Abstract:
We investigate the stability of brand and price attribute tradeoffs in choice experiments for two
product categories under conditions in which seven other attributes are systematically
included/excluded according to a master experimental design. Choice model estimates from 64
different experimental conditions were converted to willingness-to-pay (WTP) units to control for
error variance differences between conditions and allow inter-attribute utility comparisons.
Statistical analysis of the resulting 64 sets of WTP estimates for all attributes (except price, which
is used in the WTP calculations) demonstrates that even after controlling for variance
differences, there are large, systematic differences in utility estimates depending on which
attributes are included/excluded in choice experiments. Implications for future research are
noted.
