Empathy and Offending,a Study About the Role of Cognitive and Affective Empathy in Sexual Violence

Hypotheses

Our hypothesis was that individuals who have committed sexual violence have a higher degree of affective empathy than individuals with conviction for violent crimes and that both offending groups have similar levels of cognitive empathy. According to the motivation-facilitation model (Seto, 2019), sexual offending arises from the interaction of two primary components: motivating factors and facilitating factors. Motivating factors include enduring traits or predispositions that increase the likelihood of sexual offending, whereas facilitating factors increase the likelihood that a motivation is acted upon. In other words, motivation is a necessary but not sufficient condition for sexual offending because people vary in their willingness to act on those motivations (Seto, 2019). Empathy can inhibit or facilitate antisocial behavior depending on the offender’s motivational state and emotional regulation. Individuals who engage in grooming or manipulative behaviors in sexual offending may exhibit higher levels of affective empathy (Morrow, 2019), allowing them to emotionally attune to their victims to gain trust or compliance. However, this affective resonance may not translate into moral inhibition due to distorted cognitive schemas or emotional dysregulation. In contrast, violent offenders often display emotional impulsivity and poor affective empathy, reacting to perceived threats or provocations with little emotional attunement (Romero-Martínez et al., 2022). Both groups, however, may share similar levels of cognitive empathy—they may have a similar intellectual awareness of others’ emotions or intentions, but this awareness does not necessarily translate into prosocial behavior. Finally, we anticipated that individuals from the general population would score higher on both cognitive and affective empathy. When viewed through the lens of the motivation-facilitation model, high levels of empathy enhance internal barriers to offending. People without offense histories may also be more attuned to others’ emotions and more likely to engage in behaviors that reflect understanding and compassion, which might be less impaired by factors such as trauma or antisocial tendencies.

In the current study, we added a measure of impulsivity as a covariate as it is considered a facilitating factor according to the motivation-facilitation model (Seto, 2019). Impulsivity, characterized by actions without consideration of consequences, has been linked to deficits in empathy and difficulties in social interactions (Romero-Martínez et al., 2022). For example, impulsive individuals may act on immediate desires without considering the impact on others, leading to a disregard for others’ feelings and a failure to recognize the harm caused by their actions (Balcioglu et al., 2023). Relatedly, individuals with high levels of impulsivity often struggle with emotional regulation (Garofalo et al., 2018), which can subsequently impair their ability to empathize with others. For instance, individuals who exhibit impulsive behaviors may find it challenging to take the perspective of others or to be emotionally affected by others’ experiences.

Description of Participants

All participants were male. The participants from the offender groups were admitted at high secure psychiatric hospital de Rooyse Wissel in Venray, the Netherlands. These participants have committed either no offenses (N = 91, participants from the general population), a sexual (N = 33) or a violent offense (N = 30) in the past for which they were punished by the legal system. Sexual crimes consisted of sexual violence against children and sexual violence against adults (i.e., rape). No participants had been convicted for mixed offenses. In addition to their punishment, they were all imposed to compulsory treatment (i.e., treatment on behalf of the state, “TBS”). This treatment order was imposed because all participants have psychological problems that (partially) led to (a) serious (sexual) offense(s) (i.e., carrying a prison sentence of a minimum of 4 years) and because there was a high risk for reoffending.

The average age across participants was 45.3 years old (SD = 11.74, range 18–72). The Kruskal-Wallis test shows that there was no significant age difference between these groups (χ² [4] = 4.925, p = .295). There was no significant difference between the levels of education between the three groups (χ² [12] = 17.501, p = .132). For each group, 30 to 50% of the participants had completed either primary or secondary school.

Of the individuals who committed sexual offenses, 12.1% (N = 4) had an anxiety disorder, 3% (N = 1) a psychotic disorder, 6.1% (N = 2) had been diagnosed with ADHD, 18.2% (N = 6) had an autism spectrum disorder, and 30.3% (N = 10) had a substance related disorder. In addition, 66.7% (N = 4) had been diagnosed with a paraphilic disorder, 9.1% (N = 3) with antisocial personality disorder, 3% (N = 1) with borderline personality disorder, 6.1% (N = 2) with narcissistic personality disorder and 18.2% (N = 6) with a non-specific personality disorder. As for the individuals who committed violent crimes, 6.7% (N = 2) had a diagnosis of mood disorder, 3.3% (N = 1) an anxiety disorder and 46.7% (N = 14) a psychotic disorder. 10% (N = 3) had ADHD and 23.3% (N = 7) an autism spectrum disorder. 73.3% (N = 22) had a substance related disorders, 3.3% (N = 1) had been diagnosed with a paraphilic disorder, 20% (N = 6) with antisocial personality disorder, 13.3% (N = 4) with borderline personality disorder and 10% (N = 3) with narcissistic personality disorder. Finally, 46.7% (N = 14) were diagnosed with a personality disorder not otherwise defined.

