Assessing Conjoint Analysis Internal Validity: the Effect of Various Continuous Attribute Level Spacings - Archive ouverte HAL Access content directly
Journal Articles International Journal of Research in Marketing Year : 1989

Assessing Conjoint Analysis Internal Validity: the Effect of Various Continuous Attribute Level Spacings

(1) ,
1
Renée Darmon
  • Function : Author
  • PersonId : 830930

Abstract

Marketing research studies using conjoint analysis typically involve many attributes (some continuous, other discrete), at two or more levels. In the case of continuous attributes at three or more levels, the most frequent research practice has been to select equally spaced levels which encompass the relevant attribute range. This paper investigates the impact of selecting equally vs. unequally spaced attribute levels upon the quality of the recovery of a respondent's utility functions, depending upon (1) some basic underlying utility function characteristics (such as the function's range, shape, and curvature), and (2) current conjoint analysis estimation procedures. The study's methodology involves an artificial data base in which the utility functions of 64 simulated subjects for three levels of three different attributes, have been systematically varied in terms of their shapes, ranges, and curvatures. The results shows substantial differences in the various aspects of the utility functions' recovery. The implications on the internal validity of the conjoint analysis methodology as well as for the researchers using the technique and the various estimation procedures are discussed.

Keywords

Dates and versions

hal-00537869 , version 1 (19-11-2010)

Identifiers

Cite

Dominique Rouzies, Renée Darmon. Assessing Conjoint Analysis Internal Validity: the Effect of Various Continuous Attribute Level Spacings. International Journal of Research in Marketing, 1989, Vol.6 Issue 1, pp.35-44. ⟨10.1016/0167-8116(89)90045-1⟩. ⟨hal-00537869⟩

Collections

HEC CNRS
75 View
0 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More