Introduction

Advances in the technologies of media production, distribution, and access have led an increasing number of sexologically oriented social and behavioral scientists to speculate about the potential effects of pornography (Grubbs & Kraus, 2021; Wright, 2020). Following the social psychologists who pioneered the field (Donnerstein & Berkowitz, 1981; Malamuth & Check, 1985; Zillmann & Bryant, 1982), much of this research has been grounded in the concept of sexual scripting (SS) (Grubbs et al., 2019a, 2019b; Tokunaga et al., 2020). Sexual scripts are symbolically imparted guidelines for sexual behavior. They answer the questions “Who does what to whom in what kind of relationship, to what consequence?” (Gagnon & Simon, 2005, p. 206). From a SS perspective, pornography’s popularity and provision of sexual scripts make it an important source of social influence (Ezzell, 2014; Ward et al., 2016).

An emerging paradigm in the study of the impacts of pornography use is the moral incongruence (MI) approach (Grubbs & Perry, 2019; Grubbs et al., 2019a, 2019b). The MI approach attributes the influence of pornography to cognitive dissonance that arises when religious people consume a media form that their moral worldview prohibits.1 While SS research generally focuses on social learning (e.g., learning of sexually aggressive scripts) and MI research generally focuses on clinical concerns (e.g., perceived addiction to pornography), recent MI hypothesizing has been extended to topical areas traditionally conceptualized around SS principles. For instance, Perry (2018a) and Perry and Whitehead (2019) suggest that MI-induced sexual guilt, not the learning of unrealistic sexual scripts for self and partner (e.g., Wright et al., 2017; Zillmann & Bryant, 1988), may be responsible for associations between pornography use and sexual and relational dissatisfaction.

This research, alongside a close inspection of MI tenets, suggests that the MI and SS approaches can also lead to differential predictions about the nature of pornography use, religiosity, and attitude change. Using nationally representative data gathered from multiple samples of men and women in the USA across several decades, this study explores MI- and SS-derived predictions for the interplay between pornography use, religiosity, and permissive sexual attitudes. The association between pornography and sexual permissiveness is one of the most commonly studied attitude change contexts among social–psychologically oriented pornography scholars (Tokunaga et al., 2019).

Several notes about the present study’s paradigmatic orientation, data, and theorizing are in order, before proceeding. Firstly, this paper adopts an effects perspective for conceptual, organizational, and explanatory purposes. But it must be emphasized that the term “pornography effects” as used herein refers only to a field of study and interpretive orientation toward linkages between pornography use and sociosexual beliefs, attitudes, and behaviors (for an anthology of the broader field of media effects, see Oliver et al., 2020). My own views on debates about whether there is sufficient evidence to conclude that sexual media (including pornography) affect the attitudes of at least some consumers have been expressed elsewhere (e.g., Wright, 2011, 2020). The sole purpose of the present study is to examine, from an effects perspective, how data on pornography use and permissive sexual attitudes map onto MI- and SS-derived predictions. Secondly, it is readily acknowledged that the cross-sectional data employed in the present study are not sufficient in and of themselves for making causal claims. Finally, while the nature of the interplay between pornography use and religiosity on permissive sexual attitudes put forth by the present study is explicitly identified by the SS approach (e.g., Wright, 2011), the present study’s MI-derived predictions are an extrapolation from MI writings rather than a formal axiom put forth by the originators of the MI perspective (e.g., Grubbs & Perry, 2019).

Sexual Script Approach

The SS perspective is the predominant approach in the pornography effects literature for explaining correlations between the use of pornography and the expression of attitudes consistent with themes for sex depicted in pornography (Wright, 2020). The specific theoretical articulation most often applied to scripting analyses of pornography is Wright’s (2011) sexual script acquisition, activation, application model of sexual media socialization (3AM) (e.g., Braithwaite et al., 2015; Elliott, 2017; Ezzell, 2014; Foubert & Bridges, 2017; Guizzo & Cadinu, 2020; Huntington et al., 2021; Johnson et al., 2019; Rasmussen & Bierman, 2018; Rothman et al., 2021; Sun et al., 2015; Tomaszewska & Krahe, 2018; Vandenbosch et al., 2018; Ward et al., 2016). The 3AM is a critical synthesis and integration of a variety of mass communication, information processing, and behavioral theories, as well as conceptual and empirical work not tied to any particular named theory (Wright, 2020).

