Does Viewing Pornography Reduce Marital Quality Over Time? Evidence from Longitudinal Data (2016)

Printer-friendly version

Excerpt: "The current study was the first to test for directionality using nationally-representative, longitudinal data. The findings provide qualified support for the notion that more frequent pornography viewing—rather  than simply being a proxy for participants’ dissatisfaction with sex-life or marital decision-making—may negatively influence marital quality over time."

Arch Sex Behav. 2016 Jul 7.

Perry SL1.


Numerous studies have examined the connection between pornography viewing and marital quality, with findings most often revealing a negative association. Data limitations, however, have precluded establishing directionality with a representative sample.

This study is the first to draw on nationally representative, longitudinal data (2006-2012 Portraits of American Life Study) to test whether more frequent pornography use influences marital quality later on and whether this effect is moderated by gender.

In general, married persons who more frequently viewed pornography in 2006 reported significantly lower levels of marital quality in 2012, net of controls for earlier marital quality and relevant correlates. Pornography's effect was not simply a proxy for dissatisfaction with sex life or marital decision-making in 2006. In terms of substantive influence, frequency of pornography use in 2006 was the second strongest predictor of marital quality in 2012.

Interaction effects revealed, however, that the negative effect of porn use on marital quality applied to husbands, but not wives. In fact, post-estimation predicted values indicated that wives who viewed pornography more frequently reported higher marital quality than those who viewed it less frequently or not at all. The implications and limitations of this study are discussed.


Gender; Marital quality; Marriage; Panel data; Pornography; Relationship quality

PMID: 27388511

DOI: 10.1007/s10508-016-0770-y

[PubMed - as supplied by publisher]

From The Results Section of The Study

Several findings from the zero-order correlations in Table 2 are worth addressing. First, more frequent porn consumption at Wave 1 was negatively correlated with married participants’ satisfaction with their sex life and decision-making as a couple at Wave 1. While it is impossible to discern temporal precedence and directionality in this association, descriptively the correlation would suggest that, on the whole, married persons who use porn more often tend to report lower satisfaction with their sex-life and decision-making for whatever reason. Also worth nothing, while pornography consumption among married persons in Wave 1 was negatively correlated with marital quality at both waves, pornography viewing was actually more strongly correlated to the outcome measure at Wave 2 than in Wave 1. Specifically, while pornography consumption at Wave 1 is correlated with marital quality at Wave 1 (r = -.17; df = 600; p < .001), this correlation is slightly stronger at Wave 2 (r = -.23; df = 600; p < .001).

Overall, findings from the first two models strongly support the first hypothesis that pornography viewing, in its main effect, was strongly and negatively related to marital quality over time, and this effect was robust to the inclusion of controls for earlier satisfaction with sex-life and decision-making.

In order to test the second set of hypotheses, Model 3 includes an interaction term for porn viewing frequency × male in order to discern whether gender significantly moderated the link between pornography use and marital quality over time. The interaction term was significant and negative (b = -.36, p < .018; β = -.37), indicating that the negative relationship between porn use and marital quality at Wave 2 was stronger for men than for women. This supports Hypothesis 2a.

In general, those who never viewed pornography in Wave 1 reported scores higher than the mean for marital quality in Wave 2. But as pornography viewing increased in Wave 1 for the full sample, marital quality fell further below the average in Wave 2. Looking more closely, while there was a slight decline in marital quality at Wave 2 as porn viewing at Wave 1 increased, the biggest decline took place at the most extreme end of porn viewing.

Comparing married men with women, it can be observed that married men who never viewed pornography at Wave 1 reported equal or slightly higher marital quality at Wave 2 than married women who never viewed pornography. Yet as pornography viewing at Wave 1 increased for both women and men, men’s marital quality at Wave 2 declined more notably while the trend line for married women tells a different story. Women showed an initial decline in marital quality at Wave 2 as pornography viewing at Wave 1 increased similar to that of men. However, at porn viewing frequencies greater than “once a month” at Wave 1, women’s reported marital quality at Wave 2 increased and stayed relatively high. In fact, for women who viewed pornography in ranges between “2-3 times a month” to “once a day or more,” their marital quality was actually higher than those who never viewed pornography, and higher than the average marital quality for the full sample.


Scholars have often theorized that frequent pornography use can have negative effects on various aspects of marital quality. While studies have often (though not unanimously) found a negative association between porn use and relationship outcomes, in almost every instance the quantitative data have been cross-sectional, thus precluding the possibility of establishing directionality and testing for causal effects with confidence. The current study was the first to test for directionality using nationally-representative, longitudinal data. The findings provide qualified support for the notion that more frequent pornography viewing—rather than simply being a proxy for participants’ dissatisfaction with sex-life or marital decision-making—may negatively influence marital quality over time.

