Childhood Physical and Sexual Abuse as a Predictor of Reoffending: A Systematic Review

Abstract

Experiences of childhood trauma such as physical and sexual abuse are common among offender populations, and a trauma history is related to increases in the risk of criminal conduct. Previous reviews on the link between childhood trauma and reoffending are limited in scope as they only investigate adolescent populations. The aim of the current systematic review was to explore the association between childhood physical and sexual abuse and its prediction on actual reoffending rates in adult population(s). Four databases (PsycINFO, Web of Science, PubMed, and PILOTS) were searched, and the final included articles were quality assessed by two reviewers. A total of 3,151 studies were identified, and 13 met the a priori eligibility criteria. Most of the studies obtained high quality ratings. A narrative synthesis of the results suggests that childhood physical and sexual abuse are generally high among the prison population and tend to be even higher among those who reoffend. A substantial number of studies found that childhood physical and sexual abuse predict reoffending, and the majority of the studies found a positive direction for this association, albeit of a small magnitude. These findings suggest that prisons should consider trauma-informed assessment and rehabilitation of offenders.

Keywords

childhood trauma physical abuse sexual abuse reoffending systematic review

Reported rates of childhood trauma experiences in the prison population are consistently higher than the general population. For example, a large study conducted in the United States found that 44.7% of male prisoners had experiences of childhood physical abuse, and the rates of trauma were more prevalent in childhood than adulthood (Wolff & Shi, 2012). Similarly, another study conducted in the United States found that 56% of male prisoners had experienced physical abuse in their childhood (Wolff, Shi, & Siegel, 2009), whereas rates of physical abuse among a sample of men from the general community in the United States were 22% (Briere & Elliot, 2003). In terms of sexual abuse, a study by Johnson and colleagues (2006) found that 59% of the male inmates in their sample had experienced some form of childhood sexual abuse. This is not limited to male prisoners, as high rates of childhood physical and sexual abuse have also been reported among incarcerated females (Zlotnick, 1997). In comparison, the prevalence of child sexual abuse in the general population was 10.14% (Perez-Fuentes et al., 2013), although the numbers vary depending on country and gender (Singh, Parsekar, & Nair, 2014). The Ministry of Justice in the United Kingdom reported that 29% of prisoners had experienced either physical, sexual, or emotional abuse as a child (Williams, Papadopoulou, & Booth, 2012). Due to differences in the definition of trauma and abuse experiences, the prevalence rates of trauma experiences in the prison population vary across studies. However, the rates of trauma experiences among inmates seem to be consistently higher than in the general population (see Stoltenborgh, van Ijzendoorn, Euser, & Bakermans-Kranenburg, 2011).

Adverse childhood experiences (ACEs) that include childhood physical, sexual, and emotional abuse and neglect are related to increases in the risk of a range of negative outcomes later in life (Anda et al., 2006). Given the high rates of traumatic experiences among the prison population, an important factor to consider is whether ACEs are related to increases in the rates of incarceration and further reoffending. A case in point is the study by Jäggi, Mezuk, Watkins, and Jackson (2016), which found that having experienced four or more traumatic incidents was associated with a 4-fold increase of being arrested. However, this study did not limit the traumatic experiences to childhood only. The distinction between experiences of trauma in childhood versus adulthood is important for two main reasons. Firstly, it has been hypothesized that children who have experienced neglect may care less about parents’ punitiveness and therefore be more likely behave in antisocial ways in adulthood (Farrington, 2005). Secondly, it has been proposed that children who have experienced trauma might become emotionally overactive and more likely to develop a disorganized attachment style, both of which are linked to later acts of aggression (Ruth-Lyons, 1996; Smith, Leve, & Chamberlain, 2006). In fact, several studies have documented how traumatic childhood experiences are related to an increase in the risk of later acts of aggression and violence (Dutton & Hart, 1992; Sansone, Leung, & Wiederman, 2012).

The Current Review

Despite the plethora of evidence demonstrating a link between trauma and later acts of violence, to date no systematic review has investigated the association between childhood physical abuse (CPA) and childhood sexual abuse (CSA) and rates of reoffending among adults. Single studies such as one by Moore and Tatman (2016) found that ACEs significantly predicted the risk of reoffending. A limitation of this study was that risk rather than actual reoffending rates was measured. A South Korean study that looked at different types of childhood maltreatment in the prediction of self-reported reoffending found that only childhood physical neglect was related to increased likelihood of reoffending (Kim, Park, & Kim, 2016). However, other studies have found no association between childhood trauma and reoffending (Lowenkamp, Holsinger, & Latessa, 2001). These contradictory findings could be due to differences in how reoffending and trauma are measured and/or defined and what additional factors have been controlled for in the studies. In addition, several studies have been conducted on adolescent offenders and the association between trauma and reoffending. The studies on adolescents were summarized in a meta-analysis by Mallie, Viljoen, Mordell, Spice, and Roesch (2011) who found that there was a small, albeit significant association between childhood traumatic experiences and sexual reoffending but that trauma did not predict general reoffending rates. However, adolescent populations differ from adult populations in several respects (e.g., impulsivity levels; Romer, 2010), and there could be differences in outcomes of reoffending among adults.

Therefore, the current systematic review will look at the association between childhood physical and sexual abuse and actual reoffending rates (as opposed to risk of reoffending) in adult populations and seek to highlight the potential discrepancies in outcomes between the studies. The review is limited to experiences of sexual and physical abuse, as these forms of abuse are the most prevalent forms of self-reported childhood trauma among the prison populations, especially physical abuse among men and sexual abuse among females (Browne, Miller, & Maguin, 1999; Johnson et al., 2006; Wolff & Shi, 2012; Zlotnik, 1997). In addition, physical abuse has been shown to be related to later acts of both violent and nonviolent nature (Malinosky-Rummell & Hansen, 1991; Wolff & Shi, 2012) and that childhood sexual abuse is related to later acts of sexual violence, especially among men (Plummer & Cossins, 2016).

Method

Search Strategy

The current study protocol was registered on Prospero in October 2018 (ID = CRD42018112115). The systematic review was carried out using four online databases: PsycINFO (1806–2018), Web of Science (1970–2018), PubMed (1996–2018), and the National Centre for PTSD Research Published International Literature on Traumatic Stress (PILOTS; 1871–2018), as of October 2018. The search strategy was limited to studies published in the English language, peer-reviewed articles, and included studies published from the earliest availability of the databases to October 2018. Combinations of key words relating to childhood sexual and physical abuse and reoffending were used, that is (child* maltreat*, trauma, child* advers*, child* abuse) AND (recidi*, reoffend*, repeat offend*, rearrest, reconvict*, reincarcerate*, revoke*). These terms were searched as key words in the online databases. The reference lists of relevant articles were also screened for potential titles.