The model asserts that exposure to sexual media can result in the learning of novel sexual scripts [sexual script acquisition], the priming of already acquired sexual scripts [sexual script activation], and the utilization of sexual scripts to guide judgments about one’s own sexual behavior or evaluations of other people’s sexual behavior [sexual script application]. The model further proposes that scripting effects can be specific or abstract. A specific scripting effect occurs when the observation of a particular behavior affects the viewer’s attitudes about that specific behavior. An abstract scripting effect occurs when the viewer deduces the general principle or belief guiding the modeled behavior and then references this information to guide attitudes about different, but conceptually related, behaviors. In this way, the model predicts that viewing specific sexual behaviors can lead to modified cognitions of not only those behaviors but also of behaviors that were not depicted. In sum, by depicting particular sexual relationships and acts, at varying frequencies, and with varying consequences, pornography can provide consumers with cues as to what is sexually normative, appropriate, and desirable.

A “permissive” approach to sex is one that diverges from traditional sexual values (Lo et al., 1999). While there has been considerable debate about the extent and nature of certain depictions in pornography, such as sexual violence and the objectification of women (Bridges et al., 2010; McKee, 2005), there is consensus that sexual scripts in pornography diverge from traditional sexual values prohibiting behaviors such as extramarital, premarital, teenage, and same-sex sex (Bishop, 2015; Jensen, 2010, Linz & Malamuth, 1993; Perry & Schleifer, 2018; Peter & Valkenburg, 2006; Rasmussen, 2016; Vannier et al., 2014; Wright, 2013; Zillmann, 2000). From a SS perspective, attitude change toward acceptance of sexual permissiveness is a consequence of pornography consumption. That people who consume pornography are more sexually permissive than people who do not consume pornography is one of the most consistent findings in the pornography effects literature (Tokunaga et al., 2019).

While the 3AM concurs that, on the average, pornography consumers will be more sexually permissive than nonconsumers (Wright, 2013; Wright et al., 2013), it predicts that incongruity between consumers’ existing sexual scripts and the sexual scripts they encounter in pornography increases the probability that the latter “will be deemed inappropriate and discarded” (Wright, 2011, p. 356). Consistent with research on religiosity and sexual permissiveness (Ahrold et al., 2011; Murray et al., 2007), the 3AM posits that devout individuals are more likely than nonreligious persons to possess sexual scripts discordant with pornography’s permissive approach to sex. Accordingly, the 3AM suggests that “because most religions emphasize relational and procreational sexual values, high levels of religiosity may discourage the internalization of sexual scripts provided by the mass media” (p. 353). While some sexual scripting effects may occur in regard to pornography and permissiveness among the religious, the degree of influence will be less so than for the nonreligious. In sum, from a 3AM perspective, religious pornography consumers will be more sexually permissive than religious persons who do not consume pornography, but less sexually permissive than nonreligious pornography consumers.

Moral Incongruence Approach

The MI approach to pornography effects has become increasingly prominent in recent years (Grubbs & Perry, 2019; Grubbs et al., 2019a, 2019b). The essence of the MI approach is straightforward—when religious people view pornography (a behavior discouraged by the majority of religious traditions; Boulton, 2008; Hald & Mulya, 2013; Musda, 2013; Sherkat & Ellison, 1997; Thomas, 2013; Tsuria, 2017), they experience guilt, shame, and distress, and this collection of negative self-perceptions and affect leads to certain undesirable outcomes. The most common undesirable outcome identified in empirical studies is a perceived addiction to pornography independent of any objective clinical markers of an out-of-control behavior. As Grubbs et al. (2019a, 2019b) explain, in religious populations, “it seems that failure to abstain from any pornography use may be interpreted as evidence of dysregulation and compulsivity, despite the objectively small amount of use” (p. 404). Depression is another undesirable outcome that has recently been linked to MI (Perry, 2018b).

While attitude change has not been a focal point of MI research, recent efforts have expanded the approach in an attempt to explain pornography effects typically explicated using attitude change principles derived from SS theorizing. As reviewed previously, Perry (2018a) and Perry and Whitehead (2019) have suggested that MI may explain why pornography use is associated with lower sexual and relational satisfaction, rather than SS, the most common explanation (Wright et al., 2017; Zillmann & Bryant, 1988). Moreover, virtually all MI papers use the word “dissonance” (or its variants) to succinctly summarize the cognitively incongruent state of the religious pornography user. That cognitive dissonance can produce attitude change is one of the oldest social psychological principles (Festinger, 1957; Petty & Brinol, 2010).

