Intelligence and Executive Functioning: Findings and Research Needs
Low intelligence, verbal ability, and performance intelligence are all clearly associated with violent behavior. These studies further suggest that violent offenders have lower intelligence and poorer executive functioning than do nonviolent offenders.
There are fewer studies that look at specific skills, but the evidence so far suggests that violent offenders have greater deficits in planning, problem solving, attention, cognitive control and “other” cognitive abilities and executive functions (EFs). The findings are less consistent than those among studies using broader measures of intelligence and are based on a body of studies that has a variety of methodological insufficiencies. Additional research undertaken to examine specific EFs may help inform interventions for high risk children.
Although numerous reviewers have emphasized deviation scores, computed to estimate the disparity between performance IQ (PIQ) scores and verbal IQ (VIQ) scores (see Chapter 4), there were too few studies of deviation scores to draw any firm conclusions. More research should be provided to substantiate the use of deviation scores in risk assessment instruments. Given that we found very consistent negative associations between both verbal and performance IQ and violence, we cannot accept the assumption that deviation scores are a strong predictor of violence.
There were too few studies looking strictly at nonviolent offending to know whether it is associated with IQ, VIQ, or PIQ. The emerging pattern for verbal ability suggests that it is possible that the relationship is null, so further research on this point would be highly desirable. The body of work published so far suggests that poor executive function may not even be correlated with nonviolent offending. Comparisons of violent and nonviolent offenders frequently failed to ensure that the nonviolent offenders had not also committed violent acts (see above). By comparing violent offenders to a “muddy” group of nonviolent offenders, the authors of these studies probably underestimated differences between the groups. Future research should acknowledge and address this problem.
Furthermore, more multivariate studies are needed in this line of research in order to rule out confounding influences. Studies that control for school grades are badly needed. In addition, studies testing whether IQ has an indirect effect on violence via its influence on academic achievement would be needed to establish greater confidence in our understanding of the nexus between IQ, school, and violence. Although a fair number of studies has already employed controls for socioeconomic status, future research should continue this practice. The dynamics between child IQ, parenting, and child aggression are not currently known, either; studies controlling for parenting factors associated with aggression (such as parental warmth, attachment, abuse victimization, harsh parenting) or employing path analyses with indirect effects would also be of service.
In addition, the association between indicators of intelligence and EF and offending for females is still unclear. In fact, our summary calls into question any assumption that females of lower verbal ability are at risk of offending, so more research on intelligence, executive functioning, and both nonviolent and violent offending among females is needed to understand if these serve as risks for females, or if our attention should be placed elsewhere.
As a general matter, research might also be used to do a better job identifying the band of intelligence scores most where proneness to violence is highest. Individuals at the lowest levels of intelligence (e.g., the developmentally disabled) are not functioning well enough to live independently; it is more likely that these individuals would be victims, rather than perpetrators, of violence, which means that there is probably a ceiling effect where the impact of low intelligence on violent offending shifts direction. More research is needed in order to identify what this threshold might be.
Many of the studies of intelligence and EF, in particular, consist of small sample, simple correlational analyses where the independent and dependent variables were measured around the same time. Correctional samples, for example, are often very small. In many cases, statistical tests were not reported, or the reported tests were not statistically significant, even when very clear differences were seen in a visual comparison. Although there appear to be many studies, there is still quite a bit of room for longitudinal research with larger samples to ensure proper temporal order and statistical power. In addition, measures to reduce selection bias, such as taking random samples, and to reduce the likelihood that the “nonviolent” offenders have committed violent crimes (by using criminal record instead of instant offense to operationalize “nonviolent” and “violent”), would enhance this literature.
Although intelligence is thought to be a native characteristic, the idea that it cannot be influenced is untrue. Thus, future research should focus on isolating which specific intellectual and skills deficits are most influential in the etiology of violence, and then evaluate interventions to see if anything can be done to improve these skills and reduce aggression.