Consistent with previous research, this effect of pornography on marital quality applied almost exclusively to married men. In contrast, there was no evidence that frequent pornography viewing negatively influenced marital quality for women. In fact, some of the trends observed in Fig. 1 suggest that women’s marriages at Wave 2 were actually benefited by more frequent porn use in Wave 1. Following social learning or scripting perspectives, it could be that these findings simply point to the potentially negative consequences of frequent exposure to the content of pornographic material on men’s evaluations of their own relationships (Wright, 2013; Zillmann & Bryant, 1988). Yet while the general trend for men was that higher porn use led to lower marital quality, it appears that the marriages that were most negatively affected were those of married men who were viewing pornography at the highest frequencies (once a day or more). These levels of porn use were statistically extreme and may be suggestive of an addiction or otherwise compulsive behavior that could itself have a negative effect on romantic relationships, even if it were another behavior entirely besides porn use.9 Alternatively, it may be that men who engaged in the highest frequencies of pornography viewing at Wave 1 were in work situations where they were physically away from their spouses for long periods and thus the declining marital quality and higher porn use may have both stemmed from being physically apart. Future research on this topic would benefit from qualitative interview data that would help flesh out the mechanisms at work in the observed relationships.

9 Supplementary analyses were run to test for whether the men at more extreme levels of porn use were to blame for the statistically significant effect of porn use on marital quality for men. Results (available upon request) indicated the greatest difference was between those who did not view pornography at all and those who did, rather than between those who viewed pornography at moderate levels and those at more extreme levels.

In order to better frame the implications of these findings, several data limitations should be acknowledged. First, while the panel design and analysis permits the determination of temporal precedence and directionality of effect between pornography viewing in Wave 1 and marital outcomes in Wave 2, the fact that the question about porn use was not asked in Wave 2 precludes the possibility of determining whether and to what extent marital quality at T1 predicts pornography use at T2. Some research suggests that relational problems can predict pornography use (Paul, 2005; Stack et al., 2004; Willoughby et al., 2016) and it would be helpful to compare the bi-directional effects of porn use and marital quality over time. Future research would ideally make use of data that contain measures for both pornography use and marital quality at two different time periods so as to view which factor more strongly predicts the other. This data limitation also precluded the possibility of other estimation procedures that rely on changing scores, like fixed effects. Though the effects are so strong with the LDV models that fixed effects would be unlikely to change the substantive findings, these sorts of analyses would provide another test to ensure that omitted variable bias was not influencing the effects.

Second, while the pornography measure is an improvement on other measures that only ask whether a participant looks at pornography at all (e.g., the GSS), the measure does not specify the type of sexually explicit media that are used, but leaves this open for the participant to determine whether they are viewing “pornographic materials.” It could be that perhaps the difference between married women and men in the effects of porn use on marital quality is due to what types of pornography are typically consumed by either. To the extent that men consume pornography that is more likely to contain portrayals of female objectification and degradation while women are more likely to consume material that contains sensuality and intimacy, they may be influenced in their relationship behaviors and outlooks differently. Future research would thus benefit from measures that more explicitly define what sort of sexually explicit materials are being consumed and by whom. These sorts of data could help to test and elaborate on the scripting idea, that certain sorts of pornography provide scripts that consciously or unconsciously influence expectations about intimacy, sex, body images, etc. and thus influence committed romantic relationships (Willoughby et al., 2016; Wright, 2013).

A third limitation it was not possible to see whether married participants were viewing pornography alone or with their partner. As discussed above, scholars have more recently argued that pornography use, when done as a couple, can potentially benefit the relationship (Grov et al., 2011; Lofgren-Martenson & Mansson, 2010; Maddox et al., 2011; Weinberg, et al., 2010; Willoughby et al., 2016). The link between porn use and relationship outcomes may look differently for women and men to the extent that men and women engage in different use patterns. Studies find that men are considerably more likely than women to report viewing pornography alone (Maddox et al., 2011), and others find that men more often report using porn for masturbation while women are more likely to report using pornography primarily as a part of love-making (Bridges & Morokoff, 2011). While the current study was unable to test for these distinctions, to the extent that men are more likely to use pornography in isolation while women are more likely to do so as a way to build romantic intimacy, it would be unsurprising to find that married women’s relationships were benefited in some ways by greater pornography use, while men’s relationships seemed to be negatively affected by more frequent (isolated) use. Future research would ideally draw on data that would allow the researcher to control for whether the participants’ spouse also views pornography, how often, and whether they do it together.

Lastly, my analyses only included participants who were married at Wave 1 in 2006 and stayed married until Wave 2 in 2012. Thus, I do not show how pornography use may have contributed to the divorce of some couples in between Waves 1 and 2. The number of divorces among PALS participants between Waves 1 and 2 was unfortunately too few to conduct any meaningful analyses with (n < 30). However, the fact that the current study omitted persons who were divorced between Waves 1 and 2 actually makes the findings more conservative. It could be that porn consumption became so frequent that participants divorced thus leaving them out of the analytic sample. The fact that the sample only included couples who were married at both Waves means that the analysis only included couples for whom pornography had not caused a split. Future research would benefit from panel data with large enough numbers of divorces to adequately predict whether more frequent pornography consumption leads to greater likelihood of divorce over time. On a similar note, this study only included persons who were already married at Wave 1. Thus, it was not possible to see whether more frequent pornography use made persons less likely to get married at all. Future research would also benefit by testing for whether pornography viewing can diminish persons’ likelihood of entry into marriage or perhaps delay marriage entry. Or, conversely, perhaps couples’ porn use might contribute to their intimacy and thus lead to greater likelihood of marriage.