Keith E Rice's Integrated SocioPsychology Blog & Pages

Aligning, integrating and applying the behavioural sciences

Biological Factors in Crime #2

In 1980 Dan Olweus et al (1980) measured blood testosterone level in institutionalised delinquent and non-delinquent 16-year-old boys and assessed aggression using a questionnaire. High levels of self-reported physical and verbal aggression were associated with higher levels of testosterone – though the results were not statistically significant. It was also noted that those with higher levels of testosterone were likely to respond more vigorously in response to challenges from teachers and peers. John Archer (1991), in a meta-analysis of 5 studies covering 230 males, found a low positive correlation between testosterone and aggression. However, the type of participant and the form and measurement of aggression differed substantially between the studies. Angela Book, Katherine Starzyk & Vernon Quensy (2001), in a meta-analysis of 45 studies, found a mean correlation of 0.14 between testosterone and aggression – though John Archer, Nicola Graham-Kevan & Michelle Davies (2005) challenged Book, Starzyk & Quinsey’s findings on the grounds of methodological problems with the study which meant that a correlation of 0.08 was more appropriate.  James Dabbs et al (1987) measured salivary testosterone in 89 violent and non-violent criminals and found those with a history of primarily violent crime had the highest levels of testosterone whereas those with the lowest levels had committed only non-violent crime. However, the relationship between violence and higher testosterone levels was not very strong – again indicating other factors must have been involved too. However, James Dabbs et al found similar results in a similar study among 692 male prisoners in 1995. In 2001 James Dabbs et al conducted experimental studies with students and found that individuals with higher levels of testosterone were more assertive, direct and confident in their interactions with others. From this Dabbs et al speculated that this could explain the prevalence of the most cold-hearted and premeditated murders among the prisoners with the highest testosterone levels.

In a non-prison population study (1987) Ralf Lindman, Pertti Järvinen & Jan Vidjeskog found that young males who behaved aggressively when drunk had higher testosterone levels than those who did not act aggressively. Elena Kouri et al (1995) carried out a double-blind experiment in which young men playing a frustration-inducing computer game were given either testosterone or a placebo in injections over a 3-week period. Those dosed with testosterone pushed a button significantly more often to reduce the amount of cash received by a fictitious partner who was reportedly trying to reduce their cash. In 2000 Harrison Pope, Elena Kouri & James Hudson repeated the experiment almost exactly, with the differences that the testosterone and the placebo were administered over a 6-week period and with men aged between 20 and 50; they achieved very similar results.

A number of animal studies also support the role of testosterone in aggression. Eg: George Wagner, Leonard Beuving & Ronald Hutchinson (1979) found that, if a male mouse is castrated, aggression levels tend to reduce. However, if the same mouse is then given testosterone, aggression levels increase again. (See chart below.)

Graphic copyright © 2009 Nelson Thornes Ltd

Graphic copyright © 2009 Nelson Thornes Ltd

Robert Connor & Seymour Levine (1969) found that rats castrated at birth had reduced aggression as adults. However, rats castrated after puberty could be made aggressive with injections of testosterone while the rats castrated at birth could not. D A Edwards (1968) found that injections of testosterone in neonate female mice made them act with increased aggression when given testosterone as adults. However, control females only given testosterone as adults did not show this, suggesting that testosterone masculinises androgen-sensitive neural  circuits underlying aggression.

Allan Mazur & Alan Booth (1998) conducted a meta-analysis which suggests that testosterone leads men to try to dominate others and that men with higher levels of testosterone “are more likely to divorce or remain single, be arrested for offences other than traffic violations, buy and sell stolen property, incur bad debts and use weapons in fights”. However, the Mazur & Booth study also suggests that environmental experiences can act to reduce or increase levels of testosterone. From a study of 2,100 air force veterans given 4 medical examinations over a 10-year period, they found testosterone often decreased when the men were married and increased when they got divorced. These findings may be explained by the ‘Challenge Hypothesis’ put forward by John Wingfield et al (1990) which posits that testosterone secretion beyond a baseline level is caused in a primarily monogamous species such as humans by challenges to reproductive success – eg: male-to-male disputes over a woman or resources or status. This Evolutionary explanation is also an example of Reciprocal Determinism and epigentic modification.