Individuals who committed sexual offenses had fewer psychotic disorders (χ² [3] = 15.143, p = .002), and substance related disorders (χ² [3] = 13.153, p = .004). They had more often paraphilic disorders (χ² [3] = 37.293, p < .001). No diagnostic data on possible psychiatric disorders were available from the general population.

Instruments

The data necessary to answer the research questions were collected in various ways. Reports from court provide information about the type of crime participants committed. Next, participants filled out two questionnaires.

Design and Procedure

The design of this study was cross-sectional. In terms of sexual violence, only hands-on sexual offending was included in this study as hands-off sexual crimes (e.g., distributing child sexual exploitation materials) are less prevalent in patients who are admitted to a forensic hospital (Elbert et al., 2022).

After approval was obtained, we determined who was eligible to participate in the study by consultation with the psychologist who was coordinating the treatment of the patients. This consultation ensured that only those patients who met the in- and exclusion criteria were considered. In this study, exclusion criteria were an intellectual disability (full scale IQ < 70), active psychotic symptoms, or being unable to read, speak of write the Dutch language. The potential participants were invited by the research team to participate in the study. Each potential participant received both an oral and written explanation detailing the purpose, procedures, and potential benefits and risks of the study. This comprehensive explanation aimed to ensure that participants were fully informed about what their involvement would entail. To facilitate informed decision-making, participants were given a two-week window to consider their participation. During this period, they could ask questions and discuss any concerns with the research team or their psychologist. Also, permission was requested to access demographic and diagnostic information in their hospital records. Data was stored on a secure server within the hospital and was processed anonymously. 25 patients who committed violent offenses and 12 patients who committed sexual offenses declined to participate. During a one-time measurement patients were asked to complete the IRI and the BIS-11.

The general population group, consisting of a convenience sample of male individuals with low to moderate education levels, was recruited in the community through social media advertisement (e.g., Facebook) and through the researchers’ social network, targeting individuals who fit the demographic criteria. By focusing on males with lower educational attainment, the study sought to explore empathy levels across groups with relatively similar demographic profiles (with exception of offending history). Through an online survey, their data were anonymously collected.

Data Analyses

IBM SPSS statistics version 28 was used for analyzing the data. Demographic and clinical variables of the two patient groups were explored using descriptive statistics and checked for baseline differences between conditions with the Kruskal-Wallis test and Chi-square tests.

According to an a priori sample size calculation in G*Power (Faul et al., 2007), we needed 80 participants for the first hypothesis to detect medium effects (d = .30) with 80% power with alpha set at .05. For the other two hypotheses, power analyses indicated a total sample of 52 participants (n = 26 per group) to detect large effects (d = .40) with 80% power with alpha at .05.

The variables cognitive empathy and affective empathy were normally distributed. However, this did not apply to the variable impulsivity (i.e., BIS-11 total score, Shapiro-Wilk test < .05), which was strongly positively skewed. For this reason, a log transformation was performed on this variable. In this study, the scale for cognitive empathy was acceptably reliable (14 items, Cronbach’s α = .699). The affective empathy scale was also acceptably reliable (14 items, Cronbach’s α = . 682) and the impulsivity scale was also reliable (30 items, Cronbach’s α = .750).

This study had a between subject design. First, a series of descriptive analyses were conducted to examine group differences in responses to individual IRI items. Next, we used a multivariate analysis of covariance (MANCOVA) to compare affective and cognitive empathy in individuals who have sexually offended, individuals who have violently offended and the general population while controlling. For these analyses, the significance level was set at p ≤ 0.05. To further establish differences between groups in all our measures, Bonferroni post hoc correction tests were carried out. Impulsivity was considered a covariate in this study.

Empathy in Offending Groups

In our study, individuals with sexual offense histories scored higher on affective and cognitive empathy compared to violent offenders. This finding may be understood through frameworks like the integrated theory of sexual offending (ITSO; Ward & Beech, 2006), the pathways model (Ward et al., 2004), and the facilitation-motivation model (Seto, 2019). For example, the ITSO posits that sexual offending result from the interaction of neuropsychological vulnerabilities, psychological dysfunctions, and environmental factors. Unlike violent offenders, whose offenses are often driven by pervasive aggression and emotional dysregulation, sexual offenders may experience more context-specific deficits. These individuals often retain baseline capacities for empathy but suppress or distort empathetic responses toward their victims due to cognitive distortions. This suppression is situational and selective, allowing sexual offenders to display higher levels of both cognitive and affective empathy in non-offending contexts. For instance, they might show empathy towards acquaintances, family members, or even other victims of harm, while displaying a lack of empathy towards their victims (Sousa et al., 2023). Relatedly, they may retain the ability to engage in cognitive perspective-taking, which involves understanding and predicting the thoughts and emotions of others (Marshall & Marshall, 2019). This cognitive empathy allows them to navigate social interactions effectively and appear socially competent. However, their application of this ability can be selective and biased (Morrow, 2019). These distortions create a psychological buffer that allows them to disconnect their empathetic understanding from their actions during the offense. Violent offenders, by contrast, often exhibit more global empathy deficits (Saladino et al., 2021; Wang et al., 2021), stemming from chronic exposure to aggression and impaired emotional regulation systems, which diminishes their ability to connect with others’ emotional states altogether.

