Self-Assessed Effects of Pornography Use on Personal Sex Life: Results from a Large-Scale Study of Norwegian Adults

Abstract

Objectives: The study explored characteristics associated with the self-assessed effects of pornography use on one’s personal sex life. Methods: Data were collected in a probability-based sample of Norwegian adults (n = 4,160). Results: Most participants (41.3%) did not believe that pornography affected their sex life. More participants reported positive (33.2%) than mixed/negative (25.5%) effects of pornography use. Sociodemographic and sexual characteristics that predicted these self-assessments were examined. Among participants in a steady relationship, emotional intimacy and relationship satisfaction were unrelated to the self-assessed effects. Conclusions: The current study findings add to scant literature about the self-assessed effects of pornography usage.

Introduction

Pornography research of recent decades has consistently demonstrated high prevalence rates of pornography consumption, availability, and demand (Gmeiner et al., 2015; Miller et al., 2020; Regnerus et al., 2016). Further, this research has aimed to source out what effects such pornography usage may have on a variety of outcomes such as sexual satisfaction (Grubbs, Wright, et al., 2019; Wright et al., 2017), sexual risk-taking (Grubbs, Wright, et al., 2019; Harkness et al., 2015), relationship quality (Vaillancourt-Morel et al., 2019; Wright et al., 2017), and sexual aggression (Ferguson & Hartley, 2020; Wright et al., 2016). This has been done using both quantitative and qualitative methods, different operational definitions of pornography use, different research designs (e.g., questionnaire studies, experimental designs, meta-analyses), cohorts (e.g., youth, young adults, gay), outcome measures (e.g., attitudinal, behavioral, cognitive and emotional), and cultural settings (e.g., U.S.A, Indonesia, Denmark, Croatia) (Fisher & Kohut, 2020; Hald et al., 2014; Peter & Valkenburg, 2016). Generally, these studies offer mixed results (Grubbs, Wright, et al., 2019; Kohut et al., 2020; Vaillancourt-Morel et al., 2019), but point to three emerging patterns of effects.

First, research has relatively consistently suggested that there are positive associations between pornography consumption and attitudinal and sexual behavioral outcomes such as the number of sexual partners or engaging in casual sex or group sex (Harkness et al., 2015; Peter & Valkenburg, 2016). However, in this connection, it has also been found that these associations are most often modest when correcting for other relevant covariates such as sexual sensation seeking and relationship status (Hald, Kuyper, et al., 2013; Wijaya Mulya & Hald, 2014). Second, in more recent studies or theoretical models which include pornography as an antecedent of the outcome (e.g., sexually aggressive behavior), moderation by third variables (e.g., personality traits; risk of sexual aggression) is often evident or theoretically indicated (see also Hald & Malamuth, 2015; Malamuth & Hald, 2016; Veit et al., 2017). This indicates that the effects of pornography on various outcomes such as sexual aggression or attitudes supporting violence against women are most likely not uniform across consumers, but dependent on other variables such as personality traits (see also Hald et al., 2014). This is further indicated in conclusions emerging from previous meta-analytic studies in the area showing the effects of pornography to be heterogeneous across consumers (e.g., Hald et al., 2010; Wright et al., 2016). Third, the past decade of longitudinal and experimental studies has shown that effects of pornography usage or exposure may be mediated by the various individual- and content-related variables such as sexual arousal experienced during exposure. This indicates that previous relationships found between pornography exposure/consumption and outcomes studied may be both direct and indirect in nature (Hald, Malamuth, et al., 2013; Hald & Malamuth, 2015; Koletić, 2017; Wright, 2020).

Surprisingly, quantitative studies of users’ self-assessed effects of pornography consumption have been relatively sparse. The first such study was by Hald and Malamuth (2008) and included a cohort of 688 young Danish adults between the ages of 18 and 30. In their study, Hald and Malamuth found that users generally reported small, if any, negative effects but moderate positive effects of their pornography consumption on a variety of outcomes including sexual knowledge, attitudes toward sex, users’ sex life, and general quality of life. These results have later been mirrored in other studies using different age cohorts, cultural backgrounds, sexual orientation, and outcome measures. For example, using a sample of 1,333 predominately North American Men Who Have Sex with Men (MSM), Hald, Smolenski, et al. (2013) found that 97% of the sample reported positive effects of their consumption, and only 3% reported negative effects of their consumption in areas related to their sexual knowledge, attitudes, behaviors, and orientation. Similarly, using a diverse sample of 1,274 Swedish and Norwegian young adults, Kvalem et al. (2014) found that young male adults perceived a positive, albeit modest, influence of pornography use on sexual self-esteem. Using a sample of heterosexual men, Miller et al. (2018) also found greater positive self-assessed effects of pornography use than negative. Finally, compared to self-assessed negative effects of pornography use, positive effects were more prevalent in samples of Indonesian (Wijaya Mulya & Hald, 2014) and Canadian (Hesse & Pedersen, 2017) university students, and in a large-scale sample of Australian adults (McKee, 2007; Rissel et al., 2017).

