Abstract
This simulation study investigates how various kinds of utility functions are recovered by different conjoint analysis procedures and designs under error-full and error-free conditions. The results support the hypotheses that (1) noise in preference judgments may impair conjoint reliability and validity differently depending on the choice of algorithm, data collection procedure, and utility function characteristics, (2) Ordinary Least Squares yields overall the best utility function recoveries, and (3) a full design method leads overall superior results when noise is present in respondents' judgments. Implications for marketing researchers are discussed.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
