Abstract
Child sexual exploitation material (CSEM) users elicit strong negative reactions from society and people within their networks. There are symbolic and social boundaries that these individuals have transgressed, and subsequent identity work involves the negotiation of self and self-presentation. This article combines results from two studies to explore negotiation of identity, symbolic and social boundaries, and associated narratives among 103 CSEM users. One study was an anthropological ethnography with 17 months of UK fieldwork in community-based group programs, and the other involved four months of interviews in sexual offense treatment units of a US prison. Participants’ identity work had commonalities: distinguishing between acts vs identities; differentiating crimes from identities; comparing offenses to others viewed as worse; framing childhood experiences as influencing offending; and situating both offending and post-offending identities within larger society. Results are discussed in the context of debates about risk, treatment, prevention, harm, denial/downplay/minimization, and reintegration. Furthermore, we highlight how identity work occurs within potentially competing/contrasting personal, judicial, treatment, media, and societal reactions to and expectations of individuals who have committed sexual offenses. Finally, we demonstrate the methodological and analytical value of cross-disciplinary comparative qualitative research by showing similarities across participants from different countries, settings, timeframes, and interventions.
Introduction
Individuals who commit sexual crimes are among the most feared and despised types of perpetrators. Public perception is fueled by media representations, often based on reporting of extreme and rare cases such as sexual murder or kidnapping (Hebenton & Seddon, 2009; Jenkins, 1998; McAlinden, 2007; Terry, 2011; Zatkin et al., 2022). Thus, myths surrounding sexual offending exist in media reporting, most notably that those who offend are usually strangers to their victims, commit offenses solely because of deviant sexual interests, and represent a uniform risk group that cannot be reformed and are likely to reoffend (Fox, 2013; Jenkins, 1998; Katz-Schiavone et al., 2008; Lacombe, 2008; Levenson et al., 2007; McAlinden, 2006; Pickett et al., 2013; Quinn et al., 2004; Sample & Bray, 2003; Serisier, 2017; Waldram, 2009; West, 2000; Zatkin et al., 2022).
Both public fear and myths can be exacerbated by the ubiquitous and anonymous nature of the Internet, coupled with the fact that it is now in most homes and on mobile devices. The idea that strangers are responsible for abuse still pervades, but these strangers have moved to the online realm (Jewkes & Wykes, 2012). The home is commonly understood as a private sphere, while the Internet is a vast public space. This public space is brought into the private sphere, and children can thereby enter the public domain from home (Facer, 2012). Internet offending and those who perpetrate it have therefore become viewed as dangerous threats to core social values surrounding childhood innocence, the safety of the family, and privacy (Jewkes & Wykes, 2012).
Due to the harms and stigma associated with these crimes and the threats they represent, identity work is posited as central to desistance from sexual offending (Göbbels et al., 2012). Identity work refers to the process by which individuals manage both personal and social identities to control impressions of the self and representations to others (Goffman, 1963; Wagner et al., 2017). Auburn and Lea (2003) posit that narratives regarding the above by those who have committed sexual offenses “can be read as constructed to mitigate potential criticisms or labeling of the speaker as a particular sort of sex offender” (p. 295). The rejection or modification of this label represents a separation of the self from the act, and a transformation from a past to a present self (McAdams, 2006). Despite a robust literature on the role of identity work and narrative among those who have committed crimes, there has been a reluctance to include perspectives of individuals who have committed sexual offenses (Digard, 2014; Spencer, 1999). Additionally, most studies on identity work in this area have been with people who committed contact sexual offenses (e.g., Blagden et al., 2020; Borneman, 2015; Digard, 2014; Harris, 2014; Hudson, 2005; Lacombe, 2008; Victor & Waldram, 2015; Waldram, 2012), with only a handful examining narratives of child sexual exploitation material (CSEM) users (e.g., Rimer & Holt, 2023; Quayle & Taylor, 2002; Winder & Gough, 2010). CSEM is defined as illegal pictures, images, videos, films, or computer-generated material that depict a child under the age of 18 in a sexually explicit way, for example engaging in sexual activity or acts, focusing on particular bodily areas for sexual purposes, or sexual posing (Rimer, 2024).
This article combines results from two qualitative studies, one each from the United Kingdom and the United States, to explore the negotiation of identity, symbolic and social boundaries, and associated narratives among 103 CSEM users. The paper does not center around us as researchers, nor is the emphasis on how current psychological practice views these issues; rather, we focus on how participants conceptualize their behavior, how they frame their identities, how they situate CSEM usage, and the implications of all this. That said, while we do report results and perspectives from participants, we in no way condone, excuse, or justify their CSEM offending. Our work stems from the principle that in order to better comprehend why and how criminal offending happens, and to try to prevent it in the future, one must go directly to the source (which in this case is CSEM users). By doing so and moving beyond individual siloed research projects, this paper provides insight only available through comparative analysis. The research demonstrates similarities across participants from different settings, countries, timeframes, intervention programs, and engaged via multiple research methods. The article therefore provides robust findings about CSEM users, as well as highlights the value of interdisciplinary and cross-national comparative qualitative analysis for studying CSEM offending.
Literature Review: Narrative Identity, Boundary Setting, and Sexual Offending
Media representations of individuals who commit sexual crimes have had a salient effect on public perceptions (Katz-Schiavone et al., 2008; Pickett et al., 2013; Zatkin et al., 2022). Studies of public opinion in the United Kingdom (Brown et al., 2008; Viki et al., 2012), the United States (Fortney et al., 2007; Steel et al., 2022), Canada (Lam et al., 2010), and Australia (Australian Centre to Counter Child Exploitation, 2020) demonstrate alignment with media portrayals of those who commit both online and offline sexual offenses as outlined above. Yet, these representations are often based on high-profile cases and contrast available research. For example, individuals who sexually offend have been shown to have lower recidivism rates than multiple other crime types (Fortney et al., 2007; Hanson & Bussière, 1998; Hanson & Morton-Bourgon, 2005; Katz-Schiavone et al., 2008; Pickett et al., 2013; Quinn et al., 2004; Sample & Bray, 2003; Terry, 2011; West, 2000). Specific to CSEM users, studies with follow-up periods between one and 13 years have found new CSEM offense recidivism rates between 0% and 9%, and new contact offense rates between 0% and 8.2% (Seigfried-Spellar & Soldino, 2020). It is also well established that sexual offenses against children are most often committed by someone known to the victim, and not by strangers (Finkelhor, 2008; Jenkins, 1998; Jewkes & Wykes, 2012; Serisier, 2017), including production of CSEM (Gewirtz-Meydan et al., 2018; Quayle et al., 2008; Salter et al., 2021; Wolak et al., 2005a, 2012a, 2011a). Yet despite empirical research on sexual offending, inaccuracies continue to pervade society, with effects on communities, those who have committed these crimes as they attempt to reintegrate post-release and manage post-offense identities, and sexual offending prevention efforts more broadly.
As Erving Goffman noted in his seminal work on interaction, self-image, and identity, “when a face has been threatened, face-work must be done” (Goffman, 1967, p. 27). Individuals who offend engage in boundary setting and identity work to claim or attempt to retain membership in a positively constructed group (Copes et al., 2016). Identity development occurs through narratives and the active construction of a seemingly linear story – creating continuity among past, present, and future selves, across various contexts, and between the self and the larger dominant culture (Morash et al., 2020; Pasupathi et al., 2007; Syed & McLean, 2016). Narratives offer constructed accounts of who someone is, how they became that person, and who they will be (or want to be) moving forward into the future (Fivush & Marin, 2018).
Within this context, symbolic and social boundaries play an important role in identity work. Symbolic boundaries are distinctions and classifications made by individuals, which are used to define and categorize aspects of social life, including practices and people (Lamont & Molnár, 2002). When there is wide consensus on a group's symbolic boundaries, social boundaries are created that have the power to regulate, restrain, control, and punish behavior (Lamont & Molnár, 2002). Of particular importance is the permeability of boundaries, both symbolic and social, and the effects of the perceptions of these boundaries (Lamont & Molnár, 2002). If group boundaries are perceived as impermeable, it creates a sense of social competition and makes social change less likely for low-status groups (Lamont & Molnár, 2002).
