International Choice Modelling Conference, International Choice Modelling Conference 2017

Font Size: 
Modelling choice sequences when options are geographically presented.
Harmen Oppewal, Ari Pramono

Last modified: 28 March 2017



The present paper aims to develop a sequential choice model that can accommodate the situation where choice options are presented in a geographic context, for example on map.

Many decisions involve making a sequence of dependent choices instead of just a collection of independent single choices. For example, a grocery shopping trip is essentially a multiple-item choice situation in which independence among items is an implausible assumption. There are many studies that have proposed a way to model choice sequences in a Random Utility Framework, for example in the context of assortment choice (Bradlow and Rao 2000; Harlam and Lodish 1995) or more currently in the context of internet search (Moe 2006; Santos, Hortaçsu, and Wildenbeest 2012). Little effort has been made however to model choice sequences in a spatial context, although as in the case of non-spatial situations, spatial decisions often involve sequences of related choices.

While there is extensive research in spatial choice modelling (Bhat and Guo 2004; Borgers and Timmermans 1987; Fotheringham 1988), the focus in that research is typically on single independent choices rather than on choice sequences. Additionally, existing studies in spatial search are mostly conducted at aggregate levels instead of analysing individual choices. For example Miller and Finco (1995) and Miller, Reardon and McCorkle (1999) study the relationship between individual search optimality and firm spatial structure (agglomeration and competition) by using simulated data.  Chandra and Tappata (2011) use aggregate (refuelling) market data to model the relationship between price dispersion and spatial search cost.

The present study involves the modelling of sequences of choices of individual decision-makers where the alternatives are spatially distributed and the chooser is travelling across the space from one site to the other.  This means that the model needs to be able to dynamically account for the chooser’s movement as well as the sites’ locations. The model in addition needs to take into account the accumulated previous choice as choosers travel along their route.

The study is conducted in the context of tourist trip scheduling and presents participants with a trip scheduling task. The choice options for the trip include one or more options from a set of geographically distributed options as displayed on an experimentally designed interactive map. The scheduling task concerns the planning of a day trip to a city that offers various attractions including shopping, museums, and leisure park options. The paper builds on and extends existing choice models and not only accounts for the options’ quality ratings, but also their spatial location relative to each while in addition accounting for sequential state dependence. The latter is achieved through the incorporation of the individual’s choice history as suggested by Swait, Adamowicz and Bueren (2004).

To investigate and model different possible search and choice strategies, the study varies the spatial information display across several conditions, in particular, a) one where full information about option attributes is available at all times and b) a condition where some attribute information (e.g. an attraction’s star rating) is only available after accessing and interacting with the map location.

The model is estimated on responses obtained from over 200 participants who each completed several trip scheduling tasks for different hypothetical cities. The model estimates reveal how spatial and non-spatial attributes contribute to the utility of any choice sequence. By comparing model results for the different display conditions we can investigate how search and choice processes are affected by these conditions.




Bhat, Chandra R. and Jessica Guo (2004), "A Mixed Spatially Correlated Logit Model: Formulation and Application to Residential Choice Modeling," Transportation Research: Part B, 38 (2), 147-68.

Borgers, Aloys and Harry Timmermans (1987), "Choice Model Specification, Substitution and Spatial Structure Effects : A Simulation Experiment," Regional Science and Urban Economics, 17 (1), 29-47.

Bradlow, Eric T. and Vithala R. Rao (2000), "A Hierarchical Bayes Model for Assortment Choice," Journal of Marketing Research, 37 (2), 259-68.

Chandra, A. and M. Tappata (2011), "Consumer Search and Dynamic Price Dispersion: An Application to Gasoline Markets," Rand Journal of Economics, 42 (4), 681-704.

Fotheringham, A. Stewart (1988), "Consumer Store Choice and Choice Set Definition " Marketing Science, 7 (3), 299-310.

Harlam, Bari A. and Leonard M. Lodish (1995), "Modeling Consumers' Choices of Multiple Items," Journal of Marketing Research, 32 (4), 404-18.

Miller, Chip E., James Reardon, and Denny E. McCorkle (1999), "The Effects of Competition on Retail Structure: An Examination of Intratype, Intertype, and Intercategory Competition," Journal of Marketing, 63 (4), 107-20.

Miller, Harvey J. and Mark V. Finco (1995), "Spatial Search and Spatial Competition: A Probability Analysis of Basic Results from the Spatially-Restricted Theory," Annals of Regional Science, 29 (1), 67.

Moe, Wendy W. (2006), "An Empirical Two-Stage Choice Model with Varying Decision Rules Applied to Internet Clickstream Data," Journal of Marketing Research, 43 (4), 680-92.

Santos, Babur De los, Ali Hortaçsu, and Matthijs R. Wildenbeest (2012), "Testing Models of Consumer Search Using Data on Web Browsing and Purchasing Behavior," American Economic Review, 102 (6), 2955-80.

Swait, Joffre, Wiktor Adamowicz, and Martin van Bueren (2004), "Choice and Temporal Welfare Impacts: Incorporating History into Discrete Choice Models," Journal of Environmental Economics & Management, 47 (1), 94.


Conference registration is required in order to view papers.