Abstract
Recidivism among individuals who have sexually offended poses a significant public health and safety concern. It is crucial to assess the predictive validity of traditional risk factors in individuals engaged in online child exploitation. This study examines recidivism rates and risk factors among individuals involved in online child sexual exploitation, analyzing data from a sample of 228 adult males who had committed sexual and nonsexual offenses at their index crime. The findings suggest that offense-supportive cognitions (Harrell's C = .73-.75) and emotional congruence with children (Harrell's C = .77) serve as predictors for contact sexual recidivism. Consumption of child sexual exploitation material and bestiality pornography are linked to online sexual recidivism (.69 and .75, respectively) and negatively related to sexual recidivism (.29 and .32, respectively). Overall, this research contributes to a more nuanced understanding of recidivism patterns and risk factors among individuals engaged in online sexual offenses against children, emphasizing the need for tailored intervention strategies in this population.
Keywords: online sexual offending, child sexual exploitation material, recidivism, dynamic risk factors, child sexual abuse
Introduction
Between 2014 and 2022, Canadian authorities reported 15,630 incidents involving online child sexual exploitation material (CSEM, legally termed “child pornography”, see Brown, 2022) and 11,971 incidents of child sexual solicitation (legally termed “child luring”), with these numbers increasing each year (Savage, 2024). With the ongoing rise in denunciations and arrests for the sexual exploitation of children over the Internet (Canadian Centre for Child Protection, 2018; Ibrahim, 2022), concerns emerge regarding the potential for criminal recidivism and the offline escalation of offending for individuals who engage in such behaviors. Recidivism among individuals who have sexually offended represents a significant public health and safety concern, which partly explains why it is extensively studied. A recent study following 4433 Canadians over an average period of 4.5 years revealed recidivism rates of 5% for new sexual offenses of any form, 18% for violent offenses, and 24% for any type of offense (Helmus et al., 2024). Moreover, important meta-analyses of weighted average sexual recidivism rates have been conducted using large independent samples from Canada (n = 29,631) and the United States (n = 196,651). After controlling for various methodological factors, including follow-up periods, the authors estimated an average sexual recidivism rate of 9% (ranging from 8% to 11%) in the United States and 13% (ranging from 11% to 15%) in Canada (Lussier et al., 2022).
Recidivism rates vary across subgroups of offenders depending on the criminogenic characteristics they present and the type of recidivism examined (Brouillette-Alarie et al., 2018). For instance, sexual recidivism rates tend to be higher among individuals with antisocial traits, a criminal history, and deviant sexual interests, while those who do not fit this profile demonstrate lower rates of recidivism (Harris & Hanson, 2004; Hanson & Morton-Bourgon, 2005). Men who use the Internet to sexually exploit minors – whether they access, possess, distribute, or produce CSEM, or sexually solicit children – tend to exhibit lower rates of sexual recidivism (less than 6% over a 5-year follow-up period; Eke et al., 2019; Seto & Eke, 2015), despite these rates being derived from a limited number of studies with relatively small sample sizes. Similar to individuals who have committed sexual offenses offline, various risk and protective factors are likely to influence the risk of recidivism for those who have sexually offended online. Understanding these factors is crucial for targeting criminogenic needs during treatment, thereby reducing their risk of reoffending.
However, the extent to which risk factors predict sexual recidivism often depends on the type of individuals who sexually offend. For example, phallometrically measured deviant sexual interests have a strong relationship with sexual recidivism among men with child victims but slim to none in men with women victims (Hanson & Bussière, 1998; Hanson & Morton-Bourgon, 2005). This suggests that dynamic risk factors operate differently depending on the context of the offense. Given the substantial variation in the context of potential reoffenses between individuals who commit offline and online sexual crimes, it is essential to investigate the predictive validity of “traditional” risk factors in individuals who have sexually exploited children online.
Recidivism Rates of Individuals Who Sexually Offended Online
Seto et al. (2012) conducted two meta-analyses to estimate the extent of offline sexual offending among individuals who sexually offended online, focusing on their criminal histories and recidivism rates. Looking at 4464 men involved in online sexual offenses – most of whom were involved in CSEM-related offenses – findings indicated that 12% had a history of offline contact sexual offenses against children, as indicated by official records (i.e., arrests, charges, and/or convictions) (Seto et al., 2012). Acknowledging that a portion of sexual crimes goes undetected, the authors also investigated crime rates through self-revelation. They found that among a sub-sample of 523 men involved in online sexual offenses, one in two (55%) admitted to perpetrating offline contact sexual offenses against children (Seto et al., 2012). In a follow-up period ranging from 1.5 to 6 years, the authors observed that 5% of a sub-sample of 2,630 individuals with an online sexual offense reoffended with another sexual offense. Within a smaller subsample of 1,247 individuals from the same group, 2% escalated to commit an offline contact sexual offense against children, and 3% reoffended with an offense related to CSEM. Other studies specifically looking at CSEM users have reported contact sexual recidivism rates ranging from 1% to 4% over follow-up periods of up to 13 years (with an average of 4–5 years; Elliot et al., 2019; Eke et al., 2011; Faust et al., 2015; Krone et al., 2017; Seto & Eke, 2015; Soldino et al., 2019).
Although online sexual recidivism rates often surpass those of contact sexual recidivism, they remain relatively low. In a study looking at 541 CSEM users over a 4-year period, Eke et al. (2011) found that only 7% of them were subsequently convicted for a new CSEM-related offense. Other studies reported online sexual recidivism rates ranging from 0% to 9% (Eke et al., 2019; Faust et al., 2015; Goller et al., 2016; Helmus et al., 2024; Hörnle et al., 2014; Krone et al., 2017; Osborn et al., 2010; Paquette et al., 2023; Savoie et al., 2022; Seto and Eke, 2015). Recently, Babchishin et al. (2022) found that among individuals with a history of CSEM who reoffended, those who escalated with an offline contact sexual offenses exhibited more risk factors and had higher recidivism rates compared to those whose recidivism remained online. These relatively low recidivism rates among men who engage in online sexual offending against children contrast with those of individuals with a history of offline contact sexual offenses against children who, over a 20-year follow-up period, showed contact sexual recidivism rates of 13% for intrafamilial offenders and 35% for extrafamilial offenders (Hanson et al., 2009).
In contrast to CSEM users, knowledge about men who sexually solicit children online is notably limited, namely as it relates to recidivism rates. In Canada’s criminal code, this category of offenses is understood as child luring and involves the use of telecommunications to communicate with someone believed to be under 18 for the purpose of committing a sexual offense. This can involve online crimes such as the production of CSEM or invitation to sexual contact, as well as offline activities like sexual contact. Preliminary studies indicate that these individuals exhibit low rates of prior convictions. In Briggs et al. (2011), 6% of individuals who committed online sexual solicitation offenses had prior convictions of any kind, with 4% of these being specifically for sexual offenses. In Young (2005), none of the 22 individuals who committed luring offenses had a history of sexual contact with children. In terms of recidivism, Paquette et al. (2023) reported that of the 38 individuals who committed solicitation offenses in their sample, 16% reoffended with a sexual crime, while 24% and 11% reoffended with a violent and general crime, respectively. However, these rates did not consider follow-up periods, thus limiting our ability to compare them to those of other groups of individuals with a history of sexual crime.
Predictors of Recidivism Among Individuals with Online Sexual Offenses
Several obstacles arise when attempting to determine the risk factors of reoffending for individuals with online sexual offenses. These obstacles include small sample sizes, heterogeneity among offender profiles, limitations inherent to the type of data used (e.g., police vs. self-reported data), low base rates, and short follow-up periods. Consequently, the literature on predictors of recidivism among individuals with online sexual offenses is scarcer than that of individuals with offline sexual offenses. Predictors of recidivism are generally distinguished as being either static or dynamic in nature (Brouillette-Alarie et al., 2016). Static risk factors are characteristics not amenable to change via intervention, such as past criminal history, age, and gender, that help predict the likelihood of reoffending. In contrast, dynamic risk factors are changeable characteristics like substance abuse, employment status, and peer associations, which can fluctuate over time and be targeted by interventions to reduce the risk of reoffending.
Traditionally, age and history of offending behaviors have been strong static predictors of recidivism (Andrews & Bonta, 2010; Gendreau et al., 1996; Hanson & Bussière, 1998), and there is evidence that this relationship remains significant among individuals with online sexual offenses (Eke et al., 2011; Seto & Eke, 2015). For instance, Eke et al. (2011) found that the younger these individuals were at the time of the first offense or release, the more likely they would be to reoffend with a sexual offense. Individuals with online sexual offenses that have prior criminal history (sexual, violent, or general) and an important number of offenses are more likely to reoffend with a sexual offenses than those who do not (Eke et al., 2011; Seto & Eke, 2015).
Looking at dynamic risk factors, two overarching domains have consistently predicted sexual recidivism among individuals with offline sexual offenses: displaying atypical sexuality and antisocial attitudes and behaviors (Brouillette-Alarie et al., 2016, 2018; Brouillette-Alarie & Proulx, 2019; Hanson & Morton-Bourgon, 2005). Therefore, the majority of studies on risk factors in individuals with online sexual offenses have focused on the same risk factors.
Having an atypical sexuality has emerged as a significant trait among this group, especially among individuals using CSEM (Babchishin et al., 2011, 2015; Paquette & Fortin, 2021; Seto & Eke, 2017; Seto et al., 2006). Evidence suggests that atypical sexuality predicts sexual recidivism among CSEM users, as indicated by measures such as preferences for boys, self-reported interests in children, and a history of engaging in sexual offending behaviors (Eke et al., 2019; Seto & Eke, 2015; Soldino et al., 2019). Atypical sexuality is also characterized by sexual preoccupation, the use of sexuality as a coping mechanism, as well as displaying atypical sexual fantasies and preferences (Paquette, Brouillette-Alarie et al., 2022). Individuals with online sexual offenses tend to exhibit more sexual preoccupations than those with contact sexual offenses – the same being true for sexual coping strategies and atypical fantasies and preferences (Babchishin et al., 2011, 2015). In a cross-sectional study, it was found that coping using child sexual stimuli is associated with engaging in online sexual offenses against minors, whether involving the use of CSEM or soliciting them sexually (Paquette, Brouillette-Alarie et al., 2022). Furthermore, reporting sexual preoccupation and using sexuality as a coping strategy have been found to predict sexual recidivism among CSEM users (Babchishin et al., 2023). However, the predictive validity of atypical fantasies and preferences – aside from those related to children – is still unknown in this group. There is evidence that suggests that the online use of pornographic content might serve as a proxy for assessing sexual fantasies and preferences (Eke & Seto, 2023; Paquette, Brouillette-Alarie et al., 2022). If it is the case, taking into consideration the type of pornographic content consumed by the individual could be useful in assessing the risk of reoffending, particularly among individuals using the Internet to engage in sexual offending behaviors.