Research on narrative, identity, and boundaries among individuals who have committed sexual offenses has primarily focused on three areas: the techniques of neutralization used to justify or deny the crime, along with the shame associated with the sex offender label (e.g., Blagden et al., 2014, 2011), and the negotiation of identities as individuals move through criminal justice system rehabilitation/treatment and societal reentry (e.g., Hudson, 2005; Ward & Marshall, 2007). Within these foci, theories of sexual offending have emphasized the role of cognitive processes, positing that individuals must overcome internal inhibitions to both commit sexual offenses and avoid taking responsibility (Schneider & Wright, 2004). A prominent concept here for identity work, which is also a common treatment target, is “cognitive distortions,” often defined as irrational beliefs that attempt to support offending through minimizing harm and pushing blame outward (Burke et al., 2002; Ó Ciardha & Ward, 2013). For example, individuals may claim that children are not harmed, that child/adult sexual activity is consensual, or that children are denied sexual freedom by society (Waldram, 2008; Ward et al., 2007). However, some researchers are also critical of this concept, arguing it is too broad and used synonymously to also mean beliefs, attitudes, excuses, rationalizations, perceptions, denials, justifications, minimizations, and/or defenses (Maruna & Mann, 2006; Ó Ciardha & Ward, 2013). Further complicating the concept, both Maruna and Mann (2006) and Yates (2009) suggest that making excuses and denying/minimizing are cognitive processes by which all human beings attempt to maintain positive self-image and self-esteem.
Identity work among people who have committed sexual offenses can also involve either admitting to, accepting the gravity of, minimizing/downplaying, or denying their crimes. Which of these an individual communicates is then often thought to be an indicator of future risk. However, research focused on denial, minimization, risk, and recidivism has not provided conclusive agreement. Specific to CSEM, Elliott et al. (2013) found that when comparing 526 people who committed contact offenses, 459 CSEM users, and a mixed group of 143, the CSEM users demonstrated higher victim empathy and lower levels of pro-offending attitudes compared to the contact offending group. More broadly, while denial is often assumed to be an indicator of risk and thus more entrenched sexual offending identity, in their meta-analysis of 82 recidivism studies with individuals who committed multiple types of offline sexual offenses, Hanson and Morton-Bourgon (2005) found that denial had no or little relationship with recidivism, describing it as a “potentially misleading risk factor” (p. 1158). Among 6,891 individuals who committed sexual offenses of many types, Harkins et al. (2015) found that denial predicted lower recidivism irrespective of risk level. Nunes et al. (2007) analyzed three samples of 489, 490, and 73 people who committed offline sexual offenses, finding a more complicated result: denial was associated with lower recidivism among high-risk individuals, and with higher recidivism among low-risk individuals. In their study of 180 people, Harkins et al. (2010) similarly found that across multiple measures, denial predicted decreased recidivism for high-risk individuals, and those denying they were a future risk were also less likely to recidivate. However, Langton et al. (2008) contrast such results by instead finding that among 436 people who committed offline sexual offenses, higher levels of minimization predicted recidivism for higher risk individuals.
Overall, broad reviews have concluded that “the research does not convincingly demonstrate that denial is a risk factor for re-offending” (Yates, 2009, p. 183) and that “the results from more detailed individual studies suggest a slightly more complicated picture” than is possible for generalizing (Craissati, 2015, p. 399). While it may be counterintuitive to think that denial and minimization are associated with lower risk of recidivism, researchers who argue this suggest denial may indicate that individuals understand their crimes are wrong, harmful, and transgressive, and in recognizing this, they do not want to publicly admit to offending (Yates, 2009) and therefore adopt this aspect of identity.
A different area that relates to narrative identity and life experience for those who have committed sexual offenses is the connection (or not) between past abuse and later offending. The place of past abuse in an individual's understanding of their identity, and subsequent research on the notion of a “cycle of abuse,” are heavily debated. Some scholars have reported connections between abuse victimization and offending. In a meta-analysis, Babchishin et al. (2011) found that both individuals who committed Internet and contact sexual offenses reported statistically significant higher levels of childhood physical and sexual abuse compared to the general population. In another retrospective meta-analysis, Jespersen et al. (2009) similarly demonstrated that adults who committed sexual offenses were significantly more likely to have been sexually abused compared to those who did not, with a higher prevalence among people who committed crimes against children. In their meta-analysis of 89 studies, Whitaker et al. (2008) again found that those who committed sexual offenses against children were more likely to have been victimized compared to those who did not. However, they caution that “being a victim of child sexual abuse is [a] strong risk factor but is by no means the only important risk factor” (p. 541).
Other studies point to nuanced conclusions related to gender. In Australia, Ogloff et al. (2012) compared 2,759 individuals who had confirmed cases of child sexual abuse with 2,677 members of the general population matched for gender and age. Over a 45-year follow-up, they found that 99% of prior victims were not charged for a sexual offense, but that victims were 7.6 times more likely to be charged than the general population. Ogloff et al. (2012) also found that female victims were not more likely than women in the general population to have committed a sexual offense; however, male victims were significantly more likely to be later convicted than those who had not been abused in childhood (5% of male victims compared to 0.6% in the general population). For men abused at age 12 or later, this number rose to 9.2%. In their review of 47 studies, Plummer and Cossins (2018) found no evidence of a cycle for female child sexual abuse victims but did find evidence that a cycle was more likely for males where the abuse was at 12 years or older, more frequent, more serious, and/or perpetrated by someone upon whom a victim is dependent.
In contrast, in a prospective birth cohort study of 38,282 men with a history of maltreatment and/or minimum one offense, Leach et al. (2016) found that 3% of boys who were sexually abused committed a later sexual offense and that 4% of individuals who committed a sexual offense had a known sexual abuse history. The authors note that being victimized through multiple forms of child maltreatment had a significant association with sexual offending. The conclusion was that they “did not find a specific association between sexual abuse and sexual offending” and “sexually abused boys are in fact very unlikely to commit sexual offenses” (Leach et al., 2016, p. 150).
Taking all the above together, the roles played by neutralization, denial, minimization, downplay, boundaries, stigma, and childhood experience is a complicated area of both research and identity management. In the Results and Discussion, we return to all these areas and debates outlined thus far in analyzing where, why, and how they feature in CSEM user narratives and identity work. Moreover, it is important to note that all the studies cited up to this point either combined heterogeneous offense types in their samples (e.g., against adult, child, intrafamilial, and extrafamilial victims) and/or combined offline and online offenses: none focused specifically on individuals who committed sexual offenses using the Internet. This is therefore an area requiring further examination, as assuming similarity and extrapolating between offline and online contexts is not a given nor a recommendation (Eke et al., 2019; Eke & Seto, 2012; Rimer, 2021; Seto, 2013).
CSEM Users, Harm, Identity Work, and Boundary Setting
While media representations and public perception may suggest a uniform set of characteristics among CSEM users, research has consistently found that “There is no one type of Internet child pornography user, and there is no easy way to recognize an offender” (Wortley & Smallbone, 2006, p. 13). To date, the consistent finding about CSEM users is that they are almost always male, and very often White (Aslan & Edelmann, 2014; Babchishin et al., 2011; Bourke & Hernandez, 2009; Eke et al., 2011; Elliott et al., 2019; Endrass et al., 2009; Faust et al., 2015; Fortin & Proulx, 2019; Henshaw et al., 2017; Krone & Smith, 2017; Merdian et al., 2018; Middleton et al., 2005; Morgan & Lambie, 2019; Owens et al., 2016; Paquette & Cortoni, 2021; Quayle et al., 2008; Rimer & Holt, 2023; Seto, 2013; Steely et al., 2018; Taylor & Quayle, 2003; Wolak et al., 2011b, 2005b, 2012b; Wortley & Smallbone, 2012). Beyond this, and accounting for the fact that it may be partially due to selection bias (Seto, 2013) as most samples come from Europe, North America, New Zealand, and Australia, generalization is difficult (Rimer, 2024). Given variability, a quote from Jewkes (2010, p. 14) is most appropriate: Yet research from clinical psychology has found overwhelmingly that offenders (both of child pornography offences and grooming and contact offences) are in most respects a diverse and heterogeneous group … it is muted in comparison to the coverage given to individuals who fit the archetype of the “dirty old man” or social misfit.
CSEM is an abuse of the rights, dignity, and humanity of children, and the harms caused to victims by both child sexual abuse and CSEM are well known. These may differ depending on factors including life stage/age at which abuse takes place, the frequency and severity, the closeness between victim and abuser (e.g., child and parent), and the supports a victim does/does not have after disclosure (Martin, 2015; Plummer & Cossins, 2018; Zurbriggen et al., 2007). Impacts and consequences of child sexual abuse can include many potential emotional, social, physical, and personal difficulties including depression, anxiety, low self-esteem, anger, fear, aggression, dissociation, post-traumatic stress disorder, personality disorders, eating disorders, self-harming, substance abuse, intrusive thoughts, academic difficulties, problems with future relationships, and future high-risk sexual behavior (Finkelhor, 2008; Hamilton-Giachritsis et al., 2017; Lalor & McElvaney, 2010; Martin, 2015; Whitaker et al., 2008; Zurbriggen et al., 2007). In addition, victims often blame themselves for the abuse and feel guilt, shame, and embarrassment (Gewirtz-Meydan et al., 2018; Hamilton-Giachritsis et al., 2017; Hanson, 2017; Martin, 2015; Palmer, 2005; von Weiler et al., 2010). Specific to CSEM, the above impacts can be compounded by the recording and sharing of abuse. In particular, as victims know CSEM can never truly be taken offline and will continue to be viewed, the permanency of CSEM and the fact that it is not localized to where abuse took place can create a continued sense of revictimization/retraumatization, powerlessness, invasion of privacy, fear of being recognized, and an inability to have closure (Gewirtz-Meydan et al., 2018; Hamilton-Giachritsis et al., 2017; Hanson, 2017; Leonard, 2010; Martin, 2015; Palmer, 2005; von Weiler et al., 2010).