From a general perspective on dissonance reduction, a number of options are available to the religious pornography consumer. Most obviously, such a person could abandon their theological views in order to enjoy pornography without the burden of religious guilt (Perry, 2017; Perry & Hayward, 2017). But the concept of MI implies that the user’s faith maintains despite their consumption of pornography; without the maintenance of faith, moral incongruence would be theorized as transient and inconsequential rather than significantly intertwined over time with pornography use and perceived pornography problems (Grubbs et al., 2020). Adopting a pro-pornography stance seems similarly unlikely; changing one’s attitudes toward pornography could only reduce dissonance if coupled with a corresponding reduction in religiosity, but the preservation of faith in spite of pornography use is implied by the MI perspective. For example, Grubbs et al. (2018) predicted that baseline levels of religiosity-based moral disapproval of pornography (e.g., believing that viewing pornography is a “sin”) would predict perceived addiction to pornography one year later among pornography users. Hypothesizing that religiosity at present predicts an MI-related pornography outcome 12 months later presupposes the continued religiosity of the pornography consumer over time.

Conversely, it would seem to more be consistent with the MI perspective for a religious pornography consumer to reduce dissonance by increasing their adherence to other religious teachings on human sexuality. Because “most religions emphasize relational and procreational sexual values” (Wright, 2011, p. 353), this could mean, for example, becoming more antagonistic toward extramarital, premarital, teenage, and same-sex sex than would have been the case from their religious beliefs alone. Specifically, from an MI perspective on pornography use, religiosity, and dissonance reduction, it could be deduced that religious pornography consumers will be even less sexually permissive than religious persons who do not consume pornography.

To put it another way, religious pornography consumers may compensate for their use of pornography by “doubling down on their original opinions (thus becoming more extreme in the direction of their original opinions)” vis-à-vis the adoption of hardline, religiosity consistent attitudes toward nontraditional sexuality (see PytlikZillig et al., 2018, p. 81, for a general discussion of this attitudinal phenomenon). Although this prediction is derived from the present study’s interpretation of MI reasoning rather than a formal postulate put forth by the originators of MI theory, it is consistent with research and theorizing on “moral compensation” (Zhong et al., 2009) and “moral cleansing” (Tetlock et al., 2000). Theory and research findings in these traditions are consistent with the notion that religious people could become even more outspoken against permissive sexual behaviors as a way to compensate for and cleanse themselves symbolically of their morally transgressive pornography use (Green & Low, 2014; Jordan et al., 2011; Merritt et al., 2010; Sachdeva et al., 2009; Tetlock et al., 2000; Zhong & Liljenquist, 2006; Zhong et al., 2009).

Summary

A review of papers written from the SS and MI perspectives suggests the possibility of divergent predictions about the interplay between pornography, religiosity, and permissive sexual attitudes. From an SS perspective, pornography’s depiction of permissive sexuality as normative and rewarding leads to attitude change toward more acceptance of sexual permissiveness, but the preexisting traditional sexual scripts of the religious reduce the magnitude of this effect. Religious pornography consumers will be more sexually permissive than religious persons who do not consume pornography, but less sexually permissive than nonreligious pornography consumers, according to the SS approach. Conversely, the adoption of an MI perspective suggests the possibility that religious persons become more antagonistic toward permissive sexuality as a consequence of their pornography consumption, as a way to reduce the cognitive dissonance they feel for engaging in a sexual behavior prohibited by their faith tradition. It would be consistent with MI (but inconsistent with SS) if religious pornography consumers were less sexually permissive than religious persons who do not consume pornography. In sum, while the adoption of either an SS or MI approach would lead to the prediction that the religious will be less sexually permissive than the nonreligious, there is the potential for divergence in terms of the permissiveness of religious and nonreligious pornography consumers. To compare these differing perspectives, as well as to probe the permissiveness rank-ordering of the religious and nonreligious who do or do not consume pornography across all groupings, this study uses nationally representative data gathered from US adults across multiple decades via the General Social Survey (GSS).