There is evidence suggesting that individuals with online sexual offenses tend to exhibit fewer antisocial behaviors, attitudes, and traits compared to individuals with offline contact sexual offenses (Babchishin et al., 2015). A meta-analysis found that, compared to those who committed offline contact sexual offenses, CSEM users were less likely to have engaged in criminal activity, exhibit impulsive behaviors, struggle with self-regulation, or hold hostile attitudes and cognitions that support contact sexual abuse of children (Babchishin et al., 2015). However, when present, these characteristics remain strong predictors of sexual recidivism among the former (Babchishin et al., 2023; Seto & Eke, 2015). In Babchishin et al. (2023), markers of antisociality such as hostility, rejection of supervision, substance abuse, lack of concern for others, and hostility toward women were predictive of sexual recidivism among a small subsample of CSEM users.
Cognitions supporting sexual offending were also found to be associated with engaging in such behaviors, both online and offline (Paquette & Cortoni, 2021). Studies have shown that individuals who commit sexual crimes often hold beliefs that children are sexual beings, that sexual activities with children are not harmful, and that children can be intimate partners (Paquette & Cortoni, 2022; Paquette et al., 2014; Ward, 2000). They may also believe they have the right to act as they wish, view the world as a dangerous place, and see sexual crimes as inevitable. These beliefs are prevalent among those who commit crimes against children, both online and offline (Paquette & Cortoni, 2022; Paquette et al., 2014; Ward & Keenan, 1999). Additionally, online offenders often have beliefs tied to the virtual environment, such as viewing the virtual world differently from reality and considering the Internet to be uncontrollable (Paquette & Cortoni, 2022). However, the relationship between these cognitions and recidivism remains uncertain among individuals with online sexual offenses.
To the best of our knowledge, there is no study examining the predictors of recidivism among persons who committed child luring offenses. Despite results suggesting that risk factors traditionally associated with sexual recidivism in individuals with offline sexual offenses also apply to those with online sexual offenses (e.g., CSEM-related ones), caution should be warranted concerning the generalizability of these findings due to the relatively small number of studies and small sample sizes upon which they rely. Furthermore, identifying risk factors for reoffending is crucial in clinical practice, as important legal decisions rely on this process. Accurately identifying these factors enables the assessment of the level of risk posed by an individual and their criminogenic needs, which can then be used for treatment planning and case management. In this context, more studies are needed to accurately identify predictors of recurrence among individuals with online sexual offenses. This study thus aims to determine whether risk factors under the dimensions of atypical sexuality and antisociality predict recidivism in individuals who offend online against children.
Method
Ethics
The research received ethical approval from the Université de Montréal, the University Institute of Mental Health of Quebec, and the Quebec Ministry of Public Safety.
Sample
This study follows up on Paquette and Cortoni’s (2020, 2021) sample of 241 men incarcerated or under community supervision by the Correctional Services of Quebec (Canada) for a sexual offense. Because 13 individuals did not consent to a follow-up on their criminal data, the sample of the current study was reduced to 228 men. These men were assessed between June 2015 and June 2016.
To examine recidivism among different offending profiles, the sample was divided into four subgroups based on their index crime: (1) individuals who committed sexual offenses online but not offline (n = 56; 24.6%), (2) individuals who committed sexual offenses offline (contact) but not online (n = 41; 18.0%), (3) individuals who committed online and offline sexual offenses (n = 63; 27.6%), and (4) individuals who committed nonsexual offenses (the “control group”; n = 68; 29.8%).
Measures
Recidivism (Outcomes)
Recidivism information was obtained from the Canadian Police Information Centre (CPIC) files. Recidivism data were collected on October 4, 2022, and thus, 2,477 days (6 years and 9 months) after the last day of the initial data collection. For men who were incarcerated (n = 120; 52.3%), the follow-up period started on their first day of release from prison. For those who were under community supervision (n = 108; 47.3%), the follow-up period started on the day they were met for the initial data collection. For the current study, the mean follow-up period was 2,299 days (6 years and 3 months) (SD = 319.14; range = 873 to 2,790 days).
Recidivism was measured based on new convictions across four categories: online sexual, contact sexual, nonsexual violent, and nonsexual nonviolent. Online sexual recidivism encompassed convictions for crimes involving child victims committed online, such as accessing, possessing, distributing, or producing child pornography, as well as child luring and online invitation to sexual touching. Contact sexual recidivism involved convictions for sexual offenses committed offline, including sexual contact and aggression. Nonsexual violent recidivism referred to convictions for nonsexual violent crimes, such as assault or robbery. Nonsexual nonviolent recidivism, in turn, covered convictions for nonviolent, nonsexual crimes like burglary, theft, or fraud.
Risk Factors (Predictors)
The following static and dynamic risk factors reflecting atypical sexuality and antisociality were considered as potential predictors of recidivism in this sample.
Demographics
Individuals’ age at release from prison (or at data collection if met in the community) and prior convictions were collected. Prior convictions were coded dichotomously and for different types of offenses: offline sexual crimes, CSEM-related offenses, child luring, nonsexual violent crimes, and nonsexual nonviolent crimes.
Sexual Victims
The number of identifiable male and female child victims was computed. In this study, we define a victim as a child if they are aged 16 or younger, which aligns with previous research that has used this cut-off (Chopin, 2017; Chopin & Beauregard, 2019; Leclerc et al., 2009). Following the terminology used by Phoenix et al., (2017) for the Static-99R coding rules, “identifiable” victims are those specifically targeted for sexual activities by perpetrators. In the context of our study, this encompasses victims of child luring and offline sexual crimes. In contrast, victims depicted in sexual exploitation material are not classified as identifiable, unless the perpetrator produced the material themselves.
Offense-Supportive Cognitions
Three measures were used to assess the level of endorsement of offense-supportive cognitions. The Molest (Bumby, 1996) is 38-item questionnaire with a 4-point Likert scale assessing cognitions supportive of offline contact sexual behaviours with children. In the current sample, the internal consistency of the Molest, as measured by Cronbach’s alpha, was .93.
The Cognitions of Internet Sexual Offending (C-ISO; Paquette & Cortoni, 2020) is a 31-item questionnaire with 4-point Likert scales assessing cognitions supportive of the online sexual exploitation of children. The development study indicated substantial convergent validity with the Molest (r = 0.63; Paquette & Cortoni, 2020). In this study, the C-ISO had a Cronbach’s alpha of .90.
The Cognitive and Emotional Congruence with Children (C-ECWC; Paquette & McPhail, 2020) is a 12-item questionnaire with 4-point Likert scales assessing the extent to which individuals with online sexual offenses against children self-identify with children. The scale is made of three dimensions: Childlike Sense of Self, Friendship and Positive Relationships with Children, and Romantic Attraction to Children (Paquette & McPhail., 2020). In this study, the Cronbach’s alpha of the Childlike Sense of Self subscale was .78, .89 for the Friendship and Positive Relationships with Children subscale, and .78 for the Romantic Attraction to Children subscale.
Sexual Interests Towards Children
Two measures were used to assess sexual interests towards children. The Sexual Interest Cardsort (SIC) (Abel & Becker, 1979) is 75-item questionnaire with 7-point Likert scales measuring various atypical sexual interests. We used the two 10-item subscales assessing, respectively, interests towards young girls (child molestation of girls) and interests towards young boys (child molestation of boys). Among the current sample, the Cronbach’s alpha of the SIC was .92 for sexual interests towards girls and .96 for sexual interests towards boys.
The Revised Screening Scale for Pedophilic Interests (SSPI-2) (Seto, Stephens et al., 2017) is a questionnaire measuring the presence of 5 indicators of pedophilia: any boy child victims, more than one child victim, any child victims under the age of 12, any unrelated child victims, and using CSEM. The SSPI-2 is associated with sexual recidivism (AUC = .59 to .62) (Seto, Sandler et al., 2017; Seto, Stephens et al., 2017). In the current study, the SSPI-2 correlates with both the SIC child molestation of girls (r = .27, p < .01) and boys (r = .46, p < .01), and has an internal consistency of .65.
Sexual Coping and Preoccupation
Two measures were used to assess sexual coping and preoccupation. The Coping Using Sex Inventory (CUSI) (Cortoni & Marshall, 2001) is a 16-item questionnaire with 5-point Likert scales measuring the extent to which individuals use sexuality – using child, rape, and consensual content – to deal with stressful experiences and negative emotions. In this study, the CUSI demonstrated an internal consistency of .92 for the child scale, .88 for the rape scale, and .82 for the consent scale.
The Sexual Preoccupation (SP; Snell & Papini, 1989) is a 10-item questionnaire with 5-point Likert scales assessing the extent to which individuals are preoccupied about their sexuality. In the current study, the SP had an internal consistency of .90 and correlated with the CUSI (r = .48, p < .001).
Pornography/CSEM consumption
Using a bespoke questionnaire, we assessed the extent to which individuals consumed various types of pornography and CSEM. On 4-point Likert scales, participants were asked to report their lifetime frequency of use (0 = never; 1 = less than 15 times; 2 = between 15 and 30 times; 3 = more than 30 times) of five types of legal pornography (man-woman, man-man, woman-woman, bondage, group sex), two types of CSEM (child-adult, child-child), violent/coercive pornography, and bestiality material. Responses to legal types of pornography were summed in a scale, the same being true for responses to CSEM.
Analyses
Descriptive statistics were computed for the whole sample as well as subgroups differentiated by the nature of their index offense. ANOVA and Cramer’s V were used to test if significant differences existed between the subgroups. Descriptive statistics are available for both predictors and recidivism outcomes, and can be found in Table 1.
Table 1.
