Crossing the Line: Factors Associated With Escalating Pornography Use

Measuring Escalation

Pornography research has long struggled to establish a consistent taxonomy capable of classifying the vast heterogeneity of available material (Ballester-Arnal et al., 2022; Hald & Štulhofer, 2016). Classification approaches vary considerably, from broad distinctions between normophilic (e.g. vaginal or oral sex) and paraphilic content (e.g. violent, coercive sex, fetishistic; Ballester-Arnal et al., 2022) to groupings based on user demographics (Hald & Štulhofer, 2016). Although some favour granular operational classifications focusing on content complexity (e.g. performer characteristics, sexual vs. non-sexual behaviours and setting; Kohut et al., 2024), a thematic framework based on explicit content categories consumed is often more practical for studying escalation. Focusing on themes prevalent across major content platforms provides a direct means of understanding how users navigate and progress through readily available, non-illegal material towards illegal CSAM. Therefore, a clearer empirical framework for categorising and analysing how content types co-occur in user consumption patterns is essential.

Existing literature often simplifies escalation by focusing solely on viewing frequency, implicitly assuming individuals only consume what they find highly arousing. However, emerging evidence challenges this assumption, revealing a critical disconnect between arousal and viewing behaviour. Paul (2009), for instance, found that the most frequently viewed content is not necessarily the most arousing, particularly as content becomes more extreme. For example, men ranked “barely legal” pornography among their five most arousing categories, yet it did not appear among their five most frequently viewed categories (Paul, 2009). Conversely, women reported lower arousal towards content depicting ejaculations but still included it among their most frequently viewed categories (Paul, 2009). These findings suggest that desensitisation, which drives the search for novelty and extremity, compels users to engage with content despite diminished arousal. To accurately capture the desensitisation mechanism underpinning escalation towards illegal CSAM, research must integrate frequency, content type and subjective arousal when mapping consumption patterns. These findings are consistent with desensitisation-based accounts of content escalation (Knack et al., 2020), in which repeated exposure reduces arousal over time (Koukounas & Over, 1993; O’Donohue & Geer, 1985), resulting in a decoupling between subjective arousal and consumption behaviour. As such, continued engagement with less arousing content provides a theoretically grounded indicator of escalation-related processes.

Psychosocial Predictors of Escalation

Increased pornography use may precipitate content escalation, a process that fundamentally resembles tolerance, defined by the need for novel or more sexual stimuli to achieve the same level of arousal (Lewczuk et al., 2022; O’Donohue & Geer, 1985; O’Donohue & Plaud, 1991). This tolerance towards mainstream material can lead users to shift from normophilic to paraphilic content in pursuit of heightened stimulation (Krahé et al., 2011). Such progression towards extreme material is critical for understanding the consumption of illegal content such as CSAM, suggesting that some individuals consume CSAM primarily for its extremity, rather than due to a specific sexual interest in children (Merdian et al., 2018; Quayle & Taylor, 2003); a distinction with substantial implications for intervention and prevention strategies.

Sexual sensation-seeking (SSS), defined as the pursuit of novel or risky sexual stimuli (Kalichman et al., 1994) has been identified as a key psychological factor underlying this escalation process. The central hypothesis is that individuals high in SSS are more likely to seek extreme material to satisfy their heightened desire for sexual intensity (Chen et al., 2018; Sinković et al., 2013), a drive that mirrors tolerance-driven desensitisation (Bocci Benucci et al., 2024). Although findings on the relationship between SSS and pornography use are mixed (Esplin et al., 2024; Sinković et al., 2013), sensation-seeking tendencies align closely with mechanisms of desensitisation and novelty-seeking that fuel content escalation. For example, increased viewing frequency was found to predict a shift towards more hardcore content and greater openness to paraphilic and fetish content (e.g. rape or hebephilic themes; Blinka et al., 2022). Collectively, these findings indicate that SSS contributes not merely to the frequency of pornography use, but to the qualitative nature of the material sought, making it a key predictor in the escalation pathway.