To the best of our knowledge, only a couple of studies examined the self-assessed effects of pornography use on relationship quality. Using a sample of 6,463 Polish students, Dwulit and Rzymski (2019) found that reporting beneficial effects of pornography use on relationship quality was more prevalent than reporting adverse effects, while the majority of students reported no effect. A similar ratio between positive, negative, and no effects were found in a participant-informed study conducted among 430 men and women in a relationship (Kohut et al., 2017). Interestingly, there appears to be a discrepancy in the literature between self-assessed effects of pornography use on the one hand and statistical correlates of pornography use on relationship quality on the other hand. On average, self-assessed effects of pornography use are null to positive, although correlational studies suggest that among men especially, the frequency of pornography use is negatively related to relationship quality outcomes, while for women the results appear more equivocal (Veit et al., 2017).

Large population studies on the self-assessed effects of pornography use that include adult participants of all ages and not only young adults or males are missing from the literature. Further, studies specifically addressing the self-assessed effects of pornography on relationship well-being are rare (Kohut et al., 2017). This is surprising when compared to a wealth of correlational studies in the past decade on associations between pornography use and relationship- or sexuality-related correlates such as relationship closeness, sexual communication, and sexual satisfaction (Vaillancourt-Morel et al., 2019; Wright et al., 2017). Further, to date, no large-scale study has investigated the explanatory ability of a more comprehensive array of user characteristics on self-assessed effects of pornography use. According to recently proposed conceptual models (see The Differential Susceptibility to Media Effects Model, Valkenburg & Peter, 2013; Antecedents—Context—Effects Model, Campbell & Kohut, 2017), pornography use, as a solitary or joint behavior (e.g., with friends or in the context of a relationship), is related to a variety of dispositional, developmental, or social characteristics. These user characteristics could influence the context of pornography use (e.g., frequency of use, preferred pornographic content, motivation for use) and, in turn, lead to different behavioral, interpersonal, and psychological effects of pornography use. Finally, in a research field characterized by indirect measurement of pornography-related effects and (predominantly) correlational findings, which are often and unjustifiably interpreted as causal (Fisher & Kohut, 2020), self-assessment of pornography-related effects may serve as a different research perspective due to a growing societal concern about pornography use.

Dominant sociocultural narratives produced by media, policymakers and the general public often depict pornography use in terms of potentially negative effects (BBC Future, 2017; Republican National Committee, 2016; Webber & Sullivan, 2018). For example, popular media could highlight or simplify narratives about the negative effects of pornography use on some aspects of people’s sexual lives (e.g., relationship functioning), although no empirical evidence or consensus exists (Montgomery-Graham et al., 2015). In addition to growing concerns over pornography-related harms within the public domain, the majority of academic research is characterized by a negative effects paradigm, concretely a research orientation examining how pornography negatively affects people’s sexual behaviors and their attitudes toward sex (McCormack & Wignall, 2017). In such a harm-oriented climate, the possibility that pornography use may have positive or both positive and negative effect on people’s sexual lives (Peter & Valkenburg, 2016), as well as the possibility that for some users pornography use might have no effect (Weinberg et al., 2010) are rarely examined or discussed.

The current study was designed to shed light on these shortcomings in the current research in the area.

Study aims

Using a large-scale national probability sample of Norwegian adults, the current study explored self-assessed effects of pornography use on users’ personal sex life by focusing on three research questions (RQ):

RQ1: What are the characteristics of participants who report a self-assessed positive effect of pornography use on their sex life, compared to participants who report no effect of their pornography usage?

RQ2: What are the characteristics of participants who report a self-assessed mixed or negative effect of pornography use on their sex life, compared to participants who report no effect of their pornography usage?

RQ3: Among partnered/married participants, is there a link between two indicators of relationship quality (i.e., relationship satisfaction and emotional intimacy) and the self-assessed effect of pornography use?

Methods

Participants and procedure

Participant recruitment was carried out in March 2020 by e-mailing an invitation to a randomly selected sample of 11,685 Norwegians registered in Kantar’s Gallup Panel. A total of 4,160 individuals aged 18–89 years completed the survey, yielding a response rate of 35.6%. Previous response rates for Norwegian sexual behavior surveys were 63% in 1987, 48% in 1992, 38% in 1997, 34% in 2002, and 23% in 2008 (Traeen & Stigum, 2010). Nearly half of the participants (51%) completed the survey using their mobile phones.

The current study uses data from 2,487 participants who reported pornography use in the past 12 months (participants’ sociodemographic characteristics are shown in Table 1).¹ The sample consisted of 68.8% male and 31.3% female participants (seven individuals identified as “other”). Participants’ mean age was 43.3 years (SD = 15.7, range 18–87). More than half of the participants reported living in a large urban area (57.0%). Almost two-thirds of participants had a college or university education (63.3%), which is a somewhat higher percentage of tertiary-educated individuals compared to the national population. According to Statistics Norway, 29.7% of the adult population 16 years and above were college or university-educated in 2019.² About 60% of participants reported no religious affiliation. Most participants identified as heterosexual (91.6%). Over two-thirds of surveyed individuals (73.2%) were married or in a relationship at the time of the survey.

