Speculation is evident within both popular media and scientific literature regarding a possible association in the relationship between Autistic Spectrum Disorder and offending behaviour. The possible presence of antisocial and criminal behaviour in a sub-sample of individuals with Asperger syndrome has been of high interest to both psychological research and the general public. The research reviewed within this study suggests that people with Autistic spectrum disorders are potentially over-represented within the criminal justice system. However an accurate prevalence of Autistic Spectrum Disorder and Asperger Syndrome both within the general population and the criminal justice system is yet to be established as findings between studies are varied. It has been suggested that an individual with Autistic Spectrum disorder who has offending behaviour will be more likely to commit violent, sexual or arson-related crime however there is limited research to support this. Additionally it does appear that certain characteristics such as co-morbid psychiatric symptoms are important risk factors
Debbaudt (2004) described how individuals with Autistic Spectrum Disorder are seven times more likely to experience contact with the criminal justice system than the general population. Howlin (2004) argues that the perceived link between Autistic Spectrum Disorder and offending behaviour can be explained by sensational media reporting. Specific stories have highlighted the Asperger syndrome status of offenders in their reports of crime. However the apparent association with offending has been generated by sensationalised, unsubstantiated in diagnostic terms , media reports. News headlines such as ‘Autistic Boy killed baby brother’, ‘Aperger’s Boy locked up for Rape’ and ‘ A 15 year-old boy with Asperger’s Syndrome ahs been locked up for an indeterminate period for raping an 8 year-old boy and photographing the attack’ can be extremely harmful to people’s perceptions of ASD (BBC News, 2001, BBC News, 2007a, Birmingham Post 2007).
The purpose of this article is to review what literature is available relating to Autistic Spectrum Disorder and Offending Behaviour. Specifically this article will review what is meant by Autistic Spectrum Disorder, the prevalence of Autistic Spectrum Disorder within the Criminal Justice System, prevalence of offending behaviour in offenders with Autistic Spectrum Disorder, the characteristics of Autistic Spectrum Disorder that might make individuals more vulnerable to offending and the types of crime most associated with the disorder. Additionally there will be a focus on the relevant treatment and policy implications for clinicians and support staff working with this unique population.
In order to identify all known studies investigating the relationship between Autistic Spectrum Disorder and offending behaviour a search was conducted of the databases Psych Info, Academic Search Complete and Google Scholar using the key terms Autis* and Crim*, Asperg* and Offen* within the parameters of 1985 to current issues. A search was also conducted regarding News and popular media sites using the key words of Autism, Asperger, offend/offending and Crime. All of the studies are organised in Table 1.
Defining Autistic Spectrum Disorders
Before examining prevalence rates it is important to define what is meant by Autistic Spectrum Disorder. In the Diagnostic and Statistical Manual of Mental Disorders ( DSM-IV) (APA, 1994) categories of pervasive developmental disorders (PDD) are autistic disorder, Asperger Syndrome and pervasive developmental disorder not otherwise specified (PDD-NOS). The criteria (DSM-IV) for Autistic Spectrum Disorder include a core triad of impairments. These are impairments in social interaction, social communication and social imagination including restricted repetitive and stereotyped patterns of behaviour, interests and activities. There is variance in diagnostic criteria for Asperger syndrome between DSM-IV and ICD-10.
The term Autistic Spectrum Disorder is widely accepted in research and literature. The idea of the autistic spectrum is clinically useful as it illustrates the variation in symptomatology both among and within each diagnostic category. This range can include mild to very severe symptoms in each of the areas of impairment that categorise Autistic Spectrum Disorder. Those on the lower functioning end of the spectrum, such as those with ‘classic autism’, may never learn to talk using words, struggle to develop or maintain meaningful relationships and have very ritualistic/obsessional behaviour. Those on the higher functioning end of the spectrum, such as those with AS may have IQ within the average range, be verbally fluent and can be capable of leading independent lives given some support.
