Introduction

Pornography has created considerable debate in scientific, legal, and public spheres—with some even arguing that it is among the most contested moral issues of our time (Burger, 1987; Hatch et al., 2020a, b; Segal, 1990). Among these debates is: (1) whether pornography consumption should be considered an addictive or problematic behavior worthy of behavioral intervention, and (2) whether pornography consumption is negatively associated, positively associated, or uncorrelated with relationship functioning (Ley et al., 2014; Love et al., 2015; Montgomery-Graham et al., 2015; Wright, 2019). Approximately 4–6% of men and 1–2% of women report having felt addicted to pornography, which may lead them to seek help for their use (Grubbs et al., 2019). Additionally, pornography-related behaviors have often been shown to be associated with relationship distress (e.g., Crosby & Twohig, 2016; Perry, 2020). Taking preliminarily steps to determine whether existing interventions and infrastructure that is wide-reaching and affordable can reduce pornography use, reduce arguments about pornography use within a romantic relationship, and/or reduce perceptions that one’s own pornography use is problematic may be worthwhile (Grubbs et al., 2019).

Pornography Use and Relationship Well-Being

One of the most contentious areas in the pornography consumption literature is whether pornography has consistent effects on key areas of relationship functioning including relationship satisfaction, sexual satisfaction, and communication. While some have argued that pornography is a major contributing factor to marital separation and divorce (Crosby & Twohig, 2016; Schneider, 2000), the literature has produced a mixture of positive, negative, and null effects (Newstrom & Harris, 2016; Perry, 2020). Indeed, a recent study that compiled 31 measures of relationship quality across 30 independent surveys found that 44–52% of tested associations were statistically significant and were in the negative direction whereas the other 48–56% of the other tested relationships were non-significant (Perry, 2020). Aggregate effects across men and women from a meta-analysis indicate a small negative correlation between pornography consumption and relationship satisfaction (r = − 0.09, Wright et al., 2017). These aggregated studies deal with some of the limitations plaguing the pornography consumption literature: pornography use is assessed at a general level, and pornography is only assessed using a single item (Hatch et al., 2020a, b; Kohut et al., 2020). Despite these limitations, those in distressed relationships typically report using pornography slightly more often than non-distressed individuals perhaps as a way to cope with the underlying distress caused by the relationship dysfunction.

However, one of the primary limitations of these studies is individuals in relationships are recruited as opposed to both partners in the relationship (Campbell & Kohut, 2017). Some of the few studies that have recruited both partners have concluded that the effects of pornography may be positive (Kohut et al., 2018). For example, some have indicated that couples who jointly use pornography are generally closer with one another (Kohut et al., 2018, 2021; Maddox et al., 2011). Furthermore, women who use more pornography generally report greater relationship satisfaction (Bridges, 2007). These results appear to be consistent with recent longitudinal studies. Over a 20-month time period, greater solitary pornography use was indicative of decreases in men’s relationship quality; however, women’s greater solitary pornography use was indicative of increases in women’s relationship quality during the same time (Huntington et al., 2021). When both partners watched pornography together, relationship quality generally increased during the 20-month time period (Huntington et al., 2021). Finally, although greater levels of men’s pornography use are indirectly linked with lower levels of their partner’s (i.e., women’s) relationship satisfaction (Poulsen et al., 2013), greater levels of women’s pornography use are indirectly linked with greater levels of men’s relationship satisfaction (Poulsen et al., 2013). Emerging research, however, suggests that these gender differences may be a function of a similarity-dissimilarity effect (Kohut et al., 2018). Couples with discordant levels of solitary pornography use report lower closeness and inhibited comfort with sexual communication (Kohut et al., 2018). These findings have been extended in more recent research, indicating that this phenomenon only occurs when one partner is regularly using pornography and the other reports using little or none (Kohut et al., 2021).

The associations between pornography consumption and sexual satisfaction have also been frequently studied. Meta-analyses have indicated an overall small negative effect between pornography consumption and sexual satisfaction (r = − .11; Wright et al., 2017), but these effects may be dependent on gender and joint couple use. Specifically, greater pornography use has been shown to be associated with lower sexual satisfaction among males; however, greater pornography use has been shown to be associated with greater sexual satisfaction among females (Bridges, 2007; Willoughby et al., 2021). Furthermore, more joint pornography use has been associated with an increased sexual desire, greater sex quality, and more intimacy with partners (Rodrigues & Martins, 2021). Other research has indicated that more joint pornography use among couples has been associated with greater sexual satisfaction, whereas solo use has been indicative of poorer sexual satisfaction (Maddox et al., 2011). Finally, within opposite-sex couples, more male pornography use has been associated with lower sexual satisfaction among females (Poulsen et al., 2013). And, although not significant, female pornography use in the same study was positively associated with increases in male sexual satisfaction (Poulsen et al., 2013). Finally, joint pornography use is associated with greater positive communication and less negative communication (Maddox et al., 2011; Willoughby et al., 2016). More joint pornography use has also been associated with more open communication about sex (Kohut et al., 2018).

