Consistency, transitivity and inter-relationships between measures of choice in environmental preference tests with chickens

Choice experiments are a widespread method of assessing the welfare requirements of domestic animals but prior predictions about what animals should choose to maximise their welfare are not possible. It is therefore important to analyse the data from sets of choice tests to establish whether intra-set and inter-set consistency in choice occurs. Strong inter-set consistency (transitivity) would provide support for the assumption that animals are maximising a complex utility (welfare) that may map onto their subjective state. We housed 56 hens Gallus gallus sequentially in three environments, experienced as three sets (A vs B; B vs C; A vs C), over a 40-week period. At the end of each set, each chicken made six choices between the items in that set, using T-maze testing procedures. The chickens showed significantly greater within-set consistency than expected by chance (p < 0.001 for all sets). Although different chickens exhibited different environmental preferences, we also found that a high within-set consistency was associated with faster decision-making (p < 0.001). When the three sets of choice trials for each hen were compared there was some evidence that the birds made more transitive choices (indicating consistent choices between sets) than expected by chance.