Table 1.

Sociodemographic Information on Study Participants (n = 4,160).

	Study participants	Participants excluded from the analyses	Participants included in the analyses	p
	n (%)	n (%)	n (%)
Gender
Male	2,181 (52.6)	476 (28.5)	1,705 (68.8)	<.001
Female	1,967 (47.4)	1,192 (71.5)	775 (31.3)
Place of residence
Rural/small town	1,787 (43.2)	720 (43.4)	1,067 (43.0)	.840
Urban	2,354 (56.8)	940 (56.6)	1,414 (57.0)
Education
Less than college educated	1,482 (35.8)	572 (34.4)	910 (36.7)	.132
College educated	2,658 (64.2)	1,091 (65.6)	1,567 (63.3)
Religious affiliation
No	2,416 (59.5)	812 (50.1)	1,604 (65.7)	<.001
Yes	1,645 (40.5)	809 (49.9)	836 (34.3)
Sexual orientation
Exclusively heterosexual	3,815 (93.5)	1,565 (96.4)	2,250 (91.6)	<.001
Non-exclusively heterosexual	266 (6.5)	59 (3.6)	207 (8.4)
Relationship status
Single	1,076 (25.9)	410 (24.6)	666 (26.8)	.121
In a relationship/married	3,074 (74.1)	1,256 (75.4)	1,818 (73.2)
Age [M (SD)]	46.5 (17.1)	51.4 (17.8)	43.3 (15.7)	<.001

Note. p = significance of difference between excluded and included participants (χ² test, Fisher exact test in cases of small expected frequencies or t-test).

Kantar’s commercial panel, with its 46,000 members, is based on random recruitment through phone surveying in various probability samples (Gallup, 2020); self-recruitment is not possible. The panel is representative of Norway’s Internet population (Internet penetration in Norway in 2018 was 98%, see Mediennorge, 2018). Members of the panel are recruited based on a large set of sociodemographic (age, gender, education, income, and occupation) and sociocultural variables (media use, consumer habits, political affiliation, culture, and sports preferences, etc.), which enables fine-tuning to achieve representativeness. To encourage participation in this survey, Kantar rewarded participants with a small incentive.

All study procedures followed international ethical guidelines developed for market and poll organization surveys.³ The survey was also approved by the Ethical Committee of the Department of Psychology, University of Oslo.

Questionnaire

The questionnaire used in the study was composed by a group of researchers at the University of Oslo, who conducted the 2013 Norwegian Sex Study among 18 to 29-year-olds (Kvalem et al., 2014; Træen et al., 2016). The average time to complete the survey was 15 minutes. Prior to launching the survey, the questionnaire was piloted in a self-selected Facebook sample. The questionnaire contained sociodemographic questions. Attitudes toward sex and sexuality were taken from the 1996 Swedish Sex Study (Lewin et al., 2000), while body image questions were adapted from studies on breast size satisfaction (Frederick et al., 2016; Sandhu & Frederick, 2015; Swami et al., 2020). Emotional closeness with partners was measured with a visual indicator developed by Aron et al. (1992) that was also used in a recent nationwide German Health and Sexuality Survey (GeSiD, 2020). Some of the pornography use (e.g., problematic pornography use) and sexting questions were adapted from the GeSiD survey. Sexual function indicators were adopted from (Lee et al., 2016) and NATSAL-3 (Mitchell et al., 2013).

Measures

Outcome variable. The self-assessed effect of pornography use on one’s personal sexual life was addressed with a single-item indicator (“Has your pornography use had any effect on your sexual life?”), first used in the Rosser et al. (2013) study on gay men’s use of pornography. The response scale included the following options: 0 = no effect, 1 = positive effect, 2 = both positive and negative effect and 3 = negative effect. The last two categories were merged into a mixed/negative category because only 3.1% (n = 77) of participants reported a negative effect of their pornography use. A similar prevalence was found in Hald, Smolenski, et al.’s (2013) study. Similar single-item measures of net-effect have been used in other studies of self-assessed pornography use (McKee, 2007; Rissel et al., 2017). This measure was available only to participants who previously reported intentional pornography use.⁴ Additionally, to assess the discriminant validity of the outcome variable, we used a 6-item measure of problematic pornography use (i.e., “Has your pornography use ever resulted in problems in a relationship?”) with a yes/no response option (GeSiD, 2020). With scores ranging from 0 to 6, higher overall scores indicate more reported pornography-related problems. The scale demonstrated unidimensionality and acceptable reliability (McDonald’s ω = .75).

Predictor variables were grouped into four sets of indicators: sociodemographic characteristics, characteristics of pornography use, sexual well-being, and relationship well-being. Sociodemographic indicators used in the analysis were gender (0 = male, 1 = female), age, place of residence (0 = rural/small town, 1 = urban), education (0 = less than college educated, 1 = college educated), religious affiliation (“Do you currently belong to a specific religion,” 0 = no, 1 = yes), sexual orientation (0 = exclusively heterosexual, 1 = other) and relationship/marital status (0 = single, 1 = in a relationship/marriage).