Winder and Gough (2010) use the term “self-distancing” to describe the unique ways in which those who commit CSEM offenses manage identities, detach themselves from victims and the above harms caused to victims, and attempt to move away from the “dirty old man” image. CSEM users employ self-distancing to create symbolic boundaries, specifically focused on how they did not create victims, but rather looked at images. In some cases, the children are also said to have “appeared to be happy” (Winder & Gough, 2010, p. 129). Thus, for the individuals in Winder and Gough's study, when victims smiled in photos or when no force or coercion was depicted, this was claimed to be consent and affirmative participation.
CSEM users also may dissociate themselves from the label “sex offender” (Winder & Gough, 2010). In doing so, they may make clear distinctions between themselves and people who have committed contact offenses (Winder & Gough, 2010), citing that they never harmed anyone and are not a danger to children, perhaps reflecting the fact that even within this population “pedophiles” are the most reviled (Waldram, 2008). Beyond this, there are other ways in which individuals who commit CSEM offenses negotiate boundaries surrounding their post-crime identities. In particular, research has concluded that individuals motivated by a sexual interest in children are likely to be particularly risky (Eke & Seto, 2012; Hanson & Bussière, 1998; Levenson et al., 2007; West, 2000). Furthermore, those motivated by an admitted sexual interest but who are unwilling to address it are considered a greater risk to contact offend (Hanson & Morton-Bourgon, 2005; Merdian et al., 2013; Neutze et al., 2012; Seto, 2008; Wortley & Smallbone, 2006). As such, those leading treatment note that participants often distance themselves from sexual crimes that these participants feel are “worse” (Houston et al., 1995, p. 363), which has also been found with CSEM users (Winder & Gough, 2010).
CSEM users may differ in how they manage identities compared to individuals who have committed other sexual offenses, and researching this provides an opportunity for important insight, as few studies have focused on identity negotiation with this group. There are symbolic and social boundaries that these individuals have transgressed, and they must actively work to manage their post-crime identities to undergo a transformation of self which may facilitate desistance from sexual offending (Harris, 2014), or may not. In this paper, we examine the identity work of 103 CSEM users. We combine results of two studies, one each from the United Kingdom and the United States, to explore the negotiation of identity and ways in which participants draw symbolic and social boundaries in their narratives. We begin next by detailing the methodology for both this article and its underlying studies. We then move to the findings, focusing on four themes. We end with a discussion of the findings within a broader context and literature about boundaries and identity management, including around sexual offending and risk, treatment, prevention, harm, denial/downplay/minimization, and reintegration.
Methodology
The methodology for the present article combines a subset of results from two larger projects. Both authors have individually published previously from their respective projects, using distinct parts of the data about topics unrelated to this article (Rimer, 2017, 2019, 2021; Holt et al., 2024; Osuna & Holt, 2024). The authors have also published one article together using the same methodology as in the present paper (Rimer & Holt, 2023), but again about a different topic and using entirely different data from the studies. Here, through a purposive comparative analysis from Studies 1 and 2, results were combined to focus on CSEM user participants’ constructions, understandings, and presentations of identity, and the associated social and symbolic boundaries.
Methods and Participants: Study 1
Study 1 was an ethnographic and anthropological investigation which sought to answer three research questions:
How does looking at CSEM fit into participants’ lives, and how does the Internet facilitate this? How do participants’ constructions, perceptions, and actions compare to dominant cultural understandings of childhood and sexuality? What are societal and institutional efforts to normalize participants, their actions, and their perceptions about the Internet, sexuality, and childhood/children?
The project consisted of 17 months of fieldwork in community-based group programs for individuals arrested for CSEM crimes in the United Kingdom, resulting in a total sample of 81 CSEM users. Specifically, this included participant observation in 10 distinct groups over their entire programs, totaling almost 100 sessions; 31 semi-structured interviews with group members; 15 semi-structured interviews with staff from the agency offering the program; and two staff focus groups. The project received the relevant ethical approval prior to fieldwork beginning (Oxford University, Social Sciences & Humanities Inter-Divisional Research Ethics Committee, ID number SSD/CUREC2/10-28).
The groups were not labeled “treatment,” but were psychoeducational in nature. They had one 3-hour session per week involving six to nine members and two facilitators, and took place at two locations. Participants were at different stages in the judicial process, ranging from recent arrest to 18 months waiting for a trial. On rare occasions, they received sentences while partaking. Seventy-seven finished the whole program, while two left early for prison, and two left before completing by their own choice. The program was pretrial for most, voluntary, and not court mandated.
During participant observation, the researcher observed, spoke freely, and took notes, but did not change the program content or structure in any way. Interviews and focus groups were recorded and transcribed, while group participant observation was detailed in fieldnotes. Interviews were one to two hours, with a 25-question guide of five sections: background information; Internet and pornography; children and childhood; insights into offending; and current circumstances (example questions are in Rimer, 2017, 2019, 2021).
Selection of participants was not separate from the program's admissions process, and so was dependent on the administering agency. Partaking in the research was voluntary, and participants’ ability to undertake the program was in no way affected if they did not consent. Participants received an information sheet in the mail before groups began as part of their introductory package sent by the agency, and then met the researcher at the beginning and break of the first session to ask questions. Then, they decided about consent from home, away from the researcher, other group members, and agency staff. All members consented; however, the process was such that the researcher would only partake if every person in a particular group consented independently. Every participant gave separate written informed consent to have the researcher in groups and again to be interviewed. All were promised that they would not be contacted outside the group setting, and were promised confidentiality and as much anonymity as possible (e.g., full anonymity in research outputs/materials, only first names while interacting). Confidentiality was aligned with the policy of the administering agency: information was confidential unless it indicated that a child was in danger or undisclosed abuse of a child.
Participants may have had multiple motivations for being in the groups. It is possible that some were striving for change because they saw their behavior as problematic. Others may have wanted to demonstrate to other people that they were trying to change. Still others may have been aiming to project a prosocial image for court cases. After groups ended, participants could be given a letter that summarized the program and their record of attendance, if they requested this. Beyond this, the groups were not linked to court cases; in the context of the program, staff did not write reports, testify, or evaluate/assess participants for court.
Aligned with existing literature, all CSEM users in the sample were men and all but two were White. They varied considerably in age, from 20s to 70s, with a cluster in middle age. Participants’ professions were diverse and in fields including IT, civil service, defense, health, education, construction, transport, and unemployment. Fifty-eight had current or ex- spouses or partners, 19 did not, and relationship status was unknown for four. Thirty-six had children, 39 did not, and children were unknown for six. Eight discussed how they were abused as children. All were arrested for possessing or viewing CSEM, with less for distributing, and none for producing it. Regarding previous contact with the justice system, 71 were apprehended for CSEM only and sexual offenses for the first time. The other 10 had previous CSEM convictions (n = 5), concurrent charges for online grooming (n = 2), prior contact offenses against children (n = 1), a voyeurism conviction (n = 1), and an arrest for adult sexual assault (n = 1).
The analytical process for the overall Study 1 was inductive and exploratory: there was no hypothesis being tested and no predefined coding framework. Fieldnotes and transcripts were coded openly and then clustered for themes based on frequency, consistency, and outliers, focusing on participants’ statements and behaviors. Fieldnotes were used primarily for analyzing the group process, developing themes across the study, and providing short verbatim quotes. This was then expanded in interview analyses, along with an exploration of similarity in findings from both groups and interviews. In summary, the method was informed mainly by thematic analysis (Braun & Clarke, 2006; Nowell et al., 2017). However, themes were adapted in a continuous and iterative manner, as is common in anthropology.
While the above was the process for Study 1 as its own distinct project (Rimer, 2017, 2019, 2021), it is important to note that analysis for the present paper comprised a mixed approach, for which Study 1 data was analyzed in a targeted manner. The analytic process for this article with Study 1 data, along with how it relates to the process outlined in the next section from Study 2, is detailed in the section “Methods and Participants: This Article.”
