The effect of perceived environmental background on qualitative assessments of pig behaviour
Publication Type: |
Journal Article |
Year of Publication: |
2009 |
Authors: |
Françoise Wemelsfelder, Ian Nevison, Alistair B. Lawrence |
Publication/Journal: |
Animal Behaviour |
Keywords: |
animal welfare, free choice profiling, generalized procrustes analysis, on-farm welfare assessment, pig, qualitative behaviour assessment, sus scrofa |
ISBN: |
0003-3472 |
Abstract:
Qualitative behaviour assessment is an integrative methodology that characterizes behaviour as a dynamic, expressive body language (e.g. as anxious or content). Such assessments are sensitive to environmental context, which makes them informative but also vulnerable to observers’ biased views of that context. This study investigated whether and how perceived environmental background affects observers’ qualitative assessments of pig, Sus scrofa, behaviour. Fifteen growing pigs were filmed individually against a neutral background while interacting with a novel object. The footage of each pig was digitally isolated from that background and pasted against indoor and outdoor backgrounds filmed in real time. The 30 video clips thus obtained were shown to 16 observers, who were led to believe these were 30 different pigs filmed in either an indoor or an outdoor pen. Free-choice profiling was used to instruct observers in qualitative behaviour assessment, and data were analysed with generalized Procrustes analysis. Analysis of variance found a significant effect of environmental background on pig scores on the second consensus dimension (confident/content-cautious/nervous), but not on the first (playful/active-bored/lethargic). However, 95% confidence intervals and indexes for the variability attributable to environmental background, calculated for both consensus dimensions, indicated that any such effects should be relatively small. High correlations were found between indoor and outdoor pig scores on both consensus dimensions (r >= 0.95). Together these results suggest that environmental background may slightly shift, but is unlikely to seriously distort, observer characterizations of pig expression. Last, we discuss possible strategies for reducing the effect of contextual bias on qualitative behaviour assessment.