Characteristics of pornography use included four indicators: age of first exposure to pornography, frequency of pornography use, pornography realism, and use of non-mainstream pornography genre. Before asking these questions, the definition of pornography was provided: “Pornography is any type of material that depicts genitals, and clear and distinct sexual acts. Please note that nude presentations of men and women (such as those in Playboy or Playgirl, mainstream movies, etc.) that do not contain clearly depicted and explicit sexual acts are not considered pornography” (Hald, 2006). Age of first exposure to pornography was measured with the question “How old were you the first time you saw pornography?” The frequency of pornography use was measured with one item “How often have you seen pornography in the past 12 months?” Answers were anchored on an 8-point scale: 1 = never, 2 = once, 3 = a couple of times, 4 = several times a year, 5 = about once a month, 6 = about once a week, 7 = several times a week and 8 = daily. Both questions were available only to participants who previously reported intentional pornography use. Pornography realism was measured with the 4-item scale (i.e., “Pornographic sex is equal to sex in real life.”) developed by Rosser et al. (2013). A 5-point scale ranging from 1 = completely disagree to 5 = completely agree was used to anchor answers. Higher scores denoted higher perceived pornography realism (score range: 4–20). Scale reliability was satisfactory (McDonald’s ω = .80). A dichotomous indicator of the preference for non-mainstream pornography use in this study was based on the 37-item inventory and its latent structure typology developed by Hald and Štulhofer (2016). According to their study, non-mainstream (paraphilic) pornography use factor included the following categories: bizarre, bondage (BDSM), fetish (minus leather and latex fetishes), sadomasochism (SM), and violent sex (rape, aggression, or coercion). Reporting preference for any of these categories was coded as non-mainstream pornography use. Pornography realism and non-mainstream pornography questions were available only to participants reporting pornography use in the past 12 months.

Three one-item indicators of sexual well-being were used: sexual satisfaction (“All things considered, how satisfied are you with your sexual life?”), satisfaction with frequency of sexual activity (“In general, how satisfied are you with your current level of sexual activity?”) and appearance satisfaction (“How satisfied are you with your physical appearance?”). The first two items had a 5-point and the third a 7-point scale ranging from very dissatisfied to very satisfied.

Relationship well-being was assessed with two items which were filtered out for participants in a relationship and married participants. Emotional intimacy was measured with a set of overlapping circles representing the participant and his/her partner developed by Aron et al. (1992). The item wording was as follows: “These figures attempt to express how close two persons may feel. Choose the figure which best describes your relationship to your partner.” The absence of emotional intimacy was represented by a lack of overlap between the two circles, while high emotional intimacy was depicted by an almost full overlap. Higher scores denoted higher intimacy (score range: 1–7). Relationship satisfaction was measured with a single-item indicator: “All things considered, how satisfied are you with your current relationship?” Answers were indicated on a 5-point scale ranging from 1 = completely unsatisfied to 5 = completely satisfied.

Analytical strategy

First, to examine the discriminant validity of the self-assessed effect of pornography measure, differences in the mean number of reported pornography-related problems between the three outcome groups (no effect, positive effect, and mixed or negative effect) were examined. We hypothesized that participants in the mixed/negative group would report more pornography-related problems, compared to the other two groups. Differences in the number of reported pornography-related between these three subsamples were evaluated with two separate analyses of covariance. In the first, the number of pornography-related problems was included in its original form. In the second analysis, the outcome was Box-Cox transformed to adjust for the nonnormality (Sakia, 1992). Considering that men use pornography at rates much higher than women do (Grubbs, Wright, et al., 2019), the frequency of pornography use was controlled for in both analyses. Between-group difference in the number of pornography-related problems was significant (F(2, 2288.00) = 63.38, η² = .052, p < .001). Post-hoc analyses indicated that participants in the mixed/negative group (M = 0.41, SD = 0.79) reported significantly more pornography-related problems than the no effect (M = 0.06, SD = 0.28) and positive effect (M = 0.13, SD = 0.61) groups. The analysis with Box–Cox transformed number of pornography-related problems (ʎ = −1.7) also resulted in a significant between-group difference (F(2, 2288.00) = 102.39, η² = .082, p < .001). No change in post-hoc differences between the three outcome groups was observed.

To address our research questions, multinomial logistic regression was used to identify predictors of the self-assessed effect of pornography use; the no-effect group served as the reference category. Considering that the online panel used in this study was recruited by cluster-based sampling, multivariate regression models were estimated using Stata’s svy command to adjust for survey design. Estimates of standard errors were adjusted for 18 Norwegian counties and weighted for gender, age, and region. The following predictor sets were included in the multinomial regression analysis: sociodemographic characteristics, characteristics of pornography use, and sexual well-being indicators.