The Pathways Model outlines five distinct psychological pathways to sexual offending, each emphasizing different mechanisms like emotional dysregulation, deviant sexual scripts, and antisocial traits (Ward & Siegert, 2002). Offenders in pathways involving intimacy deficits or emotional dysregulation may retain their empathy capacities outside of offending situations, but their inability to manage personal emotional needs leads to maladaptive behavior, such as seeking power or connection through coercion. These offenders are likely to present higher levels of empathy than violent offenders, who often follow more antisocial pathways characterized by chronic disregard for others’ emotions and well-being (Osbourne et al., 2020). In this way, sexual offenders’ empathy deficits are more narrowly related to offense-specific contexts rather than a general desensitization to others’ suffering.

The Facilitation-Motivation model (Seto, 2019) explains sexual offending as a result of interacting motivational and situational factors. Motivation may stem from unmet emotional needs, deviant sexual interests, or negative affect, while facilitators include cognitive distortions or situational triggers. Offenders in this framework often exploit cognitive empathy to groom victims or justify their actions, indicating that cognitive empathy is intact but misapplied. Moreover, the motivation to offend often arises not from a lack of affective empathy but from an overwhelming desire to fulfill personal emotional or sexual needs, suggesting that affective empathy is not broadly impaired (Barnett & Mann, 2013). Conversely, violent offenders’ motivations may be often rooted in aggression or dominance, where empathy deficits – both cognitive and affective – are more intrinsic to their behavioral patterns.

The aforementioned text taps into potential conceptual explanations. There is, however, also a possible methodological explanation, namely the impact of social desirability because of the use of a self-report to assess empathy (Caputo, 2017; Gomes et al., 2019). Social desirability can occur when individuals respond to questions in a manner they believe will be viewed favorably by others (Perinelli & Gremigni, 2016). In the context of assessing empathy, sex offenders may overreport their empathic abilities to appear more socially acceptable (Stevens et al., 2015). This tendency may be is especially pronounced in sex offenders due to the intense social stigma (Furst et al., 2015; Lawrence & Willis, 2021; Lehmann et al., 2020). They may feel a stronger need to present themselves in a positive light to mitigate negative perceptions. This tendency could lead to inflated empathy scores in sex offenders. Consequently, their self-reported empathy scores might not necessarily accurately reflect their empathic capacities, although the extent may vary. For example, Mathie and Wakeling (2011) examined the impact of social desirable responding on self-report measures among 1,730 adult male sexual offenders. Their results indicated that the extent of socially desirable responding in their study was smaller than assumed, and that the impact on a number of self-report measures was lower than expected (Mathie & Wakeling, 2011). In contrast, violent offenders may be more likely to endorse lower empathy scores due to a lesser concern for social approval and a lesser degree of social stigma (Grossi, 2017). These differences highlight the need for complementary assessment methods, such as implicit measures or collateral reports, to obtain a more accurate understanding of empathy in offender populations.

Comparison With General Population

There were no differences between the three groups with regard to cognitive empathy. This is in contrast to our hypothesis and not in line with most studies. For example, a meta-analysis by Morrow (2019) indicated that sex offenders have lower levels of cognitive and general empathy than those in the general population but do not have significantly lower levels of affective empathy. The absence of differences in cognitive empathy between offenders and the general population may suggests that offenders possess the intellectual capacity to understand others’ emotions but may actively distort or disengage this capacity in contexts related to their offending behavior. According to the dual-process model of empathy (Blair, 2005), cognitive empathy requires conscious and effortful perspective-taking, enabling individuals to infer another person’s thoughts and feelings. For offenders, this ability may not be fundamentally impaired but rather influenced by maladaptive cognitive processes, such as self-serving justifications and schemas that minimize the harm caused to their victims (Sousa et al., 2023). These distortions allow offenders to reconcile their actions with their self-perception, reducing cognitive dissonance and moral conflict. For example, an offender may rationalize their behavior by attributing consent to the victim or downplaying the victim’s emotional suffering, thereby selectively disengaging cognitive empathy in offense-related scenarios (Dietz, 2020; Smyth et al., 2024). This selective disengagement serves as a psychological defense mechanism, enabling the offender to commit harmful acts without fully internalizing their moral or emotional consequences.

