Jul 2020

Key messages from research on child sexual abuse perpetrated by adults

University of the Sunshine Coast and University of the West of England

PDF Downloads

This paper brings together learning from existing research on child sexual abuse (CSA) perpetration by adults; CSA perpetrated by young people is common, and is discussed in the CSA Centre’s ‘Key messages’ paper on harmful sexual behaviour.

In keeping with the CSA Centre’s typology of CSA offending, this paper refers to all forms of CSA (including both intra- and extra-familial CSA, online-facilitated CSA, and group-based CSA). However, much research relates to only one type of CSA, or looks at sexual violence as a whole. 

The term ‘children’ in this paper refers to individuals under 18 years of age. 

Key messages

Who sexually abuses children?

We do not know how many adults sexually abuse children, but it is clear that there are more than official figures show, and that large numbers of individuals are engaging with CSA images.

Most contact CSA is perpetrated by a person whom the child knows and trusts.

Almost all individuals convicted of CSA are men, but perpetration by women may be more common than official data suggests.

Perpetrators come from all walks of life, and ethnicity is not a predictor of CSA perpetration.

What are the pathways to CSA offending?

There are likely to be a number of different and complex pathways into CSA offending.

Most adult perpetrators of CSA do not have extensive criminal histories or criminal lifestyles.

The great majority of victims of CSA do not go on to commit CSA themselves; however, those who commit CSA are more likely than other adults to have experienced multiple forms of abuse as children, including physical, domestic or sexual abuse, and neglect.

Social and cultural practices, as well as institutional and environmental settings, can provide a context for CSA to take place. Offences occur mostly in contexts of ordinary social interactions. Situations and locations can present individuals with opportunities to offend, and/or exert social pressure or weaken their moral restraints, leading to CSA. First offences often take place at times of depression, anxiety and stress for the perpetrator.

What interventions are effective?

Most individuals convicted of sexual offences, including CSA, are not reconvicted of further sexual offences. Few women and men convicted of online offences are reconvicted for sexual reoffences. It should be noted, however, that most CSA does not come to the attention of the authorities.

Evidence is lacking regarding the effectiveness of programmes used in the UK to address sexual offending, particularly those specific to CSA; for example, current trauma-informed programmes used in the UK and based on the Good Lives Model have not yet been evaluated. 

The Core sex offender treatment programme in the UK was replaced after it was shown not to reduce sexual reoffending; however, individuals who completed similar programmes in other countries had lower rates of sexual reoffending than comparable individuals who did not.

In the USA, Circles of Support and Accountability (CoSA) have been found to lead to fewer sexual reconvictions. Studies in other countries have shown that CoSA can help individuals released from prison to integrate into communities, which is associated with reducing reoffending.

Providing interventions for individuals at risk of committing CSA is also important, as they may struggle with their sexual thoughts for some years before committing an offence, and many seek help prior to offending.

The extent of adult perpetration of child sexual abuse

It is impossible to know the true extent of CSA committed by adults, for reasons including societal denial, secrecy and fear of the consequences of disclosure. We do know, however, that the majority of suspected perpetrators of CSA are never prosecuted: in 2017/18, only 14% of police-recorded CSA offences in England and Wales resulted in a charge or summons. We also know that police data captures only a small proportion of perpetrators; among the adults who told the 2018/19 Crime Survey for England and Wales (CSEW) that they had been sexually abused in childhood, fewer than one in four said they had reported it to the police.

The volume of CSA recorded by the police, and the number of defendants appearing in court, have increased in the past decade. This is particularly true of ‘online-facilitated’ CSA (CSA through online viewing, sharing or possessing images, and CSA through online interaction). Across the UK in 2019, approximately 400 to 450 individuals were arrested each month for online-facilitated CSA offences.

What is known about adults who commit child sexual abuse?

As CSA is a largely hidden crime, the information that we have about those who commit it is partial and mainly drawn from individuals who have been identified and convicted. 

Adults who perpetrate CSA come from all walks of life. Most contact CSA committed by adults is perpetrated by a person whom the child knows and trusts, such as family members, friends or acquaintances and, to a lesser extent, people in positions of trust or authority.

Some perpetrators of CSA may share characteristics with those who perpetrate non-sexual offences, but overall patterns of offending do not replicate those of other types of crime. While most crime is committed by young people between the ages of 16 and 20, a study in Australia of 280 males convicted of CSA found that 66% reported committing their first offence when aged between 25 and 49, and 8% when aged over 50. 