Treatments for Pornography Use

Before 2016, no randomized controlled trials investigating treatments for problematic pornography use existed (Crosby & Twohig, 2016). However, since that time, several studies have utilized various types of individual treatments to ameliorate problematic pornography use. These treatments have included Acceptance and Commitment Therapy (Crosby & Twohig, 2016; Twohig & Crosby, 2010), Cognitive Behavioral Therapy (Minarcik, 2016), Cognitive Behavioral Therapy plus Paroxetine (Gola & Potenza, 2016), and one online intervention (Bőthe et al., 2021). In an early pilot study, six men received eight 1.5-h sessions of Acceptance and Commitment Therapy (Twohig & Crosby, 2010). Five out of the six participants had notable reductions in viewing pornography, obsessive thoughts and scrupulosity, and improvements in quality of life (Twohig & Crosby, 2010). In a follow-up study in an all-male sample of members of the Church of Jesus Christ of Latter-day Saints (N = 28), those assigned to 12 one-hour sessions saw large reductions in hours spent viewing pornography (d =  − 1.77), large decreases in sexual compulsivity (d = − 1.29), and small increases in quality of life (d = 0.23) compared to the waitlist control condition; these improvements were maintained in a 3-month follow-up (Crosby & Twohig, 2016). In a final bibliotherapy study with similar authors, 19 participants (90% male) seeking help for problematic pornography use were asked to read a non-tailored self-help Acceptance and Commitment Therapy book and use a freely available coaching app. The book did not target viewing pornography specifically but targeted reducing psychological flexibility, experiential avoidance, and cognitive fusion. Despite the lack of tailoring to pornography use, participants reported large-sized reductions in weekly pornography viewing (d = − 1.72), compulsive viewing (d = 1–0.79), negative outcomes related to pornography (d = − 2.48), psychological flexibility (d = 0.97), and cognitive fusion (d = 1.65; Levin et al., 2017).

Cognitive behavioral treatments alone have seen good success in ameliorating the effects of pornography use but combining these treatments with antidepressants may be contraindicated. For example, participants receiving 12 sessions of cognitive behavioral therapy reported a 33-min reduction in pornography consumption per week—corresponding to a 94% reduction in overall pornography use from baseline to post-treatment. However, there was mixed success in participants obtaining clinically significant change in a reported addiction to pornography (Minarcik, 2016). Further, participants rarely improved in their relationship satisfaction—a secondary goal of treatment (Minarcik, 2016). When cognitive behavioral treatments have been combined with Paroxetine, this combination initially appeared to be effective in reducing pornography use (Gola & Potenza, 2016). However, new compulsive sexual behaviors started to emerge after 3 months, suggesting that cognitive behavioral treatments combined with Paroxetine may be contraindicated inviting other treatment strategies for consideration (Gola & Potenza, 2016).

An online web-based self-help aimed at reducing problematic pornography use—the Hands-off program (Bőthe et al., 2021) includes six core modules based on motivational interviewing, Cognitive behavioral therapy, and other third-wave cognitive behavioral therapies (Bőthe et al., 2021). Although this innovation was admirable, the high attrition rates (89.4%) within the treatment group make any program effects difficult to interpret. Thus, discovering alternative web-based treatment methods with higher adherence rates to reduce the potential impact of pornography use may be worthwhile.

Couple Interventions as a Treatment Option

Typically, the underlying research assumes that pornography use leads to relationship distress and rarely operates in a reciprocal or bidirectional fashion. Yet, given the small-sized links between romantic relationship and pornography-related behaviors, couples-based interventions may be an attractive tool to treat problematic pornography use because it alleviates the relationship distress that leads to the pornography-seeking behavior. This is similar to the reason why couple therapy to treat relationship distress has been able to have smaller-sized impacts on various forms of psychopathology (e.g., depression, anxiety, and substance abuse; Snyder et al., 2006; Whisman, 2007; Whisman & Baucom, 2012)—alleviating relationship conflict has positive spillover effects for individual functioning. Even individual treatments attempting to reduce pornography-related behaviors typically rely on target mechanisms of change (e.g., psychological flexibility) rather than the overarching construct (Levin et al., 2017). Thus, finding a couple-based treatment that is empirically supported, widely disseminatable, and cost-effective could be a formidable treatment option for reducing general pornography use, conflict, and perceptions of problematic use. However, to our knowledge, a couples-based intervention has yet to be leveraged to reduce pornography-related behaviors. Additionally, it is unclear how big of a problem pornography-related behaviors are for couples seeking help for their relationship.

