Heterogeneous substitution patterns based on current mode and time of departure choices in the context of uncertainty

Recent modelling approaches to jointly estimating revealed preference (RP) and stated preference (SP) have incorporated random effects or error-components to account for the panel data characteristics when observing multiple SP choices. This paper extends the panel data approach to SP data enrichment models by estimating substitution patterns that depend on the respondent’s current mode (RP). This is relevant to a choice experiment which incorporates the time-of-day choice along with the possibility of switching modes. The substitution dependence error components structure of the model is used to uncover whether a respondent perceives a change in mode or a change in departure time as the closer substitute to their current travel choice.