Descriptive Statistics and Recidivism Data Across Subgroups Based on Index Offense.
| Measures (range) | Group 1 | Group 2 | Group 3 | Group 4 | Eta2 / Cramer’s V |
|---|---|---|---|---|---|
| Online-only sexual offenses (n = 56) | Offline-only (contact) sexual offenses (n = 41) | Online and offline sexual offenses (n = 63) | Nonsexual offenses only (n = 68) | ||
| Predictors | M (SD) / n (%) | ||||
| Age at follow-up (21 – 79) | 40.25 (12.65) | 47.85 (13.16) | 45.25 (14.37) | 39.22 (11.20) | .067** |
| Criminal background | |||||
| Previous conviction for contact sexual crime (0 – 1) | 0 (.0%) | 11 (28.2%) | 10 (21.7%) | 0 (.0%) | .40*** |
| Previous conviction for child sexual exploitation material (0 – 1) | 1 (2.1%) | 0 (.0%) | 3 (6.5%) | 0 (.0%) | .19 |
| Previous conviction for child luring (0 – 1) | 0 (.0%) | 0 (.0%) | 2 (4.3%) | 0 (.0%) | .18 |
| Previous conviction for violent crime (0 – 1) | 10 (21.3%) | 8 (20.5%) | 9 (19.6%) | 35 (53.8%) | .34*** |
| Previous conviction for general crime (0 – 1) | 14 (29.8%) | 15 (38.5%) | 13 (28.3%) | 53 (81.5%) | .47*** |
| Identifiable victims | |||||
| Number of male child victims (0 – 7) | .38 (1.12) | .11 (.32) | .52 (1.32) | .00 (.00) | .073** |
| Number of female child victims (0 – 7) | 1.15 (.80) | 1.21 (1.00) | 1.34 (1.21) | .00 (.00) | .364*** |
| Deviant sexual interests and cognitions | |||||
| Molest scale (20 – 112) | 55.17 (14.51) | 54.76 (12.13) | 62.63 (18.60) | 45.88 (9.25) | .176*** |
| Cognitions of internet sexual offending scale (5 – 113) | 53.91 (12.67) | 44.10 (10.88) | 54.29 (16.98) | 43.59 (9.98) | .138*** |
| Cognitive and emotional congruence with children scale (12 – 42) | 20.79 (6.35) | 19.38 (4.49) | 22.53 (7.27) | 17.98 (4.82) | .085*** |
| C-ECWC: Childlike sense of self (5 – 18) | 6.87 (2.86) | 6.03 (1.63) | 7.29 (3.23) | 5.74 (1.34) | .067** |
| C-ECWC: Friendship and positive relationships with children (4 – 15) | 8.52 (3.06) | 8.03 (2.46) | 9.00 (3.03) | 7.75 (3.01) | .028 |
| C-ECWC: Romantic attraction to children (3 – 12) | 5.69 (2.16) | 4.89 (1.86) | 6.02 (2.88) | 4.36 (1.70) | .089*** |
| Sexual interest cardsort – child molestation of girls) (-30 – 30) | −19.33 (16.90) | −27.13 (6.59) | −18.61 (18.04) | −28.90 (4.40) | .119*** |
| Sexual interest cardsort – child molestation of boys) (-30 – 30) | −24.60 (13.33) | −29.73 (1.22) | −24.20 (13.51) | −29.18 (4.10) | .064** |
| Revised screening scale for pedophilic interests (0 – 5) | 1.09 (.51) | .71 (1.10) | 2.40 (1.25) | .00 (.00) | .548*** |
| Sexual coping and preoccupation | |||||
| Coping using sex inventory – consent | |||||
| (5 – 25) | 20.28 (4.62) | 18.58 (4.72) | 19.53 (4.47) | 18.45 (4.96) | .025 |
| CUSI – child (4 – 20) | 7.06 (4.72) | 4.66 (2.59) | 6.82 (4.81) | 4.12 (.99) | .124*** |
| CUSI – rape (6 – 30) | 8.06 (4.42) | 7.68 (4.14) | 7.76 (2.98) | 7.22 (3.38) | .008 |
| Sexual preoccupation scale (11 – 50) | 31.85 (8.72) | 27.76 (7.47) | 30.04 (7.97) | 30.43 (7.33) | .029 |
| Pornography/CSEM consumption | |||||
| Frequency of legal pornography use (0 – 15) | 8.38 (2.76) | 5.92 (3.66) | 7.96 (3.57) | 6.74 (3.08) | .075** |
| Frequency of CSEM use (0 – 6) | 2.19 (2.50) | .00 (.00) | 2.35 (2.52) | .00 (.00) | .304*** |
| Frequency of violent pornography use (0 – 3) | .36 (.53) | .22 (.58) | .51 (.74) | .12 (.33) | .073** |
| Frequency of bestiality pornography use (0 – 3) | .62 (.68) | .18 (.39) | .60 (.67) | .17 (.38) | .140*** |
| Outcomes (recidivism data) | |||||
| Online sexual recidivism (0 – 1) | 4 (7.1%) | 0 (.0%) | 4 (6.3%) | 0 (.0%) | .18† |
| Contact sexual recidivism (0 – 1) | 3 (5.4%) | 5 (12.2%) | 2 (3.2%) | 0 (.0%) | .20* |
| Nonsexual violent recidivism (0 – 1) | 6 (10.7%) | 4 (9.8%) | 4 (6.3%) | 22 (32.4%) | .30*** |
| Nonsexual nonviolent recidivism (0 – 1) | 6 (10.7%) | 8 (19.5%) | 7 (11.1%) | 37 (54.4%) | .44*** |
Note. *p < .05; **p < .01; ***p < .001; †p = .054.
Harrell’s C (Harrell et al., 1996) was employed to identify factors discriminating recidivists from non-recidivists while accounting for follow-up time. Similar to the Area under the Curve (AUC), Harrell’s C operates on a scale from 0 to 1, where a value of 1 signifies perfect discrimination and .5 that the predictor’s performance is no better than chance. Interpretively, values of .54, .63, and .71 are considered indicative of small, moderate, and large effect sizes, respectively (Rice & Harris, 2005). Harrell’s Cs are considered statistically significant when their confidence intervals (95%) do not comprise .50. They were obtained with the “survival” package in R software (Version 3.1–11; Therneau, 2020).
Predictive validity analyses were first conducted with individuals who sexually offended online in their index offense, irrespective of whether they also offended offline (groups 1 and 3 in the Sample section, n = 119) (see Table 2). Then, predictive validity analyses were conducted with individuals who sexually offended online only (group 1; n = 56), and then with CSEM users only (a subset of group 1; n = 36) (see Tables A1 and A2 in Appendix). Analyses in these subgroups were made to explore whether specific risk factors were associated to escalation patterns, that is, going from an online to an offline sexual offense in the future. However, that the small number of participants in this group introduces generalizability and standard error issues, especially when outcome measures have such low base rates. Therefore, these analyses should be considered as exploratory until they are replicated with larger samples.
Table 2.
Predictors of Recidivism in Individuals That Sexually Offended Online in Their Index Offense (Group 1 + Group 3; n = 119).
| Online sexual recidivism (n = 8 / 119) | Contact sexual recidivism (n = 5 / 119) | Nonsexual violent recidivism (n = 10 / 119) | Nonsexual nonviolent recidivism (n = 13 / 119) | |
|---|---|---|---|---|
| Harrell’s C [95% C. I.] | ||||
| Age at follow-up | .30 [.16., .43] | .37 [.17, .57] | .32 [.17, .46] | .34 [.20, .48] |
| Criminal background | ||||
| Previous conviction for contact sexual crime | .45 [.43, .48] | .42 [.22, .62] | .52 [.41, .63] | .54 [.44, .63] |
| Previous conviction for child sexual exploitation material | .63 [.47, .78] | .48 [.47, .50] | .48 [.46, .50] | .48 [.46, .50] |
| Previous conviction for child luring | .57 [.44, .70] | .51 [.50, .52] | .51 [.50, .52] | .52 [.51, .53] |
| Previous conviction for violent crime | .41 [.37, .45] | .38 [.16, .60] | .71 [.56, .86] | .71 [.58, .84] |
| Previous conviction for general crime | .38 [.34, .42] | .62 [.40, .84] | .83 [.73, .93] | .79 [.67, .90] |
| Identifiable victims | ||||
| Number of male child victims | .46 [.22, .70] | .64 [.31, .97] | .48 [.30, .65] | .49 [.34, .64] |
| Number of female child victims | .56 [.27, .85] | .22 [.00, .45] | .63 [.42, .83] | .62 [.40, .83] |
| Deviant sexual interests and cognitions | ||||
| Molest scale | .46 [.21, .70] | .75 [.63, .87] | .60 [.41, .78] | .55 [.39, .71] |
| Cognitions of internet sexual offending scale | .48 [.25, .71] | .59 [.43, .74] | .49 [.30, .69] | .46 [.30, .61] |
| Cognitive and emotional congruence with children scale | .40 [.17, .64] | .73 [.57, .88] | .49 [.29, .70] | .52 [.36, .68] |
| C-ECWC: Childlike sense of self | .55 [.34, .76] | .69 [.48, .90] | .48 [.30, .65] | .49 [.35, .64] |
| C-ECWC: Friendship and positive relationships with children | .39 [.18, .61] | .77 [.66, .89] | .54 [.34, .73] | .58 [.42, .73] |
| C-ECWC: Romantic attraction to children | .40 [.18, .62] | .55 [.32, .77] | .44 [.30, .59] | .43 [.29, .57] |
| Sexual interest cardsort – child molestation of girls | .60 [.38, .82] | .58 [.38, .78] | .42 [.23, .60] | .38 [.23, .53] |
| Sexual interest cardsort – child molestation of boys | .54 [.37, .71] | .59 [.31, .87] | .49 [.31, .67] | .45 [.30, .59] |
| Revised screening scale for pedophilic interests | .52 [.35, .69] | .37 [.14, .60] | .49 [.31, .67] | .54 [.40, .69] |
| Sexual coping and preoccupation | ||||
| Coping using sex inventory – consent | .77 [.60, .94] | .28 [.14, .42] | .58 [.23, .62] | .55 [.40, .70] |
| CUSI – child | .48 [.30, .68] | .53 [.28, .78] | .41 [.27, .55] | .37 [.26, .47] |
| CUSI – rape | .55 [.35, .74] | .51 [.24, .77] | .68 [.51, .85] | .64 [.50, .78] |
| Sexual preoccupation scale | .57 [.35, .79] | .49 [.19, .78] | .64 [.42, .86] | .58 [43, .74] |
| Pornography/CSEM consumption | ||||
| Frequency of legal pornography use | .64 [.48, .79] | .39 [.23, .56] | .46 [.26, .66] | .45 [.31, .58] |
| Frequency of CSEM use | .69 [.51, .87] | .29 [.15, .43] | .36 [.22, .50] | .37 [.23, .51] |
| Frequency of violent pornography use | .57 [.39, .75] | .50 [.30, .70] | .49 [.31, .67] | .48 [.34, .61] |
| Frequency of bestiality pornography use | .75 [.59, .90] | .32 [.16, .47] | .42 [.27, .57] | .37 [.25, .50] |
Note. Bolded values = the 95% C. I. does not comprise .50. Significance for the bold is 0.05.