The OurRelationship program is one of the most empirically supported, replicated, and cost-effective web-based relationship education programs for treating relationship distress (Doss et al., 2016, 2020; Hatch et al., 2022a, b; Roddy et al., 2018). It is a web-based self-help adaptation of Integrative Behavioral Couple Therapy and has shown to have small- to medium-sized effects on key areas of relationship well-being (e.g., communication, relationship satisfaction, and sexual functioning) regardless of demographic differences (Christensen et al., 2004; Doss et al., 2013; Georgia Salivar et al., 2020; Hatch et al., 2022a, b). Additionally, the OurRelationship programs have had significant, small-to-medium impacts on individual functioning including reductions in depression, anxiety, symptoms of insomnia, coparenting, as well as improvements in perceived health and exercise (Doss et al., 2016; Roddy et al., 2020). This pattern of findings suggests that the effects of the OurRelationship program may extend beyond relationship issues and may impact different areas of individual functioning including small-sized changes in pornography-related behaviors typically associated with relationship well-being. For instance, reducing relationship distress via the OurRelationship program may be associated with decreases in pornography-related behaviors typically associated with poorer relationship functioning (e.g., frequency and duration of pornography use), while increasing areas indicative of better relationship functioning (e.g., joint pornography use).

The Current Study

Thus, the current study has four aims. In the first aim, we examine whether couples seeking help for their relationship perceive pornography to be a problem in their relationship by examining how often they argue about pornography and whether they perceive their use to be problematic. Given the lack of research examining pornography use in couples seeking help for their relationship, no specific hypotheses were posited. Given the small-sized associations with relationship distress and pornography-related behaviors, in the second aim, we examine whether participation in the OurRelationship program is associated with change in individual, joint, and perceptions of a partner’s pornography consumption and whether these effects are moderated by gender. We expect that individual and perceptions of a partner’s pornography consumption will decrease, joint pornography consumption will increase, and that these effects will be moderated by gender. Given that males tend to view pornography more than females, we expect larger decreases for men in individual use and a partner’s perception of use. In the third aim of the current study, we examine whether participation in the OurRelationship program is associated with changes in arguments about individual, joint, and partner pornography use and whether these effects are moderated by gender. We expect that arguments about individual, joint, and partner pornography use will decrease. We do not expect these to differ reliably between men and women. Finally, in the fourth aim of the current study, we investigate whether participation in the OurRelationship program is associated with changes in perceptions of excessive individual pornography use. We expect that perceptions of excessive individual pornography use will decrease during participation in the online program and that these effects will be moderated by gender. Given that men tend to report greater levels of problematic individual pornography use (Grubbs et al., 2019), we expect that men will experience greater decreases in their perceptions of problematic use. A final exploratory moderator will also be explored throughout each of the previous aims—the moderating effect of the COVID-19 pandemic. As research in this area is still emerging and research is scant, no specific hypotheses were posited for this exploratory aim.

Method

Study procedures were approved by the Institutional Review Board at the University of Miami before the recruitment began. Study aims, hypotheses, and analyses for the current study were preregistered with the Open Science Framework (https://osf.io/jk5af/).

Participants

Couples (N = 314; 628 individuals) were randomly assigned to the OurRelationship program (OR; n = 156 couples; 312 individuals) or the OurRelationship Plus (OR+; n = 158 couples; 316 individuals) program. At baseline, participants were on average 34.03 (SD = 7.26) years old and had been together for an average of 7.26 (SD = 5.99) years. Most participants were in different-gender relationships (93.6%), and 6.4% of participants were in same-sex relationship. Most couples were married (62.1%), 19.1% were cohabiting, and 18.8% were engaged. A waitlist control group was not gathered due to the grant focus on serving more low-income couples during this period.

Most participants identified themselves as non-Hispanic White (63.7%), with fewer Black (16.7%), White Hispanic (5.9%), multiracial (4.6%), American Indian or Alaska Native (2.2%), Asian (0.8%), Native Hawaiian or other Pacific Islander (0.6%), or other (5.4%). Moreover, 11.6% of participants identified themselves as Hispanic/Latino(a). Educational attainment was reported by 86.8% of the sample, among which 6.4% reported no degree attainment, 14.7% had a GED, 17.2% had a high school diploma, 10.5% had vocational training, 24.0% had some college, 9.7% had an associate’s degree, 14.1% had a bachelor’s degree, and 3.3% had a master’s or advanced degree. Less than half of the participants were working full time (40.9%), 35.7% were unemployed, with fewer working part time (11.0%), variable employment (5.6%), or seasonal work (6.8%). Annual income was reported by 96.3% of the sample. Specifically, the sample reported making less than $5000 (22.3%), $5000–$11,999 (14.9%), $12,000–15,999 (8.9%), $16,000–24,999 (16%), $25,000–34,999 (19.8%), $35,000–49,999 (13.6%), and $50,000 and more (4.5%) in the past 12 months.