The majority of literature regarding the relationship between Autistic Spectrum Disorder and offending behaviour has focused on individuals of whom have a diagnosis or show indicators of Asperger syndrome. Langstrom et al (2009) explored characteristrics associated with violent offending among individuals diagnosed with Autistic Spectrum Disorders over a 13 year period taking information from swedish national registers. They found that criminal violence was more common among participants diagnosed with Asperger syndrome relative to participants with autistic disorder. They attributed this factor to the reasoning that those with Asperger syndrome are more likely to have unimpaired intellectual ability, better social skills. It was also concluded that those with Autistic disorder are more likely to live in supervised settings where staff or others may help to resolve conflicts or manage anger. The fact that research study examines both autistic spectrum disorder and Asperger syndrome make drawing conclusions and comparisons extremely difficult.
Asperger syndrome is defined as a condition which has many features of Autistic Spectrum Disorder however it is displayed in persons of average or above average intelligence. Gillberg (2002) describes features such as a tendency to engage in monologues or special interests around narrow interests, limited or inappropriate non-verbal communication and a compulsive need for routine. The term Theory of Mind has been used to describe the psychological dysfunction underlying the triad of impairments shared by all people with Autistic Spectrum Disorder (Wing, 1996). Theory of mind describes the ability to be able to put oneself in the position of another and be able to appreciate their feelings, wants and wishes. This inability to empathise and ‘put yourself in someone else’s shoes’ means that those with Asperger syndrome, like those with other forms of Autistic Spectrum Disorder, are unable to judge the probable actions or responses of others (Browning & Caulfield, 2011).
Many studies support the view that the majority of those with Autistic Spectrum Disorder’s and Asperger syndrome are law abiding and that very few engage in illicit criminalised behaviour. The very features of AS mean that the vast majority of this section of the population are very pedantic in their adherence to the law due to their literal thinking (Wing, 1997, Murrie et al. 2002, Myers, 2004). It may be that speculative assumptions that those with Autistic Spectrum Disorder’s or Asperger syndrome possess a greater propensity towards violent criminality is generated through media reporting and both factual and fictional film, television and radio. There are relatively few people in the public eye with an Autistic Spectrum Disorder however the reporting of rare acts of violence within the media is potentially harmful serving only to assist in the creation of inaccurate perceptions of affected individuals ( Browning & Caulfield, 2011).
An accurate prevalence of Asperger syndrome within the general population is yet to be established as findings between studies are varied. Research has indicated the prevalence of Asperger syndrome to be between 2 and 60 per 10,000 (Gillberg 1991, Wing 1996). Murphy, (2003) reported that there is evidence that 3-4 in every 100 children may develop Aspergers syndrome and there is general agreement that the prevalence of Asperger syndrome is low. Current statistics provided by the National Autistic Society estimate that the prevalence of all Autistic Spectrum Disorder’s in the UK is approximately 1% (NAS, 2008). Attempts to assess the violence and criminality rate have also been made more difficult due to the low prevalence and problems relating to gaining valid diagnosis of Asperger’s Syndrome.
It must also be recognised that additional problems in this field may stem from specific difficulties. People with Autistic Spectrum Disorder or Asperger’s syndrome within the criminal justice system may be being misdiagnosed and being given other labels such psychosis. There may be a form of societal reluctance to link specific disorders within criminality thus meaning an under-estimation of prevalence is given. Berney (2004) also considers the idea that there is an increasing unwillingness within the UK to pursue prsecution unless there is a high possibility of conviction. Data may also not be taking into account large numbers of people who do not enter the criminal justice system as they are subject to diversion or do not get charged (Hawk et al. (1993).
Some researchers have proposed that the core features of Autistic Spectrum Disorder such as abnormal development of social interaction and development may expose those on the spectrum to commit offending (Haskins and Silva, 2006). However this does not mean that having Autistic Spectrum Disorder enhances the likelihood of offending. Researchers have used a range of different strategies in order to study any possible association between Autistic Spectrum Disorder and offending.
Prevalence of persons with Autistic Spectrum Disorder in the Criminal Justice System
In order to better understand if there is any relationship between Autistic Spectrum Disorder and offending behaviour some researchers have focused on Autistic Spectrum Disorder among those who have already been or are still in the criminal justice system. Scragg and Shah (1994) considered a hypothesis that there may exist co-morbid neurological disorders in those with Asperger syndrome that may be an underlying factor of violent acts committed. Their research examined the entire male population (n=392) of Broadmoor hospital using methods such as interviewing staff and patients and screening case notes. They found a prevalence rate for Asperger syndrome of 1.5% significantly exceeding the prevalence rate of 0.36% found by Elthers and Gillberg (1993) in the general population using the same diagnostic criteria.
