Targeting Mechanisms for Problematic Pornography Use Interventions

Abstract

Evidence-based treatments for problematic pornography use (PPU) are missing, so users seek help through online communities formed around their concerns. This study investigated the effectiveness of brief interventions in (I) reducing desire for pornography use and (II) increasing self-efficacy regarding pornography use, in a non-clinical online sample (n = 711). Participants’ craving for pornography was activated and then they were randomly assigned to one of the intervention conditions: life meaning, willpower, mindfulness, awe, or active control. The results indicated that life meaning could be effective in reducing craving for pornography. Future research should investigate the potential of similar interventions with clinical samples.

Introduction

A subset of pornography users experience their pornography use as problematic. This subgroup accounts for around 3–8% of pornography users according to some estimates (Baranowski, Vogl, & Stark, 2019; Bőthe, Tóth-Király, Potenza, Orosz, & Demetrovics, 2020; Ross, Månsson, & Daneback, 2012), with higher estimates of 10–15% for young adult pornography users (Dwulit & Rzymski, 2019; Giordano & Cashwell, 2017; Kumar et al., 2021). While there is an increasing number of online self-help communities focused on problematic pornography use (PPU; Chasioti & Binnie, 2021; Osadchiy, Vanmali, Shahinyan, Mills, & Eleswarapu, 2020), there are few options for clinical treatment due to a lack of established frameworks that could guide clinicians in approaching the issue of PPU (Binnie & Reavey, 2020; Grubbs et al., 2020; for an overview of recommendations for clinicians, see Gibbons, Bernards, Bean, Young, & Wolfgramm, 2021). The present study aims to provide evidence-based recommendations about plausible cognitive-behavioral target mechanisms that can be used for the development of approaches to mitigate PPU and related concerns (either through self-help, or clinical treatment).

Pornography use

Pornography use is a ubiquitous element of sexuality in contemporary society (Paasonen, 2016), with the internet enabling vast amounts of pornographic content to be accessed daily by tens of millions of users worldwide (Ahmed, Shafiq, & Liu, 2016). The Pornhub website alone reported an average of 115 million daily visits in 2019 (Pornhub, 2019). It is estimated that pornography use has increased 4–24% during the recent lockdown periods during the COVID-19 pandemic (Mestre-Bach, Blycker, & Potenza, 2020; and Pornhub, 2020). Although accessibility to the internet introduced some generational increases in the use of pornography, pornography use remains common across different age groups (Price, Patterson, Regnerus, & Walley, 2016), with some evidence of tapering off with age (Miller, Raggatt, & McBain, 2020). There are estimates that in the United States, pornography is used by 46% of men and 16% of women during an average week (Regnerus, Gordon, & Price, 2016). When looking at young adult men (age M = 20), pornography use is frequent throughout the week, with estimates of one in five using daily or every other day (Carroll et al., 2008).

The overall individual and societal effects of pornography consumption remain a topic of debate in science and the public (Hald, Seaman, & Linz, 2014; Linz & Malamuth, 1993; Smith & Attwood, 2014). Questions involved in this debate include the links between pornography and sexual violence (Fukui & Westmore, 1994; Malamuth, Hald, & Koss, 2012; Wright, Tokunaga, & Kraus, 2016), dissolution of social norms (Wright, 2013), development in youth (Attwood, Smith, & Barker, 2018; Spišák, 2016), public health (Braun-Courville & Rojas, 2009; Harkness, Mullan, & Blaszczynski, 2015), and pornography addiction (Grubbs et al., 2020).

Compulsive sexual behaviour

Pornography addiction is related to the broader category of compulsive sexual behavior (also called sexual addiction; Carnes, 1983; Douglas Reed, 1994), which was included as a clinical disorder in the ICD-11 (Kraus et al., 2018). The diagnostic criteria for compulsive sexual behavior disorder include intense sexual craving that is difficult to control, leading to repetitive sexual behaviors that may cause distress or impairment in important areas of functioning (World Health Organization, 2019).

Problematic pornography use

There is a significant population of pornography users who experience their use as problematic or addictive (Dwulit & Rzymski, 2019; Ross et al., 2012). Such users can be said to experience self-perceived PPU, which is a term that addresses the extent to which users feel like their cravings are so strong that they have no control over their pornography use (Sniewski, Farvid, & Carter, 2018). Symptoms of PPU can map well onto the diagnostic criteria of compulsive sexual behavior disorder as PPU symptoms include an inability to control cravings for pornography, tolerance and withdrawal symptoms, resorting to pornography for emotional regulation, excessive time spent viewing pornography, a detrimental impact of pornography viewing on relationships, failed attempts at quitting pornography use and relapses (Duffy, Dawson, & Das Nair, 2016).

As indicated above, two important psychological mechanisms involved in both compulsive sexual behavior disorder and PPU include desire (craving) and self-efficacy for controlling desire. Desire for using pornography has been found to be stronger in problematic pornography users compared to non-problematic users (Antons et al., 2019), and it has been found to be related to frequency and compulsivity of pornography use (Leonhardt, Busby, & Willoughby, 2021). Self-efficacy to refrain from pornography use has been found to be lower among pornography users who use pornography more frequently, who made more (failed) attempts to reduce their pornography use, and who have higher levels of hypersexuality (Kraus, Rosenberg, Martino, Nich, & Potenza, 2017; Kraus, Rosenberg, & Tompsett, 2015). These findings suggest that strong desire for using pornography and weak self-efficacy for regulating one’s pornography use can jointly contribute to the individual experience of PPU.

Treating problematic pornography use

Overcoming self-perceived PPU is an important shared goal for many individuals, as evidenced by the number and size of online communities that have formed around this concern (Osadchiy et al., 2020). Despite an interest by many for addressing PPU, there remains a gap in high-quality treatment research in the field (Grubbs et al., 2020). A review of treatment options for compulsive sexual behavior disorder suggests that there is a lack of systematic studies with control groups in the field of treatment research, but it also points out that initial research indicated the potential effectiveness of mindfulness, cognitive-behavioral therapy, self-help, and pharmacological approaches (Efrati & Gola, 2018).

