Context Matters: Moderating Effects in the Associations between Pornography Use, Perceived Addiction, and Relationship Well-being

Abstract

While perceived addiction and religiosity have been key contextual factors in understanding the link between pornography use and personal distress, these contextual factors have been explored less in the relational literature. Using a large nonprobability convenience sample from the United States, a moderated mediation model of the association between pornography use alone and two key indicators of relationship quality (relationship satisfaction and relationship stability) was explored. Results suggested that both general and aggressive pornography use alone were associated with less relationship satisfaction and relationship stability even when accounting for a range of potentially confounding variables. Perceived addiction partially mediated these associations, while both religiosity and gender moderated them. Generally, higher religiosity and being male were linked to compounding negative associations between pornography use and lower relationship quality. Findings suggest the importance of considering both religiosity and perceived addiction as important contextual factors when studying associations between pornography use and both relational and individual outcomes.

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

¹ While there may be many reasons for these differences, it is possible that older participants were more likely to encounter technical issues while taking the survey while those with children may have been more likely to get distracted and leave the survey unfinished (given most took the survey online at home). Religious individuals may have been more likely to stop the survey due to the sexual nature of the questions. These differences should be considered when interpreting results.

² In line with suggestions from Busby et al. (2020), specific content items were used to assess pornography use rather than giving participants a specific definition of pornography. While this may have led some participants to think about non-explicit media when answering the question, the high reliability of the scale, including items specifically referencing explicit content and those not, suggests that participants were thinking about explicit and pornographic content when answering the survey items.

³ Confirmatory factor analysis results available from the first author suggested that general pornography use and aggressive pornography use items were correlated but represented distinct constructs. Given strong gender differences in previous research, measurement invariance tests were run by gender for both pornography use measures. Configural invariance was tested by first testing model fit of a CFA with both pornography and aggressive pornography scales together but run separately for men and women. This suggested good fit for both groups (male: RMSEA = .071, CFI = .987, SRMR = .020; female: RMSEA = .054, CFI = .991, SRMR = .015). Then, a collective model was run that constrained model parameters to be equal across sexes. This model also showed good fit (RMSEA = .060, CFI = .989, SRMR = .016). Metric invariance was tested by constraining factor loadings to be equal across men and women. Changes in CFI and RMSEA were examined to see if this model differed from the configural model. Based on recommendations by scholars (Chen, 2007), changes in CFI of <  .010 and changes in RMSEA of <  .015 were used as cutoffs to suggest invariance. In this case, the metric model (RMSEA = .061, CFI = .987, SRMR = .027) showed a Δ CFI = −.002 and Δ RMSEA = −.001, suggesting metric invariance.

⁴ Models were run with and without covariate variables to see if the introduction of covariate variables influenced result patterns. While effects were generally stronger without covariates, the overall pattern of results was the same in both models.