A second key predictor of escalation is Problematic pornography use (PPU), defined as a loss of control over consumption and continued use despite adverse consequences (Gibbons et al., 2020). PPU is intrinsically linked to the consumption of increasingly extreme material, with genre escalation identified as a relevant dimension of PPU severity (Ince et al., 2024; Lewczuk et al., 2021). These findings suggest that the behavioural pattern of PPU drives individuals to seek progressively more diverse and extreme content. Furthermore, PPU appears to be closely related to SSS, indicating a potential interactive pathway to escalation. Sensation-seeking tendencies are a recognised vulnerability factor for PPU (Hegbe et al., 2021), and several studies report positive associations between the two constructs (Bőthe et al., 2019). This relationship implies that the drive for novelty and heightened stimulation (SSS) may contribute to the compulsive, uncontrolled consumption that defines PPU. Together, these processes likely operate synergistically, with SSS amplifying the motivational drive for novelty, and PPU reinforcing the behavioural pattern of compulsive use, thereby accelerating the progression toward increasingly extreme and, in some cases, illegal material.

The third factor relevant to the development of problematic escalation patterns is the endorsement of myths about child sexual abuse (CSA). These myths function as offence-supportive cognitions, enabling individuals to rationalise and normalise abusive behaviour (Collings, 1997). Categories of these myths include the minimisation of harm, denial of the extent of CSA, diffusion of perpetrator responsibility, and reliance on perpetrator stereotypes (Cromer & Goldsmith, 2010). Endorsing such beliefs reduces cognitive dissonance and facilitates the perception of children as sexual beings, thereby normalising the consumption of related content. These cognitive processes are evident in pornography consumption, as individuals who endorse CSA myths are more likely to escalate their viewing patterns toward extreme material (Osbourne & Christensen, 2020). This mechanism parallels the influence of rape myths on the consumption of violent pornography (Bhuptani et al., 2024). Critically, at the extreme end of the escalation spectrum, these beliefs provide internal justification to consume illegal material. Complementing these cognitive rationalisations, the perceived or actual acceptance of CSAM viewing within one’s peer group may further reduce inhibition and reinforce the likelihood of consuming illegal material (Schonert-Reichl, 1999).

The Present Study

The present study addresses a gap in understanding early stages of escalation-related processes from mainstream pornography consumption towards extreme or illegal content. Previous research has often operationalised escalation merely as a symptom of PPU (Ince et al., 2024). Our approach focuses on thematic escalation and treats this as an independent phenomenon. Moving beyond frequency metric, we investigate how escalation clusters across different thematic content types, conceptualising escalation a behavioural pattern characterised by the continued viewing of content despite low or diminished arousal (Paul, 2009), consistent with desensitisation-based accounts of habituation to sexual stimuli (Knack et al., 2020). Importantly, this approach does not assume a single underlying motivation; rather, it captures a pattern of engagement that may arise from multiple pathways (e.g. curiosity, openness to experience), but which nevertheless reflects the behavioural conditions under which desensitisation and progression across content types may occur. We note that a definitive test of escalation would require longitudinal data to examine changes over time; our cross-sectional design allows us to characterise patterns of content engagement that could reflect early stages of escalation-related processes.

Additionally, systematically examine the psychosocial correlates of these distinct escalation patterns within a non-clinical population. Although factors like SSS, PPU and the endorsement of CSA myths have been linked to general pornography consumption, their specific relationships with content escalation patterns in the general population remains unclear. This focus enables generalisation beyond clinical populations.

To this end, our research questions are:

Participants

Participants were recruited from the university research participant platform and through social media channels. Those who participated through the university received 0.5 research credits for their participation. Our inclusion criteria were men between the ages of 18 and 35 who viewed pornography. Initially, 395 participants clicked the link to the survey. Nineteen participants were removed for falling outside the age range and 60 for indicating their gender as anything other than man, and three for not providing informed consent. A further 85 completed less than 75% of the survey and were deleted, resulting in a final sample of N = 228 (M_age = 22.65, SD_age = 3.16). This sample comprised predominantly single (54%), heterosexual (82%), students (61%) who identified as white (76%). These three percentages are based on the participants who answered these questions, but each question had 19% missing data. This likely occurred because participants were not clearly prompted to complete the final page of demographic questions; all percentages for these items are reported as valid percentages among respondents who provided data.

Materials

Procedure

This study received ethical approval from the university institutional review board. Participants were directed to the study in Qualtrics through an anonymous link. After providing informed consent, participants were presented with the questionnaires in randomised order. They had the option to skip any questions they felt not comfortable answering and could take breaks whilst completing the survey. Finally, they were asked demographic questions and presented with a debrief form that included contact information for various support organisations for people concerned about their pornography use.

