Problematic online sexual activities in men: The role of self‐esteem, loneliness, and social anxiety (2020)

1 INTRODUCTION

Since the early 2000s, the Internet has become an essential medium in both personal and professional life. One of the most popular Internet‐related activities is engagement in various online sexual activities (OSAs), for example, pornography (videos and/or pictures), searching information related to sexual behaviors, playing sexual video games, dating sex sites, and sex webcams (Ballester‐Arnal, Castro‐Calvo, Gil‐Llario, & Giménez‐García 2014; Ross, Månsson, & Daneback, 2012; Wéry & Billieux, 2016). For the vast majority of people, this use of OSAs is non‐problematic. However, for a subgroup of individuals, involvement in OSAs can become excessive and associated with loss of control and functional impairment (Albright, 2008; Ballester‐Arnal et al., 2014; Grov, Gillespie, Royce, & Lever, 2011).

It is thus essential to understand why, for a subgroup of people, the use of OSAs becomes problematic. Numerous studies have shown that problematic use of OSAs can constitute a dysfunctional coping strategy (Chawla & Ostafin, 2007; Ley, Prause, & Finn, 2014; Moser, 2011, 2013). In such cases, involvement in OSAs is likely to reflect an experiential avoidance strategy to cope with or dissociate from intolerable thoughts, bodily sensations, and emotional states (Chawla & Ostafin, 2007). Some studies have shown that between 85 and 100% of people who report excessive sexual behavior present with at least one co‐occurring psychiatric disorder (Kafka & Hennen, 2002; Raymond, Coleman, & Miner, 2003; Wéry, Vogelaere, et al., 2016). In addition, several studies have suggested that the main reasons for engaging in problematic OSAs are as a coping mechanism (with anxiety, depression, and low self‐esteem), as a distraction, or as a means of stress reduction (Castro‐Calvo, Giménez‐García, Gil‐Llario, & Ballester‐Arnal, 2018; Cooper, Galbreath & Becker,2004; Ross et al., 2012; Wéry & Billieux, 2016).

These findings are consistent with Kardefelt‐Winther’s (2014a) proposal to anchor Internet‐related disorders (such as problematic use of OSAs) in a “compensatory” framework. According to this theory, Internet use can help to alleviate a problematic situation and fulfill needs that are unachieved in real life. However, this strategy can ultimately result in various negative outcomes (e.g., professional, social, health‐related) and thus constitutes a maladaptive coping behavior. According to Kardefelt‐Winther (2014a), substantial research performed in the field of excessive Internet‐related behavior has largely focused on isolated factors (e.g., psychosocial variables) and has thus failed to test comprehensive models, including moderator and mediator effects. Such a trend has led to an overestimation of some isolated factors and an underestimation of other potentially relevant variables. For example, in a study that focused on excessive online gaming, Kardefelt‐Winther (2014b) demonstrated that the associations of loneliness and social anxiety with excessive online gaming become nonsignificant when stress was controlled for. Taking into account interactions and/or mediations between variables seems essential to improve our understanding of the problematic use of OSAs.

It appears to be important, therefore, to focus on specific risk factors (especially those associated with emotional dysregulation and maladaptive coping behaviors) that may be involved in the development of problematic use of OSAs. In particular, the role of self‐esteem, loneliness, and social anxiety—which are known to interact with each other (see below) and have been widely studied in the context of general (unspecific) problematic Internet use—has to date scarcely been studied in the field of OSAs use (or has been studied in an isolated manner, as suggested in a criticism made by Kardefelt‐Winther (2014a, 2014b)).

Several studies have, however, investigated the three above‐mentioned factors in the context of problematic online behaviors. These previous studies showed that low self‐esteem (Aydin & San, 2011; Bozoglan, Demirer, & Sahin, 2013; Kim & Davis, 2009), a high level of loneliness (Bozoglan et al., 2013; Kim, LaRose, & Peng, 2009; Morahan‐Martin & Schumacher, 2003; Odaci & Kalkan,2010), and social anxiety (Caplan, 2007; Kim & Davis, 2009) are positively related to problematic and excessive general Internet use (these studies were not focused on specific online activities). These results suggest that for individuals characterized by loneliness, social anxiety, and poor self‐esteem, a preference for online interaction progressively develops, supported by beliefs that the Internet is a safer and more reinforcing place than the offline world, which is likely to result in excessive and uncontrolled involvement (Caplan, 2007; Kim et al., 2009; Morahan‐Martin & Schumacher, 2003; Tangney, Baumeister, & Boone, 2004). Caplan (2007) focused on the role of loneliness and social anxiety in the preference for online (rather than face‐to‐face) social interaction and showed that this preference is explained by social anxiety, but not loneliness.

