Abstract
Couples who use pornography together (shared pornography use) and couples in which both members use pornography alone (concordant solitary pornography use) may have more sexual interests, preferences, and values in common than couples in which one couple member uses pornography and the other does not (discordant pornography use). From this perspective, the associations between pornography use, comfort in sexual communication, and relationship closeness should vary depending on patterns of pornography use within relationships. Several related pre-registered hypotheses were tested using an online cross-sectional survey of heterosexual dyads (N = 200) that were quota sampled to reflect the distribution of age and political affiliation of married American couples. Dyadic analysis indicated that participants who shared pornography use with their partner reported more open sexual communication and greater closeness than those that did not. Further, participants who were in relationships that were discordant in solitary pornography use reported more inhibited sexual communication and lower closeness than participants who were in relationships that were concordant in solitary pornography use. In fact, couples who were concordant in high-frequency solitary pornography use reported similar openness in sexual communication and closeness as couples in which neither partner used pornography alone. These findings have important implications for the field, because they reinforce the view that the links between pornography use and relationship health are largely a function of different patterns of pornography use within couples.
Although pornography is commonly used within relationships (Carroll, Busby, Willoughby, & Brown, 2016), and there is clear public concern about such use (Montgomery-Graham, Kohut, Fisher, & Campbell, 2015), evidence of specific relational benefits and harms of pornography use in the couple setting remains mixed. A major shortcoming in this area concerns the general failure to consider how pornography is used within relationships (Campbell & Kohut, 2017), despite evidence that the potential effects of pornography differ depending on whether it is used alone or together with a romantic partner (Maddox, Rhoades, & Markman, 2011). Such differences are reflected in research concerning pornography use and sexual communication, which suggests that any pornography use, but particularly shared pornography use, is associated with improved comfort in sexual communication (Daneback, Traeen, & Månsson, 2009; Grov, Gillespie, Royce, & Lever, 2011; Kohut, Fisher, & Campbell, 2017b). This association between pornography use and ease of sexual communication suggests a further connection between pornography use and perceived interpersonal closeness in relationships. On the basis of available evidence and relevant theory, we predicted that sexual communication would be less inhibited among couple members that were in relationships involving any pornography use. Our predictions for closeness were more circumscribed. While we anticipated that perceptions of closeness in relationships would be higher among people who used pornography together than among those that did not, we also hypothesized that closeness would be higher when both partners either used pornography alone or did not use pornography alone (i.e., concordant solitary pornography use) than when one partner used pornography alone and the other did not (i.e., discordant solitary pornography use).
Studying the context of pornography use within romantic relationships
Despite decades of empirical research, it remains difficult to generalize about the impact of pornography on romantic relationships. From one point of view, accumulating research findings, stemming largely from the study of heterosexual couples, suggest that pornography use may undermine sexual and relationship satisfaction (Morgan, 2011; Peter & Valkenburg, 2009; Stack, Wasserman, & Kern, 2004), exacerbate sexual dysfunctions (Park et al., 2016), and reduce love for, and attraction towards, romantic partners (Kenrick, Gutierres, & Goldberg, 1989; Weaver, Masland, & Zillmann, 1984; Zillmann & Bryant, 1988). At the same time, recent reviews of this literature have highlighted important conceptual and methodological limitations in this field (e.g., narrowly harm-focused research, inconsistent operationalizations of pornography use, and the reliance on non-dyadic samples to study relationship impacts; see Campbell & Kohut, 2017; Newstrom & Harris, 2016), suggesting the need to be cautious when it comes to the interpretation of research findings concerning pornography use and relationship dysfunction. This perspective is underscored by evidence that indicates that pornography users and their partners generally perceive more positive than negative effects of pornography use on their romantic relationships (Bridges, Bergner, & Hesson-McInnis, 2003; Grov et al., 2011; Hald & Malamuth, 2008; Kohut et al., 2017b; Rissel et al., 2016), as well as repeated null effect findings that challenge claims that pornography use is associated with lower sexual and relational satisfaction (Staley & Prause, 2013; Wilson & Abelson, 1973; Yucel & Gassanov, 2010), sexual dysfunctions (Landripet & Štulhofer, 2015; Prause & Pfaus, 2015), and less love for, and attraction to romantic partners (Amelang & Pielke, 1992; Balzarini, Dobson, Chin, & Campbell, 2016; see also Dermer & Pyszczynski, 1978).
