Toward an Understanding of the BDM

Predictive Validity, Gambling Effects, and Risk Attitude

Abstract

Pricing research suggests incentive-compatible evaluations when consumers’ situation-specific WTP is to be elicited. Especially, the lottery-based Becker-DeGroot-Marschak-mechanism (BDM) is recommended, as it seems to outperform other elicitation methods. In this study, the BDM was used to measure subjects’ WTP for eight shopping goods in binding purchase settings. In accordance with previous studies, the validity of elicited WTP measures was checked within subjects with respect to indicators of face and criterion validity (such as interest in buying, preference ratings, and compliance rates). In addition, this study observed real purchases of a separate validation sample measured under identical circumstances, thus assessing the predictive validity of WTPs elicited with the BDM. As a result, the BDM-based WTPs reveal a sufficient degree of internal face and criterion validity. However, the external validity in terms of predictive validity between WTP-based prediction and purchases of the validation sample seems limited. Specifically, this study found a substantial overestimation of WTP, and thus in the corresponding purchase rates in the BDM. Hence, a potential bias is indicated. However, contrary to the suggestions of earlier research, individual risk attitude or loss aversion, hence a potential gambling effect, seems not to bias BDM results or the decision whether to buy or not.