Two research questions are posed. The first research question (RQ1) asks whether, across all available GSS data, the pattern of sexual permissiveness among religious nonviewers, religious viewers, nonreligious nonviewers, and nonreligious viewers is consistent with the MI or SS approaches. The second research question (RQ2) asks whether the results for the main (i.e., RQ1) analyses are paralleled when limited to post-Internet GSS years and after individual differences are included as covariates. It has often been suggested that pornography viewed online may have differential effects from traditionally delivered pornography (Grubbs et al., 2019a, 2019b; Wright et al., 2016) and that individual differences confound associations between pornography and permissive sexual attitudes (e.g., that associations are due to men being more permissive and more likely to consume pornography or to young people being more permissive and more likely to consume pornography (Wright, 2013; Wright et al., 2013).

Method

Subjects

The GSS is an ongoing, National Science Foundation-funded, full-probability, national, computer-assisted personal interview survey examining the social beliefs and behaviors of residence-inhabiting adults in the USA aged 18 or older (The General Social Survey, 2020). The GSS undergoes institutional review board review by the National Opinion Research Center at the University of Chicago. Because the GSS adds culturally significant and socially important questions over time, some core questions are included in every GSS survey, while other questions are featured in only some surveys. To avoid respondent fatigue, not all participants in a particular survey are asked all of that year’s questions. Further, in a given year, different sets of participants are asked different sets of questions. The questions of interest to the present inquiry had to do with pornography consumption, attendance at religious services, and attitudes toward teen, premarital, extramarital, and same-sex sex. Four meta-samples were created by merging the GSS cross-sectional samples across their pertinent measures and years (i.e., teen sex sample, premarital sex sample, extramarital sex sample, same-sex sex sample). Weight variable WTSSALL was applied per GSS user instructions (GSS Codebook, 2019).

Questions were collectively asked about pornography consumption, attendance at religious services, and attitudes toward teen sex in 1986, 1988, 1989, 1990, 1991, 1993, 1994, 1996, 1998, 2000, 2002, 2004, 2006, 2008, 2010, 2012, 2014, 2016, and 2018 (N = 6,421). Questions were collectively asked pornography consumption, attendance at religious services, and attitudes toward premarital sex in 1975, 1978, 1983, 1986, 1988, 1989, 1990, 1991, 1993, 1994, 1996, 1998, 2000, 2002, 2004, 2006, 2008, 2010, 2012, 2014, 2016, and 2018 (N = 8,312). Questions were collectively asked pornography consumption, attendance at religious services, and attitudes toward extramarital sex in 1973, 1976, 1980, 1984, 1987, 1988, 1989, 1990, 1991, 1993, 1994, 1996, 1998, 2000, 2002, 2004, 2006, 2008, 2010, 2012, 2014, 2016, and 2018 (N = 8,701). Questions were collectively asked pornography consumption, attendance at religious services, and attitudes toward same-sex sex in 1973, 1976, 1980, 1984, 1987, 1988, 1989, 1990, 1991, 1993, 1994, 1996, 1998, 2000, 2002, 2004, 2006, 2008, 2010, 2012, 2014, 2016, and 2018 (N = 8,445). Table 1 presents descriptive statistics for each sample’s age, political orientation, educational achievement, sex, race, and sexual orientation.

Table 1 Descriptive statistics
Full size table

Measures

Based on whether they had viewed a pornographic (i.e., “x-rated”) movie in the last year (no/yes) and attended religious services at least once a week or not at all (i.e., either did not attend religious services at all, or attended once a week or more), participants were classified as either religious nonviewers, religious viewers, nonreligious nonviewers, or nonreligious viewers. Attendance at religious services is frequently used to index religiosity, in general (Barkan & Greenwood, 2003; Maselko & Kubzansky, 2006; Sander, 2017), and in pornography research, in particular (see Grubbs & Perry, 2019; Grubbs et al., 2019a, 2019b). Further, using weekly attendance to identify the devout is a common tactic (Borgonovi, 2008; Doran & Price, 2014; Iannaccone & Everton, 2004; McCleary & Barro, 2006; Solt, 2014). The GSS assessment of pornography consumption has been shown to yield theoretically expected associations with numerous attitudinal, behavioral, and demographic variables (see, for example, Perry & Schleifer, 2018; Wright, 2012, 2015; Wright & Bae, 2013; Wright et al., 2014).