Siponmaa et al (2001) conducted a retrospective study where a much higher prevalence rate of ASD among offenders was found in Sweden. The researchers reviewed case notes of 126 ‘mentally disordered’ offenders and found that 15% had pervasive developmental disorder, 12% had pervasive developmental disorder not otherwise specified/atypical autism and 3% had Asperger syndrome. However these results cannot be generalised to offender populations as the sample was recruited from a forensic population of offenders who have a mental disorder requiring treatment and the majority had committed violent offences.
Hare et al (1999) examined the prevalence of individuals with Autistic Spectrum Disorders within three hospitals in the UK. They found that 67.7% of the total Autistic spectrum disorder population of these prisons had Asperger syndrome. The rough prevalence estimate was 1.6% supporting the hypothesis that there is an over representation of people with Autistic spectrum disorder and in particular Asperger syndrome within forensic hospitals in the UK. In contrast Myers (2004) found very low numbers of individuals with autistic spectrum disorders in secure forensic and specialist settings in Scotland. Through examination of the number of people with learning disabilities with or without Asperger syndrome their findings indicate a prevelance estimate of 0.93% in the prison service; 0.46% in secure units and 1.39% in mental health units
It must be recognised that prevalence studies have typically been restricted to highly specialised settings and therefore highly selected samples such as those at Broodmoor prison and secure services (Scragg and Shah,1994). There also appears to be a lack of research which has investigated the prevalence of Autistic Spectrum Disorder in general populations of people who had committed crime. Additionally there is also a failure to acknowledge that the offending behaviour of their research participants could potentially be attributed to other risk factors commonly associated with offending within the general population such as co-morbid mental health issues or social circumstances.
There is large variability in prevalence rates across the discussed studies however there does appear to be some sort of association between autistic spectrum disorder and offending behaviour. Differences in types of diagnosis may account for some of the varying prevalence rates between different countries.Even specific studies involving highly specialised settings have found that a diagnosis of Autistic Spectrum Disorder had not been previously made for the majority; making this study of general populations even more difficult. . In making a diagnosis of Autistic Spectrum Disorder this can help in the understanding of reasons that crimes may have taken place and can help to decide on appropriate methods of care and management Detection of Autistic Spectrum Disorder in childhood/adolescence could provide opportunity to manage pre-disposing factors to offending behaviour. In the absence of such systems it could be argued that individuals with an Autistic Spectrum Disorder are at increased risk of developing offending behaviours.
See Table 1 for an overview of the prevalence of Autistic Spectrum Disorder found in studies within both forensic and community settings
Prevalence of Offending Behaviour in individuals with Autistic Spectrum Disorder
When reviewing the research a further method of investigating the relationship between Autistic Spectrum Disorder and offending behaviour has been to examine the prevalence of offending behaviour in individuals with Autistic Spectrum Disorder. The question of whether or not a person with Autistic spectrum disorder may be more likely to offend will provide richer information around any possible association.
Hippler et al, (2009), completed a study to examine criminal offending in 177 former patients of Hans Asperger. No increased rate of registered convictions was found compared to the general population suggesting that people with Asperger’s syndrome are not more likely to commit offences. Mourisden et al, (2008) looked at the prevalence of offending behaviours in individuals with Autistic Spectrum Disorder compared to typically developing controls using the Danish Criminal Register. Register data offers opportunity to analyse data from large populations in a historical, prospective way avoiding recall bias. Overall results found that 9% from the Autistic Spectrum Disorder group and 18% from the comparison group had convictions. This study suggests that rates of convictions are actually lower in individuals with Autistic Spectrum Disorder than in the general population. The research also indicated that criminal offences are very uncommon in childhood autism but more common in Aspergers syndrome.