An additional multinomial regression explored the role of emotional intimacy and relationship satisfaction in the self-assessed effects of pornography. The analysis was carried out in a subsample of individuals who were married or in a steady relationship at the time of the survey. In addition to all independent variables included in the first model, the analysis also included satisfaction with the relationship and emotional intimacy indicators.

Due to our merging of participants who reported mixed and those who reported a negative effect of pornography use, we carried out a robustness test in which both multivariate regression models were repeated without the negative effects (n = 77).

Less than 5% of data were missing on all but one indicator. Likely due to recall difficulties, 13.1% of information about the age at first exposure to pornography was missing. All analyses were carried out using Stata v14.1 and jamovi v1.2.18.0 (The jamovi project, 2020) statistical software packages.

Results

Self-assessed effects of pornography use

Most of the participants did not believe that pornography affected their personal sex life (41.3%). Of the rest, a positive self-assessed effect was reported by 33.2% of participants, while 25.5% reported mixed or negative effect.⁵ Using analysis of variance, significant differences in frequency of pornography use were observed between all three groups, F(2, 2338.57) = 98.43, η² = .073, p < .001). The no-effect group used pornography least frequently (M = 4.48, SD = 1.75), followed by the positive effect group (M = 5.16, SD = 1.68). Participants in the mixed/negative effect group reported the highest frequency of pornography use in the past 12 months (M = 5.65, SD = 1.65).

RQ1 and RQ2: self-assessed effects of pornography use on personal sexual life

The findings of the multinomial logistic regression assessing the characteristics of participants who reported self-assessed positive and mixed/negative effects of pornography use, respectively, in contrast to participants who reported no effect, are shown in Table 2. Older (AOR = 1.04, p < .001), female (AOR = 2.64, p < .001) and non-exclusively heterosexual participants (AOR = 1.67, p = .030) had higher odds of reporting a self-assessed positive effect, compared to no effect. In addition, more frequent pornography use (AOR = 1.39, p < .001), perceiving pornography as more realistic (AOR = 1.27, p < .001), the preference for non-mainstream pornography (AOR = 2.00, p < .001), and greater sexual satisfaction (AOR = 1.46, p < .001) all increased the odds of reporting a positive pornography effect when compared to the no effect group. The odds of reporting a mixed or negative self-assessed effect over no effect were increased by being married or in a relationship (AOR = 1.24, p = .012), more frequent pornography use (AOR = 1.40, p < .001), pornography realism (AOR = 1.12, p < .001), and preferring non-mainstream pornography material (AOR = 2.14, p < .001). Finally, older age at first exposure to pornography (AOR = 0.96, p = .012) decreased the odds of reporting a mixed or negative effect compared to no effect of pornography use on personal sexual life. According to recent guidelines, these effects should be interpreted as small in size, except for medium-sized effect for gender (Chen et al., 2010).

Table 2.

Predictors and Correlates of Self-Assessed Effects of Pornography Use on Personal Sexual Life (n = 1,919).

	Positive effect		Mixed or negative effect
	AOR	95% CI	AOR	95% CI
Age	1.04***	1.03–1.04	0.99	0.98–1.00
Gender (Female)	2.64***	1.98–3.52	1.22	0.95–1.57
Place of residence (Urban)	1.22	0.98–1.52	1.19	0.93–1.51
College education	0.97	0.73–1.30	1.33	0.93–1.91
Religious affiliation	1.02	0.80–1.30	1.04	0.73–1.47
Minority sexual orientation	1.67*	1.06–2.65	1.12	0.67–1.88
In a relationship/married	1.33	0.96–1.83	1.33*	1.08–1.66
Age of first exposure to porn	0.99	0.97–1.02	0.96*	0.93–0.99
Frequency of porn use	1.39***	1.26–1.54	1.40***	1.26–1.55
Porn realism	1.27***	1.20–1.34	1.12***	1.07–1.17
Non-mainstream porn use	2.00***	1.47–2.71	2.14***	1.59–2.88
Appearance satisfaction	1.01	0.97–1.06	0.98	0.95–1.02
Sexual satisfaction	1.46***	1.27–1.69	0.98	0.84–1.15
Satisfaction with frequency of sex	0.94	0.82–1.08	0.97	0.78–1.20

Note. Reference category is no effect of pornography use. CI: confidence interval around adjusted odds ratio (AOR).

*p < .05; **p < .01; ***p < .001.

RQ3: self-assessed effects of pornography use among partnered or married participants

Among participants who were married or in a steady relationship (n = 1,407), increased odds of reporting positive effect over no effect were associated with age (AOR = 1.03, p < .001), female gender (AOR = 3.05, p < .001), a minority sexual orientation (AOR = 2.79, p = .005), higher frequency of pornography use (AOR = 1.47, p < .001), pornography realism (AOR = 1.26, p < .001), the preference for non-mainstream pornography (AOR = 1.94, p = .002), and higher sexual satisfaction (AOR = 1.46, p = .001). The odds of a self-assessed mixed or negative effect were significantly associated with female gender (AOR = 1.29, p = .030), higher frequency of pornography use (AOR = 1.41, p < .001), pornography realism (AOR = 1.13, p < .001), and preferring non-mainstream pornographic content (AOR = 2.25, p = .002). Importantly, neither of the two indicators of relationship quality were significantly related to self-assessed pornography effects (Table 3).