Sexual permissiveness was assessed via participants’ moral attitudes (1 = always wrong, 4 = not wrong at all) toward teen sex (i.e., sex between teens aged 14 and 16), premarital sex (i.e., sex between mixed sex unmarried adults), extramarital sex (i.e., sex with someone other than the marriage partner), and same-sex sex (i.e., sex between two adults of the same sex). Evidence for the validity of these items can be found in many previous works (Davis, 2004; DiMaggio et al., 1996; Eagly et al., 2004; Evans, 1997; Harding & Jencks, 2003; Hayduk et al., 1995; Kozloski, 2010; Petersen & Hyde, 2010; Treas, 2002). Descriptive statistics for the focal measures by sample are presented in Table 1.

Analytic Approach

Analyses unfolded in the following order. First, using all available data, ANOVAs were carried out for each sample and permissiveness criterion to determine which groups were more and less sexually permissive. Second, the ANOVA for same-sex attitudes was carried out again using only heterosexual-identified participants, as the initial analysis did not differentiate participants by sexual orientation (as indicated in Table 1, the GSS sexual orientation question was not introduced until 2008). Third, the initial ANOVAs were carried out again using data from 1996 onward only. Historians have identified 1995 as an important turning point for popular internet use (Campbell, 2015; Dominick et al., 2008), and it has been suggested that pornography viewed online may have differential effects from traditionally delivered pornography (Grubbs et al., 2019a, 2019b; Wright et al., 2016). Fourth, ANCOVAs were conducted to see whether the initial ANOVA results replicated after adjusting for participants’ age, political orientation, education, sex, and race, as these have been common covariates in MI and SS research (Perry, 2018b; Perry & Whitehead, 2019; Wright, 2013; Wright et al., 2014).2 Consistent with the recommendations of Levine and Hullett (2002), effect sizes are noted with eta squared. Consistent with the recommendations of Cumming (2009, 2014), confidence intervals are used to evaluate specific differences between groups.3

Results

RQ1. The ANOVA for teen sex attitudes was significant (F(3, 6641) = 461.49, p < 0.001, eta squared = 0.17). Examination of the group means (see Table 2) indicated that the rank-ordering of sexual permissiveness regarding teen sex, from least permissive to most permissive, was religious nonviewers (M = 1.18, SE = 0.014, 95% CI: 1.158, 1.211) < religious viewers (M = 1.46, SE = 0.038, 95% CI: 1.394, 1.541) < nonreligious nonviewers (M = 1.76, SE = 0.018, 95% CI: 1.733, 1.804) < nonreligious viewers (M = 2.17, SE = 0.027, 95% CI: 2.121, 2.226).

Table 2 Groups means and standard deviations
Full size table

The ANOVA for premarital sex attitudes was significant (F(3, 8701) = 1,275.69, p < 0.001, eta squared = 0.31). Examination of the group means (see Table 2) indicated that the rank-ordering of sexual permissiveness regarding premarital sex, from least permissive to most permissive, was religious nonviewers (M = 1.92, SE = 0.016, 95% CI: 1.897, 1.959) < religious viewers (M = 2.69, SE = 0.045, 95% CI: 2.606, 2.784) < nonreligious nonviewers (M = 3.30, SE = 0.022, 95% CI: 3.257, 3.343) < nonreligious viewers (M = 3.69, SE = 0.034, 95% CI: 3.628, 3.761).

The ANOVA for extramarital sex attitudes was significant (F(3, 9024) = 396.75, p < 0.001, eta squared = 0.12). Examination of the group means (see Table 2) indicated that the rank-ordering of sexual permissiveness regarding extramarital sex, from least permissive to most permissive, was religious nonviewers (M = 1.13, SE = 0.010, 95% CI: 1.118, 1.156) < religious viewers (M = 1.39, SE = 0.026, 95% CI: 1.340, 1.441) < nonreligious nonviewers (M = 1.51, SE = 0.014, 95% CI: 1.490, 1.546) < nonreligious viewers (M = 1.83, SE = 0.020, 95% CI: 1.791, 1.870).

