Reliability of temperament tests on finishing pigs in group-housing and comparison to social tests

Abstract

If variation in behaviour is consistent within individuals and reliably associated with relevant welfare measures, this variation could aid in genetic selection, or development of management schemes designed to improve welfare. In this study, we focused on temperament tests in group-housed finishing pigs, with the objective of validating measures that are readily applicable in commercial settings and potentially related to individual differences in stress response. A total of 118 pigs in two replicate studies were housed in pens of 7–8 pigs per pen. At 24 weeks of age, animals were subjected to three tests of fear in the home pen: the human approach test (HAT), novel object test (NOT) and open door test (ODT). In each test, pigs were scored on their latency to contact the human or object, or to leave the home pen. Tests were repeated on three occasions at intervals of 3–4 days. On each test day the HAT was performed twice to compare results between different handlers. Behaviour in tests of fear was compared with two social tests in a subset of 58 pigs. Lesion scoring was performed after mixing as a measure of aggression, and a feed competition test was used to assess social status. Repeatability within-test was evaluated in a mixed model with pig as a random effect, and agreement between days, handlers and tests was evaluated by partial correlations after controlling for replicate, pen and day effects. Latency to perform all three fear tests decreased significantly over time. Correlations within-test showed significant agreement between all days for HAT and ODT, and between HAT handlers. Between tests, the HAT and ODT were correlated, and ODT and NOT tended to correlate. Comparisons between group fear tests and social tests showed that pigs which readily approach a human tended to have higher lesion scores and fewer feeding bouts in the feed competition trial. While test latencies decrease with repetition, the HAT and ODT show individual consistency over time, which suggests that these tests describe behavioural tendencies and may be useful for predicting fearful responses at slaughter.