Procedure

Participants were first directed to the OurRelationship.com website through paid advertisements on Google and Facebook, organic search, or word of mouth to complete an informed consent and an eligibility survey online. No recruitment materials indicated that the program would be used as a treatment for pornography-related behaviors. Eligible individuals were then provided with instructions to send the screening survey to their partner. Couples were eligible if they were between 18 and 64 years old; lived in the USA; married, engaged, or cohabiting with their partner for at least 6 months; had a household income within 200% of the federal poverty line; were able to read and write fluently in English; had access to high-speed internet; agreed to forgo other couple treatment in the next 6 months; and had not participated in the OurRelationship program in the past. Couples were deemed ineligible if either partner reported experiencing severe intimate partner violence in the past six months, which includes any of the following behaviors: (a) “choked me,” (b) “repeatedly punched me during a fight or beat me up,” (c) “threatened me with (or used) a gun or knife,” or (d) “physically forced me to have sex when I didn’t want to.” Couples were also ineligible if they reported feeling “quite afraid,” “very afraid,” or “extremely afraid” that the other partner would physically hurt him/her during an argument. Couples were referred to the National Domestic Violence Hotline if they were excluded due to intimate partner violence.

Eligible couples learned about the study protocol and provided verbal informed consent during the first call with their coach and were randomized to OR, an evidence-based online relationship program adapted from integrative behavioral couple therapy (Doss et al., 2016), or OR+, a revised version of OR. In both programs, couples were instructed to identify one or two of the biggest relationship problem(s) to focus on during the program. Over the course of the program, couples developed an object understanding of and solutions to the identified problem(s). The online material takes couples 6–8 h to complete and is accompanied by four 20-min telephone or video calls with a paraprofessional coach for a total of 80-min and appointments were scheduled on approximately a weekly basis (i.e., couples with no missed appointments could complete the program from baseline to post-survey in approximately four weeks). Most of the activities in OR and OR+ were the same and were completed individually except for the conjoint activity at the end of each phase where couples were instructed to come together for structured conversations. Couples in OR were instructed to complete the structured conversations before the coach call, whereas couples in OR+ were instructed to have the conversations during the coach call with the coach. We did not expect that the two versions of OR would result in different pornography outcomes. All couples were assessed at baseline (before treatment began) and again post-program completion.

Measures

Individual, Joint, and Perceptions of One’s Partner’s Pornography Consumption

Due to heavy item load in the larger study, individual, joint, and perception of pornography consumption were measured using iterations of the two highest yield single-items from the Frequency and Duration subscales of the Consumption of Pornography Scale—General (i.e., Hatch et al., 2020a, b). These subscales have been shown to be reliable (\({\omega }_{\mathrm{Frequency}}\) = 0.88; \({\omega }_{\mathrm{Duration}}\)= 0.89) and valid displaying evidence of convergent, divergent, and incremental validity (Hatch et al., 2020a, b). The frequency of individual pornography consumption was measured using the item “In the past 30 days, how often have you viewed pornography by yourself?”, the duration of individual pornography consumption was measured using the item “In the past 30 days, when you’ve viewed pornography by yourself, how long do you view it?” The frequency of joint pornography consumption was measured using the item “In the past 30 days, how often have you viewed pornography with your partner?”, the duration of joint pornography consumption was measured using the item “In the past 30 days, when you’ve viewed pornography with your partner, how long do you view it?” Next, perceptions of one’s partner’s frequency of pornography consumption were measured using the item “In the past 30 days, how often do you think your partner viewed pornography without you?” and the perception of the duration of one’s partner’s pornography consumption was measured using the item “In the past 30 days, when your partner viewed pornography by himself or herself, how much time do you believe they spent viewing it?”

Arguments Surrounding Self, Joint, and One’s Partner’s Consumption

Three items were developed for the purposes of this study to assess arguments surrounding self, joint, and one’s partner’s pornography consumption which were “In the past 30 days, my partner and I have had disagreements about the amount of time I spend viewing pornography,” “In the past 30 days, my partner and I have had disagreements about the amount of time my partner spends viewing pornography,” and “In the past 30 days, my partner and I have had disagreements about viewing pornography together,” respectively. Response options included “Not at all,” “Occasionally,” and “Frequently.”

Problematic Pornography Use

Problematic pornography use was measured using three five-point Likert-style items from the Compulsive Use subscale for self-perceived problematic pornography use (Maddock et al., 2019). Because internal consistency in the sample was poor (\({\alpha }_{\mathrm{pre}}\) = 0.53; \({\alpha }_{\mathrm{pre}}\) = 0.48), each item was modeled independently. Additionally, because each item resulted in the same outcome, in the current study, we report on only one of the items for brevity. The specific item is “I view pornography too often” as we interpreted this measure to be the most face valid.