The present study

This study investigated the effectiveness of four different interventions in reducing desire for pornography use and increasing self-efficacy related to pornography use. The interventions were selected based on approaches that were either investigated in earlier research on PPU or that were used for developing treatments in other areas of addiction research. Meaning, willpower, mindfulness, and the emotion of awe are the interventions that were compared against a neutral condition designed to control for the potential of time and distraction for reducing desire and increasing self-efficacy.

Meaning is a feeling that one’s life makes sense (comprehension aspect), that it is directed toward valuable goals (purpose aspect), and that it matters in the world (mattering aspect; George & Park, 2016a; Martela & Steger, 2016). We expected that activation of meaning could reduce desire and increase self-efficacy by providing a behavioral option of higher value compared to use of pornography (Berkman, Hutcherson, Livingston, Kahn, & Inzlicht, 2017). Meaning has been shown to be inversely related to a variety of addictive behaviors, including drinking (Copeland, Jones, & Field, 2020; Palfai & Weafer, 2006) and smoking (Thege, Bachner, Martos, & Kushnir, 2009). Furthermore, initial experimental studies have found that brief meaning interventions may reduce the appetitive quality of substance-cues such as alcohol (Ostafin & Feyel, 2019; Schnetzer, Schulenberg, & Buchanan, 2013; Shamloo & Cox, 2010).

Willpower is the ability to avoid temptations, mostly in favor of a prospective benefit (Ainslie, 2020). Willpower beliefs (Job, Dweck, & Walton, 2011) were found to be predictive of different positive outcomes (better time management, less impulsive eating and spending) in a variety of situations requiring self-regulation (Bernecker, Herrmann, Brandstätter, & Job, 2017; Bernecker & Job, 2020; Job, Walton, Bernecker, & Dweck, 2015). We expected that activation of willpower could reduce desire and increase self-efficacy by improving self-regulatory capacities of users in a situation when their craving for pornography use is activated.

Mindfulness is a state that entails a heightened awareness of the present-moment experience and an accepting attitude toward that experience (Bishop et al., 2004). We expected that activation of mindfulness could reduce desire and increase self-efficacy by providing detachment from the urge for pornography use. Mindfulness was found to be effective in reducing desire for smoking (Westbrook et al., 2013), regulating various cravings (Enkema & Bowen, 2017; Khakbaz et al., 2016; Schumacher, Kemps, & Tiggemann, 2017; Tapper, 2018), and mobilizing self-control (Friese, Messner, & Schaffner, 2012; Jenkins & Tapper, 2014). Mindfulness-based relapse prevention was found to be an effective treatment for reducing the symptoms of compulsive sexual behavior disorder such as pornography viewing (Holas, Draps, Kowalewska, Lewczuk, & Gola, 2020) and evidence from single-case studies suggests that mindfulness could be effective in treating PPU (Sniewski, 2020; Sniewski, Krägeloh, Farvid, & Carter, 2022).

The emotion of awe is experienced in the presence of something so vast that it requires cognitive accommodation (Keltner & Haidt, 2003). Although there has been no research to our knowledge that has examined awe specifically as an intervention for impulsive behaviors such as PPU, the experience of awe has been proposed as a central mechanism of the beneficial effects of psychedelic drugs (Hendricks, 2018) and initial research shows these drugs may be helpful for addiction (Garcia-Romeu, Griffiths, & Johnson, 2014). Further, awe has been found to increase feelings of interpersonal connectedness with others (Bai et al., 2017; Van Cappellen & Saroglou, 2012). We expected that activation of awe could reduce desire and increase self-efficacy by increasing feelings of connectedness in the participants, because connectedness was associated with better self-regulation outcomes among students (Datu & Yuen, 2020) and was found to increase between partners after periods of abstinence from pornography (Fernandez, Kuss, & Griffiths, 2021). In addition, interpersonal satisfaction is negatively associated with pornography consumption among men (Wright, Tokunaga, Kraus, & Klann, 2017) and relational problems are central to users’ experience of PPU (Lambert, Negash, Stillman, Olmstead, & Fincham, 2012; Sommantico, Gioia, Boursier, Iorio, & Parrello, 2021; Wéry, Canale, Bell, Duvivier, & Billieux, 2020).

Participants completed the study by filling out initial baseline measures, which included state desire and self-efficacy. Assessments of state desire and self-efficacy were also made after the cueing and intervention sections of the study. The hypotheses were that compared to control, participants in the intervention conditions would report H1) less post-intervention desire for pornography and H2) more post-intervention self-efficacy related to pornography use.

Method

Participants

Participants were recruited through an online advertisement about a study regarding pornography use that was posted on several websites related to pornography use (see Supplement). The eligibility criteria for participation in the study were the provision of active informed consent (see Design and Procedure section for informed consent details), being at least 18 years old, and having at least intermediate knowledge of English. In total, 2788 participants entered the study, of which 1047 left the study early (see Figure 1), before completing the pornography cueing section. This group was omitted from all subsequent analyses. The remaining sample of 1741 participants was predominantly male (73%) and had an average age of 27.5 years (SD = 8.6; see Supplement for demographic characteristics). Participants predominantly self-identified as Caucasian/white (72%). Most participants had a partner, being either in committed relationships (38%) or married (19%). There was also a substantial group of single participants (36%).

Figure 1. Participant flow diagram.

31 participants were excluded from furtherer analyses for reasons including: stating explicitly that they did not watch pornography in the cueing period (n = 8), having unrealistically long cueing or intervention durations (n = 4), or responding negatively to the item “Please indicate if you took the study seriously (followed the instructions and replied honestly)” (n = 19). Furthermore, there were 764 participants who either did not complete the intervention (meaning n = 164, willpower n = 213, meditation n = 151, awe n = 114, control n = 98) or quit during the post-intervention questionnaires (n = 24; this group provided sufficiently informative data to be retained in some in some of the adjunct correlational analyses). In total, complete study responses were obtained from the remaining 946 participants. This completion rate (34% of those entering the study) is typical for an online study, as previous research has shown that the average response rates of online surveys are around the same proportion (Shih & Fan, 2008).