Although groups were spaces where members were encouraged to speak candidly and without being judged, there were limitations to the research. First, in groups, the researcher was restricted to content from the program curriculum. Second, what was said could not be compared to forensic evidence, as the researcher did not have access to case files. Third, the sample was voluntary and therefore cannot be generalized (however this is not the aim). Fourth, the research took place after participants were arrested, meaning they had time to contemplate their offenses and perhaps construct accounts, thus not necessarily reflecting what they thought when offending. Finally, the timing of groups may have influenced some to change what they said in order to look more favorable for court or to align with staff expectations.
Methods and Participants: Study 2
Study 2 was based on semi-structured interviews with 101 people convicted of a sexual offense who were incarcerated in specialized mental health treatment units of a US minimum-security prison. The research questions were:
Are there distinct differences among individuals and the perceived role of pornography in their sexual offenses when compared using risk levels as measured by the STATIC, STABLE, and the overall combined score? What are the common themes among individuals’ experiences regarding pornography and how do they construct the perceived role of pornography in their offending and in their lives post-release?
After ethical approval was received from the Department of Corrections and the university ethical review board (Michigan State University, Human Research Protection Program Social Science/Behavioral/Education Institutional Review Board, ID number STUDY00002018), data was collected over a 4-month period. To recruit participants, prison mental health staff described the research to all those in the treatment units and posted informational signup sheets in common areas. Interested individuals signed up by using their prison identification numbers. The signup process and scheduling of interviews were done by prison staff. During the interviews, while auditory privacy was kept, visible privacy was not guaranteed as they took place in a mental health staff office which, for both the researcher's and participant's safety, had a large window so correctional officers could maintain visibility.
Study participation was voluntary. Due to the participants’ vulnerable status, the researcher completed a verbal informed consent process with each person (a waiver of signed consent was approved by the Institutional Review Board so there would be less potential linking of individuals to interviews). In this process, the researcher reviewed a copy of the project information/informed consent sheet and reinforced several points. She explained that she was not affiliated with the Department of Corrections, but rather, was an independent researcher granted access to participants and the physical space. Next, she indicated that due to institutional policies, there would be no voice or video recording of the responses and all interview data would be recorded manually using pen and paper. Lastly, she explained that although prison mental health staff would know who participated, they would not have access to responses, no identifying information would be gathered or kept, and participation had no impact on parole or release decisions. As with Study 1, the researcher attempted to create a space free of judgment where participants felt comfortable speaking openly. This was done through rapport building, active and empathetic listening, and when appropriate, the use of humor.
Each interview was approximately 30–45 minutes and documented manually by hand. To increase precision, a research assistant was always present. Both the assistant and researcher recorded responses and took notes. These were then compared before analysis and combined to enhance reliability. An interview guide of 30 open-ended questions was used which focused on experiences of pornography such as genres watched, age of onset, viewing frequency, duration of sessions, and effects on behavior, self, sexual relationships, and fantasies (questions are in Osuna & Holt, 2024). Demographic information was also collected from each participant (age, race/ethnicity, relationship/marital status, previous occupation, and convictions).
All participants were men. Ages ranged from 20 to 67 with a mean of 40. Most were White (n = 73). Black males were almost a quarter of the sample (n = 24), while three other people identified as Hispanic and one as Native American. Participants were serving sentences for a variety of criminal sexual conduct (CSC) offenses. In the state where the research was conducted, CSC is categorized into four degrees, each addressing different circumstances and severities of the offense. The primary concerns are victim age, vulnerability of the victim, and characteristics of the event. CSC first degree is the most serious charge, specifies victim age, and involves sexual penetration where the victim is under 13, or where the victim is 13–15 and the offense has particular circumstances (e.g., both live in the same household; or both are related; or the offender is in a position of authority; or the act involves force, violence, or injury). Second-degree CSC involves sexual contact under similar circumstances as first degree but without penetration. CSC third degree involves sexual penetration again where particular circumstances exist (e.g., there is force or coercion; or the victim is cognitively or physically impaired; or the victim is related to the offender; or the victim is 13–15). Fourth-degree CSC involves sexual contact without penetration under similar circumstances. Twenty-eight participants were serving sentences for CSC third degree, 20 for CSC second degree, 18 for CSC first degree, 15 for CSEM offenses, seven for both contact and CSEM offenses, five for indecent exposure, five for lesser sex crimes, and three for crimes with a sexual element. Specific to the 22 individuals with CSEM and CSEM plus contact offenses, all those with CSEM plus contact offenses were White (n = 7), while those with CSEM-only convictions were again majority White (n = 11). Black (n = 3) and Native American (n = 1) individuals comprised the remainder of the CSEM-only group. The mean age of these 22 people was 37.5 years, with a range of 21–58.
Study 2 included a qualitative inductive analysis to explore participants’ stories regarding the perceived role of pornography in their offenses and their views about post-release access. Coding was an iterative and multi-step process. The first step was reading through the raw interview data multiple times while identifying relevant first-level codes (Miles & Huberman, 1994). These codes were developed from the extant theoretical literature on sexual offending and pornography, as well as directly from the data, to develop an initial coding scheme. This was used to guide analysis. The next step was initial/open coding to identify and label important words or groups of words. The intermediate step between coding and analysis is constructing memos, which organize salient concepts, allow comparisons, and identify connections (Charmaz, 1999). Memos in Study 2 included references to theoretical perspectives on pornography, sexual offending, and risk. Second-level axial codes were then grouped into categories or related codes. These were used to develop meaningful themes (main ideas related to the research questions).
Study 2 had limitations similar to Study 1. First, the veracity of information could not be corroborated because the researcher did not have access to treatment or police files. Second, while the researcher told all participants that she was not part of the Department of Corrections, interviews happened in spaces used by prison staff. Thus, rapport and trust were possibly affected by the prison environment. Third, the research took place after conviction, so participants had significant time to reflect and perhaps formulate narratives surrounding their offending; thus, their responses may not have been the same if they were interviewed at other times in the offending or justice process. Fourth, it is possible that some changed their answers due to a belief that participation would help the parole process or because of social desirability. Finally, as the research was voluntary, the sample was self-selecting and not generalizable.
Methods and Participants: This Article
For the present article, we performed a comparative analysis of results from Studies 1 and 2, to explore themes surrounding identity work and boundary setting. This article stems from a purposive and intentionally restricted thematic analysis (Braun & Clarke, 2006; Nowell et al., 2017). The inductive analysis of Study 2 established findings about CSEM user identity construction and presentation, particularly post-offense, as one of the project's research questions focused on “the perceived role of pornography in their offending and in their lives post-release.” Given its research questions, this was not a foregrounded focus of Study 1, and so to methodically examine similarities and differences between samples/studies, findings and codes from Study 2 were first generated following inductive analysis of the CSEM users in the post-conviction US sample as outlined above. This then provided a framework for Study 1 to use deductive focused coding and examine if the same themes applied to the pre-conviction UK sample (all CSEM users) or not. Our process was also attentive to avoiding key pitfalls in thematic analysis as outlined by Braun and Clarke (2021), in particular ensuring themes were actual themes and not categories or topics, and thinking critically about and supporting our coding approach as opposed to performing “unjustified or incompatible ‘mash-ups’” (p. 336). The coding framework was updated as analysis by both authors continued, and included the following codes listed below the now-finalized themes:
Inherently good people: acts versus identity.
What you did is not who you are. Rejecting sex offender label. Emphasizing positive attributes. Framing behavior as “helping.” Spoiled identity. Claims to “normality.”
It could have been worse: comparing to other offenses/offenders.
Denying/minimizing harm.
“Only” viewed CSEM (no manufacturing or distributing).
No contact offending.
“Watching” versus contact.
Lack of victims (or “real” victims).
Comparing to “more serious” crimes.
Comparing to others in prison.
Sentencing disparities.
Personal measuring rod: experiences as identity.
Comparing acts to own victimization as a child.
Controlling urges.
Attraction to same age peers never developed.
Others are also attracted to children.
Identity as a reflection of larger society.
Deviant social norms.
Criminalization of sex and desire.
Objectification and glorification of sex.
Victim-centered society.
Society as criminogenic.
Stigma as a barrier to re-entry/community.
The rationale for the above coding process, reasoning for why this method of analysis was most appropriate, and justification for using Study 2 as the initial foundation, was threefold. First, analysis was around one specific set of topics and results, as opposed to open coding of both datasets and a considerably more sizeable comparison of every potential theme. Thus, the method was best suited for focusing on this within two large datasets. Second, Study 2 was used as a basis for the coding framework because, compared to Study 1, it focused more on issues of identity and the life course, it centered more on participants’ lives post-release, and it was more limited in scope and size (22 CSEM user participants over four months of fieldwork vs 81 participants over 17 months of fieldwork). This meant that Study 2 could be used as an initial manageable base to investigate the topics central to this article, as opposed to using Study 1 as the original basis, which would have required a more unstructured exploration into a much larger dataset. Study 2 was more concerned with identity and sexual offending, and so by using a purposive CSEM user sample from Study 2 to compare to Study 1, deductive focused coding was the most fitting and suitable next step to precisely explore the themes in Study 1. Third and more broadly, both datasets were appropriate for thematic analysis comparison because they were similar and complementary: they were transcripts and notes, based on interviews and observations, and stemmed from a qualitative social science orientation (albeit different exact disciplines, which we explore further in the Discussion and suggest is a strength of the research).