Robert Sapolsky (1997) takes a full Sociobiological viewpoint, arguing that it is aggressive behaviour which leads to elevated testosterone levels. A 2006 study by Jennifer Klinesmith, Tim Kasser & Francis McAndrew offers support to this view. They found that men handling a gun, associated with violence, caused testosterone levels, assessed from saliva samples, to rise 100 times more than in men constructing a game of Mouse Trap. These results correlated to real, aggressive behaviour: the men handling the gun put 3 times as much chilli sauce into a drink for another participant as the men playing the game did.

In associating elevated testosterone with increased feelings of aggression, there needs to be some caution as increased feelings of aggression do not always lead to criminal behaviour. Therefore, the association is correlational, rather than causal. Jerald Bain et al (1987) found no significant differences in testosterone levels between men who had been charged with murder or violent assault and men who had been charged with non-violent crimes such as burglary.

According to Stephanie Van Goozen et al (2007), low levels of the hormone cortisol – higher levels are most commonly associated with elongated stress – are associated with higher levels of aggression. Studies have reported low levels of cortisol in habitually-violent offenders (Matti Virkkunen, 1985) and aggressive schoolchildren (Katherine Tennes & Maria Kreye, 1985).

The biological argument implicating cortisol could be two-fold…

  1. Having low ANS arousal (and, therefore, low cortisol levels) is unpleasant so the individual deliberately indulges in aggressive behaviour to stimulate ANS activation and cortisol release
  2. Cortisol plays an an important role in mediating aggression by inhibiting the likelihood of aggressive behaviour. A study by Arne Popma et al (2007) found a significant positive relationship between testosterone and overt aggression in adolescent males with low cortisol but not in those with high cortisol levels.

Support for Van Goozen et al comes from several longitudinal studies, such as Keith McBurnett et al (2000) who assessed 38 boys annually for 4 years. The boys, aged 7 to 12, had been referred to a clinic for disruptive behaviour. The researchers took salivary cortisol measurements during the second and fourth years and found the boys with lower cortisol concentrations exhibited 3 times the number of aggressive symptoms as boys with fluctuating or higher cortisol levels. The boys with lower cortisol levels were also consistently named ‘meanest’ by their peers at sampling time. However, not all studies support Van Goozen et al. Gilberto Gerra et al (1997) even reported higher levels of cortisol in participants with higher levels of aggression.

James Dabbs, Gregory Jurkovic & Robert Frady (1991) suggest that high levels of cortisol have a mediating effect on the aggressive impulses associated with testosterone by increasing anxiety and thereby the likelihood of social withdrawal. Therefore, low levels of cortisol will not have this inhibiting effect and thus increase the likelihood of aggressive behaviour. Justin Carre & Pranjal Mehta (2011) support this, terming the relationship the Dual Hormone Hypothesis.

Interestingly Paul Bernhardt (1997) suggests that it is an inverse relationship between testosterone (high) and serotonin (low) that leads to aggression. Bernhardt notes that testosterone is associated more with dominance than aggression and that low serotonin is linked to being overly responsive to aversive stimuli. Accordingly, when the drive to dominate of a high testosterone man is frustrated, low serotonin levels make it more likely the man will respond in an aggressive way. Bernhardt’s ideas fit well with the facts that both the amygdala and the hypothalamus are associated with both serotonin and testosterone.

Bio-Psycho-Social Approaches
Eysenck’s Criminal Personality
An approach which combines both biological and psychological factors is that of Hans J Eysenck.

Having postulated (1970) that there those high in both Neuroticism (N) and Extraversion (E) – see Dimensions of Temperament – were more likely to engage in criminal behaviour, Eysenck completed his profile of the ‘criminal personality’ via his third Dimension, Psychoticism (P) (Hans J Eysenck & Sybil B G Eysenck, 1976). Thus, Eysenck’s ‘criminal personality’ is high theoretically  in P, E and N. A number of studies, such as that by B J McGurk & C McDougal (1981), have supported Eysenck’s theory. (McGurk & McDougal  compared 100 ‘delinquent’ college students with 100 ‘non-delinquent’ college students and found the former high in P, E and N – especially E and N – while the latter were particularly low in E and N.) However, other studies – eg: Raymond Cochrane (1974) –  have presented a rather mixed picture as to whether high Extraversion really is a factor in the criminal personality. But, where Psychoticism is concerned, limited research – eg: Adrian Raine, Peter Venables & Mark Williams (1995) and Hans Steiner, Elizabeth Cauffman & Elaine Duxbury (1999) – shows consistently a clear link between scoring high in Psychoticism and recidivism (frequent offending). Adrian Furnham (1984) tested 210 non-delinquents in the UK for personality, anomie (in the sense meant by Émile Durkheim,1987) and social skills. The best predictors for self-reported delinquency were Psychoticism, then Neuroticism, anomie, Extraversion and finally social skills.

