Are Utilitarian/Deontological Preferences Unidimensional?

Utilitarian versus deontological inclinations have been studied extensively in the field of moral psychology. However, the field has been lacking a thorough psychometric evaluation of the most commonly used measures. In this paper, we examine the factorial structure of an often used set of 12 moral dilemmas purportedly measuring utilitarian/deontological moral inclinations. We ran three different studies (and a pilot) to investigate the issue. In Study 1, we used standard Exploratory Factor Analysis and Schmid-Leimann (g factor) analysis; results of which informed the a priori single-factor model for our second study. Results of Confirmatory Factor Analysis in Study 2 were replicated in Study 3. Finally, we ran a weak invariance analysis between the models of Study 2 and 3, concluding that there is no significant difference between factor loading in these studies. We find reason to support a single-factor model of utilitarian/deontological inclinations. In addition, certain dilemmas have consistent error covariance, suggesting that this should be taken into consideration in future studies. In conclusion, three studies, pilot and an invariance analysis, systematically suggest the following. (1) No item needs to be dropped from the scale. (2) There is a unidimensional structure for utilitarian/deontological preferences behind the most often used dilemmas in moral psychology, suggesting a single latent cognitive mechanism. (3) The most common set of dilemmas in moral psychology can be successfully used as a unidimensional measure of utilitarian/deontological moral inclinations, but would benefit from using weighted averages over simple averages. (4) Consideration should be given to dilemmas describing infants.