Introduction

Compulsive Sexual Behaviour and Problematic Pornography Use

‘Compulsive sexual behaviour’ (CSB), also known as ‘hypersexual behaviour’ or ‘sex addiction’ is characterized by sexual behaviour that persists despite repeated efforts to control or reduce the amount of time spent on this behaviour (Gola et al., in press). These acts include fantasies, impulses, and overt behaviours, in response to a negative emotion or stressful life events (Kafka, 2010). Researchers have identified the 20–30 range of age as the period of life in which people may be most likely to experience uncontrolled sexual behaviour (Goodman, 1997).

CSB includes individual behaviour (e.g. sexual fantasies, compulsive sexual thoughts, spending a lot of time watching pornography) as well as shared behaviours (e.g. non-committed sex or ‘cruising’ for multiple sex partners; Karaga et al., 2016).

Technological advances have made it possible to access sexual content through online pornography websites or mobile apps designed to connect people for sexual purposes or to engage with extramarital affairs. The accessibility and the availability of sexual content might cause distress for those who are attracted to this content and feel unable to control their behaviour. Indeed, CSB is associated with a variety of negative feelings such as depression, shame, purposelessness and a general lower level of well-being (Karaga et al., 2016).

One of the challenges researchers of CSB encounter is the variability of the definition of the term. Despite differences in the terminology of CSB within the scientific literature, all terms seek to describe uncontrollable sexual behaviour that causes distress and maladaptive behaviours (Hagedorn & Juhnke, 2005).

Shame, Guilt, and Self-Criticism

Shame and guilt are both considered self-conscious emotions. As opposed to ‘basic’ or ‘primary’ emotions, (e.g., fear, joy etc.) self-conscious emotions (e.g., shame, pride, guilt, embarrassment, etc.) necessarily involve self-representation and self-evaluation (Tracy & Robins, 2004). Shame and guilt are both negative self-oriented emotions and both involve self-evaluation. Despite the meaningful similarities, shame and guilt are considered two distinctive emotions (see Teroni & Deonna, 2008; Baldwin et al., 2006; Ferguson et al., 1991; Sabini & Silver, 1997; Tangney et al., 2007) Although shame – as opposed to guilt – might be evoked by both immoral and moral actions (Ferguson et al., 1991; Smith et al., 2002), in this perspective we will briefly layout the differences between shame and guilt as a response to perceived moral transgression. Benedict (1946) attributes shame to public transgression while guilt mainly evoked by private transgression, but empirical results show both shame and guilt are equally evoked by either public or private incidents (Tangney et al., 1996). Yet, shame mainly focus on others view upon self while guilt mainly focusses with self-reviewing and self-judgment (Tangney et al., 2007). While the ashamed individual is self-centered, and seek to hide, the guilty individual focuses on others and seek to amend. While guilt nurtures a better and cultured behaviour, shame involves with psychological stress and cultivates destructive actions (Tangney et al., 2007). Regarding the subjective experience of shame vs. guilt, one of the main differences between these two emotions is the generalization of the negative view upon self-perception. While the sense of guilt deals with a rather specific incident or attribute of the self, sense of shame encompasses and portrays the entire personality in a negative and adverse way (Lewis, 1987). Shame has a negative impact on self-identity while guilt is associated with specific perceived negative content. While guilt says, “I have done something bad” shame conveys the message “I am a bad person” (Dearing et al., 2005).

A third interconnected term is ‘self-criticism’. While both shame and guilt are considered ‘emotions’, self-criticism at its core is a self-evaluation process (Baldwin et al., 2006). Previous approach to self-criticism has focused on the tendency to seek higher goals and the sense of disappointment and criticism when falling short of achieving them (Blatt et al., 1976). In addition to the ‘cognitive’ devaluation component of self-criticism, Gilbert et al. (2004) have addressed the emotional aspect of self-criticism where the individual treats himself or herself with disgust and hatred. Although feelings of shame and self-criticisms might imply overlapping attitudes toward the self, shame includes another component which is the experience and belief of the individual that others perceive him or her as inferior, inadequate, or flawed (Gilbert et al., 2010).

