Analyzing the effect of preference heterogeneity on willingness to pay for improving service quality in an airline choice context

In this paper, we examine different model specifications to detect the presence of preference heterogeneity in an airline choice context. We analyze the most important route connecting the Canary Islands archipelago with the Iberian Peninsula. For this purpose, we conduct a stated preference (SP) experiment to analyze individual's preference for the main attributes defining the service offered by the airlines. We find consistent specifications for multinomial logit and mixed logit models. In the first multinomial logit model, all the modelling variables enter in the utility function in linear form. The second multinomial logit model accounts for systematic taste variation including some interactions of socio-economics variables with level-of-service attributes, and the third one, considers interaction effects between cost and penalty and cost and frequency. We also consider taste variations in the preferences of passengers using mixed logit models. In these cases, we find that most of the parameters are random. To analyze the effect of preference homogeneity/heterogeneity assumptions, we obtain the willingness to pay for improving the service quality attributes under each scenario. Our results suggest that the benefit measures at the individual level are really sensitive to preference heterogeneity assumptions. We eventually obtain that the willingness to pay using mixed logit models are lower than those obtained from multinomial logit specifications. That means that preference heterogeneity needs to be considered in order to properly estimate the willingness to pay measures for service quality changes because if taste heterogeneity is important and neglected, the willingness to pay measures can be highly overestimated.