Reliability and construct validity of three temperament tests in farmed Finnraccoons (Nyctereutes procyonoides)

The associations between the Stick test, Feeding test and Subjective evaluation of the temperament were tested in 214 juvenile Finnraccoons, raised in five different housing conditions, which varied in relation to group size and composition, available area, animal density and level of enrichment. The stick test and Feeding test were carried out once, but Subjective evaluation was carried out by six assessors, in pairs. In the Stick test, 57 % of the animals explored the novel stick in a curious way, 13 % bite the stick in an aggressive way and 3 % actively avoided the stick in a fearful way. In the feeding test, 79 % of the animals ate within 30 s in the presence of human, being interpreted as confident. The animals classified as Aggressive or Curious in the Stick test ate more probably in the Feeding test than those classified as Fearful. In the Subjective evaluation of the temperament, majority of the animals were classified as confident; the percentage of Confident animals varied from 60 % to 85 % between the six assessors. The inter-assessor reliabilities between the assessor pairs doing testing at the same time varied from poor to excellent (Cohen's kappa from 0.39 to 0.85). The higher percentage of assessors that classified the animal as Confident, the more probably the animal ate in the Feeding test. The agreement between the Subjective evaluation of the temperament and the outcome of the Stick test was poor (Cohen's kappa below 0.40). We could not detect any effect of the housing condition on the results. The results show that all tests measure the same aspects of temperament, especially fearfulness. In the Feeding test and Subjective evaluation of the temperament, the Finnraccoons' response is affected mainly by their confidence towards human.