Shame, Guilt, Self-Criticism, CSB and PPU

The three subjected emotions this article is dealing with (i.e., shame, guilt, and self-criticism) are all traditionally related to CSB and pornography use. For example, perfectionism and excessive concern with mistakes – which is a major facet of self-criticism is positively correlated with CSB (Reid et al., 2012a). Moreover, Efrati and Gola (2019) have found that self-criticism mediates the link between early-life trauma and CSBD in adulthood. Reid (2010) has found shame to be the best predictor of CSB compared to other negative feelings such as sadness or anger. Similarly, Zitzman and Butler (2009) have found that pornography-using is related to increase in shame. Cavaglion (2008) as well as King (2003) have found pornography users to have a sense of guilt and shame and the tendency to blame themselves which is the actual content of self-criticism. Although shame and guilt might preserve CSBD behaviour (Walton et al., 2017) both feelings can play a role in seeking for help (Kraus & Sweeney, 2019). Gilliland et al. (2011) have found that guilt - unlike shame – is associated with motivation to change and limiting sexual behaviour. Efrati (2018) has found that the sense of shame fosters seeking for help among secular adolescents but not among religious adolescents.

Despite the similarities between shame, guilt and self-criticism, important distinctions exist as detailed above. The current study aims to investigate first the contribution of shame, guilt, and self-criticism to the variance of CSB among men and women. Since a positive contribution of self-criticism and shame has been found, the second study has investigated the contribution of shame, guilt, and self-criticism to the variance of pornography in males. Pornography is a specific modality of CSB and often puts the individual into a moral dilemma. It was further predicted that self-criticism would predict shame-and-guilt, which in turn would predict pornography consumption.

Study 1

Methods

Participants

The study included 121 participants, 63 females, 58 males. The age range of participants was 18–45 with mean age of 25.38 (SD = 5.41) for the female participants and 28.39 (SD = 4.94) for the male participants.

Measures

Demographic Questionnaire

The demographic questionnaire has included items such as: age, gender, and marital status.

Hypersexual Behaviour Inventory (HBI)

The HBI is a 19-items self-report questionnaire that was developed by Reid et al. (2011), and it assesses the experience of sexual desire. The HBI aims to assess hypersexuality via three factors: (1) the degree they are out of control (2) the degree they are used as coping strategy, and (3) the extent of adverse-consequences caused by the sexual thoughts or activities. The items scored on a 5-point Likert scale (1 = never to 5 = very often) with range of total score between 19 to 95 and score of ≥53 is considered an indicator for hypersexuality. The scale has a high internal consistency of α = 0.96 (Reid et al., 2012b) and similar results were obtained in the current study (α = 0.96).

Forms of Self-Criticizing/Attacking and Self-Reassuring Scale (FSCRS)

The FSCRS is a self-report scale developed by Gilbert et al. (2004) to assess thoughts and feelings of self-criticism. The FSCRS includes 22 self-rating statements. The items make up three components: two forms of self-criticism; inadequate self (e.g., “I am easily disappointed with myself”), hated self (e.g., “I have become so angry with myself that I want to hurt or injure myself”), and one form of self-reassurance (I am gentle and supportive with myself). The rating is made on a scale ranging from 0 (Not at all like me) to 4 (Extremely like me). After reversing the relevant items, the total score is calculated and ranges between 0 (not self-criticizing at all) to 88 (extremely self-criticizing). Cronbach’s alpha values are α = 0.90 for ‘inadequate self’ and α = 0.86 for ‘hated self’ and ‘reassured self’ (Gilbert et al., 2004). In the current study there were Cronbach’s alpha values of α = 0.89 for ‘inadequate-self’, α = 0.75 for ‘hated self’, and α = 0.86 for ‘reassured self’. The Cronbach’s alpha obtained for the entire questionnaire in the current study was α = 0.92.

Test of Self-Conscious Affect (TOSCA)

The TOSCA is a scenario-based self-report scale developed by (Tangney, 1989) and it measures proneness to pride, shame, and guilt. In this study a short version was used (Tangney et al., 2000) of 11 scenarios that measure shame and guilt. Each scenario describes a situation such as “You attend your coworker’s housewarming party and you spill red wine on a new cream-colored carpet, but you think no one notices” with two possible reactions. One reflects a sense of shame (“You would wish you were anywhere but at the party”) and the other reflects a sense of guilt (“You would stay late to help clean up the stain after the party”). For each scenario, both reactions are rated from 1 (not likely) to 5 (very likely). The scores summed up and yield two scales: shame proneness and guilt proneness. Internal consistency in the current study was α = 0.76 for the ‘shame’ scale and α = 0.61 for the ‘guilt’ scale.