In the context of OSAs, a few studies have analyzed the links between loneliness and use of pornography. For example, Yoder, Virden, and Amin (2005) found that the more time spent online consuming pornography, the greater the sense of loneliness. Other authors have also shown that problematic pornography users are lonelier than recreational users (Bőthe et al., 2018; Butler, Pereyra, Draper, Leonhardt, & Skinner, 2018). Efrati and Gola (2018) found that adolescents who displayed compulsive sexual behavior also had higher levels of loneliness and more sex‐related online activities. A recent study has also shown that a feeling of loneliness is associated with the frequency of using sexually explicit Internet material among men (Weber et al., 2018). Some studies reported a link between use of pornography and low self‐esteem, and a few suggested that problematic use of pornography was positively correlated with lower levels of general self‐esteem (Barrada, Ruiz‐Gomez, Correa, & Castro, 2019; Brown, Durtschi, Carroll, & Willoughby, 2017; Kor et al., 2014) and sexual self‐esteem (Noor, Rosser, & Erickson, 2014). Similarly, Borgogna, McDermott, Berry, and Browning (2020) demonstrated that men with low self‐esteem were especially attracted to pornography (as a way of conforming to and performing masculine role norms) and have more problematic pornography viewing. Finally, although several studies reported a high rate of social anxiety in people with hypersexual behaviors (not especially online; Raymond et al., 2003; Wéry, Vogelaere, et al., 2016), few studies have been conducted specifically in relation to OSAs. Nonetheless, some studies showed the presence of social anxiety symptoms in problematic pornography users (Kor et al., 2014; Kraus, Potenza, Martino, & Grant, 2015). Furthermore, several studies investigated the role of social anxiety in a specific population: Internet child pornography offenders. These studies reported that social anxiety is higher in online offenders than in other sexual offenders (Armstrong & Mellor, 2016; Bates & Metcalf, 2007; Middleton, Elliott, Mandeville‐Norden, & Beech, 2006), indicating that social anxiety may play a pivotal role in online offending (e.g., the Internet providing a way to explore sexuality for those who have difficulty with interpersonal interactions; Quayle & Taylor, 2003).

An important limitation of existing studies, however, is that they almost exclusively focused on online pornography, whereas a wide variety of OSAs exist (such as sex webcams, 3D sex games, searches for online/offline sexual contact, or searching for sexual information) for which these three psychological factors may not be involved in the same way. For example, it can be supposed that an individual with high social anxiety may be more comfortable with searching for online sexual partners (e.g., using specific applications). Yet, it is unlikely that all types of OSAs have the potential to become maladaptive copings, which is typically the case of an activity such as searching for sexual information. Hence, it is important to take into account the heterogeneity of OSAs when it comes to consider the psychological factors underlying problematic use.

Another important limitation of existing studies is that they do not take into account the complex interrelationships between loneliness, social anxiety, and self‐esteem. First, some authors found that people with low self‐esteem have low confidence and do not feel comfortable in social interactions, which is associated (and probably promotes) loneliness (Çivitci & Çivitci, 2009; Creemers, Scholte, Engels, Prinstein, & Wiers, 2012; Kong & You, 2013; Olmstead, Guy, O’Malley, & Bentler, 1991; Vanhalst, Goossens, Luyckx, Scholte, & Engels, 2013). Second, previous studies showed that low self‐esteem constitutes a risk factor for social anxiety (de Jong, Sportel, De Hullu, & Nauta, 2012; Kim & Davis, 2009; Obeid, Buchholz, Boerner, Henderson, & Norris, 2013). Third, some studies emphasized a link between social anxiety and loneliness (Anderson & Harvey, 1988; Johnson, LaVoie, Spenceri, & Mahoney‐Wernli, 2001; Lim, Rodebaugh, Zyphur, & Gleeson, 2016). Finally, other studies suggested that (1) self‐esteem and loneliness significantly predict social anxiety (Subasi, 2007), (2) self‐esteem (but not social anxiety) predicts loneliness (Panayiotou, Panteli, & Theodorou, 2016), and (3) the relationship between self‐esteem and loneliness is mediated by social anxiety (Ma, Liang, Zeng, Jiang, & Liu, 2014). Thus, although these variables seem to be closely related and present with complex interrelationships, they have to date never been conjointly investigated in the context of problematic use of OSAs.