While descriptive research has identified numerous potential positive and negative impacts of pornography use on relationships (Hald & Malamuth, 2008; Kohut et al., 2017b), the specific outcomes for each couple may depend on the way that pornography is used within the relationship (Campbell & Kohut, 2017; Willoughby, Carroll, Busby, & Brown, 2015). In their brief review of the existing literature, Campbell and Kohut (2017) argued that pornography use should be studied from an Antecedents-Context-Effects (ACE) perspective to clarify the ways that antecedents of pornography use (e.g., individual differences, life experiences, gender, etc.) and the context of pornography use (e.g., the frequency of use, who is using pornography within the relationship, whether or not use is solitary or shared, whether use is hidden or open, etc.) shape the ultimate effects of pornography use on couple relationships (e.g., sexual and relationship satisfaction).
Gender is an illustrative variable to consider from an ACE perspective. It is clearly established, for example, that pornography use is quite gendered, with more men reporting use and more frequent use than women (Carroll et al., 2016; Petersen & Hyde, 2010; Regnerus, Gordon, & Price, 2016). Within the ACE model, gender may be framed as an antecedent factor because it predisposes individuals to use or avoid pornography. However, within heterosexual relationships, gender can also be considered a context of use factor in this model. That is, the gender of the pornography consumer can be used to differentiate between different dyadic patterns of pornography use within heterosexual relationships (e.g., male pornography use vs. female pornography use) which may have unique correlates with various aspects of relationship functioning. To this point, most—though by no means all (see Muusses, Kerkhof, & Finkenauer, 2015)—studies of heterosexual dyads have reported negative associations between male pornography use and relationship outcomes, while also reporting positive or null associations between female pornography use and relationship outcomes (Bridges & Morokoff, 2011; Poulsen, Busby, & Galovan, 2013; Wright, Tokunaga, Kraus, & Klann, 2017; Yucel & Gassanov, 2010). From these lines of research, it appears that although women may be less predisposed to use pornography, when they do use it, their relationships may be particularly functional, or at least not dysfunctional. On the other hand, when men—who are more likely to consume pornography and consume pornography more frequently than women—use such materials, their relationships appear to be less functional.
While revealing, such conceptualizations and results are still limited because they adopt a simplistic characterization of dyadic patterns of pornography use within heterosexual relationships that only distinguishes between male and female pornography use. Patterns of pornography use within relationships, however, are more complex. Employing basic descriptive techniques, Kohut, Fisher, and Campbell (2017a) identified at least eight categorical patterns of pornography use within heterosexual dyads by considering the presence and absence of solitary pornography use by both partners within relationships alongside the presence and absence of shared pornography use. These patterns included non-use of pornography by both partners (1); solitary use by the man or woman (2 and 3); solitary use by both the man and woman (4); shared pornography use alongside solitary use by either the man or woman (5 and 6); shared pornography use alongside solitary use by both the man and woman (7); and shared pornography with no solitary use (8). Even this typology is an admitted simplification because it does not account for other context of use factors (e.g., hidden use, content of pornography used, etc.). Nevertheless, differentiating between shared pornography use and solitary pornography use is important, because shared pornography use among couple members has been found to be associated with more positive perceived impacts of pornography on the relationship (Grov et al., 2011) and better relationship functioning than solitary pornography use (Maddox et al., 2011).
Pornography use and sexual communication
Several lines of research suggest that specific patterns of pornography use within relationships may be particularly relevant for understanding the openness of sexual communication between partners (Daneback et al., 2009; Grov et al., 2011; Kohut et al., 2017b; Willoughby et al., 2015). For example, in an earlier study, Danish couples who were concordant in their pornography use (i.e., couples in which both partners used pornography) reported more open sexual communication than couples who were discordant in pornography use (i.e., one partner used pornography when the other did not), while those who were discordant in their use reported more open sexual communication than couples in which no pornography was used at all (Daneback et al., 2009). 1 Additionally, shared pornography use (i.e., both partners use pornography together) has also been found to be associated with perceived improvements in sexual communication (Grov et al., 2011). Finally, Kohut et al. (2017b), drawing from open-ended qualitative data, recently reported that many of their participants believed that using pornography had improved their sexual communication with their partner by making conversations about personal sexual likes and dislikes more open and honest. Participants further described how shared pornography use provided opportunities for couple members to discuss their sexual preferences with one another as they related to what they were viewing together.
Although these studies offer preliminary evidence that any pornography use, but particularly concordant solitary and shared pornography use, may be related to the openness of sexual communication, this evidence is also limited. Of particular concern, Daneback, Traeen, and Månsson (2009) simply asked each member of the relationship about the extent to which they used pornography and did not differentiate between solitary and shared pornography use. Instead, couples that were concordant for pornography use were assumed to represent couples who shared pornography together. Given Kohut et al.’s (2017a) findings, it is likely that a sizeable proportion of the couples in this study were concordant in solitary pornography use, but did not use pornography together. Consequently, it remains unclear whether Daneback et al.’s (2009) results best represent the correlates of shared pornography use, as the authors claim, or concordant solitary pornography use, or both.