Analyses

To operationalise escalation within each pornography category, we created an index that reflects the extent to which participants reported viewing a category despite not finding it highly arousing. This was done by multiplying the reported frequency of viewing the category (0 = never, 4 = all the time) by a reverse-coded arousal score (4 – arousal rating, where 0 = not at all arousing, 4 = very arousing). The resulting escalation score therefore increases when participants indicate relatively high use of a category that they rate as less arousing. Conceptually, this captures the degree of consumption that extends beyond an individual’s stated sexual interests, which we interpret as indicative of escalation. Scores range from 0 to 16 with higher scores indicating greater escalation. This operationalisation is grounded in the premise that escalation is reflected not only in exposure to different content types, but in continued engagement in the absence of strong subjective arousal, which is consistent with desensitisation processes. Although such patterns may also reflect individual differences in curiosity, openness, or exploratory tendencies, repeated exposure to content that is not highly arousing may nevertheless contribute to habituation over time. Further, while our cross-sectional design does not measure escalation longitudinally, this index provides a meaningful behavioural indicator of engagement with less arousing content, which may reflect early-stage escalation process.

We examined frequencies of arousal, viewing and escalation per pornography category and descriptive statistics of all predictor variables. Following this, we conducted a principal components factor analysis to examine how escalation clusters across content. Based on the results, five factors were identified and participants’ individual scores for each factor were computed. These scores were then used in a series of multiple regressions to examine whether escalation in that factor could be predicted from the independent variables in our study. All predictors were assessed for multicollinearity. Variance inflation factor (VIF) values ranged from 1.29 to 3.42, and Tolerance values ranged from 0.29 to 0.78, indicating no problematic multicollinearity across analyses.

Patterns of Escalation

Consistent with prior research (Ballester-Arnal et al., 2022), mainstream categories (e.g. vaginal and oral sex) demonstrated the highest absolute viewing frequencies and mean escalation scores, reflecting the widespread accessibility and consumption of these forms of content. However, when considering relative content escalation, or the proportion of escalation relative to the low baseline arousal and viewing of the category, more extreme and niche categories such as S&M/Bondage, fetish, teen, group sex, and violent sex showed disproportionately greater content escalation. This pattern highlights the coexistence of two distinct forms of two qualitatively distinct escalation processes: one characterised by high-frequency engagement with mainstream content, driven primarily by availability and habitual use, and another reflecting a proportional, high-severity shift toward extreme or non-traditional material, representing a meaningful qualitative change in content interest.

Importantly, even though the absolute engagement with illegal categories (e.g. Sex with children, Snuff, Sex with animals) was minimal, a subset of participants repeatedly sought out these materials despite negative emotional reactions, suggesting there may be a low-frequency, high-risk pathway with significant public health and safety implications. Together, these findings emphasise that content escalation is multidimensional and cannot be fully understood through consumption volume alone; proportional engagement with extreme content captures a critical dimension of risk and potential harm.

Our factor analysis revealed five clusters representing distinct content escalation pathways, further emphasising the heterogeneity of content escalation. The Dominance/Degradation factor captures escalation towards content involving submission and hierarchical sexual scripts, reflecting novelty-driven escalation among those habituated to mainstream content. Conversely, the Extreme/Harmful factor encompasses ethically and legally concerning content, though escalation here was not linked with CSAM involvement. The Normative cluster appears to reflect general habituation, characterised by repeated exposure to high-volume mainstream content (Hald et al., 2014). The Non-Traditional Niche and Atypical clusters suggest branching, specialised content escalation pathways, capturing engagement with niche, paraphilic, or idiosyncratic content rather than generalised sexual compulsivity. Notably, inclusion of Gay Male content in the Non-Traditional Niche factor reflects statistical rarity in a predominantly heterosexual male sample, rather than any deviance, underscoring the need to interpret clusters within demographic context. Little is known about viewing content outside one’s sexual orientation, but this might be an important avenue for future research.

Overall, these clusters suggest that content escalation is not strictly linear, but multidimensional. Although a continuum of severity exists, from mainstream to Dominance/Degradation and Extreme/Harmful, Non-Traditional Niche and Atypical patterns highlight that not all content escalation follows the more widely studied routes of dominance or direct harm, echoing research emphasising the diversity of sexual interests and the complex interplay between novelty, taboo, and individual differences in arousal (Seto, 2012). Such findings reinforce the need for nuanced frameworks that move beyond linear models of escalation and recognise the multidimensionality of sexual content engagement.

