Studies of the influence of male sex hormones on aggression and violence have had inconsistent results (for reviews see Meyer-Bahlburg, 1981; Rose, 1975; Rubin, Reinisch, & Haskett, 1981). The empirical bases are rather weak for two commonly held assumptions: that aggressive behavior in animals is determined by testosterone level and that testosterone also activates violent and aggressive behavior in humans.

The animal research gives little evidence for a direct effect of sex hormones on fighting but indicates a relation between changes in social rank or status and dominant–submissive behavior and changes in testosterone levels, at least in monkeys (Rose, Holaday, & Bernstein, 1971) and mice (Henry & Stephens, 1978). Keverne, Meller, and Eberhart (1982) found that high-rank monkeys had a 300–500% increase in testosterone when introduced into a mixed-sex group. They displayed (as expected) higher sexual activity than the low-rank animals, who had only a slight increase in testosterone. Although they did not show more aggressive behavior, the high-rank, high-testosterone monkeys were the objects of significantly less aggression by other males. When placed again in single cages, they no longer differed in testosterone level from the low-rank monkeys. Nor were there any differences between high- and low-rank monkeys when they were separately introduced to females, with no other males present. In the latter situation, the only endocrine difference observed was higher “stress” hormone (cortisol and prolactin) levels in the low-rank animals.