People don’t understand the relationship between testosterone, aggression, and crime. People hear the sensational media stating that testosterone causes crime, aggression, and anger. However, I have written numerous articles on this blog on the true nature of testosterone, what it’s really needed for and why we need it in high amounts. I’ve mused a lot on this hormone, which is one of my favorites to discuss due to the numerous misconceptions surrounding it.
Which way does causation run in regard to prisoners and their testosterone level?: heightened testosterone > aggression > violence or aggression > heightened testosterone > dominance > possibility (not necessarily, as I have written myself in the past) of violence.
People may use animal studies in support of their contention that testosterone causes aggressive behavior. However, for reasons I have discussed in the past, animal models only show avenues for future research and do not necessarily mean that this is the case for humans (as Mazur, 2006 point out). I don’t use animal studies. They’re good for future research, but to use them as evidence for causation in humans doesn’t make sense.
People may cite Dabbs et al showing that the more violent prisoners had higher levels of testosterone and therefore conclude that higher levels of testosterone drive the violent crime that they commit, however it is much more nuanced than that.
Does being a violent criminal raise testosterone or are violent people more likely to have high testosterone? Dabbs never untangles this; they just showed a correlation, which is small as evidenced by my other citations.
Testosterone is, as evidenced by numerous studies, related to dominance and dominance contests, however, during these dominance contests “a killing is rarely the outcome of a violent dominance contest” (Mazur, 2006: 25). Therefore, this throws a wrench in the testosterone-causes-crime hypothesis.
Some individuals may state that these dominance contests then lead to violence, however, as Mazur (2006) puts it: “Heightened testosterone is not a direct cause of male violence.”
Other animals assert dominance violently but we, necessarily, do not.
Mazur (2006) states that dominance contests rarely escalate to murder. Mazur also states that dominance contests also lead to increased T for the winners and decreased T for the losers, and these contests also don’t necessarily lead to murder/violent behavior. There is a feedback loop with high T causing behavior and behavior causing high T (Mazur, 2006) while this feedback loop may lead to “lethal effects” (Mazur and Booth, 2008).
It’s worth noting that Mazur seems to advocate for ‘testosterone-depressing drugs’. He concludes:
There are strong linkages between macro-level culture and the physiology of
individuals. We may find solutions to some of our social problems by altering these linkage.
Macro-level culture being white culture, black culture, Asian culture, etc.
The physiological differences are due to the preparation for dominance contests. So, his hypothesis goes, the culture of dominance among young black men with no education is why their T is so high. That low education was also associated with low education lends credence to the claim that this is changeable.
However, in his newer article on education, low testosterone and blacks he advocates for more sensible solutions (attempting an environmental change). I don’t know about you but I have big problems with using FDA/Big Pharma drugs to ‘reduce societal problems’, and it seems that Mazur has changed his view there. Mazur (2016) writes:
If high T does facilitate the high violence rate among young black men, there would be a troubling policy question of what, if anything, to do about it. Any notion of a medical or pharmaceutical fix, rather like prescribing Ritalin for hyperactivity, would reek of race-based chemical castration and should be regarded as outside the pale. However, social interventions might be workable and ethically acceptable.
I have railed against measures like this in the past, since proposing measures to attempt to ‘decrease crime through supposedly decreasing one of the main “causes”‘ is very Brave New World-ish, and I am highly against those measures. Social interventions are, in my view, the more sensible measures to undertake.
In regard to low education and testosterone, this same relationship was noticed by Assari, Caldwell, and Zimmerman (2014) where they note that testosterone was not associated with aggression in men, but low education was, which Mazur (2016) replicates, showing that blacks of the same age group with more education had lower levels of testosterone when compared to age-matched blacks. Mazur (2016) cites one study in support for his contention that education can decrease aggressive behavior (Carre et al, 2014)
The correlation is there, I agree. let’s take the middle value of .11 between Archer, Graham-Kevan, and Davies, 2005 at .08; and Book, Starzyk, and Quinsey, 2001 .14. So testosterone explains 3 percent of the the relationship with aggression. Not high at all.