Definitions

Childhood physical and sexual abuse were defined in accordance with how the two constructs are measured in Bernstein, Ahluvalia, Pogge, and Handelsman’s (1997) Childhood Trauma Questionnaire that includes a range of traumatic experiences that occurred before the age of 18 years. The current review focused specifically on childhood physical and/or sexual abuse. CPA was defined as experiencing any physical assault such as being beaten or injured by someone (e.g., “I was physically abused,” “I was beat with an object”) before the age of 18 years. Childhood sexual abuse was defined as being sexually assaulted, molested, or touched through force or persuasion (e.g., “Someone made me do sexual things against my will,” “I was touched in a sexual way against my will”) before the age of 18 years.

Eligibility Criteria

The inclusion criteria were the following: (a) original research article published in a peer-reviewed journal, (b) full-text articles available in English, (c) physical/sexual abuse present before the age of 18 years, (d) actual reoffending as outcome (as opposed to risk of reoffending), (e) reoffending measure reported, and (f) association between childhood physical and sexual abuse and reoffending made explicit. Exclusion criteria included (a) child, adolescent, or juvenile samples (<18 years) at the time their reoffending was measured, (b) non-English article, and (c) noncriminal sample.

Study Selection and Data Extraction

The lead author (M.D.) reviewed all the identified studies to determine eligibility. The studies were first screened by title and irrelevant studies were excluded, the remaining were screened by abstracts, whereby some were excluded, and then the subsequent sample of studies was screened in full text. Another independent reviewer screened a random sample of 10% of the studies by title and abstract to assess for inter-rater agreement at these stages. The last set of studies was assessed against the inclusion and exclusion criteria for eligibility independently by two reviewers (M.D. and an independent reviewer). Any disagreements between the reviewers were discussed and resolved.

Data regarding sample characteristics, trauma and reoffending measure, confounding variables, and statistics regarding the association between child physical and sexual abuse and reoffending were extracted from each study (e.g., frequencies, correlations, regression coefficients, and other effect sizes). Although studies might have looked at several childhood maltreatment measures or other measurements unrelated to abuse, only physical and/or sexual abuse were noted. Where the data were incomplete or vague, the authors of the relevant studies were contacted.

Results

A total of 3,151 publications were found across all the databases, from which 525 duplicates were removed, which resulted in a remaining 2,626 studies (see Figure 1, for a flowchart of the process). These studies were then screened by title, following which 2,378 were excluded. Two hundred forty-eight studies were then screened by abstract, whereby 225 studies were excluded. This left only 23 studies on adult populations. The 23 remaining studies were screened by reading the full text, following which 10 studies were excluded. The inter-rater reliability for assessment of the eligibility of the 23 studies was .82 (Cohen’s κ). The raters disagreed on the eligibility of two studies, which was discussed and resolved, and the final sample included 13 studies. The reference lists of the final 13 studies were also scanned to see whether other relevant studies could be included.

Figure 1.

PRISMA flow diagram depicting the different stages of finding relevant studies on the association between childhood physical abuse/childhood sexual abuse and actual reoffending in adult populations.

Quality Assessment

A quality assessment of the final sample of studies was conducted by two reviewers, one author of the current paper (M.D.) and one independent reviewer, using a 16-item quality assessment tool (Quality Assessment Tool for Studies with Diverse Designs) where each item is scored from 0 (not at all) to 3 (complete) (Sirriyeh, Lawton, Gardner, & Armitage, 2012). Items 11 and 14 were excluded as they are only relevant for qualitative studies, and the highest score possible to obtain was 42. The inter-rater agreement in the quality rating (intra-class correlation) was .85, which can be considered good agreement (Koo & Li, 2016).

The quality scores were varied across the studies (M = 28.61, SD = 6.72, range = 17–38, median = 31). The quality scores for each study are presented in Table 1. The majority of the studies gave clear descriptions of the theoretical rationale and aims of their research, and many presented the rationale, reliability, and validity of the measurements used, which resulted in higher quality scores for these areas. Due to the nature of the research question, experimental manipulation of the variables of interest (childhood abuse and reoffending) is not possible, and therefore appropriate statistical analyses would be correlational. However, some studies did not perform logistic regression analyses when this could have been appropriate and only looked at frequencies of childhood abuse and recidivism and/or correlations between these variables.

Table 1.

Overview of the Studies Including Their Statistical Analyses and Key Findings.