Eysenck (1977), using the Eysenck Personality Questionnaire and physiological measure such as electroencephalograph and galvanic skin resistance, with 156 prisoners aged from 18 to 38, found that:-

  • conmen were low in Psychoticism
  • violent offenders and those involved in property crime were low in Neuroticism and high in Extraversion – though the latter finding was not statistically significant

In their critical review of the domain, Hans J Eysenck & Gisli Guðjónsson (1989) discuss such points and attribute the heterogeneity of prison populations as a confounding variable. Cluster analyses of personality profiles in prison populations suggest, they say, 2 main types of criminals: the active type (high on P, high on E, high on N), and the inadequate (socially) type (high on P, low on E, high on N). It is the active type, which corresponds to the theoretically-expected PEN profile (high on P, E, and N) in criminal populations. Eysenck & Guðjónsson (p85) conclude that: “This is certainly a differentiation that should be borne in mind in all future studies” .

Interestingly, unlike E and N, P is not normally distributed in the general population which may tie in with the association with criminal behaviour.

Ingestants and contaminates
On the basis of 4Q/8LDon Beck (2000a, 2000b) has expressed great concern at how what we take in – nutrients, contaminates, etc – can change our biology (Upper Right quadrant) which then affects our capacity for motivational (vMEME) development (Upper Left) in dealing with the circumstances in which we find ourselves (lower quadrants).

Clare W Graves (1978/2005) demonstrated experimentally the effect of  changes in the Upper Right influencing change in the Upper Left. Increasing the amount of adrenaline in the endocrine system (Upper Right) by injection led to a shift in dominating vMEME in the individual’s vMEME stack from RED to BLUE (Upper Left).

The importance of what we ingest in relation to criminal and violent behaviour has been demonstrated by the work of Bernard Gesch et al (2002). In a small-scale study of 231 inmates between the ages of 18 and 31, Gesch et al  found that supplementing the usual junk-foodish prison meals with vitamins and minerals over 4 months contributed to a 26% decrease in minor breaches of prison rules and a huge 40% decrease in serious breaches, particularly involving the use of violence.  Reputedly, serious breaches by the prisoners returned to previous levels within a month or so of the supplementation being discontinued after funding for the study ran out.

Although Gesch is at pains to avoid the attribution of a simple cause-and-effect between nutrition and behaviour, his research demonstrates clearly that nutrition is a significant factor.

There’s a sense in which Gesch’s findings merely confirm those of Stephen Schoenthaler whose work in schools and youth detention centres – eg: Stephen Schoenhaler et al (2000); Stephen Schoenhaler et al (2009)  – has shown that replacing junk food and snacks with fresh food alternatives contributes to better behaviour, higher IQ grades and better test scores. Schoenthaler’s work has not been taken as seriously as it should have been because flaws were found in his research procedures. However, no one has been able to find any fault with Gesch’s methodology.

Switching from ingestants to contaminates, Rick Nevin (1999) has found a link between the amount of lead in the environment and the level of violent crime. He calculated the rise and fall of the presence of lead from petrol and he compared that curve to the modern history of violent crime. When the amount of lead in the environment increased, Nevin showed a corresponding rise in violent crime 2 decades later. And when the amount of lead in the environment fell, violent crime also came down about 20 years later. (See graph below.)


The findings of Jessica Wolpaw-Reyes (2012) not only support Nevin’s work but show a correlated variation state by state in the US, with those with more lead in the atmosphere having a higher rate of violent crime.

Gesch, interviewed by Dominic Casciani (2014) for BBC News comments: “Lead is a very potent neurotoxin. It has a range of effects on the brain that have been demonstrated through hundreds of different biological studies. Lead alters the formation of the brain. It reduces the grey matter in areas responsible for things such as impulse control and executive functioning – meaning thinking and planning.”

Simply lowering the ingestion of alcohol in a public (social) context seems to have the effect of reducing certain types of crime.

Vaseekaran Sivarajasingam et al (2014) reported a year-on-year fall in violent crime in the UK. They used the number of people needing treatment as a result of crime across 117 hospital Accident & Emergency departments as their measure. They correlated their findings with police statistics and British Crime Survey data for reliability. (See graph below.) They attributed this fall in violent crime, in part at least, to less alcohol.