The aim of the study was to examine whether the interventions could be effective for individuals who were experiencing heightened desire to view pornography and masturbate. We assumed that all participants would show cue reactivity, but some participants did not. As we had no control over whether participants actually viewed pornography or were undistracted during the study, we decided to include only participants who demonstrated cue reactivity in the main analysis. Cue reactivity was operationalized as an increased desire for watching pornography after the pornography cue exposure. This resulted in the exclusion of 235 participants, leaving a sample of 711 for the main analyses. This sample provided statistical power of 0.86 to detect small effects (η_p² = 0.02), and power of > 0.99 to detect medium and large effects (η_p² > 0.06).

Measures and materials

Sexual behaviour measures

Frequency of behavior

The frequency of pornography use, masturbation, and sexual intercourse was assessed with a single item with five options, ranging from “Less than once a month” to “More than once a day” (see Supplement A for all items and response options of all questionnaires).

Recency of behavior

The recency of pornography use, masturbation, and sexual intercourse was assessed with a single item with five options, ranging from “More than a month ago/Never” to “Today.”

Motivation to quit

Motivation to quit pornography use was measured using an adapted version of the contemplation ladder (Biener & Abrams, 1991; Hogue, Dauber, & Morgenstern, 2010) which had 10 response options and four anchors: “1) I have no interest in reducing my porn use at all; 4) I am somewhat motivated to reduce my porn use; 7) I am very motivated to reduce my porn use; 10) I am fully motivated to completely quit using porn.”

Compulsive masturbation with pornography

This was assessed with an adapted version of the Compulsivity subscale from the Attitudes Toward Masturbation Scale (Young & Muehlenhard, 2019), substituting the original “masturbation” with “pornography.” This is a five-item measure with a seven-point response scale ranging from “Not a reason” to “Very important reason.” Internal consistency as assessed with Cronbach’s Alpha was satisfactory (α = 0.825).

Problematic pornography use

PPU was assessed with the Problematic Pornography Consumption Scale (PPCS; Bőthe et al., 2018) which is an 18-item instrument inquiring about PPU in the past six months. It has a seven-point response scale ranging from “Never” to “All the time.” Internal consistency as assessed with Cronbach’s Alpha was satisfactory (α = 0.933).

State measures of desire for pornography use and restraint self-efficacy

These constructs were used to assess craving in the study and they were measured with two items per construct. Desire for pornography use was assessed with the items “I want to watch pornography and masturbate right now” and “My desire to view pornography and masturbate feels overpowering right now.” Self-efficacy related to pornography use was assessed with the items “At this moment, I feel I can control my urges to view pornography and masturbate” and “Right now, I could stop or distract myself from thoughts related to pornography and masturbation.” The response scale for all items was a seven-point scale ranging from “Not at all” to “Very much.” These two constructs were measured at four points during the study, with acceptable levels of internal consistency for both the desire measure (αs between 0.690 and 0.788) and self-efficacy measure (αs between 0.705 and 0.839).

Individual difference measures

Life meaning

This was assessed with the Multidimensional Existential Meaning Scale (MEMS; George & Park, 2016b), which is a 15-item instrument inquiring about the three aspects of life meaning (comprehension, purpose, mattering). It has a seven-point response scale ranging from “Strongly disagree” to “Strongly agree.” Internal consistency as assessed with Cronbach’s Alpha was satisfactory (α = 0.917).

Boredom susceptibility

This was assessed with the Boredom Proneness Scale-Short Form (BPS-SR; Struk, Carriere, Cheyne, & Danckert, 2017), which is an eight-item instrument with a five-point response scale ranging from “Strongly disagree” to “Strongly agree.” Internal consistency as assessed with Cronbach’s Alpha was satisfactory (α = 0.867).

Depressive symptoms

Depression was assessed using an adapted version (reduced response scale from seven to four) of the Major Depression Inventory from ICD-10 (MDI; Olsen, Jensen, Noerholm, Martiny, & Bech, 2003), which is a 12-item instrument with a response scale ranging from “Not at all” to “Nearly every day.” Internal consistency as assessed with Cronbach’s Alpha was satisfactory (α = 0.810).

Manipulation check

After the interventions, participants completed six items to assess whether the interventions had their intended effects. The items used a ten-point response scale ranging from “None” to “Completely” to assess the following: “How much of the following states do you experience, at this moment?” followed by the six descriptors of “a sense of meaning”; “willpower and determination”; “awareness and mindfulness”; “boredom”; “awe and wonder”; and “part of something greater than myself” (a descriptor used to assess connectedness).

Cue exposure

Before the interventions, there was a five-minute pornography cue exposure section which was designed to activate craving for pornography use. The instructions before the cue exposure repeated (information from informed consent) that the participants would be asked to browse online and watch pornographic content of their choice for five minutes. Participants were asked to refrain from masturbation during the session. There was a visible timer on the subsequent page of the study, with detailed instructions: “1. You can see the timer, which indicates how much time you have left for viewing pornography. In order to keep track of the timer, do not close this window; 2. When the timer expires, the red button will reappear which will allow you to continue the study; 3. When you access pornography, keep watching it until the timer expires. Keep track of the time; 4. You can now open a new tab or window in your browser and search for pornographic content.” When the timer expired, participants could proceed to the intervention section.

Interventions

There were five conditions to which participants were randomly assigned. The interventions were designed to last around 10–20 min. Before each of the interventions, there was a brief statement: “The next part of the study will be an exercise consisting of materials related to regulation of pornography and masturbation behavior. Follow the instructions and materials carefully.”

