Mind your language! Lessons from the application of an English published version of a Japanese horse personality instrument to a French population

Abstract

Replicability is a fundamental tenet of the scientific method and scientific reporting, but there is a preponderance to publish scientific research in English to increase international recognition, regardless of the country of origin of the research. Questionnaires are widely used to assess personality in animals. These psychometric instruments are mainly published in English but can be used all over the world in other languages. However, without safeguards relating to the translation process, the replicability of the quality of the instrument may change from its originally reported value. This study focuses on the particular issue of cross-cultural reliability of psychometric instruments used for assessing animals that have been translated from their original context. We examined the replicability of the structure of a personality scale originally used in Japanese (but reported in the English literature) on an English population (n=100), and then the reliability of the structure of a French translation of the English version with additional translational safeguards (e.g. back-translation and sense checking) on a French population (n = 159 horses). Horses were rated by 3 evaluators to also allow calculation of inter-rater reliability. We found that there was greater reliability and similarity of structure between the adapted English translation and French version of the Japanese scale, than with the originally published structure of the instrument used in Japan. These results highlight the importance of never assuming the reliability and thus validity of semantic instruments used to assess animal behaviour which have been published in a different language to that in which they were originally developed.