John Archer makes a strong argument for an evolutionary basis of sex differences in aggression. His thesis is that, starting in early childhood, sex differences in aggressive behavior exist, and, although these are mediated by social influences, they are underlain by biological variables. One of these variables is the steroid hormone testosterone. Archer's conclusions regarding testosterone and adolescence need some refinement.
As Archer himself points out, data of self-reported aggression in male adolescents do not support findings in testosterone administration studies on aggression. He argues that, although exogenous testosterone seems to enhance proneness to aggression, the rising levels of testosterone in male adolescents are not reflected in self-reported aggression measures. However, behavioral studies suggest that testosterone is a mediator of adolescent aggression. James Dabbs and colleagues showed repeatedly, in a line of studies in the nineties, associations of testosterone and violent criminal behavior. Imprisoned young males with high salivary testosterone were substantially more frequently convicted for aggressive crimes like violence and rape, and they showed more violent behavior (Dabbs et al. Reference Dabbs, Carr, Frady and Riad1995). This was also replicated in studies on women (Dabbs & Hargrove Reference Dabbs and Hargrove1997). Interestingly, in late adolescent males the hormone cortisol mediated the correlation between testosterone and aggressive behavior, which was found only in imprisoned adolescents with low cortisol levels (Dabbs & Jurkovic Reference Dabbs and Jurkovic1991).
More recently, Popma and colleagues (Popma et al. Reference Popma, Vermeiren, Geluk, Rinne, Van den Brink, Knol, Jansen, Van Engeland and Doreleijers2007) showed a correlation of testosterone and self-reported measures of violent behavior, but again, this was mediated by cortisol. Designed to investigate the relation between testosterone, cortisol, and aggression in early adolescence, their study pointed to an effect of testosterone on overt aggression only when cortisol levels were low. Confirming this, Hermans et al. (Reference Hermans, Ramsey and van Honk2008) found in an fMRI study increased activity in the hypothalamus, amygdala, and orbitofrontal cortex in response to angry facial expressions. This network of brain structures is considered vital in human reactive aggression. Importantly, activity in the subcortical part of this network, namely, the hypothalamus and amygdala, in response to angry faces proved to be related to the ratio between testosterone and cortisol.
Based on these findings one should consider taking basal levels of cortisol into account when comparing groups on aggressive behavior. Especially during adolescence, a highly stressful period (as demonstrated hormonally by marked increases in HPA-activity [Gunnar et al. Reference Gunnar, Wewerka, Frenn, Long and Griggs2009] and, behaviorally, by the onset of several stress-related psychiatric illnesses [Paus et al. Reference Paus, Keshavan and Giedd2008]), the rise of testosterone levels (alone) in boys relative to girls will not necessarily result in a relative increase of aggressive behavior.
Another issue of consideration is testosterone in early development. Bailey and Hurd (Reference Bailey and Hurd2005), for instance, have shown that prenatal levels of testosterone, reflected in the 2D:4D digit length ratio, may mediate testosterone-aggression relationships. In males, higher prenatal testosterone levels correlated with physical aggression in adulthood. Interestingly, recent evidence from testosterone administration research in humans suggests that high prenatal testosterone levels increase sensitivity to behavioral effects of testosterone in later life (van Honk Reference van Honk2009). Furthermore, during early puberty (mean age 11.9 years), clear volumetric sex differences were found in brain areas mediating aggression (i.e., amygdala, striatum, rostral anterior cingulate cortex, and superior temporal gyrus; male volume>female volume) (Peper et al. Reference Peper, Brouwer, Schnack, Van Baal, Van Leeuwen, Van den Berg, Delemarre-Van de Waal, Boomsma, Kahn and Hulshoff Pol2009). However, testosterone levels at this age could not explain these brain morphological sex differences. It might therefore be argued that a possible influence of testosterone on brain areas involved in aggression has a prenatal or early postnatal origin.
In conclusion, testosterone is unmistakably involved in human aggression and contributes importantly to sex differences in aggressive behavior. These sex differences, however, seem to originate before puberty. The relative increase of testosterone levels in adolescent boys and its relation to aggressive behavior is obscured by at least two mediators: high testosterone-sensitivity due to high prenatal testosterone levels and, especially during adolescence, levels of basal cortisol. Taking these factors into account, increased levels of testosterone enhance aggressive behavior in both adolescent boys and girls. Thus, it seems that the link between testosterone and aggression in adolescents is maintained. The here-described relations between prenatal, adolescent and adult levels of testosterone, together with results found after testosterone administration, are in support of Archer's hypothesis that sex differences in aggression are a result of sexual selection.