Table 3.

Predictors and Correlates of Self-Assessed Effects of Pornography Use on Personal Sexual Life among Participants Who are Married or in a Relationship (n = 1,407).

	Positive effect		Mixed or negative effect
	AOR	95% CI	AOR	95% CI
Age	1.03***	1.02–1.05	0.99	0.98–1.01
Gender (Female)	3.05***	2.21–4.22	1.29*	1.03–1.61
Place of residence (Urban)	1.13	0.96–1.66	1.21	0.96–1.54
College education	0.87	0.61–1.22	1.19	0.80–1.77
Religious affiliation	1.05	0.80–1.36	0.95	0.66–1.38
Minority sexual orientation	2.79**	1.43–5.46	1.13	0.90–2.40
Age of first exposure to porn	1.00	0.98–1.03	0.97	0.92–1.01
Frequency of porn use	1.47***	1.33–1.62	1.41***	1.23–1.62
Porn realism	1.26***	1.21–1.32	1.13***	1.08–1.18
Non-mainstream porn use	1.94**	1.31–2.88	2.25**	1.40–3.59
Appearance satisfaction	1.00	0.93–1.08	0.96	0.92–1.01
Sexual satisfaction	1.46**	1.19–1.79	0.99	0.83–1.18
Satisfaction with frequency of sex	0.97	0.83–1.14	1.02	0.81–1.28
Emotional intimacy	0.98	0.86–1.11	1.02	0.92–1.13
Relationship satisfaction	1.02	0.84–1.23	0.95	0.78–1.17

Note. Reference category is no effect of pornography use. CI: confidence interval around adjusted odds ratio (AOR).

*p < .05; **p < .01; ***p < .001.

Finally, both regression analyses were repeated without participants who believed that pornography negatively affected their personal sex life. This robustness test did not change the pattern of significant findings in either model—apart from age of first exposure emerging as a significant predictor (AOR = 0.96, p = .028) of reporting a mixed effect of pornography use in the model with participants who are married or in a relationship.

Discussion

Following Hald and Malamuth’s (2008) research on self-assessed effects of pornography use, this current study specifically examined characteristics of pornography users based on their own assessment of the overall effect of their pornography use on their sex life. Using a national probability sample of Norwegians and multinomial logistic regression we found that female, older, non-exclusively heterosexual, and more sexually satisfied participants believed that pornography positively affected their personal sex life, compared to participants who reported no effect of pornography use. Relative to self-assessed no effect of pornography use, individuals who reported a positive and those who reported a mixed or negative effect were characterized by a higher frequency of pornography use, increased perceived pornography realism, and the use of non-mainstream pornography. Exposure to pornography at an earlier age was a distinct characteristic of participants who reported mixed or negative effects of their pornography use. Finally, the indicators of marriage/relationship quality (i.e., relationship satisfaction and emotional intimacy) were not linked to differences in self-assessed pornography effects.

Unlike the current study, age was not related to self-assessed effects of pornography use in other relevant studies (Hald et al., 2015; Hesse & Pedersen, 2017; Wijaya Mulya & Hald, 2014). This may be due to a limited age range in other studies because these studies sampled emerging and/or young adults. The exception was a study carried out in a sample of heterosexual men (Miller et al., 2018), that found a weak negative association between age and self-assessed negative effects of pornography use.

Norwegian women were more likely to report positive self-assessed effects of pornography use than their male peers, which is the opposite of what Hald and Malamuth (2008) reported in their sample of young adults. The difference may be related to changes in pornography use and its acceptance among women in the period between 2008 and recent years (McKeown et al., 2018). It is possible that over time some women change their perception of pornography and/or become less sensitive to socially desirable attitudes about pornography use. Other possible explanations include methodological differences in the assessment of the perceived effect of personal pornography use or different mean age of female participants in the Danish study. Compared to exclusively heterosexual participants, non-exclusively heterosexual participants were characterized by more positive effects of pornography use, which resonates with findings that the majority of MSM report positive effects associated with their pornography use (Hald et al., 2015; Hald, Smolenski, et al., 2013). Underlying mechanisms are suggested in two qualitative studies, which pointed out that pornography has multiple roles for sexual minority individuals, including educational and identity-supporting, and identity-enhancing functions (Harvey, 2020; McCormack & Wignall, 2017; Nelson et al., 2014).