Almost all adults convicted of CSA offences are men; in 2016, just 2% of those proceeded against for CSA offences in England and Wales were women. But perpetration by women may be more common than official records show: among individuals reporting their experiences of CSA to the Independent Inquiry into Child Sexual Abuse in England and Wales (IICSA), 16% of those abused in residential care and 6% of those abused in other institutional contexts such as schools, sports and religious settings said that female perpetrators were involved, although in some cases these may have been other children.

Professionals’ views about the seriousness and impact of CSA committed by women differ from those of victims and survivors: professionals typically consider it to be less serious and harmful than male-perpetrated CSA, and this affects the recognition and level of support that victims and survivors receive.


Ethnicity is not a predictor of CSA perpetration, despite the attention given in the UK to the ethnicity of those involved in particular forms of child sexual exploitation. The World Health Organization reports that CSA is committed throughout the world. Among individuals prosecuted for CSA in 2017 whose ethnicity was known, 11% were from black, Asian and minority ethnic (BAME) backgrounds. Since the 2011 census found that 14% of the general population of England and Wales were from BAME backgrounds, this does not demonstrate an over-representation of perpetrators from these ethnicities.

Pathways to child sexual abuse offending

CSA is perpetrated in a range of contexts and by individuals who are impossible to distinguish from non-perpetrators, and there are likely to be a number of different and complex pathways into CSA offending.

One evidence-based theory, known as the Integrated Theory of Sexual Offending (ITSO), attempts to explain male pathways into sexual offending. It suggests that there are a number of different, interlocking factors – including ecological (i.e. social, cultural and physical environments, personal circumstances, and learning), biological (influenced by genetic inheritance and development), neuropsychological (such as motivation, emotion, memory, perception, control), and personal factors (negative/positive emotional states, thoughts about offending) – which all contribute towards a decision to commit abuse. These factors interact in dynamic, constantly changing ways to create the pathways that lead to offending, and to maintain and escalate offending. For example, the process of committing CSA influences subsequent thoughts, motivations and behaviour.

Theories have also been developed based on research to understand the pathways into CSA offending for women. The Ecological Process Model of Female Sex Offending has similarities with the ITSO, highlighting the importance of personal and life circumstances (life stressors, mental illness) in combination with situational factors (relationship difficulties) and the environments in which women live (e.g. characterised by poverty and crime), but it does not include biological and neuropsychological factors. The Descriptive Model of Female Sex Offending includes three pathways into offending:

  • Explicit Approach pathway: an intention to offend motivated by problematic goals and associated thoughts and values, where offending is planned and leads to positive emotions
  • Implicit Disorganised pathway: where women are not intending to offend and so offences are not planned but impulsive, occurring as a result of loss of control
  • Directed Avoidant pathway: women who have intimacy or fear related goals, are directed by male planning, and experience negative emotions as a result of the offence.

Childhood trauma

There is a common belief that perpetrators of CSA have themselves been victims of CSA, and that this is linked to their offending. Recent research suggests that, while most individuals who have experienced CSA do not abuse others, perpetrators of CSA are more likely than other adults to have experienced multiple forms of abuse (known as polyvictimisation) including physical abuse, neglect, domestic abuse and, in some cases, sexual abuse. There is also evidence of a higher rate of adverse childhood experiences among women convicted of sexual offences compared to non-offending women.

The ITSO suggests that these traumatic experiences increase the production of hormones associated with fear, flight and fight responses which, over time, can inhibit growth and connections in the brain. This then influences the way individuals think, interact with others, regulate and respond to emotions, and control their behaviour in response to impulses/situations. It is important to note, however, that childhood trauma is associated with a wide range of outcomes which also include increased risk of victimisation, poor health outcomes and non-sexual offending. Hence, there is no simple causal pathway from childhood trauma to the perpetration of CSA. For those who have experienced trauma, this has implications for interventions, which should be trauma-informed.

Adult experiences

Many individuals convicted of CSA report that they perpetrated their first offences at times of depression, anxiety and stress, which may have precipitated them into offending. A study of CSA committed by women in institutional settings, such as schools, showed how situational risk factors including relationship problems, stress and emotional vulnerability can affect women’s lives and contribute to their offending.

Situational and environmental factors

Situations and environments may also help to explain pathways into CSA offending, since social and institutional cultures influence levels of child protection and opportunities for CSA to occur. For example, unlike other types of crimes, there appear to be two periods that are associated with increased risk of CSA: adolescence and mid-to late 30s. CSA that starts in adulthood coincides with changes in family, work, and social circumstances that enable unsupervised access to children.