Pornography use and couple closeness
If pornography is related to more open sexual communication, then people who use pornography should be more willing to in engage sexual self-disclosures (i.e., share personal sexual interests, needs, fulfillment, etc.) with their romantic partners. Provided that sexual disclosures are not met with negative affective reactions, rejection, or open hostility by the non-disclosing partner, the willingness to engage in such disclosures may be associated with more perceived closeness, because of the process of intimacy. Intimacy, as conceptually described by Reis and Patrick (1996), is a special form of closeness involving a transactional dyadic process in which partners disclose personal self-relevant information that leaves them vulnerable and open to evaluation by their partners. In cases where partners respond to disclosures with warm positive regard, the disclosing partner feels more understood, validated, and cared for (Reis & Patrick, 1996). This may be why past research has found that the degree of reported sexual self-disclosure is positively related to relationship satisfaction (Byers & Demmons, 1999).
When it comes to the issue of pornography use within relationships, it is plausible that shared pornography use may create the grounds for intimate exchanges and ultimately closeness among couples. For some couples, shared pornography use facilitates open sexual communication and affords opportunities to learn about each-others sexual preferences (Grov et al., 2011; Kohut et al., 2017b). Further, qualitative data suggest that some pornography users report perceptions of enhanced emotional connections, bonding, closeness and intimacy stemming from shared pornography use, and such perceptions are commonly reported alongside mention of improved openness in sexual communication (Kohut et al., 2017b). Additional support for the view that shared pornography use may enhance closeness can be found in research involving couples who share other non-sexual activity and entertainment preferences, because these dyad members appear to report greater liking for one another (Boer et al., 2011; Werner & Parmalee, 1978) than members of couples who do not share activity and entertainment preferences. Similarly, research has also demonstrated that joint non-sexual recreation is associated with increased relationship quality and longevity (Aron, Norman, Aron, McKenna, & Heyman, 2000; Hill, 1988).
In addition to the effects of shared pornography use, it is also possible that concordance in solitary pornography use may be associated with enhanced closeness compared to discordance in solitary pornography use. Consider that couple members who both use pornography alone may be more likely to be similar in sexual attitudes, erotophilia, and sex drive—which all correlate with pornography use (Baer, Kohut, & Fisher, 2015; Fisher, White, Byrne, & Kelley, 1988; Kohut, Baer, & Watts, 2016; Zillmann & Bryant, 1988)—than couple members who are discordant in pornography use. Independent of specific occasions of shared pornography use, these similarities may foster a relational environment where sexual self-disclosures between couple members are met with tolerance, acceptance, or even warm positive regard. Compared to a relational context where partners are discordant in the pornography use, such responses should further feelings of acceptance, understanding, and closeness.
The current study
In the current study, we sought to replicate, clarify, and extend Daneback et al.’s (2009) findings with a method that differentiated between shared pornography use and concordant solitary pornography use. In this case, we measured each partners’ frequencies of solitary pornography use separately from their frequencies of shared pornography use with continuous scales, which captured Kohut et al.’s (2017a) eight dyadic patterns of pornography use typology with a dimensional approach. Given previous empirical evidence and the conceptual logic we have advanced, we hypothesized that the frequency of any pornography use, but particularly shared pornography use and concordant solitary pornography use (compared to concordant non-use), should be associated with more open sexual communication relative to the non-use of pornography. Based on further conceptual, logical, and empirical connections, we also predicted that concordance in solitary pornography use (or non-use of pornography) should be associated with enhanced closeness relative to discordance in pornography use. We further hypothesized that the frequency of shared pornography use should be independently associated with higher closeness. We did not have any predictions regarding gender because we could not articulate a clear theoretical reason to believe that the associations we outlined above would be further moderated by gender. However, given the gender-specific associations that have been reported in the past research involving pornography use (Wright et al., 2017), we examined the role of gender in exploratory analyses.
These hypotheses concerning the unique associations between different dyadic patterns of pornography use and open sexual communication and closeness were preregistered (see https://osf.io/kfb4z/) and tested with an online cross-sectional survey of heterosexual American couples who were preselected to match the distribution of age and political affiliation of married Americans. This study advances knowledge by examining more nuanced associations between specific dyadic patterns of pornography use and openness of sexual communication and by providing the first tests of the associations between patterns of pornography use and closeness within heterosexual couples.