The ANOVA for same-sex sex attitudes was significant (F(3, 8765) = 695.40, p < 0.001, eta squared = 0.19). Examination of the group means (Table 2) indicated that the rank-ordering of sexual permissiveness regarding same-sex sex, from least permissive to most permissive, was religious nonviewers (M = 1.38, SE = 0.016, 95% CI: 1.350, 1.414) < religious viewers (M = 1.70, SE = 0.044, 95% CI: 1.617, 1.791) < nonreligious nonviewers (M = 2.41, SE = 0.025, 95% CI: 2.367, 2.464) < nonreligious viewers (M = 2.80, SE = 0.035, 95% CI: 2.733, 2.870).

To summarize, group status was a consistent predictor of sexual permissiveness across samples and criterion variables. The nonoverlapping confidence intervals across groups support the hypothesis that the means came from different populations. The nature of the difference between religious nonviewers and religious viewers was consistent with the SS, not the MI, approach. Further, the nature of the difference between religious viewers and nonreligious viewers was consistent with the 3AM.

RQ2. All robustness checks were supportive of the main analyses. First, removing sexually diverse (i.e., LGB) participants did not impact the pattern of results for same-sex attitudes. The ANOVA was significant (F(3, 1669) = 195.64, p < 0.001, eta squared = 0.26), and the rank-ordering of the group means was in the same direction: religious nonviewers (M = 1.73, SE = 0.046, 95% CI: 1.646, 1.825) < religious viewers (M = 2.22, SE = 0.125, 95% CI: 1.979, 2.470) < nonreligious nonviewers (M = 3.17, SE = 0.054, 95% CI: 3.068, 3.280) < nonreligious viewers (M = 3.35, SE = 0.070, 95% CI: 3.216, 3.491). The only deviation was a slight overlap in the confidence intervals for nonreligious nonviewers and nonreligious viewers. If a traditional mean comparison approach were applied, the LSD test would indicate these groups were significantly different (p = 0.043), while the Scheffe test would not (p = 0.251).

Second, limiting the samples to data from 1996 onward did not impact the pattern of results. The ANOVA for teen sex attitudes was significant (F(3, 4297) = 287.69, p < 0.001, eta squared = 0.17), and the rank-ordering of the group means was in the same direction: religious nonviewers (M = 1.18, SE = 0.018, 95% CI: 1.148, 1.219) < religious viewers (M = 1.41, SE = 0.049, 95% CI: 1.318, 1.511) < nonreligious nonviewers (M = 1.76, SE = 0.022, 95% CI: 1.723, 1.811) < nonreligious viewers (M = 2.16, SE = 0.032, 95% CI: 2.097, 2.224).

The ANOVA for premarital sex attitudes was significant (F(3, 4241) = 726.89, p < 0.001, eta squared = 0.34), and the rank-ordering of the group means was in the same direction: religious nonviewers (M = 1.94, SE = 0.024, 95% CI: 1.901, 1.995) < religious viewers (M = 2.73, SE = 0.065, 95% CI: 2.606, 2.861) < nonreligious nonviewers (M = 3.41, SE = 0.029, 95% CI: 3.360, 3.474) < nonreligious viewers (M = 3.72, SE = 0.042, 95% CI: 3.639, 3.804).

The ANOVA for extramarital sex attitudes was significant (F(3, 4081) = 145.71, p < 0.001, eta squared = 0.10), and the rank-ordering of the group means was in the same direction: religious nonviewers (M = 1.11, SE = 0.015, 95% CI: 1.084, 1.143) < religious viewers (M = 1.31, SE = 0.041, 95% CI: 1.237, 1.398) < nonreligious nonviewers (M = 1.44, SE = 0.019, 95% CI: 1.406, 1.480) < nonreligious viewers (M = 1.68, SE = 0.026, 95% CI: 1.632, 1.732).

The ANOVA for same-sex attitudes was significant (F(3, 3952) = 396.97, p < 0.001, eta squared = 0.23), and the rank-ordering of the group means was in the same direction: religious nonviewers (M = 1.54, SE = 0.028, 95% CI: 1.487, 1.598) < religious viewers (M = 1.94, SE = 0.079, 95% CI: 1.792, 2.101) < nonreligious nonviewers (M = 2.82, SE = 0.036, 95% CI: 2.757, 2.898) < nonreligious viewers (M = 3.09, SE = 0.049, 95% CI: 2.997, 3.188).

