Judgement bias testing in group-housed gestating sows

Societal concerns about animal welfare have triggered the movement of gestating sows from individual stalls to group housing in many countries. Common methods of assessing sow welfare focus on overt physical ailments, and potentially neglect psychological stressors. A judgement bias task may allow researchers to evaluate an animal's subjective mental or affective state to provide a more comprehensive welfare assessment. Thus, group housed sows were trained to a spatial differentiation task to evaluate their ability to be assessed for individual judgement bias. A total of 45 sows were trained to the task across two replicates, with 24 successfully meeting the learning criteria required to be tested for a judgement bias. In the first replicate, 60% of sows displayed positive bias while 40% displayed negative biases. In the second replicate, 52% of sows displayed positive biases while 33% of sows displayed negative biases. A linear mixed effects model revealed that feed rank affected the latency to approach the ambiguous stimulus (chi(2) (1) = 9.47, p = 0.002) with more dominant animals being more likely to exhibit a positive bias. Given that all sows in the present study were group housed, as well as fed and managed similarly, these findings highlight the complexities underlying judgement bias outcomes.