The factor structure also aligns with prior research indicating that content characteristics beyond simple extremity shape engagement (Kohut et al., 2024; Sharpe & Mead, 2021). The positive association between the Non-Traditional Niche and the Extreme/Harmful clusters suggest a potentially generalised drive towards non-normative content, where individuals with an interest in one set of unconventional content are also susceptible to an interest in the most severe material (Hald & Štulhofer, 2016). Conversely, the negative relationship between the Non-Traditional Niche and Atypical Escalation might indicate distinct psychological mechanisms underlying these two forms of non-mainstream consumption, but research into such content remains limited (Park et al., 2022). An intense interest in a specific, specialised niche appears to be inversely related to the characteristics that define the more generalised Atypical pattern, reinforcing that interest can branch into specialised, independent routes (Paul, 2009)

Psychosocial Mechanisms of Escalation

Our analyses highlight cluster-specific psychosocial factors, suggesting that distinct mechanisms drive different forms of content escalation. Dominance/Degradation escalation was associated with SSS and tolerance, indicating that individuals motivated by novel sexual experiences and habituated to mainstream content may be more likely to engage with power- or submission-themed pornography, even when arousal is low. This aligns with prior research linking SSS to arousal for degradation and humiliation themes (Brown et al., 2023). These factors were specific to this content type, highlighting that content escalation in the Dominance/Degradation cluster is distinct from other clusters. Interestingly, higher endorsement of CSA myths was associated with lower likelihood of content escalation in this factor, suggesting that power-dynamics content may operate independently of offence-supportive cognitions.

In contrast, Extreme/Harmful content escalation was minimally associated with the measured variables, with relapse emerging as a negative predictor. This pattern may reflect that individuals escalating in highly extreme content may not perceive their consumption as problematic, and thus are less likely to attempt to reduce consumption. Despite correlations between CSA myths, peer acceptance, and Extreme/Harmful content escalation, these effects did not reach significance in regression analyses, suggesting overlapping variance or limited power. The weak predictive power of this model implies that other factors, such as intrinsic motivation, broader sensation- or thrill-seeking, may drive escalation towards extreme material.

Atypical content escalation, associated with relapse, might capture compulsive or difficult-to-control use, particularly towards idiosyncratic or niche content, consistent with prior research (Camilleri et al., 2021). Conversely, content escalation in the Normative and Non-Traditional Niche clusters was largely independent of measured psychosocial variables, highlighting the potential role of situational factors, such as ease of access, repeated exposure, or individual preference, rather than stable traits.

These results collectively illustrate that content escalation is a multidimensional process with multiple interacting mechanisms: novelty and habituation drive some forms, cognitive rationalisations others, and situational factors yet others, which aligns with prior research differentiating between individual differences and situational factors (Bocci Benucci et al., 2024; Hald et al., 2014). Understanding content escalation thus requires moving beyond one-size-fits-all models and considering the distinct psychological drivers of different consumption patterns.

Non-Traditional Niche Escalation and CSAM Risk

The strongest link to CSAM engagement emerged in the Non-Traditional Niche content escalation factor, which primarily loads on Hentai, sexualised non-human content, and Gay Male content. As previously discussed, in this predominantly heterosexual male sample, the inclusion of Gay Male content reflects statistically less common consumption rather than deviance per se, highlighting the specialised and non-mainstream nature of this cluster. This factor exhibited consistent correlations with participants’ self-reported behaviours, including searching for, coming across, and watching illegal material. These results suggest that consumption of highly specialised, unconventional, or fantastical content may serve as a more sensitive indicator of potential CSAM engagement than pursuit of aggressive or extreme material alone (Knack et al., 2020; Quayle & Taylor, 2003; Seigfried-Spellar & Rogers, 2013). This pattern is consistent with prior research demonstrating frequent associations between CSAM consumption and engagement with other forms of deviant or non-traditional pornography (Endrass et al., 2009; Seigfried-Spellar & Rogers, 2013). Such patterns indicate that individuals engaging with CSAM may exhibit broader proclivities toward non-traditional sexual content, reflecting fluidity across sexual interests.