Allen et al’s (2008) study investigated the prevalence of Asperger syndrome and offending within a large geographical area. A survey methodology was used to identify adults with a recorded clinical diagnosis of Asperger syndrome in the South Wales area of the UK. The services contacted included mental health teams, local health practitioners, forensic practitioners and learning disability teams. Their results were generally not supportive of there being a significant association between Asperger syndrome and offending, attributing this to methodological problems including individuals with learning disabilities skewering the data, the focus on adults and misdiagnosis.
Woodbury-Smith et al. (2006) identified that previous research had focused specifically on violent offending and suggested a need for community based research. Their reasoning was that statistical findings of research conducted within high security psychiatric or forensic settings only represented the prevalence of very serious offences therefore failing to be more representative of more generalised common-place criminality. They responded by conducting a study of offending by those with Asperger syndrome and Autistic Spectrum Disorder’s in the community. Although it must be acknowledged that the sample size was very small it was the first to look at offending patterns of individuals with Autistic Spectrum Disorder’s in the community. Their findings found that the level of offending occurring within the Autistic Spectrum Disorder group was lower than that of the neuro-typical group.
Issues with these particular studies include the difficulties that some studies have such small, unrepresentative samples that make any estimates of prevalence epidemiologically insubstantial. Consequently whilst these studies do not give a true picture, this limited evidence also suggests that individuals with ‘classic’ autistic disorder are very unlikely to offend and those with a diagnosis of Asperger syndrome are no more likely to offend than people in the general population.
Table 1 – Prevalence of Autistic Spectrum Disorder studies within both forensic and community settings
No. of participants
Ghaziuddin et al (1991)
Review of 132 published case studies
Scragg and Shah (1994)
Hare et al. (1999)
15% diagnosed with ASD
0.46% in secure psychiatric units
0.93% in the prison service
1.39% in mental health units
Hippler et al (2009)
No increased prevalence compared to the general population
Predisposing Risk Factors of Autistic Spectrum Disorder that might make a person more vulnerable to offending
In order to examine if there is any relationship between Autistic Spectrum Disorder and offending behaviour it is important to investigate whether there are any innate vulnerabilities that make it more likely that a person with Autistic Spectrum Disorder will offend. It must also be recognised that general vulnerability factors as identified in criminological literature such as low IQ, poor school achievement and Attention deficit hyperactivity disorder will still all be relevant (Farrington, 2002). More general knowledge within this area will enhance both preventative programmes and forms of treatment and rehabilitation.
Woodbury- Smith et al. (2006) completed an exploratory study where participants were recruited from different sources in the UK. These adult groups included 21 individuals with Autistic Spectrum Disorder and a history of offending, 23 individuals with Autistic Spectrum Disorder and no history of offending and a general population group of 23 people without Autistic Spectrum Disorder and no history of offending. Cognition areas known to be impaired in people with Autistic Spectrum Disorder such as theory of mind and emotional recognition were compared. Results showed that the Autistic Spectrum Disorder offenders showed a significantly greater impairment in the recognition of emotional expressions of fear, but no difference in theory of mind, executive function, and recognition of facial expressions of sadness. This finding indicates that this lack of ability to recognise fear in others may cause individuals with Autistic Spectrum Disorder to continue behaviour that is distressing to others thus possibly increasing the likelihood of offending.
Langstrom et al (2009), used data from Swedish longitudinal registers for 422 individuals hospitalised with Autistic Spectrum Disorder during a thirteen year period . Their research compared those committing violent or sexual offences with those who did not. Their results found that violent individuals with Autistic Spectrum Disorder are generally male and diagnosed with Asperger syndrome rather than Autistic Spectrum Disorder. Conclusions also found that violent offending in Autistic Spectrum Disorder is related to similar co-occurring psychopathology as found in violent individuals without Autistic Spectrum Disorder, Limitations to this study must be cautiously interpreted as the sample was based on being hospitalised and are again therefore not necessarily representative of all individuals with Autistic Spectrum Disorder in the community
There are very few reports that investigate the importance of co-morbid psychiatric factors when understanding if any relationship between Autistic Spectrum Disorder and offending behaviour exists. Newman and Ghaziuddin (2008), reviewed 37 published cases with violent behaviour and AS in respect to concurrent psychiatric disorders and found that 29.7% of the reviewed cases had a definite psychiatric disorder whilst 54.0% had a probable psychiatric disorder. Subsequently when offending occurs within an individual with Autistic Spectrum Disorder it is important to exclude other psychiatric conditions because these can independently influence the risk of offending, as it does in the general population.