Gender

Gender was assessed using the item “Are you male or female?” Participants were unable to select any other option.

Lifestyle Changes Due to the COVID-19 Pandemic

Lastly, participants responded to 18 yes-or-no questions regarding the extent to which the COVID-19 pandemic affected their life (e.g., financially, socially, and psychologically.). A composite variable was computed as the sum of each yes-or-no question and was treated as a continuous variable representing disruptions caused by the pandemic.

Data Analysis

Dyadic data could be argued to have three different levels; however, this analytical approach results in the model becoming saturated on the second level (Bolger & Laurenceau, 2013; Kenny & Kashy, 2011). Thus, program effects were modeled using two-level Bayesian multilevel models where assessments were nested within couples using the brms package in R (Bürkner, 2017). Recently, Bayesian analyses have been gaining traction due to some of their advantages over frequentist methods including seamlessly conducting informative hypothesis testing, the integration of prior information, and their focus on effect sizes to determine scientific relevance instead of p-values (Baldwin & Larson, 2017; McElreath, 2016). All models were estimated using four simultaneous chains with 1000 warm up iterations and 5000 total iterations. Models were evaluated for convergence based on \(\widehat{R}\) values, with aggregate \(\widehat{R}\) s exceeding 1.05 indicative of possible non-convergence; in the present study, no models demonstrated concerning degrees of non-convergence. Base models for each outcome included an indicator of time (0 = baseline, 1 = post-program completion), treatment type (0 = OR; 1 = OR+) and an interaction of time and the treatment type. As expected, treatment type and the interaction of time and the treatment type did not aid in the predictive capabilities of any of the models and therefore, the indicators were dropped from the proceeding analyses. To aid in model interpretation and parsimony within each aim, each moderator was examined in a separate model by including the main effect of each moderator and its interaction with time. A within-group Cohen’s d was calculated by dividing the regression coefficient of time by the baseline standard deviation of the corresponding outcome.

Finally, 17.5% of the data were missing at the post-assessment. Missing data were imputed using Blimp—a multilevel imputation program (Keller & Enders, 2018). Baseline variables that were related to study variables and missingness were identified, and 30 multiply imputed datasets were generated. Only 65% of the sample were asked questions on disruption caused by COVID-19 at baseline as these were added in the middle of the data collection shortly after the World Health Organization declared the COVID-19 outbreak a pandemic. Thus, the COVID subsample was used when examining the moderating effect of disruption caused by the pandemic.

Results

Aim 1 Do low-income couples seeking help for their relationship online report problematic pornography use? Not preregistered but included as a request from reviewers.

In the current sample, 6.1% of participants reported using pornography at least daily (NFemale = 5; NMale = 33). Furthermore, 9.2% of participants said they argue with their partner about their partner’s pornography use (NFemale = 52; NMale = 6), 6.7% of participants argued with their partner about their own pornography use (NFemale = 4; NMale = 38), and 5.3% of individuals mentioned that they spent time arguing with their partner about their joint pornography use as a couple (NFemale = 14; NMale = 19). A similar number of individuals (8.9%; NFemale = 9; NMale = 47) agreed that they viewed pornography too often.

Post-hoc Analyses

At the recommendation of reviewers, one of these subsets was further explored. Among the 6.7% (N = 42) of participants that argued with their partner about their own pornography use, 7 of these participants (all male) endorsed using pornography daily or more often, whereas the other 35 participants reported less than daily use. Further, 15 of these individuals reported arguing with their partners about their joint pornography use—all of these participants were male.

Aim 2 Is participation in OurRelationship associated with changes in individual, joint, and perceptions of one’s partner’s pornography consumption? Are these effects moderated by gender, or lifestyle changes due to the COVID-19 pandemic?

Participation in the OurRelationship program was associated with small but reliable decreases in the frequency (d = − 0.13, 95% CI [− 0.23, -0.03]) and duration of individual pornography consumption (d = -0.13, 95% CI [− 0.23, − 0.03]). Further, reliable reductions were also observed in the perception of the partner’s individual frequency (d = − 0.13, 95% CI [− 0.23, − 0.03]) and duration (d = − 0.12, [− 0.23, − 0.02]) of use. The estimates of changes in frequency (d = 0.04, 95% CI [− 0.04, 0.12]) and duration (d = 0.04, 95% CI [− 0.04, 0.13]) of joint pornography use were not reliably different from zero; the effect sizes were reliably different from the individual use pornography effects (i.e., the average effects were positive). Neither gender nor lifestyle changes due to the COVID-19 pandemic moderated the treatment effects (see Table 1).