Procedure

The questionnaires were advertised online in social networks and forums that were dedicated for finding sexual partners, dating and sex. Participants answered the questionnaires on the Internet. Participants were informed that the study investigates sex addiction and that the questionnaires will remain anonymous for research purpose. All participants signed an online informed-consent form.

Statistical and Data Analysis

The analysis of the results was performed on Statistical Package for Social Science (SPSS; IBM Corp. Armonk, NY, USA) version 20. An analysis of simple correlations was analyzed using Pearson’s r, between shame, guilt and self-criticism measures for the whole sample and for males and females separately. Finally, the contribution of shame, guilt and self-criticism measures to the variance of hypersexual behavior ratings was measured using multivariate regression analysis. Significant results of the regression models are reported following Bonferroni’s corrections (p < 0.025). Values of skewness (S = -1.08, SE = 0.22) and kurtosis (K = 0.65, SE = 0.44) have indicated a normal distribution.

Ethics

The study was approved by the Institutional Review Board (IRB, Helsinki committee) of the University. All participants signed an informed consent form.

Results

Sample Characteristics

In order to explore the epidemiology data of our sample we used a score of ≥53 as the cut-off point for hypersexual behaviour, suggesting that 30% of male are considered to have CSB compared to only 13% of females. A chi-square test of independence showed that there is significant association between gender and CSB, X2(1, N = 121) = 5.08, p < .05. Cramer’s phi effect size was found weak (φc = 0.2, p < .05). However, a t test for independent samples shows no significant difference between women and men on the HBI (M = 38.4, SD = 16.8, M = 42.8, SD = 17.7), respectively; t (119) = −1.4, p = NS). No significant difference between women and men was found for the FSCRS either (M = 32.8, SD = 14.2, M = 29.7, SD = 14.8) respectively; t(119) = 1.2, p = NS). We did find significantly higher ratings on the TOSCA for women compare to men (M = 83.4, SD = 9.1, M = 75.9, SD = 9.2), respectively; t (119) = 4.5, p < 0.01).

The Association between Self-Criticism, Sense of Shame-and-Guilt and CSB

To enable comparing of current sample (study 1) and the next sample that will be presented below (study 2), we analysed the data with and without women included. Since both analyses yielded similar results, we used data for the whole sample (men and women) in study 1 although study 2 includes men only. An initial Pearson’s correlation test indicates a positive association between CSB and self-criticism (r = 0.49) but no significant correlation between CSB and sense of shame-and-guilt. A Pearson correlation between shame-and-guilt and CSB that was calculated for men only showed a significant but weak correlation (r = 0.29). A linear regression was performed in order to predict CSB based on self-criticism. A significant regression equation was found (F (1,119) = 36.785, p < 0.01) with an R2 of 0.236 which explained 23.6% of the variance. The final mediation model was: CSB = 22.436 + (0.577*self-criticism). A Multiple linear regression analysis calculated for men only yielded similar results. A significant regression equation was found (F (2,55) =13.253, p < 0.01) with an R2 of 0.325 which explained 32.5% of the variance. While self-criticism have contributed significantly to the model (B = .601, p < 0.01), shame-and-guilt has not (B = .331, p = NS). The final predictive model for men only was: CSB = −0.2 + (0.601* self-criticism) + (0.331* shame-and-guilt).

Since there is significant distinction between shame ang guilt, our next step was to test shame and guilt separately from each other. We gathered the ‘guilt’ items and ‘shame’ items separately, into two distinctive variables. When shame and guilt were tested separately against self-criticism and CSB, a significant correlation between shame and CSB was found (r = 0.24), but no significant correlation was found between guilt and CSB. Similarly, self-criticism was significantly correlated with shame (r = 0.48) but no significant correlation with guilt. Table 1 shows correlations among questionnaires in all participants.

