Choice set awareness and ordering effects in discrete choice experiments in discrete choice experiments

Abstract

The choice experiment elicitation format confronts survey respondents with repeated choice tasks. Particularly within the context of valuing pure public goods, this repetition raises two issues. First, does advanced awareness of multiple tasks influence stated preferences from the outset, and second, even in the absence of such awareness, does the process of working through a series of choice tasks influence stated preferences leading to choice outcomes that are dependent on the order in which a question is answered? The possible motivators of these effects include economic-theoretic reasons such as strategic behavior, as well as behavioral explanations such as response heuristics and learning effects. A case study of a familiar good (drinking water quality) combines a split sample treatment of the presence/absence of advanced awareness with a full factorial design permitting systematic variation of the order in which choices are presented to respondents. A further sample division allows examination of effects arising from variation in the scope of the initial good presented to respondents. Using discrete choice panel data estimators we show that both advanced awareness and order effects exist alongside interactions with the scope of the initial good.