Study Name	Childhood Abuse Measure	Reoffending Measures	Variables Controlled for	Sample Eligibility	Quality Score	Statistical Analysis on Variables of Interest	Statistical Associations	Key Findings
Benda and Toombs (2002)	Physical and sexual abuse (4-point Likert-type scale on frequency)	New felony or parole violations after release from boot camp (official records)	Age, race, marital status, IQ, education, sentence length, reading ability, social selection, and causation model variables	No prior violent offenses, no substance abuse history, no physical or psychological problems, sentence of 10 or less years.	34/42	Cox proportional hazard analysis of survival	CPA and reoffending, β = .79, HR = 2.20. CSA and reoffending, β = 1.01, HR = 2.75	Childhood physical and sexual abuse is related to a statistically significant increase in reoffending. The effect was small in magnitude
Benda (2005)	Physical and sexual abuse measured by Childhood Trauma Questionnaire (5-point Likert-type scales)	New felony or parole violation after release from boot camp (official records)	Age, race, residence, drug and alcohol use, social network, mental health variables	No prior violent offences, sentenced for 10 or less years, no physical or psychological problems, no substance abuse problems	33/42	Cox Proportional Hazard Analysis of Survival separately for men and women	Women: CPA and reoffending, β = .45, HR = 1.56, CSA and reoffending, β = .57, HR = 1.77. Men: CPA and reoffending, β = .28, HR = 1.31, CSA and reoffending, β = .32, HR = 1.37	Childhood physical and sexual abuse were related to statistically significant increases in reoffending (of a small magnitude) for both men and women, but the effect was larger for women
Firestone (1999)	Self-reported physical and sexual abuse (dichotomous variable)	Sexual, violent and criminal reoffending (official records)	None controlled for in analysis	Convicted of hands on sexual offences among related members of family under age 16. If ever charged for sexual conviction of unrelated child they were excluded	17/42	Percentages and frequencies reported, t test	Sexual reoffending group: 28.6% CPA vs. 41.3% no CPA, 37.5% CSA vs. 38.3% no CSA. Violent reoffending group: 41.7% CPA vs. 40.2% no CPA, 35.5% CSA vs. 38.6% no CSA. Criminal reoffending group: 58% CPA versus.34.5% no CPA, 46.3% CSA vs. 35.3% no CSA	Lower frequencies of physical abuse in sexual reoffender group, and similar amounts of abuse ratings in violent reoffender groups. However, for criminal reoffending there was higher amounts of physical and sexual abuse in reoffending group.
Fowler, Cantos, and Miller (2016)	Self-reported physical abuse in childhood in reports available to probationers	Official police records of offence related to domestic violence (official records)	Age, employment status, race, education, marital status	Males who were involved in domestic violent abuse toward intimate partner. Only experiences of sexual abuse were excluded	36/42	Logistic regression of childhood abuse and reoffending	CPA and reoffending, OR = 1.86 (CI [0.68, 5.05]), HR = 1.62 (CI [0.74, 3.54])	The males who experienced physical abuse in childhood were 1.86 times more likely to reoffend in terms of domestic violence, but this finding was not statistically significant
Herbst et al. (2016)	Self-report childhood victimization assessed through Adverse Childhood Experiences Scale (ACE)	Reports of first time versus not first time being incarcerated at time of study (self-report)	Age, housing, past drug use, partner concurrency	Females who were HIV negative and incarcerated	26/42	Logistic regression of childhood sexual abuse and reoffending	CSA and reoffending, AOR = 1.41 (CI [0.94, 2.11]). 40.4% with CSA in repeat incarceration group vs. 31.1% in initial incarceration*	1.41 greater likelihood to have repeatedincarcerations if experienced CSA. However, not statistically significant
Kim et al. (2016)	Childhood Trauma Questionnaire	Self-reported reoffending rates (self-report)	Age, gender, education, living situation, mental health factors and other childhood abuse types.	Probationers	30/42	Multivariate logistic regression	CPA and reoffending, OR = 0.75 (CI [0.28, 2.02]). CSA and reoffending, OR = 1.61 (CI [0.54, 4.79]).	No significant relationship between physical and sexual abuse in childhood and reoffending.
Kimonis et al. (2010)	Child Abuse and Trauma Scale	Arrest records for 1-year follow-up (official records)	Not controlled for others	Females either in prison or substance abuse facility, excluded if active symptoms of psychosis.	32/42	Correlational analysis	CPA and reoffending, r = .01. CSA and reoffending, r = −.01.	No significant relationship between physical and sexual abuse in childhood and reoffending
Kjelsberg and Friestad (2008)	Structured interview about childhood adversities	Reoffending defined as having served at least one prison sentence before the current one (official records)	None controlled for in analysis	Randomly drawn sample of prison inmates	28/42	Percentages suffering abuse in first time and repeat sentenced	21% abuse in first time sentence, 34% abuse in recidivist group*	Statistically significant higher number of individuals in the repeat offender group had experienced childhood abuse
Levenson and Socia (2016)	Adverse Experiences Scale (ACE)	Arrest scale score, self-reported by client (self-report)	Age, gender, race, marriage status	Sex offenders in outpatient and civil commitment programs	38/42	Multiple regression analysis	CPA on total arrests, β = −.013. CSA on total arrests, β = .028. CPA on total sex crime arrest, β = −.008. CSA on total sex crime arrest, β = .090*. CPA on total nonsex arrests, β = −.016. CSA on total nonsex arrests, β = −.022	No significant association between physical and sexual abuse in childhood and total number of domestic violence arrests. The only statistically significant finding was on childhood sexual abuse predicting the total sex crime arrests
Lindsay et al. (2004)	Staff interviewed about participants background, and casenote analysis	Reports of sexual reoffending and suspicion of reoffending	Various static and dynamic risk factors in correlational model	Male sex offenders with mild to moderate intellectual disability	19/42	Correlational analysis (Spearman’s r) on evidence of reoffending and suspicion of reoffending	CSA and evidenced reoffending, r = .16 and for suspected reoffending r = .33*. CPA and evidenced reoffending, r = .12, and suspected reoffending, r = .25	Only significant association was between sexual abuse in childhood and suspected reoffending (moderate relationship)
Lowenkamp et al. (2001)	Child abuse data collected in intake interviews	Record check of reincarceration to prison facilities (official records)	Age, race, risk score, time at risk.	Felony offenders in prison	23/42	Multiple logistic regression analysis	Childhood abuse (physical and/or sexual) and reoffending, β = −.06, OR = 0.94 (CI [0.48, 1.805])	No significant relationship between child abuse and reoffending
Shepard (1992)	Child abuse information through questionnaires	Convictions for domestic assault, orders for protection due to domestic abuse and police suspicion of domestic assault (official records)	Not controlled for other variables	Domestic violence offenders in community intervention programs	22/42	Standardized canonical discriminant function coefficients	Childhood abuse and reoffending, β = .28 (no significance level reported).	Individuals with an abuse history were somewhat more likely to reoffend, but it is unclear whether this was statistically significant or not
Tikkanen et al. (2010)	Child physical abuse obtained from multiple interviews and collateral information	Reoffending data from legal records (official records)	MAOA, MAOA-L and MAOA-H genotype, heavy drinking	Male violent alcoholic offenders with violent crimes. Excluded if psychosis.	34/42	Logistic regression	CPA and reoffending, OR = 3.6 (CI [1.72, 7.49]) when controlled for MAOA. CPA and reoffending, OR = 5.3 (CI [1.77, 15.60]) controlled for MAOA-H. CPA and reoffending, OR = 2.6 (CI [0.86, 8.02]) controlled for MAOA-L	Childhood physical abuse was related to an increased rate of violent reoffending, with magnitudes ranging from medium to large depending on the genotype controlled for

Note. CPA = childhood physical abuse; CSA = childhood sexual abuse; HR = hazard rate; OR = odds ratio; CI = 95% confidence interval.

*p < .05. **p < .01.

Study Characteristics

Demographics and setting

All the studies included were published in English, and the publication dates of the studies included in the review ranged from 1992 to 2016. Overall, the total sample size across the 13 studies was 4,297 individuals. The mean age of the participants across the studies that reported mean ages was 24.8 years (SD = 4.86). Six studies included only male participants and two were only female. Overall, there were 69.2% males and 30.8% females across the studies. As can be seen from Table 2, the majority of the studies were conducted in the United States (n = 8), and the rest were conducted in Canada (n = 1), the United Kingdom (n = 1), South Korea (n = 1), Norway (n = 1), and Finland (n = 1).

Table 2.

Overview of Studies Exploring the Relationship Between CPA and CSA on Reoffending Rates.