Graphic copyright © 2014 Cardiff University Violence & Society Research Group

Graphic copyright © 2014 Cardiff University Violence & Society Research Group

As Sivarajasingam’s colleague, Johnathan Shepherd, told The Guardian’s Alan Travis: “Binge drinking has become less frequent, and the proportion of youth who don’t drink alcohol at all has risen sharply. Also, after decades in which alcohol has become more affordable, since 2008 it has become less affordable. For people most prone to involvement in violence – those aged 18 to 30 – falls in disposable income are probably an important factor….since 2008 affordability of alcohol has decreased, the real price of alcohol in both the on-trade and the off-trade has increased and UK alcohol consumption levels have decreased from 10.8 litres per capita in 2008 to 10 litres per capita in 2011. These factors may partly explain the falls in serious violence in England and Wales.”

Abdulla Badawy (2006) has provided a biological explanation for this social psychological phenomenon by showing that alcohol disrupts the metabolism of brain serotonin. Acute alcohol intake depletes serotonin levels in normal individuals. Badawy argues that, in susceptible individuals, this could induce aggressive behaviour. From fMRI scanning, Stephanie Gorka et al (2013) show that this disrupts communication between the orbitofrontal cortex and the amygdala, making regulation of the amygdala’s ’emotional responses’ more difficult.  However, Andreas Heinz et al (2014) showed that environmental factors such as early life stress, make some people more susceptible to this vulnerability. Again this shows the difficulty in trying to attribute purely biological processes to aggressive and/or criminal behaviour.

Biological factors in gender differences
There are, of course, major gender differences in both the amount and type of crime committed. Research shows consistently that men commit far more criminal actions than women. Eg: Jessica Abrahams (2013) cites females accounting for around 3% of France’s prison population, less than 4% of the UK’s and less than 6% in Germany, with a global median of 4.3%. When more serious crimes are filtered out, there is still a significant difference, with women in 2o11 still only accounting for around 25% of all criminal court cases and out-of-court disposals (warnings, cautions, etc) in the UK.

Trying to explain these gender differences on a biological basis produces evidence that is far from conclusive.

In their meta-analysis into the role of genetics Rhee & Waldman make a point of noting that gender differences in their findings are minimal. This last point stands in contrast to Tanya Button et al (2004) who, in a study of 258 twin pairs aged between 11 and 18, found that both aggressive and non-aggressive anti-social behaviour was strongly influenced by genetics. Interestingly, Button et al found that in aggressive anti-social behaviour, the genetic influence was far stronger in girls than than boys. (This gender difference was not found in non-aggressive anti-social behaviour such as truancy and lying.) Elizabeth Shirtcliff et al (2009) conducted a literature review, from which they conclude there were more instances of conduct disorder and antisocial behaviours in boys than girls, with boys having greater tendencies to Psychopathy than girls. They point out that the neurobiology involved in empathy and callousness is different in males and females, with more empathy-related neurocircuitry being active in females than males.

The ebb and flow of female hormones, particularly in the pre-menstrual interval when oestrogen levels dip and progesterone increases, have also been associated with acts of violence on occasion (Katharina Dalton, 1964). Owen Floody (1968) reviewed research on Pre-Menstrual Syndrome and found evidence to support the view that, during this time of hormonal fluctuation women, increase in irritability and hostility, and also are more likely to commit a crime.

If testosterone, associated with Psychoticism, drives impulsivity, compulsivity, ruthlessness and sexual predatoriness in males, not only do females generally have much lower levels of this androgen but Ruben Gur et al (2002) show that women generally have a far greater volume of inhibitory circuits in the lower frontal cortex. This puts most women more towards the Impulse Control end of that dimension and much less likely to commit the kind of impulsive violent crime so often associated with men.

Referring back to the work of Cases et al, if the researchers were able to make XY male mice  highly aggressive by disabling the MAOA gene on the X chromosome, this did not increase aggression in female mice. This is assumed to be because the second X on the XX of the female would compensate for the disabling of the first one. Thus, it may be that a woman’s second X may help explain why very few women display the violent impulsivity associated with the ‘warrior gene’.

A Biological theory which seemed briefly to offer a very powerful explanation for male criminality was that of the Super Male Syndrome (Avery Sandberg et al, 1961) They thought that an error in chromosome separation results in sperm cells with an extra copy of the Y chromosome, meaning that a male foetus from such a fertilisation has the designation XYY.

