A test of the generalizability of the interval scale developed for social behavior of rhesus monkeys.

Abstract

An interval scale of behavior would be useful in the study of social relations because it would allow the whole behavioral repertory to be used simultaneously and would allow greater flexibility in statistical manipulation. Maxim has developed such a scale using isolated dyads of unfamiliar monkeys. The typical social environment of monkeys, however, consists of familiar monkeys in mixed age and sex groups. This study replicates Maxim's procedure and tests the applicability of a scale developed with isolated dyads to groups of familiar monkeys. The scale developed here differed substantially from Maxim's. Behavior of monkeys in these groups was much more variable than behavior of monkeys in Maxim's isolated dyads. Behaviors at the extremes of the scale were similarly arranged in both scales, but behaviors in the central region were more labile and disordered. Distances between behaviors at the extreme of this scale were significantly larger than those in the central region whereas there were no differences in Maxim's scale. While some of these differences may be accounted for by group size and familiarity, several difficulties remain. Some behaviors may change meaning across different contexts and social relationships. The generalizability of these scales, therefore, may be limited to the behavior of groups of similar composition and familiarity.