The principal researcher of each study coded their respective dataset, meaning there was one coder per study. This is because each study was conducted separately and so had its own ethics approval with data access restrictions and corresponding informed consent processes. As such, the raw data could not be shared between this article's authors. While intercoder reliability could therefore not be measured (O’Connor & Joffe, 2020), the researchers undertook a systematic and rigorous process focused on reflexivity, communication, and teamwork to limit inconsistencies. Nowell et al. (2017, p. 4) note that in thematic analysis, there are particular practices that can be undertaken to increase trustworthiness. Below, we include and italicize those relevant to our analysis, along with our practices for each category/set of categories:
Prolonged engagement with data: we worked long term in fieldwork and with the data, as described earlier. Use of a coding framework: a coding framework was employed throughout and was updated as the analysis progressed, as outlined above. Persistent observation: we frequently communicated and met to review the coding structure and ongoing findings, which stemmed from the data. Researcher triangulation, team consensus on themes, peer debriefing, documentation of team meetings, and themes and subthemes vetted by team members: each meeting, we outlined coding decisions, shared evolving findings from the datasets, discussed, and then returned to our respective datasets. We ensured both consensus/agreement and inclusion of outliers. Each time we met and communicated, clear-cut outcomes were explicated, and aims for the next phase were decided. This procedure was repeated throughout the analytical process to enhance collaboration, communication, and the overall analysis (Campbell et al., 2013; Nowell et al., 2017).
Rather than intercoder reliability, the above reflects intercoder agreement, where potential inconsistencies are identified and coding decisions are made clear (Garrison et al., 2006). It is also worth stating that intercoder reliability measures are often illogical and inappropriate for thematic analysis: such measures originate from a positivist framework, whereas thematic analysis comes from a different epistemological and analytical position that sees researcher interpretation as both situated within a particular context and an integral part of knowledge creation, rather than a risk to “objectivity” (Braun & Clarke, 2019, 2021). The above analytical process then concluded with the complete results in this article.
The sample for this paper encompasses all 81 participants from Study 1 (all being CSEM users), and the 22 people from Study 2 who had a conviction for CSEM (CSEM-only and dual offenses). This makes the total sample size 103. We have incorporated those with dual offenses (both contact and CSEM offenses) because all those included were known to the justice system for CSEM; both samples included these individuals, as one person from Study 1 had prior dual offending convictions; and on the whole, CSEM users may have unknown or undisclosed contact offenses (Neutze et al., 2012). We therefore suggest that assuming participants in Study 1 had no contact offending history due to lack of charges, and excluding people with dual offenses from Study 2, is not strong reasoning. Instead, including every participant known to engage with CSEM provides an unusual and beneficial opportunity to compare samples from different settings, countries, timeframes, intervention programs, and engaged via multiple research methods. This can provide information about CSEM offending that moves beyond individual studies and their limitations by demonstrating commonalities among varied groups.
Limitations
This article's methodology has limitations. First, the samples were self-selected, which may skew conclusions. Second, we combined results, and there were clear distinctions between samples. Study 2 took place in a US prison, while Study 1 was in a UK community program where people were seeking help and living in society, which resulted in a setting and sample that was meaningfully different. In particular, as Study 1's intervention occurred primarily pretrial with people who were not yet convicted and Study 2's programming occurred in prison with those convicted, the nature of participants’ identity work may have been impacted or different. Yet, we found notable similar framings around identity and boundaries, suggesting that comparative research is powerful because of such differences. Third, we are unable to discern if identity work occurred as a result of intervention programming or more “naturally.” However, regardless of how this happened, it is a potentially strong force and mechanism that requires study. Fourth, we did not have access to police, case, or treatment files, nor was there follow-up in Study 1 after groups. Therefore, we only had access to the information participants wished to share and were not privy to some of the intricate details of their cases (e.g., the exact types, quantities, and severities of their CSEM collections). For this reason, and because our research is cross-jurisdictional where legal terminology may differ, we use the encompassing term CSEM; while broad, this term covers all possible material while still clearly indicating that, no matter the details for each participant’s case, everything we are discussing is illegal and involves a child under 18. As we were not able to verify the veracity of information based on police or treatment files, it may also be that participants were motivated by social desirability. However, the purpose was to understand how they framed experiences and identities. Narratives are always subjective, but they also provide rich data on the self and how people make meaning (Presser & Sandberg, 2015). Fifth, due to ethical and data access restrictions, one researcher coded each dataset. However, as outlined above, to limit inconsistencies and increase collaboration, communication, and agreement, the researchers underwent a rigorous and methodical process. Lastly, most participants were White, and all were men. While consistent with existing research, as discussed in the Literature Review, studies of CSEM users have geographical sampling biases. This limits the scope of knowledge about CSEM offending in other populations and samples (e.g., people in non-Western countries). Given this article is about identity, the experiences and beliefs of a more diverse participant base are crucial to a more encompassing understanding of CSEM offending.
Results
The paper now presents four themes found across both samples of CSEM users. We begin by detailing the various ways in which participants separated their actions from their identities. We then discuss ways in which participants compared their offenses to other crimes, both sexual and non-sexual, to show how other offenses and people were worse. The third theme then moves on to discuss how, for a minority of participants, CSEM offending was said to be connected to early experiences founded in their own childhood abuse and sexual development. Finally, the last theme demonstrates the ways in which participants connected their offending identities to features that they saw as broader societal problems and patterns. Note that throughout the presentation of results, ellipses in quotes mean that a word/words have been removed (e.g., for clarity, for redundancy), but the meaning is unchanged. Study 1 quotes derive from both interviews and participant observation, and Study 2 quotes come from interviews.
Inherently Good People: Acts Versus Identity
Participants often distinguished between actions versus identity. Through their constructions of identity around CSEM offending, participants would tell themselves and others it was not “who they were.” They made clear distinctions between their identity, or their actual selves, and the behaviors in which they engaged. They maintained that it is possible for “good” people to engage in “bad” behavior while remaining fundamentally good: That was that, and this is me. (Study 1, P73) A small area of naughtiness in a life otherwise focused on being good. (Study 1, P6) I just want people to feel that while I’ve done something bad, I’m not that person. (Study 1, P66) It is something that you’ve done, not something you were. (Study 1, P9) But now I’m a sex offender. This isn’t who I am, it's something I did. (Study 2, P32) Yes I think there is some perception, that I’m a risk … that I’m, capable of, of something, which, which to me, I would feel, would completely go against everything else in my nature. (Study 1, P77) And those that know about what's happened, they’ve all said, “well, you’ve never have looked at my child like that. I would know if you did. My child loves you to bits,” and that sort of thing. And you think, well yeah, because I’m not the same person, outside. (Study 1, P46) I want to be happy, healed, not considered a risk, because I don’t really consider myself a risk, build a new good reputation, to be trusted …. (Study 1, P66) It makes me mad that I get classified with people who are really sick. I will not go after teenagers now because I know it is wrong. I know how to say, “get the hell away from me.” (Study 2, P34; this participant recorded sexual activity with a teenager and was charged with both CSEM and contact offenses, so they may be referring to the CSEM aspects of offending, the contact aspects, or both)
The notion of boundaries between normality and monstrosity, and good and evil, similarly pointed to the ways in which those who commit these crimes are labeled as inherently bad. Yet, for these participants, what one did was said to be different from who one was. Thus, it was perhaps an attempt to reject essentialisms about evil (Waldram, 2009): And I’ve read lots of things where people say you’re born genetically wrong or, your mind's not right from birth and that's what it's all about. It's not. I was quite fine up until then, you know what I mean? Quite fine, quite normal. I consider myself normal away from the, the Internet. Just quite normal. Doing normal things. Just like anybody else. (Study 1, P69) I think to know that there are other people who on the face of it seem pretty reasonable, they’re not all monsters. They’ve just fallen by the wayside like I have, for one reason or another. (Study 1, P62) I suppose partly this is about a debate about, you know, are some people evil? Um, are, do people have the capability for evil or, do, do bad things? Um, I think some people do, maybe to some extent. However I’d hope that I wouldn’t class myself in that category. (Study 1, P66) People now question me, if they were on the fence, I am a real pervert now. I’m a monster, but I’m not a monster. I understand the severity of what I did. I have empathy but I am condemned now. Now I’m a monster and there are some who are, but I’m not. (Study 2, P2)
Finally, while there is an indisputable sexual element to CSEM use (participants were looking at and masturbating to the material), some nevertheless did not describe sexual gratification as a primary factor influencing offending. Participants live in societies where sexual attraction to children is the ultimate breaking of social norms and corresponding laws, and where such an attraction is deemed the highest level of unacceptable risk. As such, many did not claim this attraction to be a core part of their identities but rather as an action that could be separated and contrasted: Although it's a sex crime, a lot of it is entirely non-sexual. (Study 1, P6) … everyone in this group, uh, says they’re not attracted to children … but there are some that do. So, obviously to me that would be a concern for contact offending. (Study 1, P75) I never masturbated to CSEM. I would look at the pictures for stress relief. I would sit there for hours just looking. When I wanted to masturbate I would look at adult porn. (Study 2, P45) I was a normal, All American type boy. Not into kinky stuff. People that are, there should be checks and balances … I don’t have unnatural urges. (Study 2, P63)
All these distinctions meant that some of those who engaged in CSEM offending understood their actions as transitory, whereas identities were more permanent and transcended actions taken online. For these participants, actions did not wholly define identity, although they may affect identity and present the need for identity work.
