Testosterone gets a bad rep. People assume that if one has higher testosterone than average, that they will be a savage, bloodthirsty beast with an insatiable thirst for blood. This, however, is not the case. I’ve documented how testosterone is vital for male functioning, and how higher levels don’t lead to maladies such as prostate cancer. Testosterone is feared for no reason at all. The reason that people are scared of it is that of the anecdotal reports that individual A had higher testosterone when he committed crime B so, therefore, anyone who commits a crime has higher testosterone and that is the ultimate—not proximate—cause of crime. This is erroneous. There is a positive—albeit extremely low—correlation between physical aggression and violence at .14. That’s it. Furthermore, most of these claims of higher levels of testosterone causing violence is extrapolated from animal studies to humans.
Testosterone has been shown to lead to violent and aggressive behavior, largely only in animal studies (Archer, 1991; Book et al, 2001). For years, the relationship between the two variables was thought to be causal, i.e., high levels of testosterone cause violent crimes, which has been called into question over recent years. This is due to how the environment can raise testosterone levels. I have documented how these environmental factors can raise testosterone—and after these events, testosterone stays elevated.
Largely, animal studies are used to infer that high levels of testosterone in and of themselves lead to higher rates of aggression and therefore crime. However, two important meta-analyses show this is not necessarily the case (Archer, 1991; Book et al, 2001). Book et al, 2001 showed that two variables were important in seeing the relationship between aggression and crime—the time of day that the assay was taken and the age of the participant. This effect was seen to be largest in, not unexpectedly, males aged 13-20 (Book et al, 2001: 594). So since age confounds the relationship between aggression and testosterone in males, that is a variable that must also be controlled for (which, in the meta-analyses and other papers I cite on black and white testosterone is controlled for).
More interestingly, Book et al (2001) showed that the nature of the measure of aggression (self-reported or behavioral) did not have any effect on the relationship between testosterone and aggression. Since there is no difference between the two measures, then a pencil-and-paper test is a good enough index of measure of aggression, comparable to observing the behavior of the individual studied.
Archer (1991) also showed the same low—but positive—correlations between aggression and testosterone. Of course, as I’ve extensively documented since there is a positive relationship between the two variables does not necessarily mean that high-testosterone men commit more crime—since the outcome of certain situations can increase and decrease testosterone, no causal factors have been detangled. Book et al (2001) confirmed Archer’s (1991) finding that the correlation between violent and aggressive behavior was positive and low at .14.
Valois et al (2017) showed there was a relationship between emotional self-efficacy (ESE) and aggressive and violent behaviors in a statewide sample of high school children in South Carolina (n=3,386). Their results suggested that there was a relationship between carrying a weapon to school within the past 30 days along with being injured with a club, knife or gun in the past 12 months was significantly associated with ESE for specific race and sex groups.
Black girls who reported a low ESE reported carrying a weapon to school 30 days prior to the survey were 3.22 times more than black girls with a high ESE who did not report carrying a weapon to school within the past 30 days prior to the questionnaire. For black boys with low ESE, they were 3.07 times more likely to carry a weapon to school within the past 30 days in comparison to black boys with high ESE who did not carry a weapon to school in the past 30 days. White girls who reported low ESE had the highest chance of bringing a weapon to school in comparison to white girls with low ESE—they were 5.87 times more likely to carry a weapon to school 30 days prior to the survey. Finally, white boys with low ESE were slightly more than 2 times more likely than white boys with high ESE to carry a weapon to school 30 days prior to the survey.
Low ESE in white and black girls is associated with carrying a weapon to school, whereas low ESE for white and black boys is associated with being threatened. Further, their results suggested that carrying a weapon to school was associated with low ESE in black and white girls suggesting that low ESE is both situation-specific and specific to the female sex. The mediator between these things is low ESE—it is different for both black boys and black girls, and when it occurs different courses of action are taken, whether it’s through bringing a weapon to school or being threatened. What this tells me is that black and white boys with low ESE are more likely to be threatened because they are perceived to be more meek, while black and white girls with low ESE that get provoked at school are more likely to bring weapons. So it seems that girls bring weapons when provoked and boys fight.
The two meta-analyses reviewed above show that there is a low positive (.14) correlation between testosterone and aggression (Archer, 1991; Book et al, 2001). Thusly, high levels of testosterone on their own are not sufficient enough to explain high levels of aggression/violence. Further, there are race- and sex-specific differences when one is threatened at high school with black and white boys being more likely to report being threatened more (which implies a higher rate of physical fighting) while black and white girls when threatened brought weapons to school. These race- and sex-specific differences in the course of action taken when they are physically threatened needs to be looked into more.
I’d like to see the difference in testosterone levels for a matched sample of black and white boys from two neighboring districts with different murder rates as a proxy for the amount of violence in the area. I’d bet that the places with a higher murder rate would have children 1) report more violence and instances of bringing weapons to school and 2) report more harm from these encounters—especially if they have low ESE as seen in Valois (2017) and 3) the children in the high schools along with the residents of the area would have higher testosterone than the place with less violence. I would expect these differences to be magnified in the direction of Valois (2017) in that areas with higher murder rates would have black and white girls report bringing weapons to school when threatened whereas black and white boys would report more physical violence.
High testosterone itself is not sufficient enough to explain violence as the correlation is extremely low at .14. Testosterone levels fluctuate depending on the time of day (Brambilla et al, 2009; Long, Nguyen, and Stevermer, 2015) to the time of year (Stanton, Mullette-Gillman, and Huettel, 2011; Demur, Uslu, and Arslun, 2016). How the genders/races react differently when threatened in adolescence is interesting and deserves further study.