Method
Participants
In an effort to recruit a more diverse sample than is typical in pornography research—which primarily relies on convenience sampling (Newstrom & Harris, 2016)—American couples (eligibility criteria required relationships of at least 6 months duration) were quota sampled to approximate the distribution of age and political affiliation of married couples in the United States. 2 To simplify this process, we estimated the distribution of age and political affiliation for married women (between the ages of 25 and 54) in the United States (the estimates were based on 2014 data from General Social Survey; Smith, Marsden, Hout, & Kim, 2014), and sampled women who could recruit their male relationship partners. To this end, a panel service (Qualtrics LLC) was contracted to recruit couples by advertising the study to female panel members with documented relationships, from which we recorded N = 2005 unique study initiations. Although women were targeted with recruitment advertisements, a small proportion of men appear to have responded to the invitation e-mail directed to their partners. When these initial participants completed the survey (N = 518; n = 469 women; n = 49 men), they were asked to invite their romantic relationship partners to participate through an e-mail system. Participants’ low invitation referral rate reflected two factors. Many participants who initiated the study did not continue beyond the Letter of Information but were nonetheless counted as “study initiators.” The remaining participants who did not complete the survey were removed from the study during demographic screening because they failed to meet increasingly stringent eligibility criteria imposed on women as recruitment progressed. 3 Of the N = 518 participants who were asked to invite their partners to the survey, n = 244 (47.10%) were excluded from analysis because their partner did not complete the survey, n = 55 (10.62%) were excluded because they did not indicate a partner who could be contacted for data, n = 10 (2.13%) were excluded because they were in a same-sex relationship, and a further n = 9 (1.74%) women were excluded for data quality reasons (e.g., n = 6 failed one of three attention checks, n = 2 provided identical extreme responses across oppositely coded items within the same scale, and n = 1 repeated the survey more than once with non-identical responses). When participants were excluded from the target sample, quota targets were adjusted and sampling continued until all quota targets were met.
To assess sample bias, demographic and pornography use information for the first couple member to complete survey (both men and women) were compared between the excluded sample and the retained sample. The retained participants were in significantly longer relationships (M retained = 179.25 months; M excluded = 153.90), t(512) = 2.61, p = .009, and were significantly more likely to be married (retained: 95.50%; excluded 90.88%), χ2(1) = 3.85, p = .050. The retained and excluded samples did not differ in age, pornography use, race, religious affiliation, political affiliation, or gender. Demographic information for the heterosexual dyads that were retained for use in this study (N = 200) can be found in Table 1.
Demographic information for N = 200 heterosexual relationships.
Note: M: mean; SD: standard deviation.
Materials and procedure
Following an online informed consent procedure, participants completed demographic items and measures of relationship satisfaction, sexual satisfaction, closeness, sexual communication, attachment orientation, and pornography use, in that order. Participants were debriefed after completing the survey and were given token compensation. Only the measures of sexual communication, closeness, and select items concerning pornography use were analyzed in this study. The materials and procedure were reviewed and approved by the local research ethics board before study initiation.
Sexual communication
The 4-item short version of the Dyadic Sexual Communication Scale (Catania, 2011) was used to assess participants’ perceptions of the openness of sexual communication in their relationships. Participants were asked to indicate the extent that they agreed or disagreed with four items such as, “Some sexual matters are too upsetting to discuss with my sexual partner” using 6-point Likert-like scales ranging from disagree strongly to agree strongly (Cronbach’s α = .70). Items were mean averaged with reverse coding where appropriate to create scores that ranged from 1 (inhibited sexual communication) to 6 (open sexual communication). Validation information can be found in Catania (2011).
Closeness
The Inclusion of Other in the Self Scale (IOSS; Aron, Aron, & Smollan, 1992) was used to measure interpersonal closeness between couple members. This widely used single-item pictorial scale measures perceived closeness by asking participants to indicate which of a series of partially overlapping circles best describes their relationships. Scores on this scale ranged from 1 (“most distant”) to 7 (“most close”). The IOSS has demonstrated alternate form and test–retest reliability as well as evidence of convergent, divergent, and criterion validity and has moderate correlations with measures of intimacy (Aron et al., 1992).
Pornography use
Willoughby and Busby (2015) have previously recommended that attempts to measure pornography use should define “pornography” for their participants. Guided by Kohut’s (2014) findings concerning lay conceptualizations of pornography, pornography use was clearly defined for participants as, “intentionally looking at, reading, or listening to: (a) pictures or videos of nude individuals, (b) pictures or videos in which people are having sex, or (c) written or audio material that describes nude individuals, or people having sex.” These instructions also went on to explicitly exclude sexually interactive online and offline behaviors from the definition of pornography use.
Frequency of solitary pornography use was assessed with a single item, “How frequently do you use pornography while alone (i.e. without your partner)?” and participants were provided with eight response options including (1)—Never; (2)—Almost Never; (3)—Less than Once a Month; (4)—1–3 Times Per Month; 5—1–2 Times Per Week; 6—3–4 Times Per Week; 7—About Once a Day; 8—More than Once a Day. The same approach was used to assess shared pornography use (“How frequently do you use pornography together with your partner?”). Independently reported frequencies of shared pornography use were mean averaged for use in the analyses outlined below.