XYY boys experience an increased growth velocity during earliest childhood, with an average final height approximately 7 cm above expected final height. have an increased risk of learning difficulties (up to 50%, compared to an estimated average of 10% in standard XY boys) and delayed speech and language skills. Delayed development of motor skills (such as sitting and walking), weak muscle tone (hypotonia) and hand tremors or other involuntary movements (motor tics). These characteristics vary widely among affected boys and men. Behavioural and emotional problems are also possible – though many studies indicate not usually more so than standard XY boys.

In a 1965 study of difficult-to-manage, retarded men in prison Patricia Jacobs et al found a surprising number had XYY chromosome patterns – 15 in 1,000 (compared to a general rate of 1 in 1,000 boys). They attributed their aggressive, violent nature to this extra Y chromosome. W Michael Court-Brown (1968), from a 2-year study of 314 patients, recommended that those with XYY were “best hospitalised due to an increased likelihood of aggressive behaviour”. A number of infamous murderers found to have this condition – eg: Robert Peter Tait, who beat a 77-year-old woman to death in Australia in 1962 – fired media imagination, resulting eventually in the TV drama, ‘The XYY Man’ (Granada TV, 1976).

However, Alice Theilgaard (1984) found that “no single characteristic except height…has been associated with the XYY condition”. Later studies – eg: Clive Hollin (1989) – found this kind of genetic abnormality to be sufficiently prevalent in the general population to suggest there is no significant correlation between the XYY pattern and greater male aggression. Herman Witkin et al (1976) found only 12 cases of XYY syndrome in 4,000 Danish prisoners and none of the 12 were violent criminals. Even so, it is not entirely safe to dismiss the link between XYY and aggression though. Theilgaard found that XYY men, given Thematic Apperception Tests, tended to give more aggressive and fewer anti-aggressive interpretations than XY controls.

Difficulties in attributing Cause to Biological Factors
As noted several times so far, there is simply not a straight forward causal relationship between any aspect of human biology and criminality.

From the longitudinal Cambridge Study – over 30 years – of working-class boys in East London in which he had been involved, David Farrington (1995) concluded that low income, a large family, poor child-rearing techniques and parental criminality were the key predictors for juvenile delinquency. The difficulty in separating out social from biological factors is illustrated by Leo Kreuz & Robert Rose (1972) who found no difference in testosterone levels in a group of 21 young prisoners classified as either ‘fighting’ or ‘non-fighting’ while in prison. However, the 10 prisoners with histories of more violent crimes in their adolescence did have significantly higher levels of testosterone than the 11 without such a history. Accordingly, Kreuz & Rose concluded: “…within a population that is predisposed by virtue of social factors to develop anti-social behaviours, higher levels of testosterone may be an important additional factor in placing individuals at risk to commit more aggressive crimes in adolescence.”

The issue of interaction between environmental and biological factors is an important one. This is illustrated in a 2006 study by Rene Weber, Ute Ritterfield & Klaus Mathiak who used fMRI imaging to study participants in a violent video game. They found taking part in the game suppressed activity in the amygdala and the anterior cingulate gyrus – perhaps accounting for reduced empathy towards others in the game.

J Dee Higley et al (1996) reported that people with elevated levels of testosterone exhibit signs of aggression but rarely commit aggressive acts, implying that social and cognitive factors play a mediating role.

Remi Cadoret et al (1995) found that adopted children were more likely to show aggressive behaviour and have conduct disorders if either their adopted home was disrupted – eg: by marital dispute or by drug problems – or their biological parents had criminal records. However, the adopted children were at greatest risk when both these factors applied, suggesting a nature-nurture interaction of epigenetic modification.

The importance of socialisation is emphasised by Gordon Ingram (2014) who argues it is evolutionarily adaptive for children to be less physically aggressive as they mature into adults and are more capable of inflicting serious physical harm or even death. Ingram does note, though, that adults are more adept at indirection aggression – eg: malicious gossiping – which requires greater cognitive development.

Perhaps the best we can say about biological explanations for criminality is that some people may well have a predisposition to behave in a criminal manner (aggressively or non-aggressively) due to one or more biological factors. However, it may well require other factors – social/environmental and/or psychological – for that predisposition to be realised. A kind of Diathesis-Stress combination!

This lack of clarity in just how far biological factors influence criminal activity allows politicians and policy makers, to some considerable, to bend the arguments to suit their own preferences. Eg: Ron Walters (1993) puts forward the decidedly-controversial view that looking for biological causes of criminality allows politicians to avoid committing money to combat poverty and inner city problems.

Go back...


Verification Captcha (human, not robot!) * Time limit is exhausted. Please reload CAPTCHA.