Table 1 Study 1- Correlations between CSBD, Self-Criticism, Shame and Guilt
Full size table

A t test for independent sample showed that women had higher rates of shame (M = 83.4, SD = 9.1) compare to men (M = 75.9, SD = 9.2; t (119) = 4.5, p < 0 .001) but no significant difference between women and men in self-criticism (M = 32.8, SD = 14.2, M = 29.7, SD = 14.8, respectively; t(119) = 1.2, p = NS) and CSB (M = 38.4, SD = 16.8, M = 42.8, SD = 17.7) respectively; t(119) = 1.4, p = NS) .

In conclusion, the results of the first study have indicated a positive correlation between self-criticism, shame (but not guilt) and CSB. Secondly, regression analysis has shown that self-criticism have contributed to the variance of CSB and have explained 23.6% of the variance.

The lack of significant correlation between guilt to shame, self-criticism, and CSBD raises a question whether this pattern features any excessive sexual activity or that other activities such as problematic pornography use (PPU) entails moral issues and as a result a different pattern of emotional state.

Study 2

Methods

Participants

A hundred and ninety three participants were recruited for this study. Since only 7 females participated in the study, they were excluded from analysis and data analysis was based upon the male participants only: N = 186. The mean age for participants was 26.35 (SD = 4.86) age range 18–40.

Measures

Demographic Questionnaire

The demographic questionnaire has included items such as: age, gender, and marital status.

The Short Internet Addiction Test-Sex (S-IAT-Sex)

The s-IAT-Sex was developed by Wéry et al. (2016) and it measures online sexual behavior. The questionnaire was originally developed by Young (1998) as an Internet Addiction Test (IAT) and Pawlikowski et al. (2013) published the short version of the SAT (s-IAT). The s-IAT-Sex that has been used in this study is based upon the s-IAT with the replacements of the terms ‘online’ or ‘Internet’ by the terms ‘online sexual activity’ and ‘Internet sex sites’ respectively. The scale consists of 12 items with response options on a 5-point scale ranging from 1 (never) to 5 (always). The scale examines loss of control and time management with items such as “How often do you find that you stay on Internet sex sites longer than you intended?” and examines also craving and social problems with items such as “How often do you choose to spend more time on Internet sex sites over going out with others?”. Cronbach’s alpha for the whole scale in the original sample (Wéry et al., 2016) was very good (α = 0.88) and it was excellent (α = 0.91) in the current study.

Forms of Self-Criticizing/Attacking and Self-Reassuring Scale (FSCRS)

The FSCRS is a self-report scale developed by Gilbert et al. (2004) to assess thoughts and feelings of self-criticism. The FSCRS includes 22 statements and the responder rates how well a specific sentence describes him or her. The items make up three components: two forms of self-criticism; inadequate self (e.g., “I am easily disappointed with myself”), and hated self (e.g., “I have become so angry with myself that I want to hurt or injury myself”), and one form of reassured self (“I am gentle and supportive with myself”). The rating is made on a scale ranging from 0 (Not at all like me) to 4 (Extremely like me). After reversing the relevant items, the total score is calculated and ranges between 0 (not self-criticizing at all) to 88 (extremely self-criticizing). Cronbach’s alpha values are α = 0.90 for ‘inadequate self’ and α = 0.86 for ‘hated self’ and ‘reassured self’ (Gilbert et al., 2004). In the current study there were alpha values of α = 0.88 for ‘inadequate-self’, α = 0.8 for ‘hated self’, and α = 0.88 for ‘reassured self’. The Cronbach’s alpha obtained for the entire questionnaire in the current study was α = 0.89.

State Shame and Guilt Scale (SSGS)

The SSGS was developed by Marschall et al. (1994) to assess feelings of shame and guilt. The SGSS comprises of 15 items that compose three subscales: shame (e.g.,“ I feel like I am a bad person”), guilt (e.g., “I feel remorse, regret”) and pride (e.g., “I feel good about myself”) and the participants are asked to indicate on a 5-point Likert scale how well the given sentence describe them. Cronbach’s alpha in the current study was very good (α = 0.88).

Procedure

The questionnaires were advertised online in social networks and forums that were dedicated for finding sexual partners dating and sex. Participants answered questionnaires on the Internet. Participants were informed that the study investigates sex addiction and that the questionnaires will remain anonymous for research purpose. All participants signed an online informed-consent form.