Third and finally, adjusting for participants’ age, political orientation, education, sex, and race did not impact the pattern of results. Group status remained significant in the teen sex attitudes ANCOVA (F(3, 6326) = 265.73, p < 0.001, eta squared = 0.10), and the rank-ordering of the group means was in the same direction: religious nonviewers (M = 1.25 SE = 0.014, 95% CI: 1.226, 1.281) < religious viewers (M = 1.38, SE = 0.038, 95% CI: 1.312, 1.459) < nonreligious nonviewers (M = 1.73, SE = 0.018, 95% CI: 1.702, 1.774) < nonreligious viewers (M = 2.04, SE = 0.028, 95% CI: 1.994, 2.104).

Group status remained significant in the premarital sex attitudes ANCOVA (F(3, 7857) = 886.85, p < 0.001, eta squared = 0.22), and the rank-ordering of the group means was in the same direction: religious nonviewers (M = 2.01, SE = 0.017, 95% CI: 1.986, 2.051) < religious viewers (M = 2.58, SE = 0.046, 95% CI: 2.492, 2.7672) < nonreligious nonviewers (M = 3.27, SE = 0.022, 95% CI: 3.232, 3.321) < nonreligious viewers (M = 3.56, SE = 0.035, 95% CI: 3.497, 3.636).

Group status remained significant in the extramarital sex attitudes ANCOVA (F(3, 7991) = 224.85, p < 0.001, eta squared = 0.07), and the rank-ordering of the group means was in the same direction: religious nonviewers (M = 1.16, SE = 0.010, 95% CI: 1.140, 1.181) < religious viewers (M = 1.37, SE = 0.027, 95% CI: 1.319, 1.426) < nonreligious nonviewers (M = 1.51, SE = 0.015, 95% CI: 1.482, 1.542) < nonreligious viewers (M = 1.73, SE = 0.022, 95% CI: 1.691, 1.779).

Group status remained significant in the same-sex attitudes ANCOVA (F(3, 7777) = 521.975, p < 0.001, eta squared = 0.14), and the rank-ordering of the group means was in the same direction: religious nonviewers (M = 1.45, SE = 0.017, 95% CI: 1.425, 1.491) < religious viewers (M = 1.67, SE = 0.045, 95% CI: 1.591, 1.766) < nonreligious nonviewers (M = 2.45, SE = 0.025, 95% CI: 2.404, 2.502) < nonreligious viewers (M = 2.71, SE = 0.037, 95% CI: 2.638, 2.781).

Discussion

The SS perspective is the most prevalent approach in the pornography effects paradigm for explaining associations between the consumption of pornography and the expression of attitudes consistent with themes for sex depicted in pornography (Wright, 2020). The moral incongruence (MI) perspective on pornography effects has also gained prominence in recent years (Grubbs & Perry, 2019; Grubbs et al., 2019a, 2019b). A detailed reading of MI and SS literatures suggests the possibility of discrepant predictions about the effects of pornography on sexual permissiveness, one of the most studied outcomes in pornography research (Tokunaga et al., 2019). From an SS perspective, pornography’s portrayal of permissive sexuality as common and satisfying leads to attitude change toward more acceptance of sexual permissiveness, but the preexisting traditional sexual scripts of the religious lessen the magnitude of this effect. Religious pornography viewers will be more sexually permissive than religious persons who do not view pornography, but less sexually permissive than nonreligious pornography viewers, according to the SS approach. Alternatively, the adoption of a MI perspective suggests the possibility that religious persons become more opposed to permissive sexuality as a result of their pornography use, as a way to mitigate the cognitive dissonance they feel for engaging in a sexual behavior proscribed by their faith tradition (see also Green & Low, 2014; Jordan et al., 2011; Merritt et al., 2010; Sachdeva et al., 2009; Tetlock et al., 2000; Zhong & Liljenquist, 2006; Zhong et al., 2009). If this occurs, religious pornography consumers will be less sexually permissive than religious persons who do not consume pornography (it is noted for a final time that this is the present study’s deduction from reading works on MI, rather than a MI postulate formally proposed by its originators).

The present study explored these divergent possibilities using nationally representative data gathered across multiple decades from men and women in the USA. Results supported the SS approach. Specifically, across four meta-samples and four indicators of sexual permissiveness, religious nonviewers were the least sexually permissive, followed by religious viewers, nonreligious nonviewers, and nonreligious viewers. These results were robust over time and maintained after adjusting for demographics.