Table 1 Effect sizes and credible intervals
Full size table

Aim 3 Is participation in the OurRelationship program associated with changes in arguments about pornography? Are these associations moderated by gender or lifestyle changes due to the COVID-19 pandemic?

Participation in the OurRelationship program was not associated with reliable changes in pornography-related conflict about individual use (d = − 0.06, 95% CI [− 0.04, 0.13]), a partner’s individual use (d = − 0.08, 95% CI [− 0.18, 0.03]), or joint use (d = − 0.04, 95% CI [− 0.14, 0.07]). Compared to men, women reported greater decreases than men in arguments about their partner’s pornography use during the OurRelationship program (d = − 0.34, 95% CI [− 0.54, − 0.16]; see Table 1). A test of simple main effects further revealed that this effect was also different from zero for women (d = − 0.23, 95% CI [− 0.39, − 0.11]) but not for men (d = 0.11, 95% CI [− 0.02, 0.24]). Lifestyle changes due to the COVID-19 pandemic did not moderate any of the treatment effects.

Aim 4 Is participation in the OurRelationship program associated with decreases in problematic pornography consumption? Are these effects moderated by gender or lifestyle changes due to the COVID-19 pandemic?

Participation in the OurRelationship program was not associated with a reduction in self-perceived problematic pornography use (d = 0.01, 95% CI [− 0.10, 0.12]). Neither gender nor lifestyle changes due to the COVID-19 pandemic moderated the treatment effects.

Post-hoc Analyses When moderated by daily use, do those who participate in the OurRelationship program see larger reductions in the frequency and duration of individual pornography consumption, arguments about consumption, or perceptions of problematic pornography consumption?

After running the previous analyses and noting the small-sized decreases in pornography-related behaviors compared with the other pornography treatments likely due to floor effects, we wondered whether especially high users of pornography (like those seeking help for their high levels of pornography use in other treatment studies) experienced larger reductions in the frequency and duration of their individual pornography use. Thus, we used these same models but added in a treatment-by-high pornography use moderator into the model. High pornography use was operationalized as daily use on our measure of frequency. Individuals who endorsed viewing pornography daily (originally coded 5 or 6 on the measure) were scored as 1 (N = 38 individuals), and all other individuals were scored as 0. Among low pornography users, participation in the OurRelationship program was associated with non-reliable change in the frequency (d = − 0.07, 95% CI [− 0.17, 0.02]) and duration (d = − 0.08, 95% CI [− 0.17, 0.02]) of pornography consumption, arguments with one’s partner about their individual pornography use (d = − 0.03, 95% CI [− 0.13, 0.06]), and perceptions of problematic pornography consumption (d = 0.03, 95% CI [− 0.06, 0.15]). However, among high-use participants, the OurRelationship program was associated with large-sized decreases in the frequency (d = − 0.89, 95% CI [− 1.27, − 0.51]) and duration (d = − 0.91, 95% CI [− 1.30, − 0.51]) of pornography consumption and small-sized decreases in perceptions of problematic pornography use (d = − 0.34, 95% CI [− 0.62, − 0.06]) compared with the low pornography users. Further, participation in OurRelationship among high users was associated with small-sized (but non-reliable) decreases in arguments with their partner about their individual pornography consumption (d = − 0.32, 95% CI [− 0.81, 0.19]; see Fig. 1).

Fig. 1
figure 1

Figure displaying pre- to post-single-group program effects moderated by daily (i.e., high) pornography use compared with the rest of the sample (i.e., low pornography use)

Full size image

Discussion

Using a large sample of low-income relationally distressed couples in the USA, the current study sought to examine whether participation in a light-touch web-based relationship education program for relationship distress was associated with the small-sized changes in pornography-related behaviors that have typically been associated with relationship distress. Overall, problems with pornography appeared to be a rare occurrence among couples seeking help for their relationships. Couples rarely argued about their use, and only a few individuals perceived their use to be problematic.

Changes in Frequency and Duration of Pornography Use

Consistent with the literature and our expectations, couples in the program saw small (but reliable) decreases in their individual frequency and duration of use and decreases in the perception of their partner’s individual frequency and duration of use. When analyses were limited to participants with high levels of pornography use (a post hoc group selected to emulate previous trials attempting to reduce pornography-related behaviors), individuals saw the expected large-sized decreases in frequency and duration of pornography use compared with the rest of the sample. Notably, joint pornography consumption—which has not been associated with poor relationship functioning in previous studies—remained stable. The fact that gender did not moderate these findings may indicate that although men view more pornography than women, the degree of change in pornography consumption due to changes in improvements in relationship well-being is equal across gender. These findings may offer indirect support to a growing area of research in the field that suggests that gender may not moderate pornography’s impact on relationship quality but that a dissimilarity between dyadic pornography use that may be driving poorer relationship functioning—an important moderation future research could explore (Kohut et al., 2021). Next, although changes in joint pornography use were not reliably different from zero, joint use did function differentially from individual pornography use in that the coefficient was positive. Those who view pornography together tend to have better relationship well-being, sexual satisfaction, dedication, and open communication than those who report viewing pornography alone (Kohut et al., 2018; Maddox et al., 2011; Rodrigues & Martins, 2021). Thus, it would then follow that improving relationship well-being via the OurRelationship program would different associations on joint pornography use than with individual pornography use.