One plausible mechanism underlying the observed association is a boundary‑crossing psychological drive. Individuals on the Non‑Traditional Niche pathway may be attracted to sexual content that transcends conventional social and biological scripts, such as stylised Hentai, non‑human imagery, or content featuring ambiguous age‑cues or non‑consensual scenarios, making them more inclined to explore illegal material than individuals following more generalised content escalation patterns, such as Dominance/Degradation (Steel et al., 2021). Research shows that consumers of Hentai exhibit distinct affective and cognitive patterns compared with typical pornography users (Park et al., 2022). Moreover, theoretical and empirical work indicates that Hentai may act as a gateway medium, normalising sexual interest in child‑coded characters or unconventional sexual objects and thereby lowering barriers to illegal consumption (Dines & Sanchez, 2023). A further possibility is that engagement with stylised or cartoonised depictions of minors or non‑human sexual objects may represent an early‑stage paraphilic trajectory, in which content escalation follows specialised rather than purely linear patterns (Christensen & Vickery, 2023). This mechanism aligns with documented overlap between users of CSAM and wider collections of non‑traditional erotic content (Endrass et al., 2009).

Other clusters provide further nuance. Dominance/Degradation content escalation showed a modest link only to searches for teen content, consistent with its focus on power and aggression rather than explicit illegality. This could reflect the perception that younger individuals are easier to dominate than adults (Wall, 2025). Importantly, this pattern may not reflect an interest in children or the taboo/illegal, as may be captured instead by the Extreme/Harmful content escalation, but rather could represent an extension of interest in dominance and submission within a sexual context. This suggests that individuals showing content escalation in this cluster may explore boundary-pushing legal-adult or near-minor content without necessarily engaging in illegal material. In contrast, the Extreme/Harmful content escalation, despite including sexualised depictions of children, was not significantly associated with CSAM engagement. Whilst this finding seems counterintuitive, this finding relates specifically to content escalation (viewing more than arousal warrants), not viewing in general. Individuals may engage with content in this cluster in proportion to their arousal, meaning content escalation scores remain low even when exposure occurs.

Overall, these findings challenge linear models assuming progression toward extreme content predicts CSAM use. Instead, they highlight Non-Traditional Niche content escalation, characterised by highly specialised, boundary-crossing, and non-mainstream consumption, as the clearest empirically observed correlate of CSAM engagement in the present data. These results suggest that research, policy, and prevention initiatives may benefit from considering the psychological and behavioural mechanisms driving specialised content escalation, rather than relying solely on exposure to extreme or aggressive content as an indicator of risk.

Implications

The present findings contribute to the conceptualisation of pornography content escalation as multidimensional, with qualitatively distinct mechanisms. Whilst further research is needed, interventions and risk assessments might consider specialised content escalation patterns as a potential, early warning sign of content escalation to CSAM. The findings also suggest that frequency-based models of pornography use may be insufficient on their own: arousal-independent content escalation and the qualitative nature of content may be important for understanding risk trajectories. Future research should employ longitudinal and multimodal methodologies to explore causal mechanisms and situational influences. Examining motivational and inhibitory processes will be key to distinguishing between general compulsive use and specialised content escalation. Additionally, the role of illegality itself as a motivational factor warrants investigation, as some individuals may pursue content primarily for thrill or boundary-crossing rather than content themes (Steely et al., 2018). Finally, research must include more female and gender-diverse samples in pornography and CSAM research to clarify potential differences in escalation patterns.

Limitations and Future Directions

This study relied on cross-sectional, self-report data, which may be subject to recall bias and social desirability effects, particularly regarding CSAM engagement. Low base rates of illegal content engagement limit statistical power and generalisability. Future studies should incorporate longitudinal designs, experimental paradigms, and objective measures of content engagement. Our sample size was modest, particularly for a principal components analysis and multiple regression. Future studies should include larger samples to examine if the factor structure we identified, and the predictors can be replicated. Additionally, the complex interplay between social, cognitive and situational factors warrants further investigation, particularly regarding peer influences, cognitive rationalisations, and motivational mechanisms.

Conclusion

In summary, pornography content escalation may be best understood as a multidimensional, non-linear phenomenon. Distinct content escalation clusters appear to reflect different psychological mechanisms, from novelty-seeking and habituation to niche specialisation. Non-Traditional Niche content escalation was associated with CSAM engagement in the present data, providing preliminary support for models that emphasise specialised, boundary-crossing consumption as a possible pathway. These findings offer initial theoretical insights, and directions for future research, but should be interpreted with caution given the cross-sectional design.