Scale heterogeneity in discrete choice experiment: An application of generalized mixed logit model in air travel choice

Abstract

This study explores to what extent scale heterogeneity (i.e., varying standard deviation of the errors across consumers) is important in air travel choice using data from a stated preference discrete choice experiment. We used generalized mixed logit model (GMIXL) that nests scale and taste heterogeneity in the context of air travel ticket choice. We found empirical evidence of the importance of often-neglected scale heterogeneity along with pre-dominant taste heterogeneity. The findings suggests that it is important to account for various forms of heterogeneity with GMIXL to model air travel demand to identify variety of consumer segments, which has implications for airline optimal menu of ticket offerings and pricing strategies.