A Monté Carlo Simulation Study of Two Approaches for Aggregating Conjoint Data

The authors compare preference-share estimates generated by the simulation approach with those generated by a logistic model using identical data, and find that theLogistic model provides more efficient estimates than does the Simulation approach, and also proves to be robust to violations of its theoretical underlying distribution.

Abstract

The “simulation approach,” or fraction of first choice aggregation method, is widely used for aggregating conjoint data. The authors compare preference-share estimates generated by the simulation approach with those generated by a logistic model using identical data. The share estimates obtained by the two methods generally are equivalent, especially for conditions of greatest interest to marketers, i.e., when the objective is to estimate the share captured by the most preferred of several differentiated concepts. However, the logistic model provides more efficient estimates than does the simulation approach, and also proves to be robust to violations of its theoretical underlying distribution.