Results
Descriptive statistics and recidivism rates are presented in Table 1. A number of significant differences emerged among the groups. Those who had offline sexual offenses (groups 2 and 3) were on average older than individuals with online sexual offenses only (group 1) or those without sexual offenses (group 4). They were also more likely to be convicted for a prior contact sexual crime, as none of the individuals in groups 1 and 4 had any such conviction. Individuals with nonsexual crimes only (group 4) were much more likely to have prior violent and general convictions compared to the other groups. For nearly all measures of sexual victims, deviance, and cognitions, the same picture emerged: individuals with online and offline sexual crimes in the index offense (group 3) had higher scores than those with online-only sexual offenses (group 1), who, in turn, had higher scores than those with contact-only sexual offenses (group 2), who, again, had high scores than those without sexual crimes (group 4). Two measures were exceptions to this rule: (1) individuals with offline-only sexual offenses (group 2) had more identifiable female child victims than those with online-only sexual offenses (group 3); and (2) no significant differences were found for the Friendship and Positive Relationships with Children scale of the C-ECWC. Sexual coping and preoccupation measures mostly failed to significantly discriminate between the groups. Concerning pornography and CSEM consumption, those with online sexual offenses (groups 1 and 3) consumed on average more pornography and CSEM than the other two groups, who tended to be in the same range.
Recidivism outcomes were closely related to the offense type of individuals. Only individuals with online sexual offenses (groups 1 and 3) reoffended with an online sexual offense. The difference in online sexual recidivism did not reach statistical significance because of the low sample size and split of the outcome, but it was close (p = .054). The three other recidivism outcomes showed statistical differences depending on the groups. Individuals with nonsexual offenses only (group 4) did not reoffend sexually – online or offline. Contact sexual recidivism was more common in individuals with a history of contact sexual offenses (group 2). However, some individuals with online-only sexual offenses (group 1) escalated to a contact sexual offense in their reoffending pattern (5.4%). For nonsexual violent and nonsexual, nonviolent recidivisms, individuals without sexual offenses (group 4) substantially more reoffended than the other groups. Of note, the rate of nonsexual violent recidivism was quite high in individuals with contact-only sexual crimes in the index offense (group 2; 19,5%).
Results indicate some level of specialization in criminal careers: all online sexual recidivisms were committed by individuals with a history of online sexual offenses. Contact sexual recidivisms were mostly committed by individuals with a history of contact sexual offenses, although some online offenders escalated to contact offenses. Finally, violent and general recidivisms were more common in incarcerated individuals without a history of sexual crimes than in those with such an history.
Table 2 shows dynamic risk factors associated with recidivism in individuals with online sexual offenses (groups 1 and 3; n = 119). Age was negatively related to three of the four recidivism outcomes: online sexual, nonsexual violent, and nonsexual nonviolent recidivism. Prior contact sexual convictions were slightly and negatively related to online sexual recidivism, suggesting these individuals stayed in the offline rather than online space. Prior convictions for violent and general crimes were negatively related to online sexual recidivism and positively related to nonsexual violent and nonsexual nonviolent recidivism. Surprisingly, the number of identifiable female child victims were inversely related to contact sexual recidivism, suggesting that such victims may be more groomed online than offline. Measures of deviant sexual interests and cognitions, more specifically, the Molest, the total C-ECWC score, and the Friendship and Positive Relationships with Children subscale of the C-ECWC, were positively associated with contact sexual recidivism. The CUSI (consent) measure of sexual coping was positively associated with online sexual recidivism but negatively associated with contact sexual recidivism. The CUSI (child) measure was negatively associated with nonsexual, nonviolent recidivism, while the CUSI (rape) scale was positively associated with nonsexual violent recidivism. As it relates to pornography and CSEM consumption, CSEM and bestiality were positively associated with online sexual recidivism and negatively associated with nonsexual violent recidivism.
Again, a form of criminal career specialization effect was found in our sample: antisociality-related risk factors tended to predict nonsexual violent and nonsexual, nonviolent recidivism, and were often negatively related to sexual recidivism. In turn, sexual deviance-related risk factors tended to be positively associated with sexual recidivism and negatively associated with nonsexual types of recidivism.
Table A1 in Appendix comprises predictive validity analyses conducted with individuals whose index offense encompassed online-only sexual offenses (CSEM and child luring; group 1; n = 56). Then, Table A2 in Appendix contains predictive validity analyses for individuals whose sexual offenses comprised CSEM only (a subset of group 1; n = 36). Due to the low sample size and low base rates of recidivism outcomes, these analyses are highly unstable, having 95% C. I. that sometimes surpass 1.00. They should be considered as exploratory.
In subgroup analyses, very similar predictive validity patterns emerged. Even though some relationships reached significance and some did not compared to what was found in Table 2 (groups 1 and 3; n = 119), these were mostly in the same directions. Differences were found for the Molest scale, which was positively associated with all recidivism outcomes in CSEM users (n = 36), whereas it was only related to contact sexual recidivism in other groups. In addition, the SIC (girls) was positively associated with contact sexual recidivism in subgroup analyses, whereas it was not in the main predictive validity analyses (groups 1 and 3, n = 119). Finally, the CUSI (child) measure was positively associated with online sexual recidivism in CSEM users (n = 36) but not in other groups. Replication analyses with larger samples should be conducted to gain a comprehensive understanding of how these risk factors affect recidivism of individuals with online sexual offenses and no other offline crimes.
Discussion
This study aimed to report recidivism rates among men involved in online sexual offenses against children and identify risk factors associated with reoffending. Overall, the findings show recidivism rates consistent with what has been previously reported in the literature (Eke et al., 2019; Faust et al., 2015; Goller et al., 2016; Helmus et al., 2024; Hörnle et al., 2024; Krone et al., 2017; Osborn et al., 2010; Paquette et al., 2023; Savoie et al., 2022; Seto and Eke, 2015), and a certain specificity in the nature of reoffending. Patterns emerged among risk factors based on the types of recidivism predicted. On the one hand, indicators of deviant sexual interests and cognitions predicted contact sexual recidivism, while involvement in criminality was predictive of violent and general reoffending.
Recidivism Rates
Among individuals followed for online sexual offending at index crime, our findings show that a small proportion of them reoffended online, and a lesser proportion — though comparable to previous studies (e.g., Eke et al., 2019; Seto & Eke, 2015) — escalated to an offline contact sexual offense. Conversely, all individuals with offline contact sexual offenses who reoffended did so offline, and none online. This echoes a concern about the escalation of crime noted in previous studies (Babchishin et al., 2022; Fortin & Proulx, 2019). In their study, the Fortin and Proulx (2019) examined the temporal evolution of the sexual exploitation material used by CSEM users and discovered trends of escalating deviance among consumers (i.e., increasingly severe offending behaviors committed against younger children) as well as trends of improvement (i.e., less severe offending behaviors committed over time against older children). Babchishin et al. (2022) identified evidence suggesting that individuals who escalated from online to offline sexual offending were at higher risk. Our research lends support to the notion that, for a minority of individuals, there is a progression from online to offline sexual offenses. However, we must be cautious in drawing conclusions about longitudinal trends, given that only three events (i.e., history, index crime, and recidivism) were analyzed. Subsequent investigations could gain valuable insights by exploring the temporal progression of sexual offenses committed by individuals who engage in online sexual exploitation of children. Additionally, given the ongoing advancements in information technology and the widespread accessibility of the Internet and social networks, it is noteworthy that in our sample, no individuals with offline-only sexual offenses reoffended online. This outcome might suggest a preference for direct contact for those having access to children in their immediate environment, yet it could also be attributed to a substantial number of online crimes going undetected by law enforcement. It is plausible that these individuals simply evaded the scrutiny of law enforcement as it relates to their online activities.
Research indicates that individuals who committed sexual offenses against children tend to exhibit few antisocial traits – a trend that is even more pronounced among CSEM users when compared to those who engage in offline contact sexual offending against children (Babchishin et al., 2011, 2015). Antisociality is marked by violent and impulsive behaviors, as well as an overall inclination not to adhere to rules. Consequently, it is expected that individuals who sexually offend online would demonstrate lower rates of both violent and general recidivism, considering that they show lower levels of antisociality compared to other groups and that antisociality is the main predictor of nonsexual types of recidivism in individuals that sexually offend (Babchishin et al., 2016; Brouillette-Alarie et al., 2016, 2018). In this study, 10% of individuals with online-only sexual crimes at the index offense reoffended with a nonsexual crime, while this ratio increased to over 50% for individuals with only nonsexual crimes at the index offense.
Our data revealed a certain specificity regarding the nature of reoffending among participants; those with nonsexual offenses at the index crime tended to specialize in nonsexual criminal activity in the future. However, the dark figure of crime is notably high for sexual offenses compared to nonsexual ones (Abel et al., 1987; Bouchard & Lussier, 2015; Drury et al., 2020), suggesting that specialization findings could be driven by failure to detect new sexual crimes rather than true crime specialization. In the original study (Paquette & Cortoni, 2021), nearly 10% of the men reported sexual crimes for which they were never officially sanctioned.