Statistical and Data Analysis

The analysis of the results was performed on Statistical Package for Social Science version 20 (SPSS; IBM Corp. Armonk, NY, USA). An analysis of simple correlations was analyzed using Pearson’s r between shame, guilt and self-criticism measures in the whole sample and in males and females separately. Finally, the contribution of shame, guilt and self-criticism measures to the variance of Internet addiction sex ratings was measured using multivariate regression analysis. Significant results of the regression models are reported following Bonferroni’s corrections (p < 0.025). Values of skewness (S = -0.31, SE = 0.18) and kurtosis (K = -0.3, SE = 0.35) have indicated a normal distribution. The structural equation model (SEM) was tested using AMOS 26.0 with maximum likelihood estimation method to estimate the model fit and parameters.

Ethics

The study was approved by the Institutional Review Board (IRB, Helsinki committee) of the University. All participants signed an informed consent form.

Results

Sample Characteristics

In order to explore the CSB epidemiology in our sample we used the ‘pornography addiction questionnaire’ with score of >37 as a cut-off point, as recommended by the authors of the original questionnaire (Droubay et al., 2020). Scores show that out of 186 male participants, 12 participants (6.5%) can be classified having an addiction to watching pornography.

The Association between Self-Criticism, Sense of Shame-and-Guilt and Pornography Use

An initial Pearson’s correlation test has indicated a positive correlation between PPU, self-criticism and sense of shame-and-guilt (r = 0.36 and r = 0.44 respectively). A t test for independent samples also shows significantly higher levels of self-criticism for the group with PPU compare to the group without PPU (M = 40.8, SD = 16.2, M = 29.2, SD = 13.3, respectively; t(184) = 3.9,p < 0.01). Shame-and-guilt also were significantly higher among the PPU group compared with the non-PPU group (M = 40.9, SD = 9.9, M = 29.7, SD = 9.6) respectively; t (184) = 3.9, p < 0.01).

As described in study 1, since there is a significant distinction between shame and guilt, the next step was to test shame and guilt separately as two distinctive variables. When shame and guilt were tested separately against self-criticism and pornography use, both shame and guilt positively correlated with pornography-use and self-criticism. However, Steiger's (1980) Z-Test (using web tool developed by Lee & Preacher, 2013) shows that self-criticism had a stronger correlation with ratings of shame than with ratings of guilt (r = 0.737 vs. 0.616; N = 186; z = 3.158; p < 0.01). Table 2 describes associations between all questionnaires in study 2.

Table 2 Study 2- Correlations between PPU, Self-Criticism, Shame and Guilt
Full size table

A multiple linear regression was calculated to predict pornography-use based on self-criticism, shame-and-guilt. A significant regression equation was found (F (2,183) = 22.261, p < 0.01) with an R2 of 0.196 which explained 19.6% of the variance. While shame-and-guilt have contributed significantly to the model (B = 0.394, p < 0.01), self-criticism has not (B = .012, p = 0.871). The statistical parameter of tolerance was 0.36 and VIF measurers was 2.7 indicating on appropriate collinearity. The final mediation model was: pornography use = 10.9757 + (0.394*shame-and-guilt) + (0.012*self-criticism). These results suggest that the correlation between self-criticism and PPU is mediated by shame-and-guilt. A Structural Equation Modelling (SEM) has analysed the fit between mediation model (Fig. 1) and the sample data. To examine the mediation hypothesis, both direct and indirect effects were examined. The model yielded excellent fit indices (χ2[1] = 0.03, p = 0.87; NFI = 1.00; CFI = 1.00; RMSEA = 0.00). Consistent with the mediating hypothesis, after accounting for the statistically significant indirect path leading from self-criticism to pornography consumption via shame and guilt (β = 0.35, p = .01), the direct link between self-criticism and pornography consumption was no longer statistically significant (β = 0.02, p = NS), leading to the conclusion that self-criticism predicted shame-and-guilt, which in turn mediated pornography consumption, thus explaining 20% of pornography watching variance. Figure 1 shows a mediation model of pornography watching.

Fig. 1
figure 1

Mediation model analyzed by structural equation modeling technique (SEM) and postulated based on zero order correlations. The model tested direct and indirect effects leading from Self-criticism to Shame-and-guilt and from Shame-and-guilt to Pornography watching

Full size image

In conclusion, the results of the second study have indicated a positive correlation between, shame-and-guilt, self-criticism, and PPU. Secondly, a regression analysis has shown that only shame-and-guilt have significantly contributed to the variance of PPU and have explained 19.6% of the variance. A pathway analysis has indicated a significant indirect path leading from self-criticism to pornography consumption via shame and guilt.