The meaning intervention began with a brief text about authenticity which highlighted the importance of living according to one’s deepest values and finding a sense of meaning in life. This was followed by instructions which explained that the next part of the study will include a four-minute video of Steve Jobs talking about death and authentic values. After watching the video there was a text box where the participants were asked to give their impressions of the video: this served both as an attentiveness check, as well as a way to more strongly evoke the impressions they had during the video. This section was then followed by further texts about authenticity and values, where it was stressed that an authentic life consists of knowing one’s true self rather than following the rules of the group. It was highlighted to “have the courage to follow your heart and intuition.” This was followed by instructions about the final part of the exercise, where the participants would think about three values that reflect their true self, and then they would write about how they can act in accordance with these values in the next month. This intervention has been validated to induce a sense of meaning in previous research (Ostafin & Proulx, 2022; van Doornik, Glashouwer, Ostafin, & de Jong, 2021).

The willpower intervention began with a brief text about willpower which highlighted the importance of rising above challenges and temptations, and stressed that willpower can be improved and strengthened with practice “as a muscle,” which was intended to evoke unlimited willpower beliefs. This was followed by instructions which explained that the next part of the study will include a four-minute video of Roy Baumeister talking about mobilizing willpower. After watching the video there was a text box where the participants were asked to give their impressions of the video: this served both as an attentiveness check, as well as a way to more strongly evoke the impressions they had during the video. This section was then followed by further texts about the importance of applying willpower to shape one’s future, where it was stressed to remember the “willpower as a muscle” analogy. This was followed by instructions about the final part of the exercise, where the participants were asked to think and write about two situations, either from their personal experience or imagination, where applying willpower was required. After this, they were asked to write about both successful and unsuccessful ways of activating willpower in those situations.

The mindfulness intervention was adapted from a previous study (Ostafin, Bauer, & Myxter, 2012) and it began with a brief text about mindfulness which described what mindfulness is, and mentioned some of the positive effects it can have on well-being, such as rising above challenges and temptations. This was followed by instructions which explained that the next part of the study will include a ten-minute audio exercise which was a guided meditation exercise narrated by an experienced instructor in mindfulness meditation. After listening to the audio and following the instructions of the guided meditation, the mindfulness intervention was concluded.

The awe intervention began with a brief text about the “overview effect,” which is experienced by astronauts when they see the Earth from outer space (Yaden et al., 2016). There was a paragraph from a NASA astronaut about the beauty of this experience, followed by a paragraph explaining the transcendent aspects of the experience of awe. This was followed by instructions which explained that the next part of the study will include a five-minute video of a perspective departing from the Earth toward outer space. After viewing the video, the awe intervention was concluded. Similar simulations of the overview effect have already been used in previous research involving awe inductions (Chirico, Ferrise, Cordella, & Gaggioli, 2017; Stepanova, Quesnel, & Riecke, 2019).

The active control (distraction) condition began with a brief text stating that this period will serve as a distraction from the pornographic content that was watched in the pornography cueing section. It was mentioned that distraction is a commonly employed strategy when facing challenges with mental preoccupations or impulses. There was also a brief description of what the distraction period would entail: listening to a ten-minute audio chapter about a historical and legendary Roman king from ancient history (Haaren & Poland, 2012). After listening to the audio chapter, the distraction period was concluded.

Design and Procedure

Ethical approval for the study was obtained from the Ethics Committee for Psychology of the University of Groningen (PSY-2021-S-0447). The study was a single session that was administered online using Qualtrics Survey software. Participants first read an extensive information page about the study and then provided informed consent, which required that they 1) were at least 18 years old, 2) had read the information page carefully, 3) understood that the study involved watching pornographic content and refrained from participation if they were concerned that this might interfere with their well-being, and 4) agreed not to watch illegal pornographic content. Participants completed the baseline questionnaires and baseline state measures of pornography desire and self-efficacy. After this, participants completed the pornography cue procedure and the second iteration of the state desire and self-efficacy measures. This was followed by one of the five intervention conditions and the third assessment of the state desire and self-efficacy measures. After a brief mood repair, participants completed the state desire and self-efficacy measures again and were then debriefed regarding the study aims and hypotheses. Links to helpful resources about mental and sexual health were also provided.

Analysis plan

The hypotheses that the interventions would reduce desire to watch pornography (H1) and increase restraint self-efficacy (H2) were examined with an AN(C)OVA of difference scores (between pre- and post-intervention assessments of both desire and self-efficacy) with assigned condition as the between-subjects variable. Post hoc analyses included motivation to quit pornography use and intervention completion time as covariates, and the findings on the omnibus AN(C)OVA were followed by multiple comparison tests to examine which groups showed statistically significant differences from the control group. Cohen’s d was used for estimating effect sizes. Post hoc power analyses indicated good power for detecting small-to-medium effects. Correlations between a number of baseline and intervention variables were also looked into to provide additional insights into the data structure.

Missing data

The missing data were investigated by using F- and t-tests to compare the group of complete cases (completers; n = 946) with the group of dropouts who at least completed the pornography cueing section (non-completers; n = 735). It was observed that there were no significant differences between these groups in terms of age, t(1679) = 0.091, p = 0.927, gender, t(1679) = 0.832, p = 0.405, ethnicity, t(1679) = 0.448, p = 0.654, religion t(1679) = 0.136, p = 0.892, sexual orientation, t(1679) = 0.310, p = 0.756, meaning in life, t(1679) = 0.337, p = 0.736, depression, t(1679) = 1.920, p = 0.055, and boredom susceptibility, t(1679) = 0.156, p = 0.876. However, differences were observed on a number of variables related to pornography use.

