Discrete choice experiments: a primer for the communication researcher

Abstract

Experiments are widely used in communication research to help establish cause and effect, however, studies published in communication journals rarely use discrete choice experiments (DCEs). DCEs have become a mainstay in fields such as behavioral economics, medicine, and public policy, and can be used to enhance research on the effects of message attributes across a wide range of domains and modalities. DCEs are powerful for disentangling the influence of many message attributes with modest sample sizes and participant burden. The benefits of DCEs result from multiple design elements including stimulus sets that elicit direct comparisons, blocked and/or fractional factorial structures, and a wide range of analytic options. Though sophisticated, the tools necessary to implement a DCE are freely available, and this article provides resources to communication scholars and practitioners seeking to add DCEs to their own methodological repertoire.

Keywords: balanced incomplete block designs; conjoint analysis (CA); discrete choice experiments; fractional factorial designs; message evaluation tasks.

FIGURE 1

Example attribute evaluation task. Here, the evaluations are collected using semantic differential scales rather than dichotomous selection.

FIGURE 2

Example stimulus-attribute evaluation task. Here, the evaluations are collected using semantic differential scales rather than dichotomous selection.

FIGURE 3

Example task from a stimulus evaluation design. Image generated with artificial intelligence software (Midjourney, 2024).

FIGURE 4

Example stimulus-comparison set (adapted from Thrasher et al., 2018a).

FIGURE 5

Example comparison set with stimuli representing combinations of features. Image generated with artificial intelligence software (Midjourney, 2024).