Many stated choice experiments allow for varying mixtures of the number of alternatives in evaluated choice sets. The subset of alternatives shown to each sampled respondent is allocated according to the design specification in order to ensure that each alternative occurs an equal number of times across the sampled population. While this is an acceptable practice, it is not the same as establishing design rules that can explicitly account for the non-availability of one or more labelled

more »

... ives for each individual. The notion of availability sets and relevancy is not new, but despite it being a focus of revealed preference choice studies it appears to have been overlooked in the design of stated choice experiments. This paper presents a way to account for observation-specific non-availability of alternatives in the design of stated choice experiments and illustrates the method in the context of choice on main and access mode for commuting and non-commuting trips in the north-west sector of Sydney.