Atypical Sexuality: Identifiable Victims, Sexual Convictions and Sexual Interest Towards Children
The current state of knowledge suggests that sexual interest towards children is a key feature among individuals who sexually exploit children online, especially among those who use CSEM (e.g., Eke et al., 2011; Faust et al., 2015; Paquette, Brouillette-Alarie et al., 2022; Paquette and Fortin, 2021; Seto & Eke, 2015; Seto et al., 2006; Seto et al., 2012). However, in our study, pedophilic and atypical sexual interests, when measured with crime-centered measures (i.e., number of identifiable victims and prior sexual convictions), did not predict sexual recidivism online or offline. This result is counterintuitive as these factors are generally considered good predictors among men who sexually offend against children (Brouillette-Alarie et al., 2018; Hanson & Bussière, 1998; Hanson & Morton-Bourgon, 2005; Mann et al., 2010; Seto, Stephens et al., 2017). It should be remembered that identifiable victims referenced in the current study both came from offline sexual crimes and online sexual solicitation. Identifiable victims of online sexual solicitation offenders tend to be older (with a modal age of 15; Wolak et al., 2008) than children depicted in CSEM (Canadian Centre for Child Protection, 2016), suggesting a potential crime of opportunity rather than a deviant sexual motivation, such as sexual interest in young children. Hence, combining both prepubescent and pubescent victims in this measure may have impeded its accuracy in predicting sexual recidivism. Future studies should explore measuring the number of victims based on their age as a more indicative proxy for assessing pedophilic interest.
Prior sexual convictions were not associated with sexual recidivism. Previous convictions for CSEM-related offenses nearly reached statistical significance (and may do so with a larger sample) for online sexual recidivism, but other measures of prior sexual crimes had effect sizes that were either very low and/or in the opposite direction for sexual types of recidivism. This may suggest that prior sexual convictions do not have the same predictive utility in individuals that sexually offend online compared to offline, or that characteristics of the current sample precluded their predictive validity, namely in relation to the small number of participants (especially in subgroup analyses).
The lack of predictive validity of the SSPI was surprising, because the scale was updated in recent years to account for the online lives of individuals, thus making it theoretically fit for purpose for CSEM users (Seto, Stephens et al., 2017). The Sexual Interest Cardsort however proved effective in predicting offline sexual escalation among individuals whose index crimes were online sexual offenses, but it did not predict recidivism in those with contact sexual offenses. Research indicates that individuals who engage in criminal behaviors, including sexual offenses, are often reluctant to endorse the Cardsort statements due to the explicit nature of the sexual aggressions described (Holland et al., 2000; Paquette & Brouillette-Alarie, 2020). This reluctance may explain the lack of predictive validity for the group with offline sexual crimes. However, its predictive validity for individuals with online sexual crimes aligns with previous findings. Unlike those with contact sexual offenses, individuals involved in CSEM-related offenses typically exhibit more problematic sexual behaviors (Babchishin et al., 2011, 2015).
Cognitions and Emotional Identification with Children
In this study, cognitions seemed to be better predictors of sexual recidivism than measures derived from the criminal records of individuals. Indeed, this study supported the predictive validity of the Molest and the C-ECWC scales to predict contact sexual recidivism in individuals with online sexual crimes in their index offense. Even though the Molest scale was not developed as a risk assessment tool, it is logical to observe that attitudes favorable to offline sexual offending against children predict that outcome. This result is consistent with prior studies supporting the predictive validity of the Molest scale to predict sexual recidivism (Hermann et al., 2012). However, the C-ISO scale, specifically designed to measure beliefs supportive of online sexual exploitation of children, did not prove effective in predicting sexual recidivism, both online and offline. Offense-supporting cognitions and attitudes have been linked to recidivism, yet their predictive accuracy is somewhat hampered by challenges in conceptualization and measurement (see Helmus et al., 2013). Therefore, further research efforts or potential revisions of assessment tools could be warranted to bolster their usefulness in assessing risk of sexual recidivism – if such an endeavor is even desired by tool developers.
Another construct related to cognitions is the way some individuals perceive themselves in relation to children (Hermann et al., 2017). Emotional and cognitive self-identification with children has been found to predict any sexual and contact sexual recidivism (Fraser et al., 2024; Hanson & Morton-Bourgon, 2005; McPhail et al., 2013). This study supports these results, as the total score and Friendship and Positive Relationships with Children subscale predicted offline sexual reoffending against children. This scale was developed and validated to measure emotional congruence with children among both individuals with online and offline sexual offenses against children (Paquette & McPhail, 2020). However, this study does not support the predictive validity of the scale predicting online sexual recidivism. Once again, it is possible that the low baseline rates of online sexual recidivism prevented the detection of significant effects. It could also be that this component is only significant in predicting sexual recidivism among offline offenders and not among online offenders. Additional studies are necessary to better understand the relationship (or lack thereof) between this psychological construct and online sexual recidivism.
Consumption of Pornography and CSEM
A high use of bestiality pornography predicted online sexual recidivism in this study. Consistent with our findings, a study by the Canadian Centre for Child Protection (2018) revealed that 87% of criminal cases involving bestiality also involved other child sexual-related offenses, such as CSEM-related offenses, child luring, sexual incitement, or sexual assault. Online and contact sexual offending against children and engaging in bestiality also have common roots; both are associated with a history of childhood victimization (e.g., Chopin et al., 2022; Holoyda et al., 2020). Furthermore, studies have shown that some pornography consumers reported a tendency to seek a variety of diversified content, provided it becomes increasingly extreme (e.g., younger individuals, more violent) (Fortin et al., 2018; Fortin & Proulx, 2019). Meanwhile, a study conducted by Gebhard, Gagnon, Pomeroy, and Christenson (1965) reported that some individuals are “willing to try anything for kicks” (p. 168), including sexual contact with animals. If the consumption of sexual content involving animals and children is interconnected, it appears that such content is inversely related to offline contact sexual recidivism against children, as evidenced by the negative association in our study. This finding contradicts studies showing an association between bestiality and sexual abuse of children (e.g., Edwards, 2019). These conflicting results may be explained by the fact that, in our study, individuals did not engage in sexual contact with animals but rather observed such sexual activity online. For some individuals, it may be indicative of an indiscriminate interest in any types of pornographic material rather than a specific propensity to engage in sexual and physical violence. It could also translate into a curiosity for more extreme content, or even resorting to any type of content to feel excitement. This hypothesis should be examined more closely in future studies.
Moreover, our results illustrate a widely demonstrated phenomenon in criminology, namely that past behavior is generally the best predictor of future behavior (Loeber & Farrington, 2001). Consistent with this idea, past CSEM consumption was found to be predictive of sexual recidivism with CSEM. In contrast, it was found to be inversely associated with offline escalation involving contact sexual offenses. This finding aligns with the idea that online offenders, despite having more problematic sexuality, possess favorable characteristics (e.g., a more favorable sociodemographic profile and low propensity for antisocial behavior) that prevent them from committing offline offenses (Seto, 2009).
Clinical and Practical Implications
This study has important practical implications for the treatment of individuals who engage in online sexual exploitation of children. Firstly, identifying these factors provides mental health professionals and legal practitioners with clear indications of specific areas to target during therapeutic interventions to decrease risk. By incorporating these criminogenic needs into treatment programs, it can help customize therapeutic approaches to better address the specific and unique needs of individuals that sexually offend (Andrews & Bonta, 2010; Paquette, Chopin et al., 2022). Until now, there has been limited evidence for relevant dynamic risk factors in individuals who engage in online sexual exploitation of children, apart from antisocial behavior and sexual interest in children. This study contributes by expanding on atypical sexual interest, as well as cognitive and emotional factors that are important to consider. Additionally, heightened awareness of these predictive factors can enhance risk assessment and case management planning, allowing for improved post-treatment monitoring and service delivery. The low recidivism rates also highlight the importance of providing an appropriate level of intensity in treatment that matches the risk level of individuals who commit online sexual offenses, especially in the absence of additional offline offenses. Our findings also contribute to guiding public policies on the treatment of people who sexually exploit children online, emphasizing evidence-based interventions.
Limitations
Despite its contribution to scientific knowledge and clinical practice, this study has certain limitations. Similar to its predecessors, it suffers from a small sample size and baseline rates, reducing the capacity of statistical models to detect significant statistical relationships. However, the publication of recidivism rates and associated factors is crucial to advancing knowledge about online sexual offending, which can ultimately be pooled in meta-analyses, for instance, to increase statistical power and provide a clearer understanding of the phenomenon. We must also address a potential sample bias. The individuals among whom data were initially collected in the community were selected based on their attendance to treatment. Therefore, it is possible that a high-risk group, who may have reoffended during their treatment and consequently returned within the justice system, might have been excluded from our sample a priori.
Another noteworthy limitation is the nature of the measures used, as some stem from tools (e.g., CUSI, SP) developed and validated with individuals with offline contact sexual offenses and, thus, not specifically tailored to those with online sexual offenses. In the absence of validation studies with such individuals, these tools may simply not be suitable for predicting their sexual recidivism. Some predictive validity associations were nevertheless found with the Molest, C-ECWC, and CUSI.
Conclusion
This study sheds light on recidivism rates and associated risk factors among individuals involved in online sexual offenses against children. The findings reveal consistency with previous literature on overall recidivism rates, with some specificity in the nature of reoffending based on the types of recidivism. Notably, a minority of individuals with online offenses at index crime reoffended online, and those who committed offline contact sexual offenses exclusively reoffended offline. Findings indicate that cognitions supportive of sexual offending and emotional identification with children predicted contact sexual recidivism. Conversely, the use of CSEM and bestiality pornography are linked to online sexual recidivism. Furthermore, predictors related to an atypical sexuality have produced mixed findings. Notably, static factors such as the number of victims and criminal history did not demonstrate predictive validity, whereas a dynamic factor, that is, sexual interest towards girls, did show predictive validity. The study also highlights the potential progression from online to offline sexual criminal activities among a small subset of individuals.
The study offers valuable insights for mental health professionals and legal practitioners in developing targeted therapeutic interventions and risk assessments for individuals involved in online sexual exploitation of children. However, it also has limitations that should be addressed in future research. Future studies should examine the progression from online to offline offending behaviors among those who use the Internet and social media to sexually exploit children. It is crucial to focus on both static and dynamic risk factors, as well as to explore how advancing technology influences online sexual recidivism. Overall, this study enhances our understanding of recidivism patterns and risk factors in the context of online sexual offenses against children.
Appendix.
Table A1.