The current study thus aimed to fill a gap in the literature by testing a model (see Figure 1) that links low self‐esteem, social anxiety, and loneliness to OSA preferences (i.e., type of OSA performed) and ultimately to symptoms of addictive use. We hypothesized that (1) low self‐esteem is positively associated with both social anxiety and loneliness, (2) social anxiety is positively related to loneliness (mediates the role of social anxiety in the relationship between low self‐esteem and loneliness), and (3) these variables are positively associated with OSA preferences and its problematic use.

2 METHOD

3 RESULTS

4 DISCUSSION

A better understanding of the psychological factors involved in the development and maintenance of problematic use of OSAs is required, given the ubiquity of OSAs use in the general population. Despite the efforts made in this direction and the numerous studies performed in recent years, the existing literature in this field has presented important limitations. Accordingly, the aim of the current study was to test a model that linked self‐esteem, social anxiety, and loneliness to the type of OSAs performed and to the symptoms of problematic use of OSAs.

In support of our hypotheses, the present findings provided evidence for a mediational model in which low self‐esteem is associated with loneliness and high social anxiety, and in which the relationship between self‐esteem and loneliness was mediated by social anxiety. These factors are in turn associated with the use of pornography and the search for online sexual contacts, as well as with symptoms of problematic use. These findings are consistent with those of previous studies that showed that low self‐esteem is associated with loneliness (Panayiotou et al., 2016) and with higher social anxiety (de Jong, 2002; Obeid et al., 2013), that the relationship between self‐esteem and loneliness is mediated by social anxiety (Ma et al., 2014), and that problematic use of pornography is associated with low self‐esteem(Barrada et al., 2019; Brown et al., 2017; Kor et al., 2014), loneliness (Bőthe et al., 2018; Butler et al., 2018; Yoder et al., 2005), and social anxiety symptoms (Kor et al., 2014; Kraus et al., 2015). To date, these factors have been mainly studied separately and rarely in the context of OSAs. The results of the current study thus provide a better understanding of the complex relationships between these variables. Our findings, although cross‐sectional, are compatible with the view that a lower self‐esteem might constitute a risk factor for higher social anxiety and loneliness. Under such circumstances, and in accordance with the compensatory usage of the Internet model (Kardefelt‐Winther, 2014a), individuals are susceptible to display a preference for online sexuality and experience addictive use.

Furthermore, among the OSAs assessed in the present study, only two appeared to be related to problematic use: watching pornography and searching for online sexual relations. These results are in accordance with those of previous studies that showed that pornography is the most problematic OSA in men (Ross et al., 2012; Wéry & Billieux, 2016). Moreover, several previous studies emphasized that online sexual contact with other users is also a frequent activity in males and that this OSA has the potential to become problematic and engender tangible negative consequences (Daneback, Cooper, & Månsson, 2005; Döring, Daneback, Shaughnessy, Grov, & Byers,2017; Goodson, McCormick, & Evans, 2001; Wéry & Billieux, 2016). Moreover, the present results also suggest that relationship status play a role in the type of OSA use. Relationship status was not found to affect pornography use but appeared to impact the search for online sexual relations, which is consistent with the results obtained in a previous study by Ballester‐Arnal et al. (2014). This result is likely due to the fact that some OSAs—typically searching for online sexual partners—are viewed as a proof of infidelity and thus less practiced by people being in a romantic relationship (Ballester‐Arnal et al., 2014; Whitty, 2003). Our findings suggest that the use of the Internet for sexual purposes is multi‐determined and that it is essential that further research systematically take into account the specific sexual activities practiced online (for similar arguments, see also Barrada et al., 2019; Shaughnessy, Fudge, & Byers, 2017). The present results also highlight the importance of conducting research on various OSAs beyond the mere consideration of online pornography, as is often the case in this research field.

