Environmental Enrichment and Forced-Exposure Training With Pigeons: Responding to the Replication Crisis

Recent attention to the reliability of scientific literature has focused renewed attention on the role of replications. The present study aimed to evaluate 2 variables that appear to affect suboptimal choice in pigeons but have yet to be replicated. Pigeons were presented with a standard suboptimal choice task in which the suboptimal alternative led to a signaled delivery of food 50% of the time and the optimal alternative always led to food. The type of housing (enriched vs. isolated) and type of training (presence vs. absence of forced-exposure trials) was manipulated. Based on previous research, pigeons housed in isolation and trained with forced-exposure trials were predicted to make the most suboptimal choices of the four groups. Instead, birds in enriched housing and trained with forced-exposure trials chose the suboptimal alternative most frequently. A balanced 2 x 2 ANOVA found that both the main effect of housing, F(1, 8) = 4.52, p = .066, ω2 = .18, and training, F(1, 8) = 5.21, p = .052, ω2 = .22, did not reach significance, nor did their interaction, F(1, 8) = 0.37, p = .562, ω2 < .01. However, Bayes factors indicated weak evidence in support of both the main effect of housing (BF10 = 1.22) and the main effect of training (BF10 = 1.46), but not the interaction (BF10 = 0.5). The present results highlight the need to determine the degree to which initial results are reliable and generalizable prior to becoming cited and viewed as an established finding.