Predictors of Recidivism in Individuals That Sexually Offended Online Only in Their Index Offense (Group 1; n = 56).
| Online sexual recidivism (n = 4 / 56) | Contact sexual recidivism (n = 3 / 56) | Nonsexual violent recidivism (n = 6 / 56) | Nonsexual nonviolent recidivism (n = 6 / 56) | |
|---|---|---|---|---|
| Harrell’s C [95% C. I.] | ||||
| Age at follow-up | .40 [.15, .65] | .26 [.06, .45] | .29 [.17, .41] | .36 [.18, .55] |
| Criminal background | ||||
| Previous conviction for contact sexual crime | .50 [.50, .50] | .50 [.50, .50] | .50 [.50, .50] | .50 [.50, .50] |
| Previous conviction for child sexual exploitation material | .64 [.41, .87] | .49 [.47, .51] | .49 [.47, .51] | .49 [.47, .51] |
| Previous conviction for child luring | .50 [.50, .50] | .50 [.50, .50] | .50 [.50, .50] | .50 [.50, .50] |
| Previous conviction for violent crime | .41 [.35, .46] | .57 [.31, .82] | .72 [.53, .92] | .78 [.61, .95] |
| Previous conviction for general crime | .37 [.31, .43] | .68 [.39, .97] | .76 [.59, .93] | .90 [.85, .95] |
| Identifiable victims | ||||
| Number of male child victims | .40 [.23, .57] | .42 [.27, .57] | .42 [.31, .52] | .42 [.31, .52] |
| Number of female child victims | .80 [.45, 1.15] | .42 [.14, .69] | .46 [.28, .64] | .46 [.28, .64] |
| Deviant sexual interests and cognitions | ||||
| Molest scale | .56 [.20, .93] | .88 [.77, 1.00] | .59 [.30, .89] | .63 [.38, .87] |
| Cognitions of internet sexual offending scale | .53 [.20, .85] | .36 [.07, .64] | .37 [.10, .65] | .36 [.12, .60] |
| Cognitive and emotional congruence with children scale | .50 [.17, .84] | .79 [.54, 1.05] | .40 [.14, .65] | .48 [.29, .68] |
| C-ECWC: Childlike sense of self | .62 [.32, .92] | .65 [.34, .96] | .32 [.18, .47] | .43 [.23, .63] |
| C-ECWC: Friendship and positive relationships with children | .45 [.07, .84] | .84 [.71, .97] | .47 [.18, .75] | .62 [.42, .82] |
| C-ECWC: Romantic attraction to children | .54 [.24, .84] | .72 [.44, 1.00] | .40 [.19, .60] | .42 [.20, .63] |
| Sexual interest cardsort – child molestation of girls | .57 [.32, .83] | .76 [.64, .87] | .41 [.19, .63] | .41 [.18, .64] |
| Sexual interest cardsort – child molestation of boys | .48 [.25, .72] | .79 [.50, 1.09] | .47 [.30, .73] | .43 [.23, .63] |
| Revised screening scale for pedophilic interests | .57 [.40, .73] | .61 [.39, .84] | .40 [.28, .52] | .51 [.33, .69] |
| Sexual coping and preoccupation | ||||
| Coping using sex inventory – consent | .74 [.43, 1.06] | .15 [.03, .27] | .46 [.14, .78] | .34 [.16, .52] |
| CUSI – child | .56 [.32, .80] | .70 [.40, .99] | .41 [.21, .60] | .38 [.21, .56] |
| CUSI – rape | .61 [.32, .90] | .69 [.35, 1.03] | .58 [.31, .84] | .66 [.45, .87] |
| Sexual preoccupation scale | .60 [.30, .90] | .71 [.51, .90] | .57 [.28, .87] | .49 [.30, .68] |
| Pornography/CSEM consumption | ||||
| Frequency of legal pornography use | .57 [.28, .85] | .41 [.16, .66] | .49 [.22, .77] | .44 [.24, .63] |
| Frequency of CSEM use | .69 [.44, .93] | .35 [.15, .56] | .29 [.16, .41] | .28 [.15, .41] |
| Frequency of violent pornography use | .59 [.33, .84] | .67 [.40, .93] | .44 [.24, .65] | .40 [.25, .56] |
| Frequency of bestiality pornography use | .78 [.65, .91] | .63 [.38, .87] | .39 [.19, .59] | .30 [.15, .44] |
Note. Bolded values = the 95% C. I. does not comprise .50. These analyses are highly unstable due to low sample size and low base rates and should be considered exploratory.
Table A2.
Predictors of Recidivism in Individuals That Had CSEM Crimes in Their Index Offense (Subset of Group 1; n = 36).
| Online sexual recidivism (n = 2 / 36) | Contact sexual recidivism (n = 3 / 36) | Nonsexual violent recidivism (n = 5 / 36) | Nonsexual nonviolent recidivism (n = 4 / 36) | |
|---|---|---|---|---|
| Harrell’s C [95% C. I.] | ||||
| Age at follow-up | .16 [.03, .27] | .24 [.06, .41] | .26 [.13, .40] | .24 [.10, .39] |
| Criminal background | ||||
| Previous conviction for contact sexual crime | .50 [.50, .50] | .50 [.50, .50] | .50 [.50, .50] | .50 [.50, .50] |
| Previous conviction for child sexual exploitation material | .74 [.39, 1.08] | .52 [.48, .55] | .48 [.45, .51] | .49 [.34, .99] |
| Previous conviction for child luring | .50 [.50, .50] | .50 [.50, .50] | .50 [.50, .50] | .50 [.50, .50] |
| Previous conviction for violent crime | .38 [.31, .45] | .55 [.29, .81] | .75 [.56, .95] | .80 [.60, .99] |
| Previous conviction for general crime | .35 [.28, .43] | .68 [.40, .96] | .72 [.52, .92] | .89 [.82, .95] |
| Identifiable victims | ||||
| Number of male child victims | N/A | N/A | N/A | N/A |
| Number of female child victims | N/A | N/A | N/A | N/A |
| Deviant sexual interests and cognitions | ||||
| Molest scale | .65 [.57, .72] | .88 [.76, 1.00] | .74 [.54, 94] | .79 [.61, .98] |
| Cognitions of internet sexual offending scale | .50 [-.04, 1.04] | .68 [.37, 1.00] | .51 [.48, .55] | .55 [.21, .88] |
| Cognitive and emotional congruence with children scale | .65 [.18, 1.11] | .80 [.56, 1.05] | .46 [.19, .72] | .59 [.37, .81] |
| C-ECWC: Childlike sense of self | .88 [.73, 1.04] | .64 [.31, .97] | .33 [.17, .49] | .52 [.27, .77] |
| C-ECWC: Friendship and positive relationships with children | .55 [-.04, 1.14] | .83 [.68, .97] | .47 [.18, .75] | .68 [.48, .89] |
| C-ECWC: Romantic attraction to children | .59 [.16, 1.02] | .70 [.40, .99] | .44 [.23, .66] | .41 [.14, .68] |
| Sexual interest cardsort – child molestation of girls | .71 [.35, 1.06] | .72 [.57, .87] | .42 [.20, .65] | .47 [.21, .72] |
| Sexual interest cardsort – child molestation of boys | .55 [.25, .85] | .77 [.46, 1.08] | .52 [.30, .74] | .47 [.21, .72] |
| Revised screening scale for pedophilic interests | .54 [.50, .59] | .61 [.37, .86] | .45 [.40, .50] | .58 [.38, .77] |
| Sexual coping and preoccupation | ||||
| Coping using sex inventory – consent | .93 [.88, .98] | .12 [.00, .24] | .45 [.10, .81] | .36 [.11, .61] |
| CUSI – child | .72 [.53, .91] | .66 [.36, .96] | .39 [.17, .61] | .39 [.16, .61] |
| CUSI – rape | .57 [.14, 1.00] | .67 [.32, 1.03] | .38 [.09, .68] | .70 [.44, .96] |
| Sexual preoccupation scale | .70 [.29, 1.10] | .67 [.46, .88] | .67 [.44, .90] | .57 [.39, .74] |
| Pornography/CSEM consumption | ||||
| Frequency of legal pornography use | .74 [.43, 1.05] | .35 [.10, .61] | .62 [.41, .84] | .47 [.26, .67] |
| Frequency of CSEM use | .90 [.83, .97] | .66 [.47, .86] | .29 [.14, .45] | .30 [.13, .46] |
| Frequency of violent pornography use | .56 [.21, .90] | .65 [.38, .91] | .46 [.22, .70] | .42 [.22, .62] |
| Frequency of bestiality pornography use | .80 [.60, 1.00] | .35 [.11, .59] | .39 [.16, .63] | .28 [.09, .47] |
Note. Bolded values = the 95% C. I. does not comprise .50. There are no Harrell’s C analyses for the number of male and female victims because for this subsample, these variables were in fact constants. These analyses are highly unstable due to low sample size and low base rates and should be considered exploratory.