It Could Have Been Worse: Comparing to Other Offenses/Offenders
Some participants also reinforced identities by comparing their offenses to others, effectively stating that their crimes were not as bad. As described, some attested they posed little risk to children. This was brought forward when participants professed that they were not, nor were they like, people who committed contact offenses, and they would never contact offend: Never ever, um, contact, or, um, anything like that, you know. Um, chat rooms, telephones. Ugh [disgusted noise], it's totally, totally you know, abhorrent to me. (Study 1, P65) I would never want to, do anything, uh, any, any contact offenses. Um, because of the impact it would have, on other people. And on myself. (Study 1, P77) And there's thousands of people like me who would have never ever touched anybody, who have been drawn into looking at images and movies. (Study 1, P45) I don’t really fit in here because of my offense. I think the sentencing is messed up. For this case I got nine to 18 months. There is a child molester in here who got six to nine months. It's better to watch porn than to victimize someone. (Study 2, P23)
Beyond stating outrightly that they had no intention to contact offend, similar “self-distancing” (Winder & Gough, 2010) was conveyed when participants expressed disgust, disdain, or discomfort with those who committed contact offenses. This included taking issue with the label of “sex offender,” as well as not wanting to be housed with such people in prison: I’m dreading the thought that if I do go to prison I end up with sex offenders. That way I’ll be classed as one. But people who’ve actually used and abused children, don’t know that I want to be with them. (Study 1, P21) Every morning I wake up and say “pedophile.” Every night I go to bed and say “pedophile.” I know I’m not a pedophile. And now I know exactly what I’m being accused of. (Study 1, P20) I don’t understand why the registry has to be public. Why everyone who commits a sex crime has to be labeled a sexual offender. It's only gotten worse over the years. You can be a murderer and get less stigma. (Study 2, P32) There's a guy here in my cube. He is here for raping his daughter from the time she was four to 11. People like him deserve bullets not prison stays. (Study 2, P34)
“Self-distancing” (Winder & Gough, 2010) was also clear for participants who compared their online offending to other crimes that they felt were more extreme, worse, or harmful. This was most often in relation to murder, rape, or kidnapping: You set out to murder someone, you can accidentally kill someone which is manslaughter. I feel more like a manslaughterer in that sense. I didn’t set out to hurt anyone. (Study 1, P21) Isn’t it funny that people would rather say that they are an attempted murderer, than downloaded some pictures on the Internet. (Study 1, P43) You know, murderers, however fucked up it is, taking someone's life away … but that's not the worst thing. (Study 1, P54) So I’m such a horrible person that you have to monitor me for life, put on a list so everyone knows everything about me, but murderers and other dangerous offenders don’t. (Study 2, P34) Do I regret the child porn? Yes. But society is so engrossed with the sexual aspect. You have people going and taking kids and murdering them. (Study 2, P46) There are guys in here that can’t control themselves sexually, like rabbits …. If they cannot control themselves maybe they should be castrated. (Study 2, P2)
Specific to their online offenses, participants conveyed multiple techniques of “self-distancing” that attempted to convey the idea that their crimes could have been worse. This included statements aligning with Winder and Gough (2010) earlier, which focused on the fact that children appeared to be smiling or happy in CSEM, or that images were not of children engaged in sexual activity: … the only images I really liked were when the kids seemed to be happy. Hate to see a child upset. (Study 1, P32) And then you tell yourself that “oh they’re just images, they’re smiling, they’re, they look like they’re enjoying it.” (Study 1, P26) But, I think you overcome the inhibitors, “it's alright, she's smiling,” or, “it's alright” cause most of the pictures I’d look at weren’t what you’d describe as any, sort of sexual contact or anything like that. (Study 1, P30) … when you talk about children being abused, it's like, if they, if in every single clip I watched they were like crying and shouting, then yeah it would have been easy just to sort of turn it off, stop, never watch it again. (Study 1, P54) I didn’t look at anyone engaged in sexual acts. It was more like girls in bathing suits or topless, in thongs, or naked. (Study 2, P45) … if you look at it you’re kind of, uh, um, what's the word? An accessory … it's a point that has to be made, but I mean realistically the few sites I looked at, for a brief period of time, I can’t believe that I’m personally responsible for very much. (Study 1, P33) I mean the last thing I would ever want to do is, um, basically, harm somebody or be detrimental to somebody, physically, mentally, spiritually, or their property … the problem is what you’re talking about here is, accessing very free-to-access, easy-to-access images of which there's, you know 10 million copies floating around on the Internet. (Study 1, P36) And you start thinking, “Well, I’m not abusing them. This would have happened anyway, nothing I can do about it. Right, everybody else is doing it anyway, look it up, there's so many people seeding this stuff.” Um, you know, “I’m not selling this. I’m not buying this. Um, I’m not giving anybody money for this. I’m not promoting this. I’m not sending this to anybody else.” (Study 1, P43) I had one picture she sent me. That picture carries more weight than if I had intercourse. I got seven-and-a-half years. I see a difference between post/sexually mature individuals and children. Children victims are more deviant. (Study 2, P61)
As can be seen, across both samples, CSEM users often compared their offenses and behavior to other people, other actions, other harms, and specific features of CSEM. This appeared to be an attempt to signify that what they had done was not as bad as what they could have done, or as bad as what others have actually done.
Personal Measuring Rod: Experiences as Identity
A minority of participants discussed offending and identity in terms of their childhoods. One way this manifested was those connecting their online offending to childhood experiences of abuse. Echoing research that connects prior victimization and later sexual offending (Babchishin et al., 2011; Jespersen et al., 2009; Ogloff et al., 2012; Whitaker et al., 2008), some participants described their CSEM usage as related to childhood abuse. For example, in Study 1, P15 described offending as “in some sort of way, I wanted to make sense of what happened to me as a child.” Another participant discussed how abuse “may have had an influence that I don't know about … I thought I’d put it out of my mind but I haven't, cause every now and then it always comes back,” and how during a particular group exercise, “all I could think about was me as a child” (Study 1, P26). In Study 2, P6 similarly stated, “I didn't see child porn as victims of rape. I thought they enjoyed it. Because I enjoyed it as a kid.”
Also in the minority but related to early life experiences, a few participants described their online offending as linked to a sexual preference, often starting in teenage years and coupled with shame. While this was not necessarily related to abuse, it nevertheless represented another way in which participants connected their offending to childhoods: I’ve known that's been an issue since I’ve been about 16-17 really. Um, but I wasn’t in the position to be able to sort of, deal with that … I would shut myself off, bury my head in the sand and get so stoned and so high that, didn’t, didn’t think about that. (Study 1, P19) … and then all the lads were getting older and sort of, starting to fancy girls and I wasn’t. And, I, I, I seemed to be stuck in the 13 to 16 age range. (Study 1, P45) Sexual preference is something that's there. That conversion stuff for people who are gay is bullshit. It's about controlling the preference and not seeking it out. You live with that hidden shame. You know it's wrong but you are still attracted. (Study 2, P45)
For the participants above, CSEM offending was therefore said to be connected to early life experiences. For some, this was their own childhood abuse. For others, it was sexual development and preference that did not seem to progress after a certain age.
