The Potential of Conjoint Analysis for Communication Research

Abstract

Surveys on choices oftentimes rely on Likert-type scales. However, they prevent researchers from identifying the relative importance of different stimulus features for the choice decision. Here, it is shown that conjoint analysis offers a variety of advantages emerging from its use of questions that force the subject to trade off conflicting attribute levels. This study discusses some problems of Likert-type scale measures and shows when conjoint analysis may solve some of these problems. Differences between measurements are illustrated in a uses and gratifications study, assessing convergent validity on Likert-type scale measures with traditional conjoint analysis and choice-based conjoint analysis. Results show that subjects prefer “cognitive needs” by Likert-type scale measures against “tension release needs” by both conjoint analyses. The different outcome is independent from linguistic variation and from dual-process models of information processing. Implications for communication research are discussed.

Keywords:

• Conjoint Analysis
• Likert Scales
• Research Methods
• Validity

Acknowledgments

This article is based on a study that was part of the master’s thesis by Elena Pelzer. I would like to thank Jens Woelke and Armin Scholl for their insightful suggestions regarding this research and Benjamin O. Turner and W. P. Moerje for comments on an earlier version of this paper.