Footnotes
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 iDs
Sarah Paquette https://orcid.org/0000-0003-3276-5641
Sébastien Brouillette-Alarie https://orcid.org/0000-0002-7620-8786
References
- Abel G. G., Becker J. V. (1979). The sexual interest card sort. Unpublished manuscript. [Google Scholar]
- Abel G. G., Becker J. V., Mittelman M., Cunningham-Rathner J., Rouleau J. L., Murphy W. D. (1987). Self-reported sex crimes of nonincarcerated paraphiliacs. Journal of Interpersonal Violence, 2(1), 3–25. 10.1177/088626087002001001 [DOI] [Google Scholar]
- Andrews D. A., Bonta J. (2010). Rehabilitating criminal justice policy and practice. Psychology, Public Policy, and Law, 16(1), 39–55. 10.1037/a0018362 [DOI] [Google Scholar]
- Babchishin K. M., Dibayula S., McCulloch C., Hanson R. K., Helmus L. M. (2023). ACUTE-2007 and STABLE-2007 predict recidivism for men adjudicated for child sexual exploitation material offending. Law and Human Behavior, 47(5), 606–618. 10.1037/lhb0000540 [DOI] [PubMed] [Google Scholar]
- Babchishin K. M., Hanson R. K., Blais J. (2016). Less is more: Using Static-2002R subscales to predict violent and general recidivism among sexual offenders. Sexual Abuse, 28(3), 187–217. 10.1177/1079063215569544 [DOI] [PubMed] [Google Scholar]
- Babchishin K. M., Hanson R. K., Hermann C. A. (2011). The characteristics of online sex offenders: A meta-analysis. Sexual Abuse, 23(1), 92–123. 10.1177/1079063210370708 [DOI] [PubMed] [Google Scholar]
- Babchishin K. M., Hanson R. K., VanZuylen H. (2015). Online child pornography offenders are different: A meta-analysis of the characteristics of online and offline sex offenders against children. Archives of Sexual Behavior, 44(1), 45–66. 10.1007/s10508-014-0270-x [DOI] [PubMed] [Google Scholar]
- Babchishin K. M., Lee S. C., Eke A. W., Seto M. C. (2022). Temporal order of sexual offending is risk-relevant for individuals with child sexual exploitation materials offences. Sexual Offending: Theory, Research, and Prevention, 17(e7229), 1–22. 10.5964/sotrap.7229 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bouchard M., Lussier P. (2015). Estimating the size of the sexual aggressor population. In Blokland A., Lussier P. (Eds.), Sex offenders: A criminal career approach (pp. 351–371). John Wiley & Sons. [Google Scholar]
- Briggs P., Simon W. T., Simonsen S. (2011). An exploratory study of internet-initiated sexual offenses and the chat room sex offender: Has the internet enabled a new typology of sex offender? Sexual Abuse, 23(1), 72–91. 10.1177/1079063210384275 [DOI] [PubMed] [Google Scholar]
- Brouillette-Alarie S., Babchishin K. M., Hanson R. K., Helmus L.-M. (2016). Latent constructs of the static-99r and static-2002r: A three-factor solution. Assessment, 23(1), 96–111. 10.1177/1073191114568114 [DOI] [PubMed] [Google Scholar]
- Brouillette-Alarie S., Proulx J. (2019). The etiology of risk in sexual offenders: A preliminary model. Sexual Abuse, 31(4), 431–455. 10.1177/1079063218759325 [DOI] [PubMed] [Google Scholar]
- Brouillette-Alarie S., Proulx J., Hanson R. K. (2018). Three central dimensions of sexual recidivism risk: Understanding the latent constructs of Static-99R and Static 2002R. Sexual Abuse, 30(6), 676–704. 10.1177/1079063217691965 [DOI] [PubMed] [Google Scholar]
- Brown S. J. (2022). Assessing the risk of users of child sexual exploitation material committing further offences: A scoping review. Journal of Sexual Aggression Advance online publication, 30(1), 1–24. 10.1080/13552600.2022.2104394 [DOI] [Google Scholar]
- Bumby K. M. (1996). Assessing the cognitive distortions of child molesters and rapists: Development and validation of the MOLEST and RAPE scales. Sexual Abuse, 8(1), 37–54. 10.1007/BF02258015 [DOI] [Google Scholar]
- Canadian Centre for Child Protection . (2016). Child sexual abuse images on the internet. A Cybertip.ca Analysis. https://protectchildren.ca/pdfs/CTIP_CSAResearchReport_2016_en.pdf [Google Scholar]
- Canadian Centre for Child Protection . (2018). “Bestiality” as reflected in Canadian case law. CanLII Authors Program, 266. https://content.c3p.ca/pdfs/C3P_BE_CanadianCaseLaw_en.pdf [Google Scholar]
- Chopin J. (2017). La gestion des liens entre les crimes sexuels de prédations : Repenser ViCLAS sous la perspective du paradigme situationnel. [Thesis]: Université de Lausanne. [Google Scholar]
- Chopin J., Beauregard E. (2019). Sexual homicide of children: A new classification. International Journal of Offender Therapy and Comparative Criminology, 63(9), 1681–1704. 10.1177/0306624X19834419 [DOI] [PubMed] [Google Scholar]
- Chopin J., Fortin F., Paquette S. (2022). Childhood victimization and poly-victimization of online sexual offenders: A developmental psychopathology perspective. Child Abuse & Neglect, 129(1), Article 105659. 10.1016/j.chiabu.2022.105659 [DOI] [PubMed] [Google Scholar]
- Cortoni F., Marshall W. L. (2001). Sex as a coping strategy and its relationship to juvenile sexual history and intimacy in sexual offenders. Sexual Abuse, 13(1), 27–43. 10.1177/107906320101300104 [DOI] [Google Scholar]
- Drury A. J., Elbert M. J., DeLisi M. (2020). The dark figure of sexual offending: A replication and extension. Behavioral Sciences & the Law, 38(6), 559–570. 10.1002/bsl.2488 [DOI] [PubMed] [Google Scholar]
- Edwards M. J. (2019). Arrest and prosecution of animal sex abuse (bestiality) offenders in the United States, 1975-2015. Journal of the American Academy of Psychiatry and the Law, 47(3), 335–346. 10.29158/jaapl.003836-19 [DOI] [PubMed] [Google Scholar]
- Eke A. W., Helmus M., Seto M. C. (2019). A validation study of the child pornography offender risk tool (CPORT). Sexual Abuse, 31(4), 456–476. 10.1177/1079063218762434 [DOI] [PubMed] [Google Scholar]
- Eke A. W., Seto M. C. (2023). Correspondence of child age and gender distribution in child sexual exploitation material and other child content with age and gender of child sexual assault victims. Sexual Abuse, 35(3), 375–397. 10.1177/10790632221108951 [DOI] [PubMed] [Google Scholar]
- Eke A. W., Seto M. C., Williams J. (2011). Examining the criminal history and future offending of child pornography offenders: An extended prospective follow-up study. Law and Human Behavior, 35(6), 466–478. 10.1007/s10979-010-9252-2 [DOI] [PubMed] [Google Scholar]
- Elliott I. A., Mandeville-Norden R., Rakestrow-Dickens J., Beech A. R. (2019). Reoffending rates in a UK community sample of individuals with convictions for indecent images of children. Law and Human Behavior, 43(4), 369–382. 10.1037/lhb0000328 [DOI] [PubMed] [Google Scholar]
- Faust E., Bickart W., Renaud C., Camp S. (2015). Child pornography possessors and child contact sex offenders: A multilevel comparison of demographic characteristics and rates of recidivism. Sexual Abuse: A Journal of Research and Treatment, 27(5), 460–478. 10.1177/1079063214521469 [DOI] [PubMed] [Google Scholar]
- Fortin F., Paquette S., Dupont B. (2018). From online to offline sexual offending: Episodes and obstacles. Aggression and Violent Behavior, 39(1), 33–41. 10.1016/j.avb.2018-01.003 [DOI] [Google Scholar]
- Fortin F., Proulx J. (2019). Sexual interests of child sexual exploitation material (CSEM) consumers: Four patterns of severity over time. International Journal of Offender Therapy and Comparative Criminology, 63(1), 55–76. 10.1177/0306624X18794135 [DOI] [PubMed] [Google Scholar]
- Fraser J. M., Babchishin K. M., Helmus L. M. (2024). Emotional congruence with children: An empirical examination of different models in men with a history of sexually offending against children. Sexual Abuse, 36(5), 546–571. https://doi.org/10.1177.10790632231172160 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gebhard P. H., Gagnon J. H., Pomeroy W. B., Christenson C. V. (1965). Sex offenders: An analysis of types. Harper & Row. [Google Scholar]
- Gendreau P., Little T., Goggin C. (1996). A meta-analysis of the predictors of adult offender recidivism: What works. Criminology, 34(4), 575–608. 10.1111/j.1745-9125.1996.tb01220.x [DOI] [Google Scholar]
- Goller A., Jones R., Dittmann V., Taylor P., Graf M. (2016). Criminal recidivism of illegal pornography offenders in the overall population—a national cohort study of 4,612 offenders in Switzerland. Advances in Applied Sociology, 06(02), 48–56. 10.4236/aasoci.2016.62005 [DOI] [Google Scholar]
- Hanson R. K., Bourgon G., Helmus L., Hodgson S. (2009). The principles of effective correctional treatment also apply to sexual offenders: A meta-analysis. Criminal Justice and Behavior, 36(9), 865–891. 10.1177/0093854809338545 [DOI] [Google Scholar]
- Hanson R. K., Bussière M. T. (1998). Predicting relapse: A meta-analysis of sexual offender recidivism studies. Journal of Consulting and Clinical Psychology, 66(2), 348–362. 10.1037/0022-006X.66.2.348 [DOI] [PubMed] [Google Scholar]
- Hanson R. K., Morton-Bourgon K. E. (2005). The characteristics of persistent sexual offenders: A meta-analysis of recidivism studies. Journal of Consulting and Clinical Psychology, 73(6), 1154–1163. 10.1037/0022-006X.73.6.1154 [DOI] [PubMed] [Google Scholar]
- Harrell F. E., Lee K. L., Mark D. B. (1996). Multivariable prognostic models: Issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Statistics in Medicine, 15(4), 361–387. 10.1002/(SICI)1097-0258(19960229)15:4<361::AID-SIM168>3.0.CO;2-4 [DOI] [PubMed] [Google Scholar]
- Harris A. J. R., Hanson R. K. (2004). Sex offender recidivism: A simple question (user report 2004-03). Ontario: Public Safety and Emergency Preparedness Canada. [Google Scholar]
- Helmus L., Hanson R. K., Babchishin K. M., Mann R. E. (2013). Attitudes supportive of sexual offending predict recidivism: A meta-analysis. Trauma, Violence, & Abuse, 14(1), 34–53. 10.1177/1524838012462244 [DOI] [PubMed] [Google Scholar]
- Helmus L. M., Eke A. W., Farmus L., Seto M. C. (2024). The CPORT and risk matrix 2000 for men convicted of child sexual exploitation material (CSEM) offenses: A predictive accuracy comparison and meta-analysis. Criminal Justice and Behavior, 51(1), 3–23. 10.1177/00938548231208194 [DOI] [Google Scholar]
