Figure 4.

Conditional inference tree on first principal component: cat's positive interactions. each node represents a dichotomous split based on a demographic factor that distinguishes between lower-responding and higher-responding observations for the first principal component. As an example, node 1 shows that having more than 2 cats generally was favorable and associated with higher scores for cat's positive interactions with the child; this was especially true for cats in Europe (node 22) that were females (node 23). When having 2 cats or less that are 6 years of age or less, nodes 7 and 8 show that for cats living with no more than one other cat and no older than 6 years, and obtained as a feral or from an ad, having children age 9–12 at home was associated with the cat's positive interactions whereas having children age 6–8 was not. Statistical tests are noted at each node.