Murphy’s (2003) study examined in the setting of a high-security psychiatric hospital whether male patients with AS could be distinguished from patients with personality disorder or schizophrenia on the basis of admission and neuropsychological details. This study adds to the review of research on Autistic Spectrum Disorder and offending behaviour as exploratory comparisons found that patients with Asperger syndrome were less likely to have a history of alcohol or illicit substance abuse and had lower index violence rations. Qualitative observations in this study suggested that whilst circumstances varied, offending behaviour tended to be associated with particular problems such as externalised misplaced blame, difficulty understanding consequence of their actions and difficulty with prospective taking.
Murphy (2006) compared in-patients with schizophrenia, personality disorder and Asperger’s syndrome in Broadmoor high security psychiatric hospital. The research indicated that those with AS were less likely to have histories of substance abuse or alcohol misuse and their offences had lower violence ratings than those of other patient groups.
Palermo, (2004) examined the relationship between complex developmental disorders and delinquency by analysing three adult patients with Pervasive Developmental Disorder whose admission to hospital was precipitated by criminal behaviour. The patients all shared similar difficulties including struggling to understand social cues, showing mannerisms ad prosodic oddities and all had the presence of a major psychiatric illness whilst meeting the diagnostic criteria for Asperger syndrome. Their findings found that the reported behaviours resulted from co-morbid psychopathology and not as a direct consequence of a developmental disorder.
Howlin (2004) suggests that the innate levels of empathy may be associated in particular with violent and sexual offending. Additionally Haskins and Silva, 2006. discuss how the illegal behaviours of a person with Asperger syndrome are predominantly associated with either a deficient theory of mind or an intense preoccupation with a narrow interest. The term ‘weak central coherence’ refers to the detail focused processing style that is proposed to characterise autism spectrum disorders (Happe and Frith, 2006). A ‘persistent preoccupation with parts of objects’ is one of the diagnostic criteria for autistic disorder in current practice (DSM-IV, 1994) Subsequently it has been proposed that this weak central coherence may make it difficult for an individual with Autistic spectrum disorder to think about the consequence of their actions.
To confirm factors that predispose a person to commit a particular type of crime, further large scale, longitudinal prospective studies are needed.
See Table 2
Table 2 – Factors mediating offending in Autistic Spectrum Disorders in key studies
Factors mediating offending in AS
externalised misplaced blame
difficulty understanding consequence of actions
difficulty with prospective taking
Barry-Walsh and Mullen (2004)
narrow focus and preoccupation with a special interest
struggling to understand social cues
Lack of empathy
innate levels of empathy
Happe & Frith (2006)
weak central coherence
Haskins and Silva, (2006)
deficient theory of mind
An intense preoccupation with a narrow interest.
Woodbury-Smith et al. (2006)
significantly greater impairment in the recognition of emotional expressions of fear
Newman and Ghaziuddin (2008)
Co-morbid Psychiatric factors
Allen et al (2008)
Lack of concern and awareness
Impulsivity and mis-interpretation of the rules
Types of Crime most associated with the disorder
It is difficult to ascertain the types of offending behaviour in individuals with Autistic Spectrum Disorder due to the lack of valid epidemiologic studies. Clinical studies within the case study literature have also explored the hypothesised relationship between Autism and Offending behaviour. The offense type most commonly described within these studies tends to be sexual offences, arson and violent offenses. Studies of single cases and referred samples can provide indications of the types of offending which may be found in people with Autistic Spectrum Disorder (Baron-Cohen, 1988, Barry-Walsh & Mullen, 2004, Haskins & Silva, 2006, Murrie et al. 2002, Schartz-Watts 2005).