Notably, the two activities retained in our model (watching pornography and searching for online sexual relations) further support the view that structural characteristics of OSAs are important in explaining their potential problematic usage. Indeed, the anonymity offered by the Internet makes it a privileged place to explore sexuality outside social judgment (Cooper, Scherer, Boies, & Gordon, 1999). Along the same lines, our results could be explained by the online disinhibition phenomenon, that is, the decrease in concerns about the presentation of oneself and the judgment of others (Suler, 2004). On the whole, the physical distance and anonymity offered by the Internet generates a sense of security that increases comfort during virtual intercourse with potential partners (Daneback, 2006). Indeed, several studies reported that individuals with these characteristics tend to prefer online over offline social interactions (Caplan, 2007; Lee & Cheung, 2014; Steinfield, Ellisonthose, & Lampe, 2008; Valkenburg & Peter, 2007). These previous results are consistent with the social compensation hypothesis (Kardefelt‐Winther, 2014a), which suggests that people with poor social skills are especially prone to developing a preference for online interactions; the current study suggests that this might also valid in matters of sexuality. It is thus possible to speculate that in the initial stages, the use of OSAs will efficiently increase self‐esteem and mitigate social anxiety and loneliness. Such an effect has, for example, been suggested by Shaw and Gant (2002), who found that engagement in online chat leads to a decrease in loneliness and depressive symptoms and an increase in self‐esteem and perceived social support. However, with time and potential maintenance of the behavior, it is possible to expect that the use of OSAs may become indispensable and cause negative consequences (Caplan, 2007), resulting in compromised self‐esteem and increased isolation and social anxiety. Crucially, continuing to use the Internet for sexual behavior implies an avoidance of real‐life mating situations, which is likely to further reinforce the phenomenon of sexual avoidance.

The current study presents some limitations. First, the sample was relatively small and self‐selected, and its composition and representativeness limit the generalizability of the results. Nonetheless, the sample size (N = 209) can be considered adequate for the path analyses used here, ensuring satisfactory statistical power (Bentler & Chou, 1987; Kline, 2005; Quintana & Maxwell, 1999). Second, we have not included measures of offline sexual behaviors, which implies that the interpretation of our findings based on the online disinhibition hypothesis remains speculative. Third, the present study was only conducted in males, while realizing future studies also involving women is required. Indeed, previous studies emphasized gender difference in OSAs use preferences (e.g., women tend to prefer interactive OSAs such as sexual chat whereas men tend to prefer OSAs including visual content such as pornography, see Green, Carnes, Carnes, & Weinman, 2012; Cooper et al., 2003; Schneider, 2000). Future studies involving both genders are thus required to extend the present findings. Fourth, it might be that some alternative explanations not addressed in the current paper explained the patterns of association found. For example, the moral incongruence theory (Grubbs & Perry, 2019) posits that some users think OSAs are wrong (e.g., at a religious or moral levels), but perform them anyway, which ultimately promote emotional symptoms and diminish self‐esteem. Future studies should thus be conducted in order to test these alternative theoretical frameworks. Fifth, our study was based on self‐reported measures and may be limited by response and recall bias. Finally, the study used a cross‐sectional design that did not allow us to test the model in time. This latter point is of importance because it would also have been very conceivable to test the hypothesis that excessive OSAs use predicts loneliness and low self‐esteem. Longitudinal studies are thus required to confirm the hypotheses developed in our discussion and to ascertain the role of the study factors in the development and maintenance of problematic use of OSAs.

Despite its limitations, this study contributes to knowledge on the relationships between self‐esteem, loneliness, and social anxiety in problematic use of OSAs in men. Regarding these results, improved self‐esteem and reduced symptoms of loneliness and social anxiety would constitute sound targets for psychological interventions in people who are experiencing dysfunctional and impaired use of pornography or searches for online sexual contacts.