Changes in Arguments About, and Levels of, Problematic Pornography Use

Contrary to our expectations, participation in the OurRelationship program was not reliably associated with changes in arguments about pornography use or changes in the perception of problematic pornography use for the average participant. However, those with high pornography use at baseline reported small-sized improvements in these domains. Also, gender and lifestyle changes due to the COVID-19 pandemic rarely moderated treatment effects. Indeed, the only treatment effect that gender moderated was that women appeared to see larger reductions in pornography-related conflict about their partner’s use when compared with men and the effect for women was different from zero. Consistent with the theoretical underpinnings of the OurRelationship program (i.e., Integrative Behavioral Couple Therapy), this might suggest that women may be accepting men’s pornography use; however, men’s frequency and duration of pornography use did change during the program, indicating that acceptance may have simultaneously occurred with their partner changing their frequency and duration of use. Another explanation might be that this sample has lower base rates of problematic pornography use and therefore fewer arguments about pornography with their partner compared with a sample of those seeking help specifically for problematic pornography (e.g., Bőthe et al., 2021; Crosby & Twohig, 2016; Twohig & Crosby, 2010). Indeed, base-rates of arguments surrounding pornography consumption within this sample is already low. The modal response to the arguments about pornography being that individuals argued with their partner “Not at all” and the modal response to problematic pornography item “I view pornography too often” was “Strongly Disagree.” Thus, perceptions of problematic use and arguments surrounding pornography consumption were infrequent to begin with, resulting in floor effects for the general sample. A final interpretation of these findings may be that the general absence of reductions in conflict coupled with the observed reductions in frequency and duration of use might speak to the private and self-regulating nature of pornography use (Cashwell et al., 2017; Katehakis, 2016). Because solitary use may be sensitive and stigmatized, the behavior may never be explicitly discussed within the relationship. However, this does not prevent previously established attitudes or cultural beliefs about pornography (e.g., moral incongruence) from contributing to relationship distress. Along this line of reasoning, some have characterized problematic pornography use as affect dysregulation (Cashwell et al., 2017; Katehakis, 2016) and more recent research appears to bolster this argument (Hatch et al., 2022a). Indeed, partners simply might not interpret or understand the arguments they are having as being proximally associated with their pornography-related behaviors due to the sensitive nature of the topic. Future research could test whether negative attitudes about pornography or if pornography is being used to regulate affect are changed during participation in the OurRelationship program.

Next, although there were no reliable decreases in arguments concerning pornography consumption generally, women reported reliable decreases in arguing about their perceptions of their partner’s pornography consumption. This is corroborated by past research which has shown that women’s reports of their partner’s problematic pornography use are related to poorer relationship satisfaction (Stewart & Szymanski, 2012; Szymanski et al., 2015). It would then follow that among different-gender couples, if women’s appraisal (i.e., perceptions) of their partner’s pornography use is accurate, then improving relationship satisfaction through online relationship education should reduce women’s perceptions about their partner’s pornography use, reducing arguments surrounding men’s use. Another explanation is that OurRelationship may foster honesty within the couple. This may encourage men to be more forthcoming regarding their pornography consumption, which past research has indicated may buffer against some of the negative effects associated with men’s use of pornography (Resch & Alderson, 2014).

Strengths, Weaknesses, and Directions for Future Research

This study has several strengths. First, although the effects were small, they were expected: participating in web-based relationship education is associated with small-sized decreases in the frequency and duration of pornography use. Rather than targeting pornography consumption directly, this study operated from the theoretical position that relationship distress is associated with small-increases in general pornography-related behaviors which is empirically consistent with the surrounding literature. Thus, reducing this distress would lead to these small-sized reductions. Indeed, the OurRelationship program is not designed to reduce the frequency or duration of pornography use; however, this does appear to be a spillover effect from the online program. Second, recent online studies attempting to reduce problematic pornography use have demonstrated a significant difficulty in retaining participants and getting them to participate in treatment with only 11% of participants in an online intervention completing all content. However, in the current study, 64.3% of couples completed all the online content. It is possible, therefore, that involving a romantic partner may operate to increase accountability to the program. Third, once moderated by high pornography use, the effect sizes became small to large and most were different than zero. Online relationship education may be a promising, indirect way to reduce pornography-related behaviors caused by relationship distress, but more research is needed to confirm these post hoc findings. Finally, to date, this is the largest treatment study trying to reduce pornography-related behaviors.