Mawson et al, (1985) case study focused on a 44 year old man cataloguing a series of ‘strange’ violent behaviours including dropping a firework into a girl’s car, assaulting a crying baby at a railway station by putting his hand over its mouth to stop the noise, following a female teacher, getting close to her at every possibility. The authors speculate that there is an association between Asperger syndrome and violent behaviour however they do not put the offences in context of the man’s diagnosis and significant impairments including sensory, theory of mind and inability to read social situations. Palermo’s, (2004) case studies behaviours respectively included threatening to kill a police officer, threatening to burn down a family member’s house and touching a prepubescent boy at a playground.
Several case reports associated with Arson and Asperger syndrome have appeared in the literature. Mouridsen et al’s, (2007) study looked at the prevalence of offending behaviours in individuals with Autistic Spectrum Disorder compared to typically developing controls and found that only ‘arson’ statistically separated Asperger syndrome cases from the comparison group. Haskins and Silva (2006), discuss a range of particular cases highlighting specific issues that might make a person with Autistic Spectrum Disorder have heightened risk for engaging in criminal behaviour. One reviewed case history involved a young man who was accused of starting a fire in his apartment to obtain insurance money. This fire killed his young daughter and nearly killed his wife. Additionally Sionmaa et al, (2001) reported that 10 (63%) of 16 crimes of Arson were perpetrated by individuals with AS. Barry-Walsh and Mullen, (2004), reported on two cases that had a history of fire-starting. Reasons for the offence of arson vary within each study however it is hypothesised that a major reason includes the narrowed fixations that some people with Autistic Spectrum Disorder may have and the lack of understanding of consequences and reasoning. It could be suggested that Arson may be more specifically related to behaviours seen in Autistic Spectrum Disorder .
Epidemiological studies indicate that people with AS do commit sexual offences (Hare, 1999; Murphy 2003) but there is evidence that the rates of sex offending and child sex offences in particular are lower than in general populations ( Hare, 1999, Elvish, 2007). It has been suggested that individuals with Asperger syndrome display criminal behaviour due to their sexual preoccupations. Cases include that of a 21 year old male who had a history of stealing cotton lingerie and masturbating whilst holding women’s night-dresses (Chesterman & Rutter, 1993) and a young male with Asperger syndrome who had a history of recurrent sexual offences including touching the privates of young women; watching women in toilets and making obscene phone-calls (Milton et al, 2002).
It appears that when individuals with autistic spectrum disorder do commit crimes they are less likely to involve alcohol, substance misuse and/or drugs (O’Brian & Bell, 2001). Additionally the offences of individuals with autistic spectrum disorder did not generally provide any direct gain to the individual. Wahlund & Kristiannsson (2006) investigated 35 male offenders with diagnosis of ASD or anti-social behaviour referred for psychiatric assessment in Sweden. It was found that those with autistic spectrum disorder were less likely to be intoxicated and did not use knives or guns as frequently as the personality disorder group. It was also hypothesised that individuals with autistic spectrum disorders may offend at an earlier age than other ‘neuro-typical’ offenders as difficulties in social interaction will start to be marked in early adulthood and youth resulting in a higher risk of offending behaviour.
Recommendations to support individuals with ASD with Offending Behaviour
There must be recognition that a number of people with Autistic Spectrum Disorder are involved with the criminal justice system whether that be as victims, witnesses or perpetrators of crime. Professionals working within the CJS must be educated regarding the unique diversity that may be found within people with Autistic Spectrum Disorder in order to improve the likelihood that these individuals will be treated more appropriately. Tiffin & Nadkarni (2010) discuss how specialised assessments can identify risk factors associated with violent behaviour even if social and communication problems consistent with an Autistic Spectrum Disorder are present. However there is recognition that these tools compliment but do not replace structured clinical assessment. Subsequently a formulation of risk that recognises complexity but accepts limitations enables useful management plans to be made.
The treatment of vulnerable groups within the criminal justice system has been focused on within specific research and it must be recognised that the process of arrest, questioning and trial may be more difficult for a person with Asperger syndrome then others. Difficulties may include misinterpreting what they hear, being unable to function effectively in unpractised, unfamiliar environments and using words without fully understanding their meaning (Mayes, 2003, Barry-Walsh and Mullen, 2004) Subsequently these difficulties may be considered to be the individual being un