from inserting variables into the preference estimation equation that control for individual consumer heterogeneity. Previous research in international marketing has reported significant main effects that explain international airline preferences (Bruning, 1997), and interaction effects of product attributes and consumer characteristics upon choice (Cordell, 1991; Johansson, 1985). Greene (1997), however, suggests that by adding individual dummy variables, significant efficiencies are gained in estimating cross-sectional demand and customer preference models. The interaction effects could provide us further insights about more appropriate segmentation schemes in international markets. Also, controlling for individual heterogeneity in estimating models could clarify the unique impacts of main and interaction variables. In the empirical portion of our paper, we employ ordered probit analysis, as opposed to multinomial probit, or logit, to test main, interaction and individual heterogeneity hypotheses, and to estimate our international air carrier preference model. The remaining sections of the paper are organized as follows. Section two reviews the literature dealing with product and consumer related factors typically involved in choice decisions as well as the literature dealing with various interaction effects among the major factors. Section three presents the study methodology. We describe the data collection and sampling components, our conjoint experiment, and our procedure for estimating the ordered probit model. The fourth section presents the study results 85