Abstract:
|
Empirical demand functions are critical in many areas of marketing. Many times, these demand functions are estimated using survey-based methods like conjoint analysis. With any survey-based method, the quality of the estimated model is dependent on the quality of the information obtained from the respondents. We develop a model that classifies and reweights respondents based on the statistical information contained in the respondents' survey questionnaires. We illustrate our model using simulated data and several data from empirical conjoint studies. Our approach provides an automated way of determining which respondents are relevant for model estimation. Importantly, we show that reweighting the respondents based on the quality of their information can lead to substantial different managerial insights than by treating each respondent equally in model estimation.
|