The Effect of Learning and Fatigue on Preferences and WTP in a Choice Experiment

This paper addresses the issues of how learning and fatigue effects related to the repeated choices potentially affect respondents’ stated preferences in a Choice Experiment setting. In a case study concerning food quality attributes in chicken breast filet, the results show that a learning effect is indeed present in a sequence of 16 choice sets, whereas there is no evidence of fatigue effects. This conclusion is based on an observed reduction in the error variance for the last 8 choice sets relative to the first 8 choice sets. An analysis incorporating choice set specific scale factors reveals that the main part of the observed learning process takes place in the first choice set. The overall preference structure is found to differ significantly between the two sequences of choice sets. This suggests that the learning effect leads to changes in overall preference structure. Turning to a comparison of attribute WTP estimates, the results indicate that respondents refining and significantly adjusting their stated values upwardly for two of the four quality attributes considered in the survey. If the main objective of a Choice Experiment survey is to obtain policy advice in terms of WTP, then allowing for learning effects can be essential.