Study Name	Country	Study Design	Number of Subjects	Mean Age (SD)	Gender	Institutional Setting	Trauma Measure	Reoffending Measure	Follow-Up
Benda and Toombs (2002)	The United States	Prospective cohort study	480	24.8 (10.2)	Males	Boot camp for adults	Self-reported physical and sexual abuse	Parole violations and arrest after release (any)	3 years prospectively
Benda (2005)	The United States	Prospective cohort study	600	25.9 (6.1), m, 24.1 (4.3), f	Mixed (50% male)	Boot camp for adults	Childhood Trauma Questionnaire (CTQ): physical and sexual abuse	Parole violations and arrest after release (any)	5 years prospectively
Firestone et al. (1999)	Canada	Retrospective cohort study	251	Adult ≥ 18	Males	Sexual Behaviours Clinic	Self-reported childhood physical abuse	Sexual, violent or criminal reoffending after index offence	12 years postrelease
Fowler et al. (2016)	The United States	Retrospective cohort study	228	33.6	Males	Adult probation services	Prison reports of childhood physical abuse	Domestic violence official reoffending records	3 years post completion of probation
Herbst et al. (2016)	The United States	Cross-sectional	521	Single: 31.8 (8.8) Repeat: 34.9 (9.2)	Female	Prison	Childhood Experiences (ACE) Scale: childhood physical and sexual abuse	History of repeat incarcerations	No follow-up
Kim et al. (2016)	South Korea	Cross-sectional	183	40.1 (11.8)	Mixed (83.6% male)	Probationers	Childhood Trauma Questionnaire: physical and sexual abuse	Self-reported reoffending	No follow-up
Kimonis et al. (2010)	The United States	Retrospective cohort study	266	Adults ≥ 21	Females	Prison and substance abuse treatment facility	The Child Abuse and Trauma Scale: physical and sexual abuse	Official arrest data	1 year
Kjelsberg and Friestad (2008)	Norway	Cross-sectional	260	Adults ≥ 18	Mixed (86.5% male)	Prison	Self-report through interview: sexual and physical abuse	History of repeat incarcerations	No follow-up
Levenson and Socia (2016)	The United States	Cross-sectional	740	Adults ≥ 18	Mixed (93.5% male)	Outpatient and civil commitment programs	Adverse Childhood Experiences Scale (ACE): Physical and sexual abuse	Number of arrests (multiple arrests versus single arrest): Self-reported	No follow-up
Lindsay, Elliot, and Astell (2004)	United Kingdom	Cross-sectional	52	35.6 (nor reported)	Male	Treatment facility for people with intellectual disability	Childhood sexual and physical abuse	Evidence and suspicion of sexual reoffending	No follow-up
Lowenkamp et al. (2001)	The United States	Cross-sectional	442	29.6 (8.45)	Mixed (72% males)	Prison	Self-reported childhood physical and sexual abuse	Reincarceration to prison	No follow-up
Shepard (1992)	The United States	Retrospective cohort study	100	31.9	Males	Community	Self-reported childhood physical abuse	Records of reoffending on convictions for domestic assault	5 years
Tikkanen et al. (2010)	Finland	Retrospective cohort study	174	32.5	Males	Psychiatric Unit	Self-reported childhood physical abuse	Official records of reoffending	11.7 years

Note. CPA = childhood physical abuse; CSA = childhood sexual abuse.

The settings in which the studies were carried out were diverse. Only four studies were conducted in a prison facility (Herbst et al., 2016; Kimonis et al., 2010; Kjelsberg & Friestad, 2008; Lowenkamp et al., 2001). The rest were either outpatient/community settings, probation setting, or psychiatric/treatment facility units.

The studies also differed in their inclusion and exclusion criteria for participants, where some reported more stringent criteria than others. For example, there were differences in the severity of participants’ prior offences. Two studies (Benda, 2005; Benda & Toombs, 2002) only included participants with no prior violent offences and with a sentence of 10 or less years, whereas one study specified that the participants had to be violent offenders (Tikkanen et al., 2010). Four studies required that the participants had to have an IQ equal to or above 70 (Benda, 2005; Benda & Toombs, 2002; Kimonis et al., 2010; Tikkanen et al., 2010), while one study only included individuals with an intellectual disability (IQ < 70; Lindsay, Elliot, & Astell, 2004). Four studies excluded participants who had psychosis (Benda, 2005; Benda & Toombs, 2002; Kimonis et al., 2010; Tikkanen et al., 2010).

Trauma predictor measure

Of all the included studies, nine measured and reported on both physical and sexual childhood abuse, whereas the other four studies only reported CPA. Five studies used standardized questionnaires that were valid and reliable to measure childhood abuse, such as the Childhood Trauma Questionnaire or the Adverse Experiences Scale. The rest of the studies relied on interviews, case note analysis, or collateral information to gather information about the participants’ trauma history.

Mode of reoffending

Eight studies operationalized reoffending as any new felony after release (Benda, 2005; Benda & Toombs, 2002; Firestone et al., 1999; Kim et al., 2016; Kimonis et al., 2010; Levenson & Socia, 2016; Lowenkamp et al., 2001; Tikkanen et al., 2010), two studies looked at domestic violence specifically (Fowler, Cantos, & Miller, 2016; Shepard, 1992), one measured sexual reoffending (Lindsay et al., 2004), and two studies operationalized reoffending retrospectively as any prior incarcerations (Herbst et al., 2016; Kjelsberg & Friestad, 2008). Ten studies used official records to measure reoffending status (Benda, 2005; Benda & Toombs, 2002; Firestone, 1999; Fowler et al., 2016; Kimonis et al., 2010; Kjelsberg & Friestad, 2008; Lindsay et al., 2004; Lowenkamp et al., 2001; Shepard, 1992; Tikkanen et al., 2010). Three studies measured reoffending by self-report (Herbst et al., 2016; Kim et al., 2016; Levenson & Socia, 2016).

Study design and follow-up lengths

As can be seen in Table 2, seven studies reported a follow-up period, with the minimum being 1 year and the longest being 12 years (M = 5.8, SD = 4.3). Only two studies were prospective cohort studies (Benda, 2005; Benda & Toombs, 2002) and their follow-up periods were 3 and 5 years, respectively; whereas, the other five studies were retrospective cohort studies. The studies that did not include follow-up periods were all cross-sectional by design.

Analyses to explore trauma and reoffending link

As can be seen in Table 1, the majority (n = 8) of the studies performed some type of multiple regression (e.g., logistic hierarchical regression or Cox proportional regression) that includes odds ratios for risk of reoffending. Three studies reported correlational analyses, and two studies only reported the frequencies of trauma and reoffending.

Prevalence of Childhood Physical and Sexual Abuse

As can be seen from Table 3, 10 studies reported the prevalence rates of childhood physical and sexual abuse in their samples (Benda, 2005; Firestone et al., 1999; Fowler et al., 2016; Herbst et al., 2016; Kim et al., 2016; Kjelsberg & Friestad, 2008; Levenson & Socia, 2016; Lowenkamp et al., 2001; Shepard, 1992; Tikkanen et al., 2010). The weighted mean prevalence rate of childhood physical and sexual abuse was calculated by giving higher weights to larger sample sizes. The weighted mean of the prevalence rates of CPA was 32.2%, and the weighted mean for the prevalence rate of childhood sexual abuse was 33.8%. The median percentage of physical abuse prevalence was 24% (range 11.3–56%), and the median percentage of sexual abuse was 28% (range 11.3–56%). It is noteworthy that the study by Lowenkamp, Holsinger, and Latessa (2001) did not report separate values for childhood physical and sexual abuse and combined both in their calculations.