There were differences in relationship status, t(1679) = 3.590, p = 0.001, indicating that completers were more often single (39%) than non-completers (31%) and were also more often in “other” relationship status (1.9% v. 0.5%), such as long-distance relationship or having multiple partners. Completers had less recent sexual intercourse (M = 2.13, SD = 1.20) compared to non-completers (M = 2.33, SD = 1.25), t(1679) = 3.371, p = 0.001, d = 0.162. Completers also had less frequent sexual intercourse (M = 1.87, SD = 0.90) compared to non-completers (M = 1.98, SD = 0.91), t(1679) = 2.465, p = 0.014, d = 0.121, which corresponds to them being single more often. Motivation to quit pornography use was higher in completers (M = 4.10, SD = 2.84) compared to non-completers (M = 3.79, SD = 2.72), t(1679) = 2.271, p = 0.023, d = 0.111. PPU score (PPCS) was higher in completers (M = 44.8, SD = 21.3) compared to non-completers (M = 41.6, SD = 18.9), t(1679) = 3.282, p = 0.001, d = 0.164. Desire cue reactivity was higher in completers (M = 1.30, SD = 1.46) compared to non-completers (M = 1.11, SD = 1.54), t(1679) = 2.543, p = 0.011, d = 0.126.

In cases with large amounts of missing data, it can be advised to avoid multiple imputation and proceed with a complete case analysis and pay attention to the possible shortcomings of such an approach (Jakobsen, Gluud, Wetterslev, & Winkel, 2017). Complete case analysis can be unbiased and even more appropriate than multiple imputation in cases when the likelihood of observing a missing case is not directly related to the outcome variable (White & Carlin, 2010). We looked into the correlation between the missing part of the dependent variables (post-intervention assessments) and post-cue assessments, and these indicated high collinearity for desire assessments (r = 0.574, p < 0.001) and self-efficacy assessments (r = 0.522, p < 0.001). Then, the association between the dichotomous variable indicating whether a case had missing data and the post-cue assessment (part of the dependant variable) could be investigated instead, and no relationship was found (r = 0.025, p = 0.299 for desire; r = −0.012, p = 0.625 for self-efficacy). This gave some basis to presume that the missingness mechanism was not related to the missing part of the dependent variables.

With these considerations in mind, it was decided to proceed with complete case analysis as the main analysis of the study, and then additionally repeat the main analyses on imputed data, using a multiple imputation procedure based on multivariate models that include all the variables from the study design. The imputation procedure was conducted in R with the statistical package “MICE” (Van Buuren, 2012). The imputation method was random forest, which is a method robust to potential interdependencies in the data (Wulff & Jeppesen, 2017). There were 50 imputation cycles, and passive imputation was used to impute the derived variables (difference scores). The results from the imputed data were used to provide an additional perspective for interpreting the obtained results.

Results

Preliminary analyses

We conducted a number of preliminary analyses in order to help with interpretation of the main analyses. Analyses on the baseline variables indicated no differences between the five conditions in terms of age, F(4, 706) = 0.598, p = 0.664, gender, X²(8, 706) = 3.852, p = 0.870, relationship status, X²(16, 706) = 12.392, p = 0.717, sexual orientation X²(12, 706) = 10.571, p = 0.566, compulsive masturbation with pornography, F(4, 706) = 0.594, p = 0.667, problematic pornography use, F(4, 706) = 1.101, p = 0.355, meaning, F(4, 706) = 0.765, p = 0.548, boredom, F(4, 706) = 0.337, p = 0.853, depression, F(4, 706) = 0.504, p = 0.733, post-cue desire, F(4, 706) = 0.822, p = 0.511, or post-cue self-efficacy, F(4, 706) = 0.218, p = 0.929. This suggests that the results are not a function of these variables. There were differences in motivation to quit pornography use, F(4, 706) = 3.574, p = 0.007, η_p² = 0.019, with willpower condition having the highest (M = 4.87, SD = 3.01) and mindfulness condition the lowest (M = 3.48, SD = 2.49) motivation to quit pornography use.

Even though there were no differences in the overall study completion times between the five conditions, F(4, 706) = 0.319, p = 0.865, we also assessed how much time it took for participants to complete each of the interventions in the possibility that the effects of the interventions on desire and self-efficacy were related to their duration. There were differences in the average intervention duration, F(4, 706) = 100.58, p < 0.001, η_p² = 0.362, with willpower intervention taking the longest (M = 16.39, SD = 8.15) and awe intervention taking the shortest (M = 5.93, SD = 1.36) time to complete.

Hypothesis testing

As discussed above, the hypotheses analyses were conducted on the sample of participants who demonstrated an increase in pornography desire after the cue exposure. Among these participants, a paired-samples t-test indicated a significant increase from pre-cue desire (M = 2.16, SD = 1.32) to post-cue desire (M = 4.03, SD = 1.58), t(710) = 42.832, p < 0.001, d = 1.284, as well as a significant decrease from pre-cue self-efficacy (M = 5.92, SD = 1.41) to post-cue self-efficacy (M = 5.19, SD = 1.63), t(710) = 13.043, p < 0.001, d = 0.479.

H1) desire for pornography use

The initial results of an ANOVA with post-intervention reduction of desire indicated significant differences between the five conditions, F(4, 706) = 4.515, p = 0.001, η_p² = 0.025. The meaning intervention was more effective in reducing desire than the control condition, t(311) = 2.855, p = 0.005, d = 0.319. The willpower intervention was also more effective in reducing desire than the control condition, t(272) = 3.413, p < 0.001, d = 0.428. Neither the mindfulness intervention, t(323) = 0.065, p = 0.948, d = 0.007, or the awe intervention, t(346) = 1.309, p = 0.191, d = 0.134, demonstrated more reduction of desire than the control condition.

When motivation to quit pornography use and intervention duration were added as covariates to the analysis, the results of an ANCOVA with post-intervention reduction of desire again indicated significant differences between the five conditions (see Table 1). In order to investigate the hypotheses, we looked into comparisons of each of the interventions with the control condition. The meaning intervention was more effective in reducing desire than the control condition. The willpower intervention was not more effective in reducing desire than the control condition, and neither was the mindfulness intervention or the awe intervention.

