| This article uses a case study involving Hawaii food baskets to show that although a choice based conjoint experiment can elicit respondents’ most preferred alternative, this “preferred” option may not be one that respondents are willing to purchase. Thus, a choice based experiment that involves hypothetical product selection may predict different behavior, depending on the type of questions asked in the survey. This study shows that follow-up questions in a conjoint survey may serve an important role in improving model fit and the comprehensiveness of behavioral prediction. |