Planned analyses
We analyzed the associations between solitary pornography use, shared pornography use, and sexual communication/closeness separately using a linear mixed modeling approach to the Actor-Partner Interdependence Model (Kenny, Kashy, & Cook, 2006). These models included couple members’ independently reported frequencies of solitary pornography use as well as their mean averaged reports of their frequency of shared pornography use. Specifically, openness in sexual communication and closeness were predicted with participants’ own solitary pornography use (i.e., actors’ solitary pornography use), their partners’ solitary pornography use, the interaction between couple members’ solitary pornography use, each couples’ shared pornography use, and gender. 4 No random effects were included in these models. Residuals were estimated by treating gender as a repeated-measure nested within dyads. All pornography use measures were standardized for centering purposes. Note that in these models, all participants were used as both “actors” and “partners.”
In the context of these models, support for our predictions concerning sexual communication would be indicated by positive associations with actors’ solitary pornography use, partners’ solitary pornography use, and shared pornography use. Although an interaction term was included in the model predicting openness in sexual communication to test for synergy between actors’ and partners’ solitary pornography use, the significance of this predictor was not essential for testing the hypothesis, because positive associations for both actors’ and partners’ solitary pornography use would indicate that concordant solitary pornography users report more open sexual communication than concordant non-users of pornography. With respect to closeness, a positive interaction between actors’ and partners’ solitary pornography use would indicate that concordant solitary pornography users (and non-users) report more closeness than discordant pornography users. Therefore, support for our predictions concerning closeness required a significant positive interaction between actors’ and partners’ solitary pornography use, as well as significant positive association for shared pornography use.
We did not expect interactions between gender and the other components of these models or interactions between altogether different components within these models (e.g., actor by shared used interaction). However, after the initial models were tested, we considered all other possible two-way and three-way interactions between predictors in subsequent steps, observing changes in -2 log likelihood ratio statistics. Models were estimated with maximum likelihood to allow us to test changes in model fit between nested models. The data and syntax for these analyses (https://osf.io/fe4cw/) can be found one the Open Science Framework.
Results
Summary statistics for the focal measures under study can be found in Table 2. Notable positive correlations were apparent between partners’ reports of openness in sexual communication (r = .66, p < .01), closeness (r = .64, p < .01), frequency of solitary pornography use (r = .38, p < .01), and frequency of shared pornography use (r = .76, p < .01). All independent and dependent variables were subjected to gender comparisons using paired t-tests. Significant differences were found for reports of closeness, t(200) = 2.41, p = .017, d = .17, and solitary pornography use, t(200) = 2.41, p = .000, d = 61. Men reported higher closeness (mean (M) = 5.61; standard deviation (SD) = 1.49) and higher frequencies of solitary porn use (M = 3.01; SD = 2.01), than women (M closeness = 5.39; SD closeness = 1.61; M porn. = 1.77; SD porn. = 1.19).
Summary of the correlations, means, and standard deviations of the focal variables.
Note. M: mean; SD: standard deviation.aMen report significantly higher closeness than women, p < .05.
bMen report significantly more frequent solitary pornography use than women, p < .001.
*Signifies significant correlations, p < .05; **Signifies significant correlations, p < .01
In the initial model, sexual communication was predicted by standardized frequencies of actors’ and partners’ solitary pornography use, their interaction, the standardized frequency of the average shared use, and gender. All further two-way and three-way interactions were considered (e.g., interactions with gender, interactions between solitary and shared use, etc.) but none significantly improved model fit. The association with gender did not contribute significantly to the prediction of sexual communication (p = .636), and gender was not found to interact with the other components of the model, so it was removed from the model without significantly worsening model fit, χ2(1) = 0.22, p = .639. In the retained model, there was a significant association between openness in sexual communication and shared pornography use, b = .22, t(200) = 2.76, p = .006, indicating that sexual communication was more open among participants with higher frequencies of shared pornography use. Although we anticipated positive associations, openness in sexual communication was negatively related to actors’ pornography use, b = −.12, t(260) = −2.12, p = .035, and unrelated to partners’ pornography use, b = −.07, t(279) = −1.26, p = .210. Importantly, the negative actor effect was further qualified by a significant positive interaction between actors’ and partners’ solitary pornography use, b = .16, t(200) = 2.30, p = .022. Tests of simple slopes indicated that when actors’ never used pornography alone, their sexual communication was more inhibited when their partners’ used pornography alone, z = −2.33, p = .020; however, when actors engaged in solitary pornography use more than once a day, their sexual communication was more open when their partners’ also used pornography alone, z = 1.97, p = .049.