When compared to the general population, affective empathy showed a significant difference, with individuals who committed sexual offenses scoring significantly higher. Affective empathy, the capacity to emotionally resonate with others, operates as an automatic and involuntary response to observed emotional states (Morrow, 2019). While it might seem counterintuitive, research suggests that many sexual offenders retain or even exhibit heightened affective empathy in general, as they are not devoid of the capacity to feel for others. Instead, their empathy deficits are often specific to their victims rather than being generic (Sousa et al., 2023). This may result from underlying psychological vulnerabilities, such as emotional dependency or unmet intimacy needs, which are common among sexual offenders. These vulnerabilities may make offenders more emotionally attuned in non-offending contexts, as they seek emotional connection and validation. Moreover, offenders’ ability to empathize outside of offense-specific situations can reflect the compartmentalization of their empathetic capacities, allowing them to respond emotionally in general while avoiding the distress of acknowledging their victims’ pain.

Strengths and Limitations

This study has several strengths. The role of empathy in sexual offending is an important topic in treatment programs, and efforts have been made to consider not only risk factors but also personal factors in preventing sexual offending. However, the role of empathy in sexual offending is not extensively studied, and the results of previous research have been inconsistent. This study contributes to the knowledge about factors that may potentially play a role in sexual offending. Another strength of our study is that it examines the relationship between cognitive and affective empathy in three groups, including a non-offending comparison group. However, there are several limitations that need to be considered when interpreting the results of this study. First, the sample size is relatively small and the participants self-reported their levels of empathy. There are various reasons why participants may provide biased or inaccurate responses. One reason is social desirability bias, where individuals tend to present themselves in a positive light and report socially desirable behavior, particularly among individuals who committed sexual offenses who have been involuntarily clinically hospitalized. However, the scores on the IRI were no very high. Self-justification is another potential reason. Individuals who committed sexual offenses may try to justify their behavior by reporting empathy. Manipulation is a third reason to consider. Some are skilled manipulators and may report empathy as a strategy to gain the trust of others or to appear favorable to professionals involved in assessing their behavior. A related limitation is the fact that we only had one instrument to assess our main outcome variable. In future studies, it is advisable to include several complimentary tools, for example combining several self-report and observer-reports with qualitative data to get a comprehensive understanding of empathy and it’s relation with risk factors for sexual violence.

Another limitation of the research is the fact that the group of individuals who committed sexual offenses is heterogeneous, consisting only of hands-on offenders and no distinction wase made between pedophilic and other sexual offenses. This lack of differentiation could impact the generalizability of the research findings. Relatedly, in our study, we did not distinguish between empathy for the victim(s) of their crime(s) and empathy for non-victims, such as family members and acquaintances. Also, we did not have data on demographic information (with exception of age) and possible (history of) psychiatric disorders in the general population group which might impact the comparability of the three groups.

Lastly, a third limitation is that the participants who had offended are exclusively men who were admitted to a high secure hospital. This restriction may affect the generalizability of the research results to a broader population. Further research with larger and more diverse samples, including subgroups of individuals who committed sexual offenses, as well as objective measures of empathy, would be beneficial in advancing our understanding of this complex topic. A new questionnaire called the Affective and Cognitive Measure of Empathy (ACME) scale (Vachon & Lynam, 2016), may be helpful in making these distinctions more effectively.

Implications

The findings of the study may have several implications for research and practice. Firstly, the higher scores on affective and cognitive empathy in individuals with sexual offense histories compared to violent offenders may suggest that sexual offenders’ empathy deficits may be more selective and situational, rather than pervasive. This supports the need for targeted therapeutic interventions that address cognitive distortions and selective empathy deficiencies specific to their offending behavior rather than building perspective-taking abilities from scratch. Secondly, the lack of differences in cognitive empathy among all three groups implies that intellectual understanding of others’ emotions is not fundamentally impaired in offenders, highlighting the importance of focusing on affective empathy in rehabilitation programs. Lastly, the significant difference in affective empathy between sexual offenders and the general population suggests that while sexual offenders may retain higher baseline levels of emotional resonance, their empathy may be directed away from their victims due to self-serving biases. This underscores the necessity for tailored empathy training that helps offenders develop a more consistent and genuine application of empathy, reducing recidivism rates and enhancing rehabilitation outcomes. Overall, these findings highlight the complexity of empathy in offending behavior and emphasize the importance of nuanced, individualized treatment approaches.

Concluding Remarks

There is still much to learn about the link between empathy and sexual offending. In this study, affective and cognitive empathy was increased in individuals with sexual offense histories compared to violent offenders. But it remains to be seen whether these findings are replicated when using different measurement approaches. The finding that the general population does not show greater levels of empathy than the offending groups suggests that empathy, both affective and cognitive, is complex and influenced by various factors beyond just the presence of offending behavior.