On September 26, 2026 PsyPost reported on a meta-analysis mapping which personality traits and attachment styles actually track with problematic pornography use — and, more usefully for anyone in this industry, which ones track with ordinary use. The paper, "Personality, Attachment, and Pornography: A Meta-Analysis" (DOI 10.1177/00936502241287834, Communication Research), was authored by Mehdi Akbari of Kharazmi University with Shiva Jamshidi, Zahra Sadat Hosseini, Sonay Sheikhi, Rezvaneh Asadi Asadabad, Mahshid Zamani and Paul J. Wright. It pools 51 studies published between 2005 and 2024, covering 45,808 participants with an average age of 38, and splits the analysis into two distinct outcomes: general internet pornography use and problematic use marked by distress, perceived loss of control and damage to daily life or relationships.

Why It Matters

The commercially relevant finding here is the one the headline buries: the traits that predict *problematic* use are largely not the traits that predict use at all. Sexual compulsivity and impulsivity load onto the distressed minority; sensation-seeking, openness and extraversion load onto the ordinary majority. That distinction is the empirical basis for a case the industry has struggled to make with evidence — that adult consumption is a normal-distribution behaviour with a clinically distressed tail, not a uniform pathology. Expect this paper to be cited in exactly that argument, and expect it to be cited by the other side too, since sexual compulsivity being the top correlate is a real finding about a real subgroup. For product and platform design, the attachment result is the operational one. If anxious attachment predicts problematic use because content removes rejection risk, then the features that make platforms stickiest — parasocial intimacy, always-available responsiveness, simulated relationship continuity — are precisely the features most likely to deepen the pattern in the vulnerable subgroup. That is not a hypothetical concern in 2026: it describes the AI-companion category's core value proposition, and it lands the same month California's Adam's Law (SB 1119) imposed risk-assessment and covered-harm duties on companion-chatbot operators. A regulator building a "covered harm" record now has a 45,808-participant meta-analysis pointing at the mechanism. Sexual-wellness and therapy-tech platforms — Blueheart, Lover, AASECT-certified teletherapy — get the inverse opportunity: conscientiousness and self-control as protective factors is a screening and intervention target, not just a finding.

The strongest correlate of problematic use was sexual compulsivity — an intrusive, repetitive urge to engage in sexual behaviour — followed by impulsivity, the tendency to act without weighing consequences. Neuroticism correlated positively with both general and problematic use, consistent with content being used as an affect-regulation strategy. Sensation-seeking and narcissism correlated with overall consumption. Extraversion and openness also correlated with general use but were not statistically significant for problematic or addictive patterns — a distinction the authors are careful to preserve.

The protective side is arguably the more interesting finding. Conscientiousness and self-control both corresponded with lower consumption, general and problematic alike; agreeableness tracked with lower overall use, which the authors link to offline relationship success reducing the pull of compensatory online activity. On attachment, anxious attachment correlated with problematic use — the authors' reading being that content offers sexual stimulation without exposure to interpersonal rejection.

Two moderator results sharpen the picture: the trait–use associations held steady across gender, despite men consuming more in absolute terms, and held steady across age groups. Whatever these vulnerabilities are, they are not demographically local.

The limitations are stated plainly and matter. This is observational and correlational, so causal direction is unresolved — heavy use may erode self-control as easily as low self-control drives heavy use, and an unmeasured third variable such as a mood disorder could drive both. Most included studies drew on Western industrialised samples, definitions of "internet pornography" varied across the 51 studies, and small per-trait sample sizes (narcissism among them) blocked deeper comparisons.

Sources

Update — 2026-09-27

Initial entry — story first created.

Update — 2026-10-05

A second meta-analysis, from the same Iranian lab, looks at the emotional side. On October 4, 2026, PsyPost reported on a larger companion study: "Psychological correlates of problematic pornography use: A systematic review and meta-analysis of measurement and conceptualization differences," by Sonay Sheikhi, Shahram Mohammadkhani and Mehdi Akbari, published in Addictive Behaviors Reports (DOI 10.1016/j.abrep.2026.100741; online August 2026). Akbari and Sheikhi are also co-authors of the personality and attachment paper above. The new study pools 97 articles covering 111 independent samples and 121,728 participants (mean age about 27.5, about 46% women, almost all data collected online, published 2012–2025).

The variables most strongly associated with problematic pornography use (PPU) were craving and sexual shame, each with an average correlation of about 0.54. Smaller associations appeared for depression, loneliness, emotion dysregulation and moral disapproval of pornography. The methodological finding is more useful. How PPU is measured changes the results. Scales built around dysregulation (loss of control) tracked more closely with impulsivity and craving, while scales built around distress tracked more closely with shame and moral incongruence. The authors stress that prediction intervals were wide and often included values near zero, so the pooled estimates are average tendencies, not effects that apply to everyone.

Why it matters, alongside the earlier paper: Taken together, the two studies describe "problematic use" as at least two different things: a self-regulation problem, and a guilt-and-values problem in which people distressed by their use are not necessarily using more. That split matters for the expanding market of "porn-addiction" apps, blockers and coaching programs. Products built on shame-based framing may increase the very variable most associated with distress. It also matters for policymakers who treat any heavy use as harm.

New Sources