Prevalence of Reoffending

Ten studies reported the rate of reoffending in their samples (Benda, 2005; Benda & Toombs, 2002; Firestone et al., 1999; Herbst et al., 2016; Kimonis et al., 2010; Kjelsberg & Friestad, 2008; Lindsay et al., 2004; Lowenkamp et al., 2001; Shepard, 1992; Tikkanen et al., 2010). The weighted mean percentage rate of reoffending across the 10 reported studies was 51.7%. For the studies that reported different types of reoffending (e.g., sexual, violent, criminal), the percentages were combined to calculate the weighted mean. The median percentage of reoffending was 40% (range: 34–62.3%). It is noteworthy that the separate rate of sexual reoffending was of lower prevalence rates such as only 4.8% sexual reoffending in the Firestone et al. (1999) study.

Associations Between Childhood Physical and Sexual Abuse and Reoffending

Tables 1 and 4 give an overview of the key findings related to childhood physical and sexual abuse and reoffending. There were some conflicting findings regarding the association between childhood physical and sexual abuse and reoffending. As can be seen in Table 4, across the different types of analyses, there were six studies that found a significant relationship between childhood physical and sexual abuse and reoffending (Benda, 2005; Benda & Toombs, 2002; Firestone et al., 1999; Herbst et al., 2016; Kjelsberg & Friestad, 2008; Tikkanen et al., 2010). The majority of these found a positive association between childhood abuse and reoffending, in which experiences of childhood abuse were associated with increases in reoffending rates. Only the study by Firestone et al. (1999) found that rates of CPA were lower among those who committed sexual reoffences, while those who performed criminal reoffences (general nonviolent and nonsexual) had higher rates of childhood physical and sexual abuse. However, this study purely reported frequencies and did not control for confounding variables. The results in this study should therefore be interpreted with caution.

Table 3.

Summary of the Prevalence Rates of Childhood Physical and Sexual Abuse in the Entire Sample of Each Study That Reported Prevalence Rates.

Study	Abuse Type	Prevalence, % (n)	Sample Size
Benda (2005) ^a	CPA CSA	56 (336) 56 (336)	600
Fowler, Cantos, and Miller (2016)	CPA	12.2 (29)	228
Herbst et al. (2016)	CSA	37 (193)	521
Firestone et al. (1999)	CPA CSA	17.1 (43) 19.1 (48)	251
Kim, Park, and Kim (2016)	CPA CSA	24 (49) 12 (25)	206
Kjelsberg and Friestad (2008)	CPA	29 (65)	260
Levenson and Socia (2016)	CPA CSA	42 (311) 38 (281)	740
Lowenkamp, Holsinger, and Latessa (2001) ^b	CPA/CSA	11.3 (50)	439
Shepard (1992)	CPA	24 (24)	100
Tikkanen et al. (2010)	CPA	29 (51)	174

Note. CPA = childhood physical abuse; CSA = childhood sexual abuse.

^a Prevalence was combined for males and females to total scores. ^bSample size is smaller than total sample size due to missing numbers in abuse measure.

Table 4.

Summary of the Relationships Between CPA and CSA and Reoffending.

Study	Abuse Type	Significant Association Present	Direction of Association	Magnitude of Effect	Quality Ranking
Regression studies
Benda and Toombs (2002)	CPA	Yes	+	Small	3
Benda and Toombs (2002)	CSA	Yes	+	Small	3
Benda (2005)	CPA	Yes	+	Small	5
Benda (2005)	CSA	Yes	+	Small	5
Fowler et al. (2016)	CPA	No	+	Small	2
Herbst et al. (2016)	CSA	No	+	Small	9
Kim et al. (2016)	CPA	No	−	Small	7
Kim et al. (2016)	CSA	No	+	Small	7
Tikkanen et al. (2010)	CPA	Yes	+	Moderate–large	4
Correlational studies
Kimonis et al. (2010)	CPA	No	+	Small	6
Kimonis et al. (2010)	CSA	No	−	Small	6
Levenson and Socia (2016)	CPA	No	−	Small	1
Levenson and Socia (2016)	CSA	Yes	+	Small	1
Lindsay et al. (2004)	CPA	No	+	Small	12
Lindsay et al. (2004)	CSA	No	+	Small	12
Lowenkamp et al. (2001)	CPA/CSA	No	−	Small	10
Shepard (1992)2	CPA	Not reported	+	Moderate	11
Frequency studies
Firestone et al. (1999)	CPA	Yes	+	N/A	13
Firestone et al. (1999)	CSA	Yes	+	N/A	13
Kjelsberg and Friestad (2008)	CPA	Yes	+	N/A	8

Note. “+” Indicates abuse predicts higher degrees of reoffending/positive correlation and “−” indicates abuse predicts lower degrees of reoffending/negative correlation. Reoffending refers to actual reoffending rates (not suspected or risk). Rating of relative quality ranking: 1 (highest quality) to 13 (lowest quality). CPA = childhood physical abuse; CSA = childhood sexual abuse.

As can be seen from Table 4, the majority of the studies (n = 9) only found associations of a small magnitude (see Chen, Cohen, & Chen, 2010, for an interpretation of odds ratios), regardless of whether the findings were statistically significant or not. The study by Benda (2005) found greater effect sizes for women (CPA: 1.56, CSA: 1.77) than for men (CPA: 1.31, CSA: 1.37), although these can all be considered effects of small magnitude. Only one of the studies that reported a regression analysis (Tikkanen et al., 2010) found a relationship between CPA and violent reoffending of a moderate to large magnitude. They found that having experienced CPA was related to an increase of 2.6–5.3 times in the rates of reoffending violently, depending on the genotype controlled for.

Confounding Variables

The studies that used some form of regression analysis differed in terms of what variables they controlled for. All the studies, with the exception of Tikkanen et al. (2010), that used regression analysis controlled for age in the analysis. Three of these studies controlled for drug and/or alcohol use (Benda, 2005; Herbst et al., 2016; Tikkanen et al., 2010). Three studies controlled for some mental health variables such as social anxiety (Benda & Toombs, 2002), depression (Benda, 2005), and at least one psychiatric disorder such as depression or post-traumatic stress disorder (Kim et al., 2016). The study by Kim, Park, and Kim (2016) was the only study that controlled for various other childhood trauma experiences (e.g., neglect).