Identity as a Reflection of Larger Society
The final theme speaks to participants’ connecting of offending and identity to what they saw as broader societal trends and problems. In Sykes and Matza's (1957) seminal work on techniques of neutralization, they identified five rationalizations or justifications that allowed individuals to engage in crime while still claiming membership to the non-criminal dominant culture. These play an important role in identity work and can be employed after the commission of offenses when individuals must account for their norm-violating behavior (Copes & Williams, 2007; Maruna & Copes, 2005; Morris & Copes, 2012). They can also play an important role in the construction of non-deviant and non-criminal identities, and in reclaiming a non-offending positive identity that aligns with cultural norms and expectations (Gathings & Parrotta, 2013). One of these neutralizations which is focused on macro-level society, condemnation of the condemners, sees individuals blame others for their actions. Among some participants, they sometimes did this by blaming society for depicting children and teenagers in a sexual manner, as well as by commenting on the ubiquitous nature of sex both in popular culture and online: Because the big question underlying all this is how much is natural human behavior, and how much is perversion, how much is to do with society makes people sick? Or society, society itself is sick? … I think society is sick. Um, and I think it, it makes people sick. (Study 1, P33) Um, this country's got one of the strongest sort of, anti-pedophilic sort of views, let's say for want of a better word. Um, but yet it's got the highest teenage pregnancy. You know. The number of kids having sex is … it's totally bizarre sort of, you know, and we have a very very sexualized society as well. Um, so there's, a sort of floppy sort of concept of childhood in that sense. (Study 1, P36) You look at MTV. When I was younger, they played music. Now it's all sex and drama. It's dysfunctional and toxic behavior. There is going to be an epidemic of young boys exposed to this stuff. Kids are impressionable. You want a society that's reasonable and upstanding, but we have the opposite. (Study 2, P61) If you look at the sexual industry, they make billions of dollars. But then you lock people up for a sexual crime. (Study 2, P46)
In connecting their identities to broader society, some participants also spoke about their perceptions of post-offense life, stigma, and the competing demands by society (e.g., to reintegrate but also be punished in an ongoing manner). The identity work in which these individuals must engage was constructed as integral to reintegration into the community, yet society was also viewed as punitive and unforgiving of offenses: I feel like I am not allowed to be living. (Study 1, P71) Everything that you were is gone. (Study 1, P5) But you get a drunkard, gambler … a druggy, whatever. They’re all addictions. But people seem to forgive them if they do something like that, but they don't forgive us. (Study 1, P69) Um, it's a horrible word [offender] … Because it's a word that I’m never gonna escape from. Even when I’m finished my sentence, whatever that may be, I’m still gonna be an offender. (Study 1, P26) You tell a guy you can’t live with your kids, even though his crime wasn’t against kids. Even though being with your family can lead to less recidivism. If you don’t have a crime against children, why can’t you be around your own kids? Lots of guys have to be tagged like a dog for the rest of their lives. You are turning people into lepers. There is a stigma that creates more criminal behavior. We’ve created criminals. We’ve criminalized everything and then you don’t treat them, you say good luck. (Study 2, P61) Society wants you to do something bad and get caught before they want to prevent you from doing it. (Study 1, P10) I don't know what's available to help people before they commit an offense … I wish there had been something, or something that, beforehand, stopped me and said, “OK, I’m, there's something wrong, I really need to get help on this.” (Study 1, P41) … it's like, you can only come on this program once you’ve been arrested or charged, you can’t seek help before it. (Study 1, P54) I need to find out information. I need a bit of support, bit of guidance. And everywhere I turn, people tell me come back when I’ve been charged or, or after I’ve been convicted. (Study 1, P24) There needs to be more research on this. People on the other side of the fence need to see we aren’t the only ones with this problem. There needs to be more programs and support. (Study 2, P10)
Overall, for the participants quoted in this theme, offending itself, offending prevention, and the ability to move on after committing a crime were said to be connected to societal ills and norms. Participants particularly noted that some of these demands and norms appear, to them, to be contradictory or competing.
Discussion
Those with a sexual interest in children have been found to be one of the most negatively evaluated groups (Jahnke et al., 2015). Thus, identity work is necessary for individuals who have violated these normative boundaries and committed CSEM offenses. These individuals navigate managing agency and accountability for their harmful behavior while simultaneously trying to distance themselves from the identity of someone who is sexually attracted to children (Winder & Gough, 2010). CSEM users have transgressed core symbolic and social boundaries, and corresponding laws, and they have engaged in behavior that has serious consequences for victims. Their resultant identity work therefore involves the negotiation of self and self-presentation in trying to attain a balance of accountability and distance.
Criminological research has included considerable focus on post-offense identity among particular types of people who have committed crime; however, little is known about the patterns and functions of identity work among CSEM users, with the other major study having been conducted by Winder and Gough (2010). Our paper builds upon Winder and Gough (2010) to explore CSEM users’ identity and boundary work, patterns of “self-distancing,” how these are explained, and how they might function for these individuals. In utilizing two studies about CSEM users through cross-national comparison, the present paper demonstrates commonalities across different settings, countries, timeframes, intervention programs, and conviction statuses.
Participants’ identity work had common features including distinguishing between acts vs identities (or what one did vs who one was); separating crimes from identities; comparing offenses to other crimes and people viewed as worse; framing offending as influenced by childhood experiences of abuse and sexual development; and situating both offending and post-offending identities within larger society by decrying societal and childhood sexualization, stigmatization and ongoing punishment, and societal focus on retribution over prevention. In unpacking these themes and patterns, we discuss multiple perspectives about functions and implications, all of which may be pertinent but to different CSEM users.
Interpreting CSEM Users’ Identity Work and Its Multiple Implications
Participants’ identity work could be interpreted through multiple, sometimes incongruent, lenses. In many ways, much of the identity work by CSEM user participants can be considered as forms of denial, downplay, minimization, rationalization, distortion, and/or attempts at justification. Identity work among those who commit CSEM offenses has been framed in these ways before (Winder & Gough, 2010; Winder et al., 2015). Statements made by participants can be understood as types of denial that exist on a continuum and can include minimizing impacts on and harms to victims, externalizing blame, and/or refusing to acknowledge offending or sexually abusive behaviors (Jenkins, 1990; Marshall et al., 1999; Schneider & Wright, 2004; Trepper & Barrett, 1989; Yates, 2009). CSEM users who are denying, downplaying, rationalizing, minimizing, or attempting to justify their offending could therefore potentially reach harmful and damaging conclusions including that what they have done is not wrong; that they do not need to change; that reoffending would be acceptable; and/or that they are not accountable for harms including ongoing feelings of powerlessness, violation, trauma, revictimization, and lack of closure for victims who know their images are online and will potentially be forever viewed by others (Gewirtz-Meydan et al., 2018; Martin, 2015; Palmer, 2005). These conclusions would be especially damaging given the further serious and sometimes lifelong impacts and consequences of child sexual abuse and CSEM discussed in the Literature Review. If continuing to deny harms of CSEM, participants are then invalidating the experiences of those affected, while also potentially being more open to reoffending (for an article focused specifically on Study 1 participants’ perceptions of harm, CSEM as victimless, and victim empathy, see Rimer, 2019). This may be why, when considering implications, Winder and Gough (2010) assert that “'self-distancing” is amenable to challenges and reconsiderations within treatment, which is indicative of an acceptance of responsibility.
From a different analytical perspective, for some CSEM user participants, identity work may serve as a way to make sense of what they have done, who they were, and what they will be (or want to be). Fundamental shifts in identity are critical to desistance from crime generally (Digard, 2014; Liem & Richardson, 2014; Rocque et al., 2016; Soyer, 2014) and may serve a similar role for CSEM use. Such identity work could potentially be transformative in terms of changing future behavior. For example, some literature on risk and recidivism has argued that denying, excuse making, and minimizing are processes that all human beings undertake when trying to maintain self-esteem and a positive self-image (Maruna & Mann, 2006; Yates, 2009). Similarly, denial might be a sign that some individuals are aware of the harm, wrongfulness, and transgression of their sexual crimes, and therefore they do not want to admit to offending (Harkins et al., 2010; Yates, 2009). In such cases, denial may serve important functions such as “making sense” through narrative, relational, and self-constructive processes (Blagden et al., 2014). For some, this may offer a way to make sense of past behaviors and move toward a post-offense identity. In this sense, drawing symbolic boundaries could potentially be indicators of a desire for change among some CSEM users. Denial can tell us much about an individual's view of the world and themselves, providing rich information that is not necessarily an obstacle to treatment (Schneider & Wright, 2004).
