Assessing the impact of school marketing: conjoint choice experiments incorporating availability and substitution effects

In this paper we illustrate the idea that conventional conjoint choice models are not well suited to the modelling of parents' school choices for the purpose of assessing the impact of school policies on school choice. We apply an extended logit model formulation to capture the differential competition between and among elementary schools of different or the same denomination. Aggregate models are compared with models estimated separately for different religious groups and it is found that these segment-based models perform best if substitution and availability effects are included.