Great evidence against the testosterone-causes-aggressive-behavior hypothesis are data on the Yanomami. About 50 percent of Yanomami men meet their deaths by other Yanomami men. So the Yanomami must have testosterone levels through the roof, right? Wrong. De Lima et al (2015) write:
We observed that Yanomamis present lower levels of testosterone (414 ng/dL) in relation to other ethnic groups (502/512 ng/dL), but still within normal limits (350-1000 ng/dL).
(Note that these values for “normal limits” changed, going into effect at the end of July.)
The Yanomami with an extremely high murder rate with nowhere near a modern society have T levels on the lower end of our range. So….. The Yanomami example is direct evidence against the assertion of testosterone directly causing crime, as some people assert (it is even evidence against an indirect cause). The evidence of the Yanomami having testosterone levels near our lower range is direct evidence against the testosterone/crime hypothesis. Clearly, other variables drive the high violence rate in this society that are not testosterone. More interestingly, these people have had little contact with Western societies, and their T levels are still low compared to ours despite constantly being vigilant for threats from other Yanomami.
Most dominance contests do not end violently in the first-world, there is numerous evidence to attest to this fact. So with the low correlation between testosterone and aggression (Book, Starzyk, and Quinsey, 2001; Archer, Graham-Kevan and Davies, 2005; Book and Quinsey, 2005), along with dominance contests rarely ending in murder/violent crime, then there are way more factors influencing these phenomena.
So the feedback loop goes: Testosterone rises in expectation of a challenge which then, after the dominance contest (which doesn’t always necessarily lead to violence), it affects both individuals differently depending on whether or not they won or lost that dominance contest and these values then persist over time if the dominance contests continuously end up the same.
Let’s say, for argument’s sake, that testosterone is a large cause for aggressive behavior in lower-educated blacks, what should be done about it? Mazur cites evidence that behavioral interventions seem to work to decrease violent behavior during certain circumstances (Carre et al, 2014), so that is a good way to lower violence in populations that have low education.
So heightened testosterone does lead to dominance which then facilitates a dominance contest between two individuals which does not necessarily lead to crime and aggressive, violent behavior (this outcome is rare in dominance contests among “higher primates” [Mazur’s words]) so, therefore, while testosterone does facilitate dominance contests, it rarely leads to violence in our species. Therefore, testosterone does not cause aggressive behavior and crime, but it does cause dominance which, for the most part, do not always result in violent, aggressive, murderous behavior.
I’ve shown that Mazur replicated other analyses that show that testosterone and aggressive behavior are related to lower education. Testosterone wasn’t associated with aggressive behavior in Assari, Caldwell, and Zimmerman’s (2014) study, and, as Mazur (2016) replicates, low education was. So one way to end this relationship is to educate people, as shown by Carre et al (2014), and with this education, crime will begin to fall. Heightened testosterone is not a direct cause of male violence.
(Note: I also believe that other factors such as sleep and depressed nutrition play a factor in crime, as well as racial differences in it. See Birch, 1972; Liu et al, 2003; Liu et al, 2004; Walker et al, 2007; Galler et al, 2011, 2012a, 2012b; Spratt et al, 2012; Gesch, 2013; Kuratko et al, 2013; Waber et al, 2014; Raine et al, 2015; Thompson et al, 2017 for more information, I will cover this in the future. I’m of course not daft enough to believe that no genetic differences between individuals/populations are the cause of a lot of crime between them, however, as I have laid out the case in regard to testosterone and MAOA numerous times, these two explanations for both individual differences in crime as well as racial differences in crime leave a lot to be desired. Other genetic factors, of course, influence these differences, however, I am only worried about refuting the popular notions of ‘testosterone and MAOA, the ‘warrior gene” cause crime. The relationship is a lot more nuanced as I have provided mountains of evidence for.)