More frequent pornography use was associated with increased odds of reported positive, as well as mixed or negative self-assessed effects of pornography use—compared to no effect. These findings are concordant with previous studies (Hald & Malamuth, 2008; Miller et al., 2018; Wijaya Mulya & Hald, 2014). The latter finding is not surprising, considering that a number of negative effects of pornography use have been explored in the literature (Grubbs, Perry, et al., 2019). Furthermore, as mainstream media tend to emphasize problems associated with pornography use in terms of “pornography addiction” or “public health problems” (Ley et al., 2014; Ley, 2018), some individuals may ascribe their sexual difficulties to pornography use. It is important to note that the effect size for the association between the frequency of pornography use and self-assessed effects of pornography use (either positive or mixed/negative) was small. The link between more frequent pornography use and positive self-assessed effects contradicts the common notion that frequent pornography use is inherently perceived as problematic and related to adverse psychological outcomes. Our findings suggest that the frequency of pornography is not a risk per se. This is compatible with findings from Bőthe et al.’s (2020), according to which the majority of both low- and high-frequency pornography consumers were found to be non-problematic users with only a small percentage of high-frequency users (3–8%) identified as having problematic usage.

The perceived realism of pornography predicted both positive and mixed/negative self-assessed effects of pornography use. Although both effect sizes were small, the self-assessed positive effect was slightly larger in size. The positive association between perceived realism and positive self-assessed effects of pornography use was also found in other studies (Hald & Malamuth, 2008; Kvalem et al., 2014; Wijaya Mulya & Hald, 2014). The dual role of perceived pornography realism might reflect different perspectives found in the literature. On the one hand, it is possible that higher levels of perceived realism reflect internalized unrealistic expectations (Wright & Arroyo, 2013), which would negatively affect the person’s sexual life (Paul, 2005). On the other hand, perceived realism could underline pornography’s function as a source of information about a variety of sexual options, which may enhance personal sex life (Hesse & Pedersen, 2017).

We also observed the dual role of non-mainstream pornography. The use of non-mainstream content was associated with both positive and mixed/negative self-assessed effects of pornography use, with effect sizes being roughly equal. Leonhardt et al. (2019) suggest that the preference for non-mainstream pornographic material (i.e., paraphilic pornography) is congruent with the need for novelty-related excitement and pleasure. Compared to mainstream material, non-mainstream content is likely associated with factors such as sexual openness and seeking sexual novelty. Importantly, as indicated in a recent paper by Hald et al. (2018), some non-mainstream content, particularly soft sadomasochism, bondage, and dominance has been increasingly normalized in popular novels such as Fifty Shades of Gray (see Sexualities journal special issue from December 2013), so that its use may not indicate any underlying paraphilic interest. In cases characterized by a true paraphilic pattern of arousal, the link between non-mainstream pornography and its self-assessed effect on a person’s sex life may reflect challenges to the enactment of the paraphilic interest in real life (i.e., difficulty in finding a partner willing to engage in paraphilic sex acts; see Štulhofer et al., 2010). Further, the fact that the type of pornography was associated with self-assessed pornography effects more strongly than the frequency of its use has clinical relevance and should be explored in more detail in future studies. Most of the recent literature on pornography use does not include indicators of the type of content preferred and/or used, which is among major limitations in the field (Kohut et al., 2020).

No major change in the significance of indicators predicting positive and mixed/negative self-assessed effects of pornography use, relative to no effect, was observed in the subsample of participants who were married or in a relationship. We also found that emotional intimacy and relationship satisfaction were unrelated to the self-assessed effects of pornography use on user’s personal sex life. This finding seems to be somewhat in contrast to a recent meta-analytical study that observed a small but significant negative association between pornography use and relationship satisfaction (Wright et al., 2017). However, the meta-analysis did not address participants’ perceptions about how pornography use affected their sexuality and found adverse effects only in male participants. Further, a previous Norwegian dyadic study found that couples who used pornography had a better erotic climate compared to couples who did not use pornography (Daneback et al., 2009). Consequently, we encourage the development of studies specifically focusing on relationship quality and measures related to self-assessed effects of pornography consumption. In addition to measures of frequency of pornography use, assessing partners’ acceptance of pornography and dyadic patterns of pornography use (i.e., shared and/or solitary use) may help disentangle how and to what extent pornography use relates to relationship quality (Kohut et al., 2017; Maas et al., 2018; Vaillancourt-Morel et al., 2020).

To summarize, a minority of participants in the current study reported self-assessed negative effects of pornography use on their sex life. This finding is consistent with previous studies of self-assessed effects of pornography use in younger adults across gender and sexual orientation (Hald et al., 2015; Hald, Smolenski, et al., 2013; Hald & Malamuth, 2008; Hesse & Pedersen, 2017; Kvalem et al., 2014; Miller et al., 2018; Rissel et al., 2017; Wijaya Mulya & Hald, 2014). The ratio between the reported null, positive, and negative self-assessed effects of pornography use among participants who were married or in a steady relationship was also similar to the findings from previous studies (Dwulit & Rzymski, 2019; Kohut et al., 2017).

Results from our robustness check analyses showed no substantial change in user characteristics when the negative effect group was excluded. Although this suggests a certain degree of similarity between the mixed effect group and the negative effect group—relative to no effect group—the current study’s findings should not be interpreted in terms of self-assessed negative effects, specifically. Since self-assessed negative effects are repeatedly reported by a minority of individuals, a more nuanced approach (e.g., participant-informed, see Kohut et al., 2017) should be employed to examine which aspects of sexual life these individuals perceive to be negatively affected by their pornography use.