Many treatment programs require acceptance of responsibility as a condition for admission and successful completion (McGrath et al., 2010). Treatment goals often include the acceptance of deviance or criminality, admission of crimes, and the notion of taking responsibility, therefore accepting consequences that follow and being the “bigger person” (Levenson, 2011, p. 348). Yet, some studies have also shown that neither denial nor admission of guilt are reliable predictors of risk or recidivism (Hanson & Bussière, 1998; Hanson & Morton-Bourgon, 2005; Harkins et al., 2015), or that they predict risk and recidivism only for specific types/groups of individuals who commit sexual offenses (Harkins et al., 2010; Nunes et al., 2007). Thus, programs (e.g., treatment), practices (e.g., risk assessments), and policies (e.g., requirements of treatment success) that center admission of guilt or acceptance of responsibility may be targeting the wrong areas for some individuals, or at worst, may hinder the identity work necessary to turn away from crime. For some individuals, perhaps it may be more beneficial for treatment success that they have a sense of awareness about the psychological and situational factors which led to the offending and focus on responsibility for future actions (Ware & Mann, 2012). Overall, while denial/downplay/distortion/minimization/justification has been examined extensively in people who have committed contact sexual crimes, the role it plays in CSEM users’ identity work is less understood. Therefore, we suggest that nuance is needed to discern CSEM user identity work specifically, including how it may function differently for different people.
Also regarding behavior change, there has been debate over the function of CSEM and whether users have a sexual preference for children. In our research, a minority of participants did describe a sexual preference for children, while a majority denied a sexual attraction to children and separated this from their core identity. Speaking to this, while there is an undeniable sexual factor to CSEM use, some scholars argue there is a need to distinguish between a sexual preference versus an interest. Some CSEM users may have a primary sexual preference for children (Seto et al., 2006), and for individuals with such a preference, medical researchers have also proposed exclusive and non-exclusive forms, with some individuals indicating an interest in children as well as people of other ages, and others having a sole sexual preference for children (Hall & Hall, 2007). However, research also supports the idea that others accessing CSEM are not necessarily primarily attracted to children, or that viewing the material triggers a sexual interest not realized beforehand where “over time the individual may become increasingly interested in child pornography, become attracted to images of increasing severity, and become desensitized to the harms that victims experience” (Wortley, 2012, p. 193). Related, the Internet provides an opportunity for people to look at abusive material in private, perhaps to indulge “curiosity,” where they may not have taken the risk beforehand (Jewkes & Sharp, 2003; Rimer, 2017; Seto, 2013). In addition, research has also described some CSEM user collections as indiscriminate whereby they have wide-ranging pornographic and CSEM interests (Elliott & Beech, 2009; Krone, 2004; Wolak et al., 2005b), a finding we also report in a previous publication where some participants described collected/viewing sexualized material of many different kinds (Rimer & Holt, 2023). The key point here is that an interest can perhaps develop and change, and Internet use can be instrumental in this process for some CSEM users. As such, treatment providers could consider that some individuals who deny a sexual preference for children or challenge the idea that it is part of their core identity may not be doing so out of deceit, but may instead be reflecting an interest instead of a preference, therefore also highlighting the need for an individualized approach. However, it must be acknowledged that treatment providers may find it difficult to differentiate between sincerely held beliefs vs attitudes that support sexual offending behaviors, and in some cases, both may be present (Blagden et al., 2020).
Finally, beyond the individual functions of, implications of, and treatment considerations for CSEM user identity work, it is important to also consider how broader systemic criminal justice policy and practice intersects with this same identity work. Previous research has found that creating a positive identity can increase protective factors (de Vries Robbé et al., 2015; Perrin et al., 2018) and that maintaining identification with a social group or network is associated with lower risk of recidivism (Farmer et al., 2012). At the same time, researchers have argued that criminal justice policies/practices focused on shame and punitiveness may have the opposite effect. For example, sexual offender registries that are made public can potentially drive these individuals further underground (Lieb et al., 2011; Long, 2009; West, 2000) and make them less willing to register (Brown et al., 2008; Long, 2009). Many researchers have suggested that public notification does little to aid community protective behavior, but conversely, may increase recidivism by hindering post-release reintegration through stigmatization and isolation (Brown et al., 2008; Burchfield & Mingus, 2014; Levenson et al., 2007; Long, 2009; McAlinden, 2007; Petrunik & Deutschmann, 2008; Seto, 2013; Terry, 2011; Tewksbury & Lees, 2006; Vess, 2009; Waldram, 2012; Ward & Salmon, 2011; West, 2000). Some participants reflected such points when discussing their belief that they are punished in an ongoing way, and that society is more concerned with retribution than prevention or rehabilitation.
Evident in these debates, and regardless of the motivations or functions behind participants’ identity work, there appears to be an inherent conflict between some of the identity work expected by treatment programs and criminal justice professionals versus that which responds to social/societal reactions to offending. In their situating of offending and post-offending identities within larger societal norms, some CSEM user participants suggested that identity work occurs within competing and contrasting personal, judicial, treatment, media, and societal reactions to and expectations of individuals who have committed sexual offenses. Put another way, the identity work required by treatment, psychoeducational, and parole programs (e.g., accepting one's risk, admitting to crimes, admitting deviance, ongoing disclosure) can be incompatible with the identity work CSEM users may do in response to societal reactions to sexual offending and which are found in this article (e.g., claiming lack of deviance, lack of risk, that one is a good person regardless of actions, that one's offending could have been worse, that one has the ability to be “normal,” trying to demonstrate one is not “that bad” or a “monster”). This tension may help explain some of the patterns presented in this paper and why participants represented themselves in the ways they did.
Future Directions and the Value of Interdisciplinary Comparative Qualitative Research
Beyond results, this article also makes a methodological contribution by demonstrating the value of comparative cross-national qualitative research, particularly with understudied and hard-to-reach populations such as CSEM users. While one may suggest that comparing samples from two studies is problematic because they are different, we argue the opposite: the research is powerful because of the differences. Despite differences in countries (the United Kingdom and the United States), settings (community and prison), conviction statuses/timeframes (pretrial to incarceration), interventions (community group and prison treatment programs), and research methods (ethnography and interviews in Study 1 and interviews in Study 2), we still found striking similarities in participants’ identity work, narratives, and framing of boundaries. Unconventional thematic analysis brought forth findings and discussions not possible in a siloed, one-sample piece of work: our research demonstrates consistency across a larger group of people than would be possible in one study, and moreover, across a more diverse group. The fact that participants and their identity work have some consistency provides insight into offending through this consistency. The method therefore advances knowledge of offending, as the more that is known about consistency across populations, the closer researchers can be to broader understanding.
This paper and our previous article that uses the same method (Rimer & Holt, 2023) outline a roadmap through which qualitative thematic analysis can be done comparatively using a mixed inductive/deductive approach. Future research on CSEM offending using such a method could explore different areas not covered here. As noted, participants described competing and contrasting identity work required by different people, systems, and parts of their lives. Future comparative research could aim to differentiate and disentangle forms of CSEM user narratives and identity work to try to better understand if particular aspects relate to particular expectations/demands, from which areas, and subsequently, which may be most amenable for desistance (e.g., by targeting in treatment). A second area of future research could focus on cultural context. While caution is needed when attempting to generalize national “cultures,” there nevertheless may be differences in sexual offending behavior, justice system responses, and/or treatment modalities between countries (or regions within countries) that could impact and be impacted by CSEM users’ identity work. Future research could compare cross-national samples while emphasizing the impact of cultural context on narratives, identity work, and symbolic and social boundaries for individuals who have sexually offended online.
Conclusion
To prevent and respond to CSEM offending, we must study the narratives and perceptions of individuals who have engaged in these offenses. The current paper examined symbolic and social boundaries that these individuals have transgressed, subsequent identity work, and the ways in which this involves the negotiation of self and self-presentation. This process occurs within often competing personal, judicial, treatment, media, and societal reactions. Regardless of settings, countries, timeframes, interventions, and conviction statuses, we found commonalities in the stories and identity work of CSEM user participants as they attempted to situate their offenses within their identities. For some, these commonalities may represent attempts to deny, downplay, minimize, rationalize, or justify their harmful behavior, while for others, they may be an attempt to negotiate a post-offending identity away from crime. We have presented multiple and at times incompatible perspectives to consider for studying this type of offending, toward the ultimate goal of reduced victimization and rates of child sexual abuse. As most criminological scholarship on these topics has focused on offline sexual crimes or non-sexual offenses, understanding how CSEM users negotiate identity and symbolic and social boundaries is an important step in expanding this consequential research area.
Footnotes
Acknowledgments
The authors thank the anonymous reviewers and editor for their helpful feedback during the review process. The authors would also like to thank all participants for taking part in the studies and Tom Holt and Sarah-Ann Burger for commenting on previous versions of this article.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Study 1 was supported by the Commonwealth Scholarship Commission in the United Kingdom (Commonwealth Scholarship to Jonah R. Rimer), the Social Sciences and Humanities Research Council of Canada (Doctoral Fellowship to Jonah R. Rimer), and the Royal Anthropological Institute (Sutasoma Award to Jonah R. Rimer). Study 2 was supported by the Michigan State University School of Criminal Justice (Start-Up Research Funds to Karen Holt). Funders were not involved in the research or manuscript preparation.