In institutional contexts, victims and survivors giving accounts of their abuse to IICSA noted that power, authority and reverence bestowed upon religious institutions and the individuals working within them meant that the conduct of perpetrators was not questioned. Similarly, in residential care settings, perpetrators could take children off-site and have unquestioned contact with children on-site, as well as easy access to locations where they could perpetrate CSA.

A meta-analysis of studies internationally found that men who committed CSA imagery offences had more access to the internet and less access to children compared with those whose CSA offences involved direct contact with children. Situational factors (e.g. awareness that a guardian is nearby) have also been shown to influence the severity and duration of contact CSA. 

For the most part, CSA offences occur during ordinary social interactions, when a potential perpetrator, a suitable ‘target’ and the absence of capable guardians combine to make it possible for the crime to occur. This explains why the majority of victims know and trust the individuals who perpetrate CSA against them, and why CSA occurs in locations and at times when adults have ‘routine’ access to children.

Offending is most likely to take place where the risk of apprehension is low and effort minimal. Empirical studies have shown that, like individuals who commit other types of crime, adult perpetrators of CSA make rational decisions during the act of committing the offence. Situations and locations can present individuals with cues that prompt them to commit CSA, exert social pressure, weaken their moral restraints and generate an emotional arousal leading to CSA.

This highlights the importance of situational crime prevention strategies to prevent CSA, recognising that children are located within a set of interlinking environments of family, peers, school, community and wider society. This approach is used in contextual safeguarding, which was developed to improve safeguarding services for adolescents at risk of child sexual exploitation and has since been adopted more widely throughout the UK. The Lucy Faithfull Foundation’s Eradicating Child Sexual Abuse (ECSA) prevention toolkit provides a resource to guide practitioners and policymakers through prevention strategies based on the place-particular problems children face in a given location. No situational strategies have yet been evaluated to assess their impact on offending, however.

Pathways between types of child sexual abuse

The evidence indicates that there is no clear pathway from online-facilitated CSA to contact CSA, although it is important to recognise that existing research is based on individuals who have been identified or convicted, which is only a proportion of those committing CSA.

Studies have shown that individuals convicted of online-facilitated sexual offences are less likely to be subsequently convicted of a contact sexual offence than those originally convicted of contact sexual offences, with just 2% committing a contact sexual offence in a follow-up period of 1.5 to six years. Men convicted of CSA imagery offences who were not antisocial (e.g. had no prior offences or problems with supervision), had limited access to children, and had psychological barriers to committing contact sexual offences were found to have a low risk of committing contact offences.

Likelihood and predictors of sexual reoffending

It is widely believed that adults convicted of CSA have an enduring high risk of sexual reoffending, but reoffending studies – although not CSA-specific – show that most individuals convicted of sexual offences are not reconvicted of further sexual offences. In England and Wales, for example, 7.5% of those released from prison between 2002 and 2012 had been convicted of another sexual offence by October 2015. It should be noted, however, that most CSA does not come to the attention of the authorities.

In that study, the reconviction rate for offences of any sort over the same period was considerably higher at 38%. However, there is some evidence that, in comparison with those convicted of sexual offences against adults, individuals convicted of CSA offences are more likely subsequently to commit only sexual offences rather than other offences.

Long-term international studies suggest that most adult males convicted of sexual offences no longer have a significant risk of sexual reoffending after 10 years (having a similar risk to those with histories of other crimes), and only a small proportion remain at risk after 15 years. This applies for all risk levels; even in the highest risk group, four-fifths of individuals in a 20-year study were never reconvicted for another sexual offence.

Research evidence also indicates that women convicted of sexual offences reoffend at significantly lower rates than men, with an average sexual reoffending rate of around 3%, which may partly reflect that women are unlikely to be identified and reported. This suggests that distinct policies and procedures are needed to assess and manage the risks from women and from men.

A review of the international literature concluded that the following factors were related to sexual reoffending (not specifically CSA offences):

  • sexual preoccupation and sexual interests supporting offending
  • offence-supportive attitudes
  • emotional congruence with children (i.e. feeling more comfortable around children than adults, thinking of children as friends, and being emotionally attracted to children) and a lack of emotionally intimate relationships with adults
  • lifestyle impulsivity, general self-regulation problems and poor cognitive problem-solving
  • resistance to rules and supervision, grievance/hostility and negative social influences.