- Hermann C. A., Babchishin K. M., Nunes K. L., Leth-Steensen C., Cortoni F. (2012). Factor structure of the bumby RAPE scale: A three-factor model. Criminal Justice and Behavior, 39(7), 869–886. 10.1177/0093854812436802 [DOI] [Google Scholar]
- Hermann C. A., McPhail I. V., Helmus L. M., Hanson R. K. (2017). Emotional congruence with children is associated with sexual deviancy in sexual offenders against children. International Journal of Offender Therapy and Comparative Criminology, 61(12), 1311–1334. 10.1177/0306624X15620830 [DOI] [PubMed] [Google Scholar]
- Holland L. A., Zolondek S. C., Abel G. G., Jordan A. D., Becker J. V. (2000). Psychometric analysis of the sexual interest Cardsort questionnaire. Sexual Abuse, 12(2), 107–122. 10.1177/107906320001200203 [DOI] [PubMed] [Google Scholar]
- Holoyda B., Gosal R., Welch K. M. (2020). Bestiality among sexually violent predators. The journal of the American Academy of Psychiatry and the Law, 48(3), 358–364. 10.29158/JAAPL.003941-20 [DOI] [PubMed] [Google Scholar]
- Hörnle T., Tetal C., Wössner G. (2024). Reoffending after convictions related to child sexual exploitation material: Data from the German Federal Central Criminal Register. Child Abuse & Neglect, 153(106806), 1–12. 10.1016/j.chiabu.2024.106806 [DOI] [PubMed] [Google Scholar]
- Ibrahim D. (2022). L’exploitation sexuelle des enfants et la violence sexuelle à l’égard des enfants en ligne au Canada : un profil statistique des affaires déclarées par la police et des accusations portées devant les tribunaux, 2014 à 2020. Juristat, no 85-002-X. [Google Scholar]
- Krone T., Smith R. G., Cartwright J., Hutchings A., Tomison A., Napier S. (2017). Online child sexual exploitation offenders: A study of Australian law enforcement data. Report to the Criminology Research Advisory Council. https://handle.westernsydney.edu.au:8081/1959.7/uws:43731 [Google Scholar]
- Leclerc B., Proulx J., Beauregard E. (2009). Examining the modus operandi of sexual offenders against children and its practical implications. Aggression and Violent Behavior, 14(1), 5–12. 10.1016/j.avb.2008.08.001 [DOI] [Google Scholar]
- Loeber R., Farrington D. P. (Eds.), (2001). Child delinquents: Development, intervention, and service needs. Sage Publications, Inc. 10.4135/9781452229089 [DOI] [Google Scholar]
- Lussier P., McCuish E., Proulx J., Chouinard-Thivierge S., Frechette J. (2022). The drop in sexual recidivism in Canada. A meta-nalaysis of sex offender recidivism rates over an 80-year period, 22(1), 125–160. 10.1111/1745-9133.12611 [DOI] [Google Scholar]
- Mann R. E., Hanson R. K., Thornton D. (2010). Assessing risk for sexual recidivism: Some proposals on the nature of psychologically meaningful risk factors. Sexual Abuse, 22(2), 191–217. 10.1177/1079063210366039 [DOI] [PubMed] [Google Scholar]
- McPhail I. V., Hermann C. A., Nunes K. L. (2013). Emotional congruence with children and sexual offending against children: A meta-analytic review. Journal of Consulting and Clinical Psychology, 81(4), 737–749. 10.1037/a0033248 [DOI] [PubMed] [Google Scholar]
- Osborn J., Elliott I. A., Middleton D., Beech A. R. (2010). The use of actuarial risk assessment measures with UK internet child pornography offenders. Journal of Aggression, Conflict and Peace Research, 2(3), 16–24. 10.5042/jacpr.2010.0333 [DOI] [Google Scholar]
- Paquette S., Brouillette-Alarie S. (2020). Sexual offenders against children’s patterns of sexual interests. In Akerman G., Perkins D., Bartels R. (Eds.), Assessing and managing problematic sexual interests: A Practitioner’s guide. Routledge. 10.4324/9780429287695-1 [DOI] [Google Scholar]
- Paquette S., Brouillette-Alarie S., Seto M. C. (2022). Pornography use, offense-supportive cognitions, atypical sexual interests, and sexual offending against children. The Journal of Sex Research, 59(6), 792–804. 10.1080/00224499.2021.2023450 [DOI] [PubMed] [Google Scholar]
- Paquette S., Chopin J., Fortin F. (2022). Child sexual exploitation material users, one size fits all for? Exploring tailored clinical dimensions based on cognitive and behavioural criminogenic factors. CrimRxiv. https://www.crimrxiv.com/pub/412bwf5m [DOI] [PubMed] [Google Scholar]
- Paquette S., Chopin J., Fortin F. (2023). Crimes sexuels en ligne. délinquants et victimes: théorie, recherche et pratique. Les Presses de l’Université Laval. [Google Scholar]
- Paquette S., Cortoni F. (2020). The development and validation of the cognitions of internet sexual offending (C-ISO) scale. Sexual Abuse, 32(8), 907–930. 10.1177/1079063219862281 [DOI] [PubMed] [Google Scholar]
- Paquette S., Cortoni F. (2021). Offence-supportive cognitions, atypical sexuality, problematic self-regulation, and perceived anonymity among online and contact sexual offenders against children. Archives of Sexual Behavior, 50(5), 2173–2187. 10.1007/s10508-020-01863-z [DOI] [PubMed] [Google Scholar]
- Paquette S., Cortoni F. (2022). Offense-supportive cognitions expressed by men who use internet to sexually exploit children: A thematic analysis. A Thematic Analysis. International Journal of Offender Therapy and Comparative Criminology, 66(6-7), 647–669. 10.1177/0306624X20905757 [DOI] [PubMed] [Google Scholar]
- Paquette S., Cortoni F., Proulx J., Longpré N. (2014). An examination of implicit theories among francophone child molesters. Journal of Sexual Aggression, 20(2), 182–196. 10.1080/13552600.2013.798689 [DOI] [Google Scholar]
- Paquette S., Fortin F. (2021). A peek in their mind? A preliminary identification of behavioural predictors of implicit theories among men who sexually exploit children and adolescents. International Journal of Offender Therapy and Comparative Criminology, 67(6-7), 591–617. 10.1177/0306624X21101352 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Paquette S., McPhail I. V. (2020). Construction and preliminary validation of the cognitive and emotional congruence with children (C–ECWC) scale. Psychological Assessment, 32(8), 739–751. 10.1037/pas0000832 [DOI] [PubMed] [Google Scholar]
- Pheonix A., Fernandez Y., Harris A. J. R., Helmus M., Hanson R. K., Thornton D. (2017). Static-99R coding rules. Revised – 2016. Public Safety Canada. https://www.publicsafety.gc.ca/cnt/rsrcs/pblctns/sttc-2016/index-en.aspx [Google Scholar]
- Rice M. E., Harris G. T. (2005). Comparing effect sizes in follow-up studies: ROC area, Cohen’s d, and r. Law and Human Behavior, 29(5), 615–620. 10.1007/s10979-005-6832-7 [DOI] [PubMed] [Google Scholar]
- Savage L. (2024). Online child sexual exploitation: A statiscal profile of police-reported incidents in Canada, 2014 to 2022. Juristat, (85-002-x). https://www150.statcan.gc.ca/n1/pub/85-002-x/2024001/article/00003-eng.htm. [Google Scholar]
- Savoie V., Quayle E., Flynn E., O’Rourke S. (2022). Predicting risk of reoffending in persons with child sexual exploitation material offense histories: The use of Child Pornography Offender Risk Tool in a Scottish population. Sexual Abuse, 34(5), 568–596. 10.1177/10790632211047190 [DOI] [PubMed] [Google Scholar]
- Seto M. C. (2009). Assessing the risk posed by child pornography offenders. Communication presented at the global symposium for examining the relationship between online and offline offenses and preventing the sexual exploitation of children. University of North Carolina. [Google Scholar]
- Seto M. C., Cantor J., Blanchard R. (2006). Child pornography offenses are a valid diagnostic indicator of pedophilia. Journal of Abnormal Psychology, 115(3), 610–615. 10.1037/0021-843X.115.3.610 [DOI] [PubMed] [Google Scholar]
- Seto M. C., Eke A. W. (2015). Predicting recidivism among adult male child pornography offenders: Development of the Child Pornography Offender Risk Tool (CPORT). Law and Human Behavior, 39(4), 416–429. 10.1037/lhb0000128 [DOI] [PubMed] [Google Scholar]
- Seto M. C., Eke A. W. (2017). Correlates of admitted sexual interest in children among individuals convicted of child pornography offenses. Law and Human Behavior, 41(3), 305–313. 10.1037/lhb0000240 [DOI] [PubMed] [Google Scholar]
- Seto M. C., Sandler J. C., Freeman N. J. (2017). The revised screening scale for pedophilic interests: Predictive and concurrent validity. Sexual Abuse, 29(7), 636–657. 10.1177/1079063215618375 [DOI] [PubMed] [Google Scholar]
- Seto M. C., Stephens S., Lalumière M. L., Cantor J. M. (2017). The revised screening scale for pedophilic interests (SSPI-2): Development and criterion-related validation. Sexual Abuse, 29(7), 619–635. 10.1177/1079063215612444 [DOI] [PubMed] [Google Scholar]
- Seto M. C., Wood J. M., Babchishin K. M., Flynn S. (2012). Online solicitation offenders are different from child pornography offenders and lower risk contact sexual offenders. American Psychological Association, 36(4), 320–330. 10.1037/h0093925 [DOI] [PubMed] [Google Scholar]
- Snell W. E., Papini D. R. (1989). The sexuality scale: An instrument to measure sexual-esteem, sexual-depression and sexual-preoccupation. The Journal of Sex Research, 26(2), 256–263. 10.1080/00224498909551510 [DOI] [Google Scholar]
- Soldino V., Carbonell-Vayá E. J., Seigfried-Spellar K. C. (2019). Spanish validation of the child pornography offender risk tool. Sexual Abuse, 33(5), 503–528. 10.1177/1079063220928958 [DOI] [PubMed] [Google Scholar]
- Therneau T. (2020). Package “survival” (version 3.1-11). [Computer software; ]. [Google Scholar]
- Ward T. (2000). Sexual offenders’ cognitive distortions as implicit theories. Aggression and Violent Behavior, 5(5), 491–507. 10.1016/S1359-1789(98)00036-6 [DOI] [Google Scholar]
- Ward T., Keenan T. (1999). Child molesters’ implicit theories. Journal of Interpersonal Violence, 14(8), 821–838. 10.1177/088626099014008003 [DOI] [Google Scholar]
- Wolak J., Finkelhor D., Mitchell K. J., Ybarra M. L. (2008). Online “predators” and their victims: Myths, realities, and implications for prevention and treatment. American Psychological Association, 63(2), 111–128. 10.1037/0003-066x.63.2.111 [DOI] [PubMed] [Google Scholar]
- Young K. (2005). Profiling online sex offenders, cyber-predators, and pedophiles. Journal of Behavioral Profiling, 5(1), 1–18. https://netaddiction.com/articles/Pedophile.pdf. [Google Scholar]
