Modeling Service Trade-Offs in Air Itinerary Choices

Abstract

The application of a mixed logit approach using stated-preference survey data to the development of itinerary choice models is described. The models include the effects on itinerary choices of airline, airport, aircraft type, fare, access time, flight time, scheduled arrival time, and on-time performance. The empirical results demonstrate the importance of explicitly accounting for traveler preference heterogeneities by using segmentation by trip purpose, interaction effects involving frequent flier status, and random parameter specifications. Explicitly including preference heterogeneity by using the mixed logit specification results in significant statistical improvements and important coefficient differences as compared with using a standard fixed-parameter logit model. The calculated marginal rates of substitution show the relative importance that travelers assign to key service variations among itineraries. All service features that were included in the model had significant values to travelers, and the values were affected, as would be expected, by the traveler's frequent flier status. Although current reservation and ticketing services provide information to prospective travelers on most of these itinerary features, most services do not report on-time performance, which, however, can be an important selection criterion for travelers.