Some factors had a more limited evidence base but are likely to be related:

  • hostility toward women
  • callousness/lack of concern for others
  • dysfunctional coping.

Although mental illness, poor victim empathy, poor social skills, perceptions of inadequacy and lack of motivation for treatment have long been believed to be related to sexual reoffending, research has found no such link. Similarly, although denial of offending is often used as a reason to exclude adults from interventions, evidence suggests that it not related to sexual or violent reoffending.

On the other hand, family and community support is an important factor in desistance. For example, individuals convicted of sexual offences were less likely to reoffend if they received support on release from prison.

Interventions to reduce reoffending

Most interventions to prevent CSA are aimed at individuals (primarily men) who have been convicted of sexual offences, rather than those who have inappropriate thoughts or concerning behaviours. Moreover, most interventions focus on sexual offending generally, rather than being specific to CSA. The ECSA website contains a list of interventions and evaluations of them.

Sex offender treatment programmes

Much attention has been focused on sex offender treatment programmes (SOTPs). These are delivered in prisons and other establishments, as well as in the community (e.g. via probation services). Usually delivered to small groups of individuals, SOTPs typically involve a range of activities believed to reduce the likelihood of reoffending, such as challenging and changing offence-supportive beliefs, improving problem-solving, and developing appropriate coping strategies. Individuals are assessed for their suitability for attendance, with restrictions (such as exclusion of anyone who denies their offence) varying between service providers.

The Core SOTP programme delivered in England and Wales to individuals convicted of any sexual offences, including CSA, was discontinued in 2017 after an evaluation found that those who completed it had a sexual reoffending rate of 10% compared to 8% for a comparison group who did not complete treatment. It was replaced by programmes (e.g. ‘Horizon’ and ‘Kaizen’) developed using the desistance literature, the Good Lives Model – a strengths-based approach that focuses on assisting individuals to develop and implement meaningful life plans that are incompatible with future offending – and trauma-informed approaches. However, these programmes have not yet been evaluated.

The most recent review of SOTP effectiveness, in the UK (including the Core programme evaluation) and internationally, highlighted some factors found to increase programme effectiveness. These included:

  • a qualified licensed psychologist being consistently present
  • supervision by psychologists of the staff delivering the programme
  • participants rating the quality of the programme highly
  • programme delivery in groups, with a focus on managing sexual arousal
  • not using the polygraph (see below).

Management in the community

A number of strategies have been implemented around the world for the management in the community of adults convicted of sexual offences, with most research attention directed towards the USA’s registration/notification requirements and residency restrictions. These studies show that, rather than reducing reoffending rates, such measures may in fact increase reoffending by restricting the individual’s ability to obtain employment, housing and other aspects of building a socially positive life.

In the UK, each police force has a Management of Sex Offenders and Violent Offenders (MOSOVO) team which manages the convicted individuals in its policing area, according to the risk they pose. Some of these individuals are subject to Multi Agency Public Protection Arrangements (MAPPA), with different levels of monitoring and intervention depending on their risk level.

One comparative study has assessed MAPPA’s impact on reoffending. It found small reductions in one-year reoffending and in serious reoffending among a group of MAPPA-supervised individuals compared to those who did not receive MAPPA supervision. However, a more recent report noted that the setting of MAPPA levels was not always consistent, well-communicated or underpinned by robust assessment and regular reviews.

One aspect of managing individuals convicted of a sexual offence involves the use of the polygraph. Commonly referred to as a ‘lie detector test’, the polygraph works on the assumption that telling lies causes more arousal or stress (i.e. increased respiration, cardiovascular activity, and sweating) than telling the truth does. Since 2014 in England and Wales, it has been a legal requirement that all individuals convicted of sexual offences deemed to be at high risk of reoffending undergo polygraph testing following release into the community, to assess their compliance with licence conditions and risk. The polygraph is not currently used in this way in Scotland and Northern Ireland.

While the use of the polygraph is controversial, research has shown that its use enables professionals to elicit more information. In a two-year evaluation, polygraph testing by police in England was found to increase the likelihood that individuals revealed risk-relevant and in-depth information, resulting in more police action to protect the public. As noted above, however, treatment programmes that use the polygraph have been found to be less effective than those that do not.