Despite these strengths, the current study was not without some shortcomings. First, the study lacked a control group, limiting our ability to infer cause. Additionally, without a control group, it was not possible to test the possibility that the OurRelationship program prevents increases in problematic pornography consumption. Thus, including a control group in future studies would be worthwhile. However, recent findings indicate that pornography consumption is stable over a six-month period, suggesting that these decreases in pornography-related behaviors are not likely due to chance alone (Esplin et al., 2021). Second, most participants were low-income individuals that rarely argued about pornography and had low levels of problematic pornography use which limits our ability to generalize to other populations. However, when the OurRelationship program has not had an income requirement, effect sizes have been larger by approximately a small-sized effect (i.e., d = 0.10–0.20; Doss et al., 2016, 2020; Roddy et al., 2020). Thus, recruiting a sample that might include higher-income participants, we might see a small-sized increase in the effect sizes. Third, contrary to expectations, we did not reduce arguments about, or perceptions of, problematic use in the whole sample. However, once moderated by high baseline levels of pornography consumption (i.e., daily), we saw that program participation was associated with small- to large-sized reductions in pornography-related behaviors. Fourth, due to the limited number of questions that were able to be asked in the trial, we were only able to use single-item measures of many of the constructs. However, these items came from scales that have been shown to be reliable, valid, and reproducible even between vastly different populations (Esplin et al., 2022; Hatch et al., 2020a, b). Additionally, these one-item measures revealed associations consistent with what is currently found in the literature when assessing general pornography consumption. Therefore, we expect the findings from these one-item measures to be valid. Fifth, participants were only able to identify as either male or female. Assessing gender more broadly in this field that has typically assessed gender as operating in a binary fashion would likely be a fruitful endeavor for future research. Sixth, we were unable to determine whether pornography use was one of the core issues in the relationship or the nature of pornography-related conflict in the current study. Given the growing popularity of the moral incongruence model, we also did not assess for levels of religiosity—a theoretically relevant moderator of these effects. Exploring the nature of these arguments in a qualitative fashion (and whether they were related to moral incongruence) would be an interesting pursuit for future research. Finally, we did not utilize traditional measures of relationship distress in the current study. While we have seen that participation in the current version of the program did predict decreases in relationship distress in other manuscripts, these analyses were excluded in the current study. However, other studies using this same sample have reported medium-sized decreases in relationship distress (d = 0.76; Le et al., 2022). Determining whether these effects might be attributable to decreases in relationship satisfaction would be a worthwhile pursuit for future research.

Clinical Implications

The current study fits in within the purview of the current literature—third-wave cognitive behavioral therapies (i.e., Acceptance and Commitment Therapy and Integrative Behavioral Couple Therapy) targeting underlying mechanisms seem to change pornography-related behavior. However, there are three unique takeaways from this study that are worthy of discussion. First, for clients that are involved in a relationship with significant amount of distress, targeting relationship distress in treatment appears to reduce the frequency and duration that an individual views pornography. Indeed, if pornography-related behaviors are a function of relationship distress, it may dissipate once it is gone. Second, involving a romantic partner in treatment could be worthwhile especially when considering adherence to the underlying treatment model. Finally, in the overall sample, the effect sizes in the current study are much smaller than those reported in previous individual treatments that have been tailored to treat problematic pornography use in the general population; however, we would caution clinicians from assuming that individual treatments are superior to those involving couples especially if the post hoc findings are confirmed. The OurRelationship program was not aimed at reducing pornography-related behaviors, and the intervention was brief. Even with these important constraints, program participation was associated with decreases several pornography-related behaviors and adherence was high. Thus, more research into couple interventions for reducing pornography use is indicated.

Conclusions

This study sought to investigate how participation in the OurRelationship program was related to individual pornography consumption, joint pornography consumption, perceptions of one’s partner’s pornography consumption, arguments regarding pornography consumption, the perceptions of problematic pornography consumption, and whether these were moderated by gender or lifestyle changes due to COVID-19. We found that couples participating in the OurRelationship program reported changes in individual pornography consumption and perceptions of one’s partner’s pornography use overtime. In addition, women reported reliable decreases in arguments about pornography consumption compared with men and women’s decreases in these arguments were different from zero. These findings suggest that pornography consumption that is a symptom of relationship distress may be reduced by web-based relationship education. Future research should seek to incorporate a control group which would strengthen the evidence in support of the web-based relationship education to treat pornography-related behaviors.