The paper illustrates a method to detect and estimate interaction effects in household decision-making and tests the representative member hypothesis by studying how preferences of single members influence family choices. We performed a stated choice experiment concerning hypothetical choices among alternative residential locations. We found that the rankings of the attributes for the family and for each member were the same. The estimated coefficients for each member were statistically different from family ones, which demonstrates the fallacy of the representative member hypothesis. We estimated single members’ relative power at the attribute level and extended a model developed in the literature to three-member groups.