At the suggestion of reviewers, we followed-up with a polynomial regression model to further examine concordance-related effects. To this end, squared actors’ and partners’ solitary pornography use were added to our existing model; however, these variables did not improve the fit of the model χ2(2) = 2.43, p = .297, and were not retained. 5 To further describe the nature of the concordance-related effects, a response surface analysis (RSA; see Shanock, Baran, Gentry, Pattison, & Heggestad, 2010) was applied to our retained model. 6 This analysis employs regression coefficients to determine how agreement and disagreement between two independent variables (e.g., actors’ and partners’ solitary pornography use) are related to a dependent variable. Of particular interest in this analysis, the a 4 coefficient, which reflects the curvature along the “line of incongruence,” can be used to determine whether the degree of discrepancy between actors’ and partners’ pornography use is related to openness of sexual communication. As can be seen in Figure 1(a) and 1(b), this analysis 7 indicated that sexual communication was more inhibited when partners were discordant in their solitary pornography use than when they were concordant in their solitary pornography use or non-use, a 4 = −.16, t(400) = 2.30, p = .022, a 3 = −.05, t(400) = −.89, p = .375. A further significant curvilinear effect along the “line of congruence,” a 2 = .16, t(400) = 2.30, p = .022, a 1 = .17, t(400) = .98, p = .328, indicated that sexual communication was somewhat more open among participants who were in couples that were concordant in either no solitary pornography use or high-frequency solitary pornography use than among participants who were in couples that were concordant in moderate-frequency solitary pornography use.
(a) Response surface analysis plot that depicts the associations between concordant–discordant solitary pornography and sexual communication by plotting openness in sexual communication (vertical axis) as a function actors’ (x-axis) and partners’ (y-axis) frequencies of solitary pornography use (in SD units) and their interaction for cases in which there is no shared pornography use between partners. (b) Depicts the same relationships for cases in which partners share pornography use more than once a day. Note that Figure 1(b) is essentially a vertical translation of Figure 1(a), reflecting more open sexual communication among participants who reported high-frequency shared pornography use compared to no shared pornography use. Further, in both plots, openness in sexual communication was higher among participants who were in couples who were concordant in no solitary pornography use (front-most corner) and among high-frequency concordant solitary users (back-most corner) than among moderate-frequency concordant users (intersection of curvilinear lines that join opposite corners). In turn, moderate-frequency concordant solitary pornography users reported more open sexual communication than participants who were in couples where one member reported a high frequency of solitary pornography use while the other reported no solitary pornography use (left- and right-most corners).
Closeness was predicted using the same basic model and once again additional two-way and three-way interactions did not appear to improve the prediction. Gender also failed to contribute significantly to the prediction of closeness (p = .138), and it was not found to interact with the other components of the model, so it was removed from the model without significantly worsening model fit, χ2(1) = 2.20, p = .138. In this retained model, as predicted, there was a significant association between closeness and shared pornography use, b = .33, t(200) = 3.11, p = .002, such that closeness was higher with increasing shared pornography use. Similar to the analysis involving sexual communication, closeness was negatively related to actors’ solitary pornography use, b = −.17, t(268) = −2.23, p = .027, but also negatively related to partners’ solitary pornography use, b = −.34, t(282) = −4.29, p = .000. Consistent with our hypothesis concerning concordance–discordance in solitary pornography use, these negative actor and partner effects were qualified by a significant positive interaction between actors’ and partners’ solitary pornography use, b = .22, t(200) = 2.36, p = .019. Tests of simple slopes indicated that when actors’ never used pornography alone, they reported lower closeness when their partners’ used pornography alone, z = −4.43, p < .001; however, when actors engaged in solitary pornography use more than once a day, their reported closeness was not related to their partners’ solitary pornography use, z = 1.185, p = .236.
We again followed-up with a polynomial regression model by adding squared actors’ and partners’ solitary pornography use to this model. These variables did not improve the fit of the model χ2(2) = 0.58, p = .748, so they were not retained. RSA was applied to our retained model to further describe the concordance-related effects (see Figure 2(a) and 2(b)). The results of this analysis 8 indicated that closeness was lower among discordant solitary pornography users, a 4 = −.22, t(400) = 2.37, p = .018, particularly when an actor’s use was lower than their partner’s, a 3 = .17, t(400) = 2.25, p = .025, than among concordant pornography users or non-users. Further, closeness was also lower among participants who were in couples with concordant moderate-frequency solitary pornography use compared to couples who were concordant in no solitary pornography use or high-frequency solitary use of pornography, a 2 = .22, t(400) = 2.37, p = .018, a 1 = −.01, t(400) = −.04, p = .965.