Discussion

Synthesis of Findings

This systematic review explored childhood trauma, specifically physical and sexual abuse, and its association with actual reoffending among adults. The 13 identified studies were quite diverse in terms of setting, participant population, and reoffending measures used. The studies also controlled for multiple different predictors in the regression models and did not report unadjusted odds ratios, rendering a meta-analysis inappropriate for the current review (see Norton, Dowd, & Maciejewski, 2018; Schmidt & Hunter, 2004). Although aggregate estimates of the pooled effect sizes could not be performed, the results offer some tentative suggestions on how childhood physical and sexual abuse are linked to reoffending. Firstly, the identified studies reported relatively high prevalence rates of childhood abuse, with CPA weighted mean being 32.2% and childhood sexual abuse being 33.8%. This is consistent with the studies showing that prison populations generally have higher rates of trauma experiences than the general population (King, Steenson, Shannon, & Mulholland, 2017; Perez-Fuentes et al., 2013; Wolff & Shi, 2012).

Secondly, the studies that only reported frequencies of abuse experiences (n = 2) found that the rates of abuse experiences were higher among the offenders who reoffended compared to those who did not. For example, Kjelsberg and Friestad (2008) found that those who had reoffended reported significantly higher amounts of CPA (34%) compared to those who had not reoffended (21% reported abuse). Firestone et al. (1999) found that there were higher degrees of childhood physical and sexual abuse among those with general reoffending, whereas he did not find significant differences in these forms of trauma among those who committed violent and sexual reoffences. These studies did not, however, control for confounding variables and the correlations between these trauma experiences and the reoffending behavior could be due to other trauma-related factors. For instance, Herbst et al. (2016) controlled for several confounding variables. Although they found that reoffenders reported significantly higher amounts of childhood sexual abuse compared to those who did not reoffend (40.4% vs. 31.1%), they did not find a significant association between sexual abuse and reoffending when controlled for other confounding variables. Childhood sexual abuse could be related to increases in other adverse outcomes related to reoffending, such as increases in substance abuse (Widom & White, 2006; Wilsnack, Vogeltanz, Klassen, & Harris, 1997) and poor emotion regulation (Coyle, Karatzias, Summers, & Power, 2014; Sharma, Prakash, Sengar, Chaudhury, & Singh, 2015).

A third finding was that although only 4 of the 11 regression and correlational studies found statistically significant relationships between physical and sexual abuse and reoffending, most of the studies were in a positive direction. This suggests that having experienced these forms of childhood abuse seems to be associated with increases in reoffending rather than decreases. However, the majority of the studies (n = 9) found only small associations between childhood physical and sexual abuse and increases in reoffending, when controlling for other variables. Only two studies found moderate or moderate-to-large effects of childhood abuse and reoffending (Tikkanen et al., 2010, and Shepard, 1992, respectively). Interestingly, both these studies looked at CPA only and its relation to violent forms of reoffending (i.e., domestic assault in Shepard, 1992, and general violent reoffending in Tikkanen et al., 2010). It could be that there is a stronger predictive power of CPA and violent reoffending, compared to other forms of reoffending (e.g., nonviolent, minor felonies).

Some methodological differences between the studies could account for the different outcomes in relation to childhood abuse and reoffending. For example, Herbst et al. (2016) did not find that childhood sexual abuse was a significant predictor of reoffending, but they focused solely on female inmates who had volunteered to partake in a program on empowerment to reduce intimate partner violence. In addition, they controlled for dynamic variables such as unstable housing and current drug use, which were significant predictors of reoffending. Studies have shown that dynamic factors tend to be better predictors of reoffending that static, historical variables (Andrews & Bonta, 2010). It could therefore be that these factors had a stronger predictive power for their reoffending patterns rather than their childhood sexual abuse experiences.

There were also differences in how reoffending was defined and measured, which could account for the variations in the results. For example, Fowler, Cantos, and Miller (2016) who also found no significant association between childhood abuse and reoffending, focused solely on intimate partner violence as a measure of reoffending rather than more general reoffending measures. The mechanisms involved in intimate partner violence might differ from the mechanisms involved in other forms of violent and nonviolent reoffending. It is also noteworthy that although the study by Kim et al. (2016) did not find a significant relationship between physical and sexual abuse and reoffending, they did find that neglect—another form of childhood trauma, positively predicted reoffending in their sample.

Limitations and Strengths of the Studies

Several limitations regarding the studies should be noted. The majority of the studies included (n = 11) relied exclusively on retrospective reports of reoffending, either through official records, databases, or by self-report. Most of the studies used official records (e.g., police databases) to verify reoffending status (n = 10), but some studies such as Kim et al. (2016) relied on the offenders own self-reports of reoffending, and no corroborative measures were used. Other studies defined reoffending historically, such as previous history of incarceration or multiple arrest history, and did not have a measure of whether the arrest history was violent, sexual, or nonviolent in nature (e.g., Kjelsberg & Friestad, 2008). Similarly, the rates of childhood abuse were largely measured retrospectively by self-report questionnaires and the accuracies of these reports could be questionable and susceptible to various biases. Only about half of the studies (n = 6) performed more advanced analyses such as logistic regression to look at the predictive power of physical and sexual abuse on reoffending rates. These also controlled for other confounding variables, which was not done in the studies that did not perform these types of analyses. The other studies relied solely on correlational analysis or only included percentages of trauma experiences in the offender groups.

Limitations and Strengths of the Systematic Review

Several limitations of the review should also be noted. Firstly, the review included studies on all types of reoffending such as violent, nonviolent, and sexual reoffending. It could be argued that this is problematic, because the base rates of reoffending depend on the type of criminal act committed. For example, sexual recidivism base rates (i.e., rate of sexual reoffending among sex offenders) tend to be quite low and also lower than other forms of reoffending (Durose, Cooper, & Snyder, 2014; Fazel & Wolf, 2015). For example, a meta-analysis by Hanson and Bussiere (1998) found a base rate of 10–15% for sexual recidivism. Thus, the probability that someone will reoffend sexually is lower than for other types of reoffending. Moreover, the manner in which trauma influences reoffending could differ depending on what type of trauma experience and what type of reoffending is investigated. For example, physical abuse could be a stronger influence on violent reoffending, whereas sexual abuse could be a stronger predictor of sexual reoffending. Although the studies that found larger effect sizes looked at CPA and violent forms of reoffending, there were too few studies on this link to draw clear conclusions about this association.

Lastly, the current review did not search for gray literature and only included peer-reviewed articles. Although only including peer-reviewed articles allows for a quality assurance of the studies included, the disadvantage of not investigating the gray literature is the susceptibility of losing important studies due to the file-drawer problem where nonsignificant findings are not published.