Table 1. Reduction of desire.

	n			F (df)	p	ηp^2
ANCOVA	711			2.530 (4, 706)	0.039	0.014
Intervention v. Control	n	M	SD	T (df)	p	d
Control	183	1.627	1.366	–	–	–
Meaning	130	2.056	1.368	2.732 (311)	0.007	0.313
Willpower	91	1.950	1.573	1.642 (272)	0.101	0.219
Mindfulness	142	1.696	1.370	0.444 (323)	0.655	0.050
Awe	165	1.910	1.490	1.867 (346)	0.063	0.197

Note. Larger estimated marginal means (M) indicate larger decreases of desire.

When we conducted the main analysis with the dataset with imputed data, the results of an ANCOVA with post-intervention reduction of desire did not indicate significant differences between the five conditions (see Supplement).

H2) self-efficacy related to pornography use

The initial results of an ANOVA with post-intervention increase of self-efficacy indicated no significant differences between the five conditions, F(4, 706) = 1.457, p = 0.214, η_p² = 0.008. The meaning intervention was not more effective in increasing self-efficacy than the control condition, t(311) = 1.426, p = 0.155, d = 0.163. The willpower intervention was more effective in increasing self-efficacy than the control condition, t(272) = 2.235, p = 0.026, d = 0.282. Neither the mindfulness intervention, t(323) = 0.677, p = 0.499, d = 0.071, or the awe intervention, t(346) = 0.998, p = 0.319, d = 0.104, demonstrated more increase of self-efficacy than the control condition.

When motivation to quit pornography use and intervention duration were added as covariates to the analysis, the results of an ANCOVA with post-intervention reduction of desire indicated no significant differences between the five conditions (see Table 2). In order to investigate the hypotheses, we looked into comparisons of each of the interventions with the control condition. The meaning intervention was not more effective in reducing desire than the control condition, and neither was the willpower intervention, the mindfulness intervention or the awe intervention.

Table 2. Increase of self-efficacy.

	n			F (df)	p	ηp^2
ANCOVA	711			1.016 (4, 706)	0.398	0.006
Intervention v. Control	n	M	SD	T (df)	p	d
Control	183	0.738	1.501	–	–	–
Meaning	130	1.008	1.516	1.551 (311)	0.121	0.178
Willpower	91	1.118	1.736	1.751 (272)	0.080	0.234
Mindfulness	142	0.912	1.513	1.023 (323)	0.305	0.115
Awe	165	0.901	1.657	0.976 (346)	0.330	0.103

Note. Larger estimated marginal means (M) indicate larger increases of self-efficacy.

When we conducted the main analysis with the dataset with imputed data, the results of an ANCOVA with post-intervention increase of self-efficacy did not indicate significant differences between the five conditions (see Supplement).

Manipulation check

The manipulation check indicated that the interventions were generally effective in activating the intended constructs, although with some overlap between conditions. The meaning and willpower conditions showed statistically significant (p < 0.05) higher activation compared to the other conditions in terms of “a sense of meaning” (M = 6.66, SD = 2.61; M = 6.71, SD = 2.29), “willpower and determination” (M = 6.49, SD = 2.57; M = 7.08, SD = 2.42), and “awareness and mindfulness” (M = 7.48, SD = 2.48; M = 7.46, SD = 2.41), but they largely overlapped with each other. The mindfulness condition showed statistically significant higher average activation of “awareness and mindfulness” (M = 6.65, SD = 2.71) compared to the control condition (M = 6.04, SD = 2.62). The awe condition showed statistically significant higher activation of “awe and wonder” (M = 6.69, SD = 3.01) compared to the other conditions, as well statistically significant activation of “part of something greater than myself” (M = 5.58, SD = 3.15) compared to the control condition (M = 4.21, SD = 2.76). There were no statistically significant differences in the activation of “boredom” across the conditions, although the control condition had the highest average activation (M = 5.83, SD = 2.86).

Correlations

For further exploration, we looked into correlations between baseline assessments of variables related to pornography use and assessments of desire and self-efficacy at pre-cue, post-cue, and post-intervention time points of the study (see Table 3). These analyses are provided separately for male (n = 523) and female (n = 170) participants from the main analysis. Problematic pornography use, compulsive masturbation with pornography, motivation to quit pornography use, boredom, and depression were positively related to desire for pornography use and negatively related to self-efficacy related to pornography use. In the analysis with the complete analysis sample (n = 711), male gender was found to be positively related to problematic pornography use (r = 0.35), compulsive masturbation with pornography (r = 0.27), and motivation to quit pornography use (r = 0.21), but not to boredom (r = −0.03) or depression (r = −0.1). Comparisons of baseline and post-mood repair assessments of desire (M = 2.11, SD = 1.22; M = 2.01, SD = 1.23) and self-efficacy (M = 5.94, SD = 1.39; M = 6.15, SD = 1.34) suggest that the participants left the study in a similar state as when they entered the study.

Table 3. Correlations of baseline and intervention variables.

	Males (n = 523)											Females (n = 170)
Variable	1.	2.	3.	4.	5.	6.	7.	8.	9.	10.	11.	1.	2.	3.	4.	5.	6.	7.	8.	9.	10.	11.
1. PPCS^a	–											–
2. ATMS^a	.74	–										.61	–
3. MQPU^a	.59	.57	–									.32	.21	–
4. Boredom	.45	.37	.30	–								.31	.27	.15	–
5. Depression	.47	.40	.28	.67	–							.29	.22	.11	.66	–
6. Life Meaning	−.12	−.12	.07	−.51	−.44	–						−.18*	−.07	−.03	−.62	−.54	–
7. Desire pre-cue	.55	.45	.20	.20	.23	−0.02	–					.47	.32	.10	.13	.08	.05	–
8. Desire post-cue	.57	.49	.33	.25	.27	−0.04	.71	–				.44	.38	.07	.08	.11	.01	.65	–
9. Desire post-inter.	.36	.28	.08	.12	.08	.01	.53	.52	–			.30	.25	.01	.16*	.14	.01	.49	.59	–
10. Efficacy pre-cue	−.45	−.42	−.27	−.28	−.27	.10*	−.48	−.42	−.29	–		−.30	−.16*	−.29	−.25	−.14	.12	−.26	−.22	−0.17*	–
11. Efficacy post-cue	−.48	−.44	−.35	−.25	−.23	.03	−.47	−.65	−.34	.58	–	−.46	−.39	−0.18*	−.17*	−.05	.08	−.40	−.50	−.36	.40	–
12. Efficacy post-inter.	−.32	−.30	−.21	−.17	−.12	.05	−.33	−.33	−.54	.48	.47	−.37	−.35	−.20	−.17*	−.12	.11	−.30	−.41	−.60	.44	.55

