Search based internet surveys: Airline stated choice

Abstract

Stated preference (SP) experiments are becoming an increasingly popular survey methodology for investigating air travel choice behaviour. Nevertheless, some evidence suggests that SP experiments do not mirror decisions in real markets. In this paper we introduce a novel survey methodology that aims to make air travel surveys more consistent with real world settings, with the aim of obtaining more realistic results. The survey is modelled on the interface and functionality of an online travel agent (OTA). As with a real OTA, many ticket options are presented. Sort tools allow the options to be reordered, search tools allow options to be removed from consideration, and a further tool allows attributes to be hidden and shown. Extensive use of these tools is made by the 462 respondents, with the captured data revealing some attribute preferences at the individual level, and significant heterogeneity of preference across individuals. A traditional SP component was also completed by the respondents. Mixed multinomial logit models were estimated on data from both the traditional SP and OTA components, with the later exhibiting greater willingness to pay (WTP) heterogeneity.