Several strengths should be noted. Firstly, the current review is the first to look at childhood physical and sexual abuse in the prediction of actual reoffending as opposed to risk of reoffending. The risk assessment tools only estimate the likelihood that someone might reoffend, based on various predictor measures (e.g., gender, substance abuse). However, these tools do not measure whether the person actually committed a new criminal act. Although the risk measures, such as the HCR-20 and Risk Matrix 2000, have been found to be valid and reliable, they are not perfect measures of reoffending (Douglas & Belfrage, 2014; Tully & Browne, 2015). Secondly, this was the first review on the association between childhood physical and sexual abuse and reoffending in adult populations. Most previous studies focus on adolescent/juvenile offenders, where the trauma experiences are more recent than in adult samples. Furthermore, although we cannot discern the isolate effect of these types of trauma experiences on reoffending, as the different studies controlled for different confounding variables, the inclusion of studies controlling for different variables can give insight into what factors are important in predicting reoffending. For example, Tikkanen et al. (2010) controlled for heavy alcohol use that has been found to be a strong dynamic risk factor in the prediction of reoffending and still found that there was a main effect of physical abuse on violent reoffending.

Implications

The findings of the current review offer a tentative suggestion that childhood physical and sexual abuse are related to increases in the odds of reoffending, although of a small magnitude. This finding, in addition to the fact that high numbers of prisoners have trauma experiences in their childhood, points to a need to consider trauma in the rehabilitation and interventions for prisoners. Although the association between childhood physical and sexual abuse was generally small, these forms of abuse could be related to increases in other variables such as substance abuse and poor self-regulation, which in turn can increase the risk of reoffending. If this is the case, interventions and programs targeting trauma could still be of importance, due to its potential widespread impact on other maladaptive behaviors. Future research should explore the mechanisms in which trauma experiences contribute to the increased risk of reoffending.

Implications of the Review for Practice, Policy, and Research

Implementing routine assessment of rates of childhood trauma among prisoners due to high prevalence rates.

Tailoring trauma-informed interventions for prisoners to support mental health and potential rehabilitation into the community.

Policy guidelines for prisoner rehabilitation and treatment of mental health should address the high prevalence rates of trauma in this population and highlight the need for these issues to be targeted in rehabilitation.

Future research should explore the potential widespread impact of childhood trauma on other maladaptive behaviors (e.g., substance abuse) that could be linked to reoffending.

Conclusion

In summary, the main findings of the review can be catalogued as follows: There were high rates of childhood physical and sexual abuse experiences among the prison populations generally, and the rates of abuse tended to be higher among those who reoffended. Lastly, there was a trend in the positive direction that childhood physical and sexual abuse predicted reoffending, although the majority of the studies did not reach statistical significance. In addition, the majority of the associations were small in magnitude. Only two studies found effect sizes in the moderate to large magnitude, and these looked at CPA and violent reoffending outcomes.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Donncha Hanna

References

Anda

R. F.

Felitti

V. J.

Bremner

D. J.

Walker

J. D.

Whitfield

Perry

B. D.

… Giles

W. H.

(2006). The enduring effects of abuse and related adverse experiences in childhood: A convergence of evidence from neurobiology and epidemiology. European Archives of Psychiatry and Clinical Neuroscience, 256, 174–186.

Andrews

D. A.

Bonta

(2010). The psychology of criminal conduct. New Providence, NJ: Anderson.

*Benda

B. B.

(2005). Gender differences in life-course theory of recidivism: A survival analysis. International Journal of Offender Therapy and Comparative Criminology, 49, 325–342.

*Benda

B. B.

Toombs

N. J.

(2002). Two preeminent theoretical models: A proportional hazard rate analysis of recidivism. Journal of Criminal Justice, 30, 217–228.

Bernstein

D. P.

Ahluvalia

Pogge

Handelsman

(1997). Validity of the childhood trauma questionnaire in an adolescent psychiatric population. Journal of the American Academy of Child & Adolescent Psychiatry, 36, 340–348.

Briere

J. B.

Elliott

D. M.

(2003). Prevalence and psychological sequelae of self-reported childhood physical and sexual abuse in a general population sample of men and women. Child Abuse and Neglect, 27, 1250–1222.

Browne

Miller

Maguin

(1999). Prevalence and severity of lifetime physical and sexual victimization in incarcerated women. International Journal of Law and Psychiatry, 22, 301–322.

Chen

Cohen

Chen

(2010). How big is Odds Ratio? Interpreting the magnitude of Odds Ratios in epidemiological studies. Communications in Statistics: Simulation and Computation, 39, 860–864.

Coyle

Karatzias

Summers

Power

(2014). Emotions and emotion regulation in survivors of childhood sexual abuse: The importance of “disgust” in traumatic stress and psychopathology. European Journal of Psychotraumatology, 5, 1–9.

10.

Douglas

K. S.

Belfrage

(2014). Interrater reliability and concurrent validity of the HCR-20 version 3. International Journal of Forensic Mental Health, 13, 130–139.

11.

Durose

M. R.

Cooper

A. D.

Snyder

H. N.

(2014, April). Reoffending of prisoners released in 30 states in 2005: Patterns from 2005 to 2010 (Report No. NCJ 244205). U.S. Department of Justice. Retrieved from https://www.bjs.gov/content/pub/pdf/rprts05p0510.pdf

12.

Dutton

D. G.

Hart

S. D.

(1992). Evidence for long-term, specific effects of childhood abuse and neglect on criminal behaviour in men. International Journal of Offender Therapy and Comparative Criminology, 36, 129–137.

13.

Farrington

D. P.

(2005). Childhood origins of antisocial behaviour. Clinical Psychology & Psychotherapy, 12, 177–190.

14.

Fazel

Wolf

(2015). A systematic review of criminal recidivism rates worldwide: Current difficulties and recommendations for best practice. PLoS One, 10, 1–8.

15.

*Firestone

Bradford

J. M.

Greenberg

D. M.

McCoy

Larose

M. R.

Curry

(1999). Prediction of recidivism in incest offenders. Journal of Interpersonal Violence, 14, 511–531.

16.

*Fowler

D. R.

Cantos

A. L.

Miller

S. A.

(2016). Exposure to violence, typology, and recidivism in a probation sample of domestic violence perpetrators. Child Abuse & Neglect, 59, 66–77.

17.

Hanson

R. K.

Bussiere

M. T.

(1998). Predicting relapse: A meta-analysis of sexual offender recidivism studies. Journal of Consulting and Clinical Psychology, 66, 348–362.

18.

*Herbst

J. H.

Branscomb-Burgess

Gelaude

D. J.

Seth

Parker

Fogel

C. I.

(2016). Risk profiles of women experiencing initial and repeat incarcerations: Implications for prevention programs. AIDS Education and Prevention, 28, 299–311.

19.

Jäggi

L. J.

Mezuk

Watkins

D. C.

Jackson

J. S.

(2016). The relationship between trauma, arrest, and incarceration history among Black Americans: Findings from the national survey of American life. Society and Mental Health, 6, 187–206.

20.

Johnson

R. J.

Ross

M. W.

Taylor

W. C.

Williams

M. L.

Carvaial

R. I.

Peters

R. J.

(2006). Prevalence of childhood sexual abuse among incarcerated males in county jail. Child Abuse and Neglect, 30, 75–86.

