Abstract/Summary

Understanding individual behavioural differences could enhance welfare research. Many methods for assessing fear to make inferences about ‘personality’ have been proposed but not validated for goat kids. The study's primary aims were to 1) investigate individual-level test-retest reliability of Novel Object (NO) and Familiar Person (FP) tests; 2) establish the effects of testing environment (Modified Home (MH) or Unfamiliar Testing (UT) arenas) and weaning method (Gradually (GW; n=18) or Abruptly (AW; n=17)), and 3) test for associations between repeatable behaviours and production metrics (e.g. growth). A secondary aim was to assess interobserver reliability. One AW pen and one GW pen were used for each testing environment (MH/UT). Four NO tests were conducted, two pre-weaning (25d and 29d), and two post-weaning (63d and 67d); FP tests were conducted the day after each. Tests lasted 180 s (+90 s habituation for NO tests) and were recorded. An observer blinded to weaning treatment scored all videos, and another scored 1 kid’s test for each timepoint/pen. One kid from MH and seven from UT environments were removed from testing and subsequent analysis due to distress. Intraclass coefficient calculations indicated good interobserver reliability (W1=0.670, p<0.001). For FP tests ‘bipedal stance’ (W3=0.379 p<0.001) and ‘stand still-look’ (W3=0.378 p=0.010) and for NO tests bipedal stance (W30.234 p=0.006) and ‘latency until contact’ W3=0.202 p=<0.001 showed test-retest repeatability. MH and UT environments were compared using Kendall’s W (calculated for each behaviour separately), the coefficients were compared with Mann-Whitney U tests, which found no impact of testing environment (p=0.579). GLMMs assessed the effect of external factors pen, test type, and weaning treatment; concordance coefficients between individual kids’ tests were the dependent variable with test type, kid age and pen as independent variables. The model of best fit was selected on the basis of lowest Akaike Information Criterion (AIC); test type (p=0.184) did not predict individual consistency, but pen did (p=0.002). Further GLMMs found no effect (p>0.05) of ‘personality’ indicators on weight gain, milk intake/day, nor relationship with weaning treatment, but a significant effect of bipedal stance on milk feeds/day (p=0.04) was identified. Results indicate that ‘bipedal stance’ in both tests, ‘stand still-look’ in FP tests, and ‘latency to interact’ in NO tests have some repeatability, therefore could be ‘personality’ indicators. Statistical analysis found that testing environment did not affect fear among kids tested, but in practice the unfamiliar environment resulted in more distress-related removals therefore a modified home pen is recommended.