Research Article: A multidimensional measure of animal ethics orientation – Developed and applied to a representative sample of the Danish public

Date Published: February 7, 2019

Publisher: Public Library of Science

Author(s): Thomas Bøker Lund, Sara Vincentzen Kondrup, Peter Sandøe, Matthew Parker.


We present a questionnaire-based measure of four animal ethics orientations. The orientations, which were developed in light of existing empirical studies of attitudes to animal use and ethical theory, are: animal rights, anthropocentrism, lay utilitarianism, and animal protection. The two latter orientations can be viewed as variants of animal welfarism. Three studies were conducted in Denmark in order to identify the hypothesised orientations, evaluate their concurrent validity, and report their prevalence and relevance in animal-related opinion formation and behaviour. Explorative factor analysis (Study 1) and confirmative factor analysis (Study 2) successfully identified the four orientations. Study 2 revealed good measurement invariance, as there was none or very modest differential item functioning across age, gender, living area, and contrasting population segments. Evaluation of concurrent validity in Study 2 found that the orientations are associated with different kinds of behaviour and opinion when the human use of animals is involved in the hypothesised directions. In Study 3, a representative population study, the animal protection orientation proved to be most prevalent in the Danish population, and as in study 2 the four orientations were associated with different behaviours and opinions. Remarkably, the animal protection orientation does not lead to increased animal welfare-friendly meat consumption, the main reason for this being non-concern about the current welfare status of farm animals. We argue that the developed measure covers a wide range of diversity in animal ethics orientations that is likely to exist in a modern society such as Denmark and can be used in future studies to track changes in the orientations and to understand and test hypotheses about the sources and justifications of people’s animal-related opinions and behaviours.

Partial Text

To discover whether the four animal ethical orientations could be identified in a convenience sample of university students, we developed a pool of potential scale items (in Danish) for each of the four anticipated orientations set out in Section 1.3.

The aims were to find out whether the factorial structure identified in Study 1 could be replicated in a diverse study population, to assess measurement invariance across socio-cultural subpopulations, and to test concurrent validity. In addition to examining invariance across socio-demographic segments we examined it in two groups that are expected to be culturally and attitudinally remote: namely “meat avoiders” (i.e. vegans, vegetarians and semi-vegetarians) and people employed in the meat production sector (principally farmers, meat factory workers, butchers, and political-strategical/consultancy work in the meat sector). The recruitment of the study sample was designed so that these two groups were included along with an intermediate population of the general public in Denmark.

This study aimed: to confirm that the factorial structure identified in Study 1 and Study 2 could be replicated in a representative sample of Danes; to evaluate whether it is acceptable to use summated scores to construct the latent variables; to report scores on the four animal ethics orientations across socio-demographic segments in the Danish population; and to examine whether the four orientations are associated with different types of animal-related behaviour and opinion.

We set out to develop a measure of animal ethics orientations based on the hypothesised existence of four orientations. The orientations were identified and confirmed in three studies where explorative and confirmatory factor analysis was undertaken. We found that the orientations can be identified using 12 questionnaire items. At most, these items show very modest signs of DIF bias across a range of socio-demographic factors, implying measurement invariance. Further, we found that the four orientations can reasonably be constructed into variables by summing the raw scores of the factor-specific items. This makes the four orientations easy to reproduce and facilitates cross-study comparison of mean scores [40].




Leave a Reply

Your email address will not be published.