21.

*Kim

E.Y.

Park

Kim

(2016). Type of childhood maltreatment and the risk of criminal reoffending in adult probationers: A cross-sectional study. BMC Psychiatry, 16, 294–301.

22.

*Kimonis

E. R.

Skeem

J. L.

Edens

J. F.

Douglas

K. S.

Lilienfeld

S. O.

Poythress

N. G.

(2010). Suicidal and criminal behaviour among female offenders: The role of abuse and psychopathology. Journal of Personality Disorders, 24, 581–609.

23.

King

Steenson

Shannon

Mulholland

(2017). Prevalence rates of childhood trauma in medical students: a systematic review. BioMed Central, 17, 159.

24.

*Kjelsberg

Friestad

(2008). Social adversities in first-time and repeat prisoners. International Journal of Social Psychiatry, 54, 514–526.

25.

Koo

T. K.

M. Y.

(2016). A guideline for selecting and reporting intraclass correlation coefficients for reliability research. Journal of Chiropractic Medicine, 15, 155–163.

26.

*Levenson

J. S.

Socia

K. M.

(2016). Adverse childhood experiences and arrest patterns in a sample of sexual offenders. Journal of Interpersonal Violence, 31, 1883–1911.

27.

*Lindsay

W. R.

Elliot

S. F.

Astell

(2004). Predictors of sexual offence recidivism in offenders with intellectual disabilities. Journal of Applied Research in Intellectual Disabilities, 17, 299–305.

28.

*Lowenkamp

C. T.

Holsinger

A. M.

Latessa

E. J.

(2001). Risk/need assessment, offender classification, the role of childhood abuse. Criminal Justice and Behaviour, 28, 543–563.

29.

Malinosky-Rummell

Hansen

D. J.

(1991). Long-term consequences of childhood physical abuse. Psychological Bulletin, 114, 68–79.

30.

Mallie

A. L.

Viljoen

J. L.

Mordell

Spice

Roesch

(2011). Childhood abuse and adolescent sexual re-offending: A meta-analysis. Child Youth Care Forum, 40, 401–417.

31.

Moore

M. D.

Tatman

A. W.

(2016). Adverse childhood experiences and offender risk to re-offend in the United States: A quantitative examination. International Journal of Criminal Justice Sciences, 11, 148–158.

32.

Norton

E. C.

Dowd

B. E.

Maciejewski

M. L.

(2018). Odds ratios—Current best practice and use. JAMA Guide to Statistics and Methods, 320, 84–85.

33.

Perez-Fuentes

Olfson

Villegas

Morcillo

Wang

Blanco

(2013). Prevalence and correlates of child sexual abuse: A national study. Comprehensive Psychiatry, 54, 16–27.

34.

Plummer

Cossins

(2016). The cycle of abuse: When victims become offenders. Trauma, Violence, & Abuse, 19, 286–304.

35.

Romer

(2010). Adolescent risk taking, impulsivity, and brain development: Implications for prevention. Developmental Psychobiology, 52, 263–276.

36.

Ruth-Lyons

(1996). Attachment relationships among children with aggressive behaviour problems: The role of disorganized attachment patterns. Journal of Consulting and Clinical Psychology, 64, 64–73.

37.

Sansone

R. A.

Leung

J. S.

Wiederman

M. W.

(2012). Five forms of childhood trauma: Relationships with aggressive behaviour in adulthood. The Primary Care Companion for CNS Disorders, 15, 1–13.

38.

Schmidt

F. L.

Hunter

J. E.

(2004). Methods of meta-analysis: Correcting error and bias in research findings. Thousand Oaks, CA: Sage.

39.

Sharma

Prakash

Sengar

K. S.

Chaudhury

Singh

A. R.

(2015). The relation between emotional intelligence and criminal behaviour: A study among convicted criminals. Industrial Psychiatry Journal, 24, 54–58.

40.

*Shepard

M. F.

(1992). Predicting batterer recidivism five years after community intervention. Journal of Family Violence, 7, 167–178.

41.

Singh

M. M.

Parsekar

S. S.

Nair

S. N.

(2014). An epidemiological overview of child sexual abuse. Journal of Family Medicine and Primary Care, 3, 430–435.

42.

Sirriyeh

Lawton

Gardner

Armitage

(2012). Reviewing studies with diverse designs: The development and evaluation of a new tool. Journal of Evaluation in Clinical Practice, 18, 746–752.

43.

Smith

D. K.

Leve

L. D.

Chamberlain

(2006). Adolescent girls’ offending and health-risking sexual behaviour: The predicting role of trauma. Child Maltreatment, 11, 346–353.

44.

Stoltenborgh

van Ijzendoorn

M. H.

Euser

E. M.

Bakermans-Kranenburg

M. J.

(2011). A global perspective on child sexual abuse: Meta-analysis of prevalence around the world. Child Maltreatment, 16, 79–101.

45.

*Tikkanen

Ducci

Goldman

Holi

Lindberg

Tiihonen

Virkkunen

(2010). MAOA alters the effects of heavy drinking and childhood physical abuse on risk for severe impulsive acts of violence among alcoholic violent offenders. Alcoholism: Clinical and Experimental Research, 34, 853–860.

46.

Tully

R. J.

Browne

K. D.

(2015). Appraising the Risk Matrix 2000 static sex offender risk assessment tool. International Journal of Offender Therapy and Comparative Criminology, 59, 1–14.

47.

Widom

C. S.

White

H. R.

(2006). Problem behaviours in abused and neglected children grown up: Prevalence and co-occurrence of substance abuse, crime and violence. Criminal Behaviour and Mental Health, 7, 287–310.

48.

Williams

Papadopolou

Booth

(2012, March). Prisoners’ childhood and family backgrounds. Results from the Surveying Prisoner Crime Reduction (SPCR) longitudinal cohort study of prisoners (Research series 4/12). London, England: Ministry of Justice.

49.

Wilsnack

S. C.

Vogeltanz

N. D.

Klassen

T. R.

Harris

T. R.

(1997). Childhood sexual abuse and women’s substance abuse: National survey findings. Journal of Studies on Alcohol, 58, 264–271.

50.

Wolff

Shi

. (2012). Childhood and adult trauma experiences of incarcerated persons and their relationship to adult behavioural health problems and treatment. International Journal of Environmental Research and Public Health, 9, 1908–1926.

51.

Wolff

Shi

Siegel

J. A.

(2009). Patterns of victimization among male and female inmates: Evidence of an enduring legacy. Violence & Victims, 24, 469–484.

52.

Zlotnick

(1997). Posttraumatic stress disorder (PTSD), PTSD comorbidity, and childhood abuse among incarcerated women. Journal of Nervous and Mental Disease, 185, 761–763.