^a PPCS: Problematic pornography use; ATMS: Compulsive masturbation with pornography; MQPU: Motivation to quit pornography use. p < 0.01.

* p < 0.05.

Discussion

The main research question inquired whether four brief interventions (meaning, willpower, mindfulness, awe) would be effective in improving the self-regulation of activated urges for pornography use in a non-clinical adult sample. This was investigated by comparing the effects of each of the interventions with the active control condition. We examined two elements of compulsive behavior that are central to a number of disorders, including strong desire and low self-efficacy for controlling the desire. We hypothesized that after desire for pornography and masturbation was elicited by viewing pornography online, the interventions would be more effective than the control condition in terms of reducing desire and increasing self-efficacy. The results of the analyses provided mixed support for some of the hypotheses.

The results showed a statistically significant omnibus test for the reduction of desire for pornography measure (i.e. desire after the pornography cue minus desire after the intervention). Post hoc tests showed that the meaning and willpower interventions, but not the mindfulness or awe interventions, were more effective in reducing desire compared to the control group. However, when intervention duration and motivation to quit pornography use were added to the analysis as covariates (because the five conditions differed on these variables, with the willpower condition having the highest levels on both variables), only the meaning intervention indicated effectiveness in reducing desire compared to the control group. The effect of the meaning intervention on reducing desire was in the range of small-to-medium effects.

The results showed a statistically non-significant omnibus test for the increase of self-efficacy measure (i.e. self-efficacy after the intervention minus self-efficacy after the pornography cue). Post hoc tests showed that the willpower intervention, but not the meaning, mindfulness, or awe interventions, was more effective in increasing self-efficacy compared to the control group. When intervention duration and motivation to quit pornography use were added to the analysis as covariates (because the five conditions differed on these variables, with the willpower condition having the highest levels on both variables), no intervention indicated effectiveness in increasing self-efficacy compared to the control group.

The meaning intervention was somewhat effective in enhancing self-regulation of urges for pornography use, which is in line with some research. The reduction of desire for pornography use in the meaning intervention corresponds to the reduction of desire for alcohol consumption found in previous research (Ostafin & Feyel, 2019; Schnetzer et al., 2013; Shamloo & Cox, 2010), while the lack of increase of self-efficacy related to pornography use in the meaning intervention does not correspond to the positive association between meaning and self-efficacy found in alcohol use (Marsh, Smith, Piek, & Saunders, 2003). These findings suggest that there may be a protective effect of meaning when it comes to pornography use, a proposition which would be in line with previous research that found purpose in life to be related with positive treatment outcomes for substance and behavioral addictions (Çevik, Ciğerci, Kılıç, & Uyar, 2020, Martin, MacKinnon, Johnson, & Rohsenow, 2011; Waisberg & Porter, 1994). It could be argued that the meaning intervention evokes a future-orientation in participants (by leading them to think about long-term goals and aspirations), which could be a mechanism which underlies the effect of this intervention. The activated urges toward using pornography, which can be seen as a hedonic short-term goal in this context, would subside due to simultaneous activation of valuable long-term aspirations and goals, which are evoked by the meaning intervention. This kind of process would occur because in cases when competing behavioral options are activated, the one with a higher relative value is likely to be executed (Berkman et al., 2017); this would suggest that meaning can activate constructs which are of a higher relative value compared to the fulfillment of urges for pornography use.

The willpower intervention seemed initially effective in both reduction of desire and increase of self-efficacy for pornography use, but after motivation to quit pornography use and intervention duration were added to the analyses as covariates, these effects of the willpower intervention disappeared. These results do not align with some of the previous research that found willpower to have various positive effects on self-regulation (Bernecker et al., 2017), particularly in situations of high self-regulatory demand (Job et al., 2015), but they do suggest that motivation to quit pornography use and time spent being distracted from tempting stimuli could be useful components of effective self-regulation when it comes to pornography use.

The mindfulness and awe interventions were ineffective in enhancing self-regulation of pornography urges. We expected beneficial effects of the mindfulness intervention, as previous mindfulness research found it to be effective in reducing desire for smoking (Westbrook et al., 2013), eating (Schumacher et al., 2017), and consuming drugs (Khakbaz et al., 2016). The present findings are not in line with previous research where mindfulness has been shown to be a promising intervention both generally when it comes to the regulation of cravings (Enkema & Bowen, 2017; Tapper, 2018), as well as specifically when it comes to treating compulsive sexual behavior disorder and PPU (Efrati & Gola, 2018; Holas et al., 2020; Sniewski et al., 2022). The reduction of desire for pornography use or increase of self-efficacy related to pornography use did not occur in the awe intervention, and these findings do not have an equivalent in addiction research to be compared to. It is likely the case that the hypothesis of enhancing self-regulation via increased sense of connectedness activated through an awe induction was too intricate to be tested with the current study design.

