Choice models are used to get an insight into the likely effects of policy measures regarding parking facilities. Usually, the well-known Multinominal Logit model is estimated given observed parking choice behaviour. The result is an array of parameters which represents the weights of the explaining variables for all observed individuals. However, it may be quit possible that within the group of respondents, subgroups can be distinguished. For example, the parking choice behaviour of older people may differ from that of younger people. In this case it is logical to estimate two choice models: one array of parameters for subgroup 1 and one array for subgroup 2. The software package NTELOGIT provides facilities to check whether two subgroups of respondents need to be distinguished. This framework is used to model parking choice behaviour of visitors of the main shopping centre of Veldhoven, a municipality in the south of The Netherlands. (A)
Abstract