Circles of Support and Accountability

As noted previously, individuals are less likely to reoffend if they have support on release from prison. Circles of Support and Accountability (CoSA), where a small ‘circle’ of trained volunteers provide support to a core member – the individual convicted of a sexual offence, e.g. a CSA offence – who has been released into the community, were developed in Canada and have been adopted in a number of countries including the UK. Although CoSA programmes around the world are based on similar principles, there are differences in their operation.

The only randomised control trial to date was conducted in the USA with 50 core members assessed as having a moderate risk of reconviction of another sexual offence: none of them was convicted of a new contact sexual offence in a six-year period, compared to four of 50 men in a control group.

In other countries including the UK, CoSA programmes tend to recruit lower-risk individuals, making it more difficult to assess effectiveness using reoffending rates. Those who take part in CoSA have to meet a number of criteria, such as agreeing to regular meetings, which may make them less likely to reoffend regardless of the support they receive from CoSA. Evaluations of CoSA programmes in the UK and overseas have shown that they can assist core members to integrate into communities – for example, through positive role modelling and pro-social support which is related to desistance.

Interventions for those at risk of committing child sexual abuse

The multiple pathways into offending show how difficult it is to distinguish those who commit CSA from those who do not, unless they request support or help. Prevention efforts are hampered by the erroneous belief that perpetrators of CSA are ‘sick’ or noticeably different to the rest of the population, as risk in apparently ‘normal’ individuals may be overlooked as a result.

In recent years, interventions targeting individuals at risk of perpetrating CSA have attracted increased attention and resources, partly because sexual interest in children is one of the best predictors of sexual reoffending against children. These interventions are important, as a Canadian study of men convicted of sexual offences reported that they had struggled with their sexual thoughts for almost a decade on average before their first arrest. Research also shows that many individuals who define themselves as sexually attracted to children will actively seek help, such as by engaging in mentoring programmes.

Concerns have been raised that the use of the term ‘paedophilia’ can confuse understandings of CSA. The term ‘paedophilia’ is used to describe a sexual interest in pre-pubescent children, but it is frequently applied to those who commit offences against older children. Paedophilia has a distinct clinical definition and criteria for diagnosis that includes both intense and persistent sexual interests in children and the existence of distress, dysfunction, and/or acting on the interests. Nevertheless, it is important to note that not all CSA offences are committed by paedophiles, not all people who have a sexual interest in children are paedophiles, and not all paedophiles commit sexual offences. In addition, not all individuals who commit CSA offences have a sexual interest in children.

A further indication of the need for interventions targeting those at risk is that individuals who recognise a sexual attraction to children, and/or who have been or are being investigated for sexual offences, have an elevated suicide risk. In England and Wales in 2017/18, 29 of the 56 apparent suicides following detention in police custody involved individuals who had been detained for alleged sexual offences, and 13 of these were related to CSA imagery. 

Some organisations and practitioners in the UK – including the Lucy Faithfull Foundation, Safer Lives, the Safer Living Foundation and StopSO – have developed activities and interventions for a range of individuals who may be at risk of committing a sexual offence, or are being investigated or prosecuted for sexual offences. Most of this work has yet to be evaluated, however.

Stop It Now! originated in the USA and now has affiliated organisations in the UK and Ireland and the Netherlands. Besides information and resources, it offers a confidential helpline and support for anyone with concerns about CSA. Although the extent to which these reduce the perpetration of CSA is unknown, an evaluation of the Netherlands and UK/Ireland helplines found that they enabled people with sexual thoughts, feelings or behaviours towards children to better understand and manage their behaviour, and informed them about how to protect children from risk of harm.

Prevention interventions for the general population

Comparatively little research attention has been given to ‘primary prevention’ interventions, targeted at the general population to reduce the likelihood of CSA offending. An international review identified significant gaps in the evidence base for interventions aimed at preventing violence against women and girls. In respect of sexual violence (but not specifically CSA), it found good evidence that the following preventative interventions are effective, when well designed and executed:

  • community activism to shift harmful gender attitudes, roles and social norms in the general population through multi-year intensive community activities
  • school-based interventions to prevent dating violence or sexual violence, with more effective approaches being longer and focused on transforming gender relationships.

There was also “good but conflicting” evidence for the effectiveness of intensive interventions with men and boys alone, but good evidence that brief bystander interventions have no effect. The limited evidence in relation to social marketing campaigns, ‘edutainment’ and digital technologies suggested that they do not change violent behaviour but may raise awareness and influence attitudes; they were felt most likely to be useful as part of multi-component interventions.