Modeling social influence using sequential stated adaptation experiments: A study of city trip itinerary choice

摘要

This paper introduces a model that captures the effect of social influence on decision makers' choice behavior in the context of sequential stated adaptation experiments, in which a decision maker first chooses the alternative he/she personally likes best from an experimentally varied choice set and then, after being informed about the choice of a certain social network member, chooses again from the same choice set. Social influence represents the effect of the choices of one or more social network members in the same choice situation on the choice behavior of a decision maker. It is assumed to depend on a decision maker's socio-demographics and personality traits, and the strength of the relationship with the social network member. The concept of choice consistency is introduced to reflect a respondent's tendency to stick with his/her original choice in completing the sequential stated choice tasks. By accounting for this effect, a less biased estimate of social influence is obtained. Similarly, the model accounts for the case that the first choice of the respondent and the social network member's choice are identical by introducing a correction factor to further reduce the bias in the estimated social influence. To access the validity of the model, the choice of city trip itinerary was chosen as an example. Decision makers' personality traits and their relationship with the social network member were treated as latent variables. Therefore, first, exploratory and confirmatory factor analysis was carried out to investigate the reliability and validity of the constructed scales. Then a sequential estimation approach was employed to avoid confounding between decision makers' preferences towards attributes and the effects of social influence and choice consistency. In the first stage, the taste parameters were estimated and then in the second stage the composition of social influence effect and choice consistency effect were estimated simultaneously. Results show that the model reproduced the observed data well. Social influence has a modest but significant effect on decision makers' choice behavior; decision makers' socio-demographics and personality traits, social network type as well as the relationship with social network members all have significant effects on social influence. The effect of choice consistency is also significant.