(a) Response surface analysis plot that depicts the associations between concordant–discordant solitary pornography use and interpersonal closeness by plotting openness in sexual communication (vertical axis) as a function actors’ (x-axis) and partners’ (y-axis) frequencies of solitary pornography use (in SD units) and their interaction for cases in which there is no shared pornography use between partners. (b) Depicts the same relationships for cases in which partners share pornography use more than once a day. Once again, Figure 2(b) is essentially a vertical translation of Figure 2(a), reflecting more closeness among participants who report high-frequency shared pornography use compared to no shared pornography use. Further, in both plots, closeness was higher among participants who were in couples who were concordant in no solitary pornography use (front-most corner) and among high-frequency concordant solitary users (back-most corner) than among moderate-frequency concordant users (intersection of curvilinear lines that join opposite corners). In turn, these moderate-frequency concordant solitary pornography users reported higher closeness than participants who were in couples where one member reported a high frequency of solitary pornography use while the other reported no solitary pornography (left- and right-most corners). In this case, closeness was especially low among participants who reported no solitary pornography use who had a partner that reported high-frequency solitary pornography use (left-most corner).
Discussion
This study sought to examine the associations between solitary pornography use, shared pornography use, sexual communication, and closeness, in a sample of 200 heterosexual American couples. The results confirmed our predictions about shared pornography use by indicating that the frequency of shared pornography use was positively associated with both the openness of sexual communication and closeness. Our hypotheses regarding solitary pornography use were only partially supported. Sexual communication was more inhibited, and closeness was lower, among both solitary pornography users and their partners when one partner used pornography alone and the other did not. However, the reported openness of sexual communication and closeness was similar between participants who were in couples where both partners used pornography frequently as well as participants who were in couples where neither partner used pornography.
Initially, we expected that any pornography use should be associated with more uninhibited sexual communication; however, this was clearly not the case. Although unexpected, our finding that sexual communication was more inhibited among couples that were discordant in solitary pornography use is consistent with recent evidence that dyad-level communication quality may be negatively associated with differences between partners’ reported frequencies of pornography use (Willoughby et al., 2015). We were also admittedly surprised to find that concordant non-users of pornography reported such high levels of openness of sexual communication given pornography’s tendency to reduce sexual anxiety (Morrison, Bearden, Harriman, Morrison, & Ellis, 2004; Wincze & Caird, 1976), and improve tolerance of sexual variety (Wright & Randall, 2014; Zillmann & Bryant, 1988), in addition to pornography’s known associations with erotophilia (Fisher et al., 1988). In retrospect, however, it appears possible that our rationale for the associations between closeness and concordance in solitary pornography use and non-use may also apply to sexual communication. If partners believe that they share sexual likes, or even dislikes (e.g., pornography), they may feel more comfortable with sexual self-disclosures than if they believe that their sexual interests are dissimilar. In fact, couple members’ intolerance of pornography may even be a subject of open sexual communication in some couples. This may be why concordance in no solitary pornography use was associated with such high openness in sexual communication. This explanation also further clarifies the similar pattern of results across sexual communication and closeness. If partners are more willing to self-disclose sexually, it should enhance the process of intimacy and feelings of closeness between partners.
These concordance-related findings bring to mind the literature concerning similarity and attraction. It is relatively well established that similarity in attitudes (Byrne, 1971; Montoya & Horton, 2013), activities, and entertainment preferences (Boer et al., 2011; Werner & Parmalee, 1978) are related to higher intra-dyadic liking. From this perspective, concordance/discordance in pornography use may either signal an additional dimension on which partners are similar or different, or given the correlations between pornography use and sexual attitudes, erotophilia, and sex drive (Baer et al., 2015; Fisher et al., 1988; Kohut et al., 2016; Zillmann & Bryant, 1988), may indicate a host of other dimensions on which partners are similar or different. In either case, similarity or dissimilarity in pornography use may be associated with a range of relationship functioning variables—including sexual communication and closeness—even if pornography itself has no direct causal impact on such endpoints.
It is tempting to believe that the correlates of shared and concordant solitary pornography use that are reported here are driven by the same theoretical mechanisms. Just like concordant solitary pornography use, partners who share pornography use may believe that they share more things in common with one another or believe that they can expect more sexual acceptance from one another because of their perceived similarity in sexual attitudes, sex drive, and so forth. On the other hand, the statistical independence of these two effects suggests that other factors may be at play as well. Of particular relevance to this issue, the incremental increase in the openness of sexual communication and closeness as a function of the frequency of shared pornography use may provide a clue about one specific additional mechanism that may further explain the correlates of shared pornography use. As we mentioned previously, qualitative research has suggested that shared pornography use may serve as a catalyst for sexual communication by providing opportunities for sexual discussions to take place (Kohut et al., 2017b). Consequently, it appears possible that the frequency of shared pornography use may be related to openness of sexual communication and closeness, in part, because it may simply increase the frequency of sexual discussions. Although we are clearly limited by the cross-sectional nature of the data, we are more inclined to believe that shared pornography use may have a causal impact on these outcomes than concordant solitary pornography use, though longitudinal and experimental research is needed to confirm this suspicion.