When looking into the missing data analysis, some inferences can be made about the causes of the study dropout. There were no differences in demographic characteristics between the study completers and non-completers. However, the differences in other variables suggest that the study was to some degree more likely to be completed by participants who have problems with their pornography use (indicated by higher PPCS) and who are motivated to resolve these issues (indicated by higher motivation to quit pornography use), because they might be suffering from some perceived negative effects related to their pornography use (i.e. being single more often, having less frequent sexual intercourse) or feel like they are having struggles with controlling their drives toward pornography use (indicated by stronger cue reactivity). Such an interpretation would imply that the dropout from the study can be partially explained by non-completers being less interested or not sufficiently intrinsically motivated to do an online study about PPU, while completers were more engaged because of personal relevance of the topic. When it comes to developing treatments, a bias in the sample toward more of the problematic behavior in question is not necessarily problematic, because there are findings in substance use research which propose that treatment interventions could be more effective for those who show more severe symptoms of the addiction (Blow et al., 2009; Fernandez et al., 2019; Roy-Byrne et al., 2014).

Strengths and Limitations

This study aimed to provide findings that would contribute to opening the way for filling the existing gaps in the field of self-help and treatment research on PPU. The different interventions used in the study were standardized in terms of time and required effort, which enabled a direct comparison of the obtained outcomes in terms of usefulness for controlling the underlying PPU mechanisms (desire and self-efficacy). The sample size provided good statistical power for detecting any effects that the interventions might have had. Another strength of the study was the use of the pornography cues relevant to the participants’ experience, because they were allowed to freely select the pornographic content that they found arousing, anonymously and in privacy.

Limitations that prevent generalizing the inferences

Ideally, the sample would be representative of the population of PPU treatment-seekers. The missing data analysis indicated that the study completers had somewhat higher levels of PPU, but this was still below the expected level of PPU in the clinical target population: the participants from the analysis had a PPCS score of M = 45, SD = 21, while the cutoff score indicating clinically significant PPU is 76 (Bőthe et al., 2018). This makes it difficult to generalize the findings to that target population, although a clinical cutoff might be too high when dealing with self-perceived PPU. It could also be the case that the intervention effects were attenuated, because it is common that the interventions have larger effects with samples that have higher levels of problematic behavior in question (Blow et al., 2009; Fernandez et al., 2019; Roy-Byrne et al., 2014). This makes it difficult to propose with confidence that willpower, mindfulness, and awe are ineffective for treating PPU and provides a basis to expect that the effects of meaning could be even stronger in samples representative of the target population.

The main analysis was carried out with only the complete cases, which means that the imputation of data for the dropout participants primarily served as a reference point. While complete case analysis can be appropriate sometimes, multiple imputation is the generally favored option that helps with reducing biases in cases when there are a lot of missing data. However, in this study the pattern of missing data was gradual dropout, which indicated that the participants were dropping out throughout the study, most likely due to loss of interest or tiredness. The concern with dropout could be that the participants with the highest levels of PPU tended to drop out the earliest, because there are indications from studies with adolescents that those who are the most vulnerable to pornography use drop out the earliest from online studies about pornography use (Štulhofer et al., 2021).

Limitations that might have influenced the (null) findings

The interventions were not of the same duration (the willpower intervention was the longest, while the awe intervention was the shortest), which introduced some interference of completion times with the intervention effects. The manipulation check indicated that there was large overlap between the meaning and willpower interventions, suggesting that it was difficult to completely distinguish these two constructs with the present design. Participants had to browse for pornographic material themselves, which was in one aspect a strength of the study (because it allowed the participants to look for pornographic content that they would find arousing), but it was also a limitation because participants had to spend a part of their cueing period browsing and might have had difficulty finding desired material on time and completing the cueing section appropriately. However, the inclusion of only those participants who showed post-cue increase of desire should have controlled for this issue.

The mindfulness intervention was a general guided meditation which was not specifically designed to address urges for pornography use; a guided meditation with a focus on observing and accepting the experience of desire might have been more effective. Studies which found benefits of mindfulness for treating compulsive sexual behavior disorder (Holas et al., 2020) and PPU (Sniewski, 2020) used multiple mindfulness sessions over extended periods of time, while this study used mindfulness as a single-session intervention. The manipulation check suggested that this intervention was probably too short to develop familiarity with the practice and experience its beneficial effects thoroughly.

The analyses with the imputed data indicated no overall differences between the conditions for either reduction of desire or increase of self-efficacy. Individual group comparisons indicated effectiveness of the willpower intervention in both reducing desire and increasing self-efficacy; we suspect that these results occurred as a consequence of the influence of motivation to quit pornography use and intervention duration (the covariates in the analyses) in the multiple imputation procedure which preceded the analysis. While all of this could mean that there were no prominent effects of interventions once the bias of using only complete cases for analyses was reduced, it is also plausible that the imputation model which included all the variables from the study was too broad. Devising a theoretically informed and concise imputation model would be challenging given the current lack of systematic studies in the field of PPU treatment research.

Conclusion

The findings suggest that meaning interventions can be effective in improving self-regulation of pornography use. This improved self-regulation would come about as a result of decreased desire for pornography use. The proposed targeted mechanism by which meaning could have such a beneficial effect is orientation toward goals or aspirations of higher value than short-term hedonic goals: when two contrasting goals are simultaneously salient, the one bearing higher relative value is the one that determines the course of action taken (Berkman et al., 2017). This means that the activation of internally valuable constructs could make it easier to overcome temptations toward pornography use. Willpower priming, mindfulness meditation and awe induction did not have enhancing self-regulatory effects compared to the active control condition, or the effects were so small that they could not be detected in the obtained sample.

Improvements that are suggested for subsequent research in the field of PPU treatment research are recruitment of treatment-seeking random samples and adaptation of treatment interventions to the topic of PPU. To follow up on this study, future research in the field could investigate how the positive effects of meaning on self-regulation of pornography use come about (e.g. long-term orientation, simultaneous higher value activation). Such investigations should aim to provide useful guidelines for both practicing clinicians as well as individuals looking to self-treat their (self-perceived) PPU.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Declaration of interest

The authors report there are no competing interests to declare. This study was pre­registered at AsPredicted.org as entry #67580.