Pornography-related indicators—frequency of use, perceived pornography realism, and the use of non-mainstream pornography use—did not distinguish between the positive and the mixed/negative effects groups, when compared to participants in the no effect group. This suggests that the link between the self-assessed effects of pornography use and the context of its use is far from being uniform. For example, the interplay of personality characteristics (such as hypersexuality, paraphilic interest, mood disorders, shame and self-stigmatization, or low self-esteem) might moderate the link between pornography use and its perceived outcomes (see Bőthe et al., 2020). A future examination of self-assessed effects should take into account intertwined relationships between pornography use and antecedents that might determine the context of pornography use, as well as the effects of its use (Campbell & Kohut, 2017; Valkenburg & Peter, 2013).

Self-assessment of pornography use effects may serve as a minor departure and a possible corrective for the research area in which effects of pornography use are overwhelmingly assessed using statistical correlates of pornography use. With its limited exploratory aims, the current study identified a few discordances with previous correlational evidence (i.e., relationship satisfaction was unrelated to self-assessed effects of pornography use), as well as with public health hazard thesis (i.e., more frequent pornography use was related to positive effects of pornography use). This should prompt future researchers to thoroughly consider how the effects of pornography use are measured (Fisher & Kohut, 2020; Kohut et al., 2020) and interpreted.

In the current concern-driven climate over pornography use, our findings highlight the importance of exploring users’ self-assessed effects of pornography use. Self-assessment of pornography effects does not operate solely on personal and interpersonal levels but is embedded in a broader sociocultural context (Alves & Cavalhieri, 2019). Although mainstream media and policymakers have been particularly vocal in emphasizing negative aspects of pornography uses (Ley, 2018; Nelson & Rothman, 2020), the evidence about the self-assessed effects of pornography use does not support this discourse. To ignore or overlook the evidence about positive and neutral self-assessed effects of pornography use, the behavior is portrayed as inherently problematic and harmful—which may induce shame and self-stigmatization among some users.

Considering that almost all information about the self-assessed effects of pornography use was collected in sexually permissive Western countries, we would encourage researchers to conduct similar studies in countries with more traditional and sexually restrictive perspectives on sexuality (Rowland & Uribe, 2020). Different social organizations of sexuality regimes are likely to influence not only pornography use as such but also ways in which the behavior is self-perceived and self-evaluated.

Study strengths and limitations

Key strengths of the current study are probability-based sampling and high power. However, several limitations should also be noted. Firstly, the operational definition of pornography used in the current study is at odds with recent recommendations (Kohut et al., 2020). Unlike the recommendation, our definition focused solely on explicit depictions of sexual behavior (i.e., hardcore pornography), which is in line with how pornography was usually defined in studies of self-assessed pornography effects conducted among Scandinavian adults. It is not clear how the inclusion of nudity (i.e., erotica or softcore pornography) might change participants’ self-assessment of their pornography use, if at all. Secondly, due to the group’s small size and the analytical strategy employed in the current study, a detailed examination of the characteristics of participants who reported negative effects of pornography use was not feasible. It would be important to address this issue in future large-scale studies, perhaps by using a different operationalization of the construct. Thirdly, self-reported information about a (potentially) controversial topic should always be carefully considered for possible biases. Hald and Malamuth (2008) suggested that self-assessed reports about personal effects of pornography use may be affected by ingrained optimism or the tendency to perceive personal outcomes in a more positive manner than they objectively are. In addition to the use of self-reported measures, which is a typical limitation in research focusing on sexual behavior, we measured self-assessed effects of pornography use by a single-item indicator. Although there are more comprehensive measures, such as the Pornography Consumption Effects Scale (Hald & Malamuth, 2008; Miller et al., 2019), they are not suitable for large-scale national studies of sexuality and sexual health. Finally, our measure of self-assessed effects of pornography use was limited to intentional pornography users. The inclusion of unintentional pornography users would render direct comparison to self-assessments of intentional pornography users difficult. At the very least, excluding unintentional users (e.g., potential victims of forced exposure to pornography) might have resulted in an underestimation of reported negative effects of pornography use. On the other hand, participants included in our analyses were more likely to be men, younger, not religious, and non-exclusively heterosexual. If these sample differences affected our findings, they likely resulted in an overestimation of the reported positive effect of pornography use (see Hald, Smolenski, et al., 2013).

Conclusions

This study indicates that the majority of pornography users in Norway perceive the effect of their pornography use on sex life as either positive or neutral. Only a minority of participants reported mixed and even fewer indicated negative effects of pornography use. Our findings point to the sociodemographic, pornography-related, and sexual characteristics underlying self-assessments of positive or mixed/negative effects. Given the current research in the field, the finding that emotional intimacy and relationship satisfaction were unrelated to self-assessed effects of pornography use across gender is noteworthy.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Funding Statement

This work was financially supported by the Department of Psychology, University of Oslo.