Discussion

The results of the first study have shown an association between CSB, shame and self-criticism but no association between CSB and guilt. The results of the second study have shown an association between pornography-use, self-criticism shame-and-guilt. Despite the association between PPU and self-criticism, regression analysis has revealed that shame and guilt (but not self-criticism) have contributed to the explained variance of pornography use.

These allegedly contradicting results might be explained by arguing that CSB entails self-criticism and shame while pornography watching is mainly associated with feelings of shame-and-guilt. This conclusion is supported by the structural equation model which has shown that the association between self-criticism and pornography-watching is entirely mediated by shame-and-guilt.

Although it is hard to compare the two studies, based on the data it is quite convincing that while CSB and pornography watching both entail self-criticism and shame, pornography-watching involves also a sense of guilt that alongside with shame takes dominance over self-criticism. Despite the long-debated issue whether CSB should be labelled as a behavioural addiction (Sassover & Weinstein, 2020), it is quite agreeable that CSB shares major elements with addiction such as the uncontrollable action and the adverse result. Addictive behaviours, in general, involve shame but not guilt. For example, O’Connor et al. (1994) have found that individuals with substance use disorder felt more shame and less guilt compared to control participants. The distinction between shame and guilt appears to be relevant to behavioural addictions as well. For example, Shim (2019) also distinguished guilt from shame and found that young individuals that are addicted to smartphone-use are characterized with a sense of shame. Similarly, shame, significantly more than guilt, predict severity of problematic gambling (Yi & Kanetkar, 2011), compulsive buying (Yi, 2012), compulsive eating and impulse regulation in general (Sanftner et al., 1995). Regarding CSB, Adams and Robinson (2001) emphasize the need to differentiate between shame and guilt and claim that what drives the addictive cycle in CSB is the shame and not the guilt.

Pornography on the other end, by its very nature, regardless of addictive usage, is subjected to a negative perception of an improper act and perceived a social vice (McKee, 2006). This attitude might characterize pornography watchers as well, who believe that viewing pornography is wrong, and subsequently tend to feel guilty about their usage of pornography (Perry, 2018). Grubbs et al. (2019) also found that many individuals who watch pornography strongly condemn pornography watching and that this moral incongruence cause emotional stress. Moreover, they have found that individuals who perceive their action as immoral and wrong tend to perceive this very act as unregulated and uncontrollable. Based on Efrati (2019), condemning pornography might also lead to suppressing the urge for watching and in turn might actually increase this actual behaviour. Thus, as already suggested by Droubay et al. (2020) it might be that the sense of guilt that accompanies pornography use causes users to pathologize their condition and perceive themselves as addicted. Hence, the feeling of guilt that accompanies pornography use might give rise to actual or perceived addiction, shame and self-criticism.

Limitations

First and foremost, the main limitation of the current paper is the different tools that were used for two different samples. Despite the convincing conclusions we deduced from the combined data, we suggest repeating this research with evaluating the same group for PPU as well as for CSB. Additionally, since moral incongruence leads individuals to perceive their desire as excessive and addictive, it is of value to add a measurement tool that it is independent as possible of subjective judgment.

Both studies have used self-rating questionnaires on the Internet hence there is a possibility of inaccuracies in responses. In addition, both studies have included small sample sizes and there were potential biases of the samples. The first study has included an equal number of women and men whereas in study 2 there were only men. Finally, since both studies are cross-sectional it is impossible to make any statement on causality, for example that people watch pornography due to feelings of shame and guilt caused by being self-critical. An alternative explanation that the effect of watching pornography on shame is mediated by self-criticism has not been tested.

In Conclusion

The results of these studies indicate that there should be a distinction between CSB and PPU regarding the role of guilt, shame, and self-criticism. While CSB is characterized by feeling of shame and self-criticism like other addictive or compulsive behaviours, PPU is mainly characterized by feelings of guilt which is essentially an emotion aroused in response to an interpersonal transgression (Estrada-Hollenbeck & Heatherton, 1998). Those distinctions are important for clinicians to understand better their patients’ emotional distress.

Data Availability Statement

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.