What conclusions should be drawn from these results for the SS and MI perspectives? For theorizing about pornography, religiosity, and sexual permissiveness, they suggest that an SS approach may be preferable to an MI approach. The persistent rank-ordering of religious viewers as more sexually permissive than religious nonviewers leaves little leeway for alternative interpretation on this front. The 3AM-specific finding that religious viewers were less sexually permissive than nonreligious viewers provides additional support for the SS perspective. Perhaps, if religious nonviewers and religious viewers had been indistinguishably ranked as the least permissive (even if intermittently), a case could have been made for some effect of MI. But this was not the case. Given the present study’s results on sexual permissiveness, as well as the fit between SS predictions and pornography effect findings in other areas of sexuality, the present findings support the continued application of SS to research on pornography and sexuality (Tokunaga et al., 2019, 2020; Wright et al., 2016, 2017).

It would be misguided, however, to conclude that MI does not result when the religious view pornography. Grubbs and Perry (2019) rightly note that “by definition, the contrast between religious individuals’ reported moral disapproval of pornography and their reported use rates suggests that many religious people are likely experiencing some moral incongruence around their actions” (p. 31). Second, it would be imprudent to conclude that the MI approach is invalid. Clearly, the MI approach is adroit at predicting perceived addiction to pornography (Grubbs et al., 2019a, 2019b; Wright, 2019; see also Perry et al., 2021, for MI consistent findings on sexuality and happiness). Third, it would be unwise to conclude that MI-induced cognitive dissonance does not work in tandem with SS dynamics to produce certain pornography effects. For example, the link between pornography and sexual dissatisfaction has already been shown via path analysis to be predictive from SS variables (Wright et al., 2019, 2021). However, both SS (Wright et al., 2018) and MI (Perry & Whitehead, 2019) inspired papers have found that pornography-sexual dissatisfaction linkages are stronger for the more religious. In sum, while the results of the present study were in alignment with SS and not MI, there are sufficient data in certain areas of pornography effects research to clearly support the continued exploration of MI-derived predictions. Furthermore, it must be remembered that the most specific instantiation of MI—the Pornography Problems due to Moral Incongruence model, or PPMI—has the goal of explaining perceived addiction to pornography and similar self-reported pornography-related problems, not attitude change following pornography use (Grubbs et al., 2019a, 2019b).

Limitations and Future Directions

Subsequent research can expand upon the present study in several ways. First, analyses could be stratified by faith tradition to see if an MI indicative attitudinal doubling down on pornography promoted but religiously discouraged sexual attitudes does occur for some religious groups. Second, permissive sexual behaviors could be studied to see whether the SS consistent and MI inconsistent attitudinal findings of the present study replicate onto behavioral indices. Third, efforts could be made to enhance internal validity, as the present data were from a cross-sectional survey. The present study could be replicated experimentally, for example. Religious and nonreligious persons could be identified via pretest, assigned to either view or not view pornography, and then presented a questionnaire embedded with items tapping sexual permissiveness. While the religious would not choose to view pornography in such an experimental design, it is still possible that an MI dynamic may evidence itself from mere exposure alone. In addition to experimental replications, the present findings could be replicated with a multi-wave survey design to more definitively address questions of temporal sequencing (but see Wright, 2021b, for a review of longitudinal survey studies suggesting the downstream effects of pornography use). Finally, studies that propose MI explanations for findings previously theorized in the SS tradition could measure whether MI is actually associated with cognitive dissonance surrounding pornography use, and whether this dissonance is more predictive of the outcomes under study than sexual script mediators.

Footnotes

1While MI could arise from nonreligious worldviews (e.g., political views oppositional to pornography), MI research has primarily focused on religiosity as the source of MI (Grubbs & Perry, 2019; Grubbs et al., 2019a, 2019b).

2For a critical treatment of the use of covariates in pornography effects research, see Wright (2021a).

3This approach is more conservative than the conventional p < 0.05 method for examining group differences (Schenker & Gentleman, 2001). As noted by Cumming (2009), “when 95 per cent confidence intervals (CIs) on independent means do not overlap, the two-tailed p-value is less than 0.05 and there is a statistically significant difference between the means. However, p for nonoverlapping 95 per cent CIs is actually considerably smaller than 0.05” (p. 205).