Does porn cause anxiety?

The honest version has three parts, and only the first is a finding: there is a link, it is small, and nobody knows which way it runs.

What has been measured. The only meta-analysis to pool it put the correlation between pornography use and anxiety at about r = 0.16 across six studies. That is a real association whose confidence interval excludes zero, and it is also a few percent of the variance, from six observational studies. The link was significant among the cohorts measuring problematic or self-described addictive use, and was not shown in the single cohort measuring ordinary use — one cohort, which settles nothing in either direction. A separate review that read 40 studies on anxiety and compulsive sexual behaviour declined to pool them at all, and concluded the relationship was unclear.

That shape will be familiar. What the mental-health measures track is distress about use rather than amount of use: people who use frequently without distress and people whose use is out of control are distinguishable groups, and it is the second group that differs. Frequency on its own is a poor indicator of anything.

Direction is unknown, and anxiety-first is a live explanation. Using porn to take the edge off an anxious evening would produce exactly the correlation that has been found, and the meta-analysts name that possibility — pre-existing anxiety leading to use as a way of coping — as a major concern in their own limitations. The one study designed to separate "these people differ from those people" from "this person changed" followed more than four thousand adults across a year, and found the link was almost entirely the former: a stable difference between people, with the within-person paths small and, to the authors' own surprise, pointing the other way. A second cohort of several thousand young men found small effects running in both directions at once.

One exploratory detail, offered as exploratory: the correlation was stronger in samples with a mean age under 25 than in older ones.

The better-supported claim in this area is not about anxiety at all. Moral disapproval and religiosity predict feeling addicted to pornography even after controlling for how much someone actually uses, and that interaction shows up for pornography and gambling but not for tobacco or illicit drugs. Feeling bad about your own use is well documented. "Porn causes anxiety" is a different sentence, and a weaker one.

Which of the two you are in matters, because it changes what there is to work on. Shame is a concealment emotion — it points at withdrawal and hiding — and it predicts poorer behaviour-change outcomes than guilt does, which stays on the behaviour and points at action.

If anxiety is the thing you actually want gone, this is worth saying plainly: stopping has not been shown to fix it. The one meta-analysis of treatment for problematic use found no significant improvement in anxiety, on two studies and twenty participants between them — which establishes nothing except that nobody has shown an effect yet. Anxiety is treatable in its own right, by routes that have nothing to do with any of this.

A last note on where this evidence comes from. Two of the six pooled studies are built around moral disapproval rather than use, and both of the studies whose removal most changed the result were religiously conservative samples. That is a reason to handle the number carefully. It is not a licence to tell you that your anxiety is really guilt, because nobody can read that off a survey, least of all about you.

Where this comes from

Read the full lesson: What the research says about the actual harms →