Despite the growing body of evidence indicating gender differences in the associations between pornography use and relationship functioning (Wright et al., 2017), our analyses did not indicate gender-specific associations. This may be because most of the research involving gender-specific associations has relied on non-dyadic samples (Campbell & Kohut, 2017), but even among studies that have recruited intact dyads, few or none have examined the interactional effects of each partners’ pornography use. In fact, when we re-analyzed our models without the interaction between actors’ and partners’ pornography use, we too found gender-specific associations in the current sample (e.g., male use was associated with lower closeness among men while female use was associated with higher closeness among women, etc.). Importantly, however, these gender-specific associations were entirely accounted for by the interaction between actors’ and partners’ pornography use, which in turn, was not moderated by gender. Why might this be? Because men are more likely to use pornography than women (Carroll et al., 2016), the probability that female users’ partners also use pornography is not the same as the probability that male users’ partners also use pornography. Indeed, in this sample, 82% of female solitary pornography users had a partner that also used pornography alone, while only 53% of our male solitary pornography users had a partner that used pornography alone. In other words, gender appears to be partially confounded with dyadic patterns of pornography use in heterosexual relationships, with female pornography users more likely to be in relationships with concordant solitary pornography use than male pornography users. These findings offer preliminary evidence that previously established gender-specific correlates of pornography use might be better conceptualized as the result of differing dyadic patterns of pornography use within relationships. If corroborated by future research, this perspective might also be usefully employed in the study of pornography users within same-gender relationships, where existing gender-specific correlates of pornography use may not apply.
Finally, our post hoc analyses indicated that the frequency of concordant solitary pornography use was related to openness of sexual communication and closeness in a curvilinear fashion. Specifically, openness of sexual communication and closeness were lower among moderate-frequency users than among non-users or high-frequency users. Importantly, these findings were not simply an artifact of our retained models, because the inclusion of quadratic actor and partner effects did not improve the models’ predictive utility. These curvilinear associations were not anticipated and it is difficult to understand why they occurred and what their interpretive value might be. One possibility is that these curvilinear patterns are simply the result of a sampling idiosyncrasy and further replications may find more sizeable quadratic effects that change the nature of these relationships. On the other hand, it is quite possible that these patterns are meaningful, and there is something fundamentally different about moderate-frequency concordant solitary pornography use compared to concordant non-use and high-frequency concordant solitary pornography use. At present, we are unsure what the difference might be, but we encourage others to speculate about this issue as we endeavor to resolve it ourselves. Independent of the reasons behind these patterns of associations, it is important to note that the existence of this pattern qualifies but does not undermine our interpretation of the comparative benefits of concordant solitary pornography use over discordant solitary pornography use. Although it is true that moderate-frequency concordant solitary pornography users in the current sample reported more inhibited sexual communication and lower closeness than concordant non-users and concordant high-frequency users, moderate-frequency concordant solitary pornography users still reported more open sexual communication and more closeness than discordant solitary pornography users and their partners.
As is typical, this research is subject to several limitations, including a relatively small sample size which may limit the generalizability of our findings, the single-item assessments of pornography use which may be attenuating the strength of some of the associations we present, and the correlational nature of the findings which makes it difficult to infer causal directions. Future research should seek to replicate these findings with additional and alternative measures of pornography use, sexual communication, and closeness and consider whether pornography use can help to establish intimacy when both partners use pornography (e.g., Is one partner’s sexual self-disclosure more likely to be met with more positive regard if both partners use pornography? Does the disclosing partner feel more understood, validated, and cared for?, etc.). Additionally, further work in this area should explore how patterns of pornography use are related to other indices of relationship functioning (e.g., sexual satisfaction, relationship satisfaction, commitment, relationship dissolution, etc.), because these associations are likely not limited to openness in sexual communication and closeness. Finally, longitudinal research examining associations between patterns of pornography use and relationship functioning variables should investigate whether patterns of pornography use predict stable differences in relationship functioning or change in relationship functioning over time. Such research can also explore potential bidirectional effects between pornography use and these outcomes.
Footnotes
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by an Insight Grant (PI: Dr W. Fisher) provided by the Social Sciences and Humanities Research Council.