Development of Qualitative Behaviour Assessment (QBA) in farmed fox species and Finnraccoons

Measures of emotional states have not gained much attention in farmed fox species (blue foxes and silver foxes) and Finnraccoons. This study aims to start the development of Qualitative Behaviour Assessment (QBA), to be later included in to the on-farm welfare assessment scheme, WelFur. To collect the potential QBA terms, two separate expert opinion surveys were carried out. Based on the questionnaire surveys, a fixed list of 26–27 terms (depending on the species) was generated for each species. To test the observer agreement and to investigate potential latent factors describing animal welfare, eleven persons assessed behavioural expressions of the fox species and Finnraccoons from 13 to 15 1–2 min video recordings from commercial farms. The consistency of the use of terms was studied by using Kendall’s coefficient of concordance and the data reduction was carried out by using Principal factor analysis. In all species, the inter-and intra-rater reliabilities of the individual terms were mainly moderate (Kendal W 0.4–0.7). They were interpreted sufficient for proceeding to factor analysis. The factor solutions differed between the species. In blue foxes, the analysis produced factors including all four components of the commonly found model of emotional dimensions describing bipolar valence (positive vs negative) and arousal (low vs high). In silver foxes, negative low arousal dimension was missing and in Finnraccoon, the clear low arousal factor was not divided into positive and negative valence. Besides these dimensions, a factor indicating attentiveness or interaction (with human) was found in both fox species and an incoherent factor in Finnraccoons. The present results show that, the behavioural expressions of farmed fox species and Finnraccoons can be described by using a large variety of terms, and fixed list could be generated out of these terms. The observer agreement in the scoring of the individual terms from videos is at least moderate and the selected terms constitute logical dimensions of animal welfare in the data reduction analysis. The further development of the method requires testing of the vocabulary on farms. • Diverse terms describe behavioural expressions of fox species and Finnraccoons. • A fixed list of QBA terms could be generated out of the terms suggested by experts. • The observer agreement of the terms was mainly moderate in video data. • Factor analysis produce logical, but slightly different, dimensions in all species. • Components of valence and arousal are described in the dimensions of emotions.