← chaosfactor.xyz

Does Porn Create Desire?

8,400 people rated whether porn exposure would make a man actually pursue 19 acts — from vanilla sex to bestiality and child abuse. With a twist: half rated a hypothetical man, half rated themselves. And every act has an independently measured taboo score.

Three Porn & Incentives surveys + the Taboo Ratings survey · analysis June 2026

TL;DRPeople believe porn corrupts other men — at almost exactly the rate they insist it wouldn't move them.

r = 0.71the "porn would change him, not me" gap grows with how taboo the act is — tiny for sex, maximal for bestiality & child abuse

39% vs 12%believe CSAM exposure makes a man more likely to assault a child — vs men's own self-prediction

r = −0.82porn is believed to channel mild desires toward the partner, taboo desires toward strangers

r = −0.39folk beliefs about which porn changes behavior are anti-correlated with what the BKS data actually shows

The shape of folk belief is clean: porn is thought to be a strong motivator for mild and moderately kinky acts (cunnilingus, spanking, choking — believed effect ≈ +1 or more), to do little for vanilla sex itself, and to do almost nothing — or even backfire — for the most taboo acts. But that calm assessment only applies to the self. Asked about a hypothetical other man, the same population assigns porn far more power, and the more taboo the act, the bigger the discrepancy. The most charged case: 39% believe CSAM would make a man more likely to assault a child; first-person, that's 12%, with a plurality saying it would make them less likely.

01The experiment

Three parallel surveys (n≈2,800 each; respondents each rated a random subset of acts, ~110–390 ratings per act-cell):

v1 & v2 — third person: "Assume a single man, who previously had never seen pornography, suddenly got a thumb drive with terabytes of videos of [X]. Would this make him more or less likely to attempt to go find someone new to do [X] with?" Each act also asked in a married frame ("…initiate [X] with his wife?"). Scale: −3 "Much less likely" → +3 "Significantly more likely."
v3 — first person: "Assume (over the course of a few months) you are forced to watch a hundred hours of videos of [X]. This happens around times where you're already kinda aroused. Would this make you more or less likely to attempt…" — branched by actual relationship status.
Taboo Ratings (n=6,426, separate sample): every act rated 0–5 for tabooness — giving us a measured stigma axis instead of guesswork (sex = 0.1, spanking = 0.9, pegging = 2.1, scat = 4.2, child sexual assault = 4.8).

These are beliefs, not measured effects of porn. Nobody was shown anything; this is a study of folk causal models — what people think porn does. That's exactly what makes the third-person/first-person comparison meaningful: same acts, same scale, different target.

02The act map: what porn is believed to do

Believed effect of porn by act, single vs married frames

Reading from the bottom: porn is believed to be a real motivator for oral sex, anal, spanking, choking — the "mainstream kink" band, all around +0.7 to +1.3. For vanilla sex itself, the believed effect is near zero (people apparently think watching sex doesn't make men pursue more sex — consistent with a substitution intuition). And at the taboo end, beliefs flip: scat, bestiality and CSA hover at zero or below in the single frame. One eyebrow-raiser: about a third of raters think gay porn would make a straight man more likely to seek sex with men (mean +0.18) — a folk belief in porn bending orientation.

03Belief falls as taboo rises — for "him" and for "me"

Believed effect vs measured tabooness, third vs first person

Against the measured taboo axis, both curves decline: porn is believed to motivate the mildly kinky, not the extreme. But the curves are offset — the third-person curve sits above the first-person curve everywhere except vanilla sex, and first-person predictions actually go negative at high taboo: men say being saturated in bestiality or CSA imagery would make them less likely to do anything of the sort (revulsion/aversion), while crediting other men with susceptibility.

04"Porn would change HIM — not me"

The classic third-person effect, act by act. Every act except vanilla sex shows a positive gap, and the gap scales with taboo at r=0.71:

Third-person minus first-person believed effect vs taboo

For low-taboo acts, people think porn affects other men about as much as themselves (sex: gap ≈ 0). For bestiality and child sexual assault the gap exceeds a full scale point. The pattern reads as stigma logic, not causal logic: the more morally horrifying the act, the more porn's power gets projected outward onto the hypothetical other man.

The CSA case, in numbers: third person — 39% "more likely," mean +0.18. First person — 12% "more likely," 40% "less likely," mean −0.91. Whatever the true effect of such material is, the public's model of other people and the public's model of themselves cannot both be right.

The first-person vignette ("forced to watch a hundred hours… while aroused") actually describes heavier exposure than the third-person one ("got a thumb drive"). If anything that biases against our finding — the self gets the stronger stimulus and still reports a smaller effect — but the wording difference is a real confound to keep in mind.

05Ask the wife, or find a stranger?

Comparing the married frame ("initiate with his wife") to the single frame ("go find someone new"), against taboo (r=−0.82):

Married minus single believed effect by taboo

For vanilla and mild acts, porn is believed to channel desire toward the partner (blowjobs +0.54, cunnilingus +0.47). Around taboo ≈ 2 the buffer disappears, and for genderplay, rapeplay, pee and scat it reverses — people believe a porn-saturated husband takes the weird stuff outside the marriage. A folk map of what men are expected to be able to ask their wives for.

06Who believes porn is powerful?

Mean believed effect by politics, porn use, and gender

Women credit porn with more power than men (mean believed effect 1.03 vs 0.66) — the people the vignettes aren't about believe most strongly in their effects.
Liberals believe more than conservatives (0.56 → 0.79 across the politics scale) — the opposite of the anti-porn-crusader stereotype, at least in this very-online sample.
Porn use is U-shaped: people who rarely watch (0.88) and people who watch far more than average (0.86) both believe porn changes behavior more than average users do (0.64). Abstainers and power users agree the stuff is potent.

07Folk belief vs what the data shows

The Big Kink Survey lets us compute, for each fetish, how much commoner that arousal is among daily porn users than among never-users (men). Plotting believed-effect against that reality:

Believed effect vs BKS porn-use prevalence ratio

The correlation is negative (r=−0.39). Bestiality is the single most porn-use-linked arousal in the BKS — 11× commoner among daily porn users — and the public believes porn does nothing for it (and scat, the other big "porn does nothing" belief, is ~2× linked). Spanking and high heels, believed to be among porn's strongest effects, show some of the weakest real gradients. Folk intuition tracks moral alarm, not the empirical porn-fetish association.

Why only 11 acts? The reality axis exists only for things the BKS measures arousal to — and the BKS is a kink survey, so it has no item for vanilla acts (sex, blowjobs, cunnilingus, anal, squirting), and no clean counterpart for tickling↔ a few others. So the 8 vanilla/unmatchable acts from the earlier charts simply can't be placed on this one. Matches marked * use the nearest fetish category (choking→brutality, pee/scat→secretions, heels→clothing, rapeplay→nonconsent, bimbo→mental-alteration, tickling→sensory); the rest are direct. And both axes are correlational — the BKS gradient could be "people with the fetish seek the porn" rather than porn creating it — so this compares one belief against one association, not against ground truth.

08Bonus: the Taboo Map

The companion survey is a useful artifact on its own — 139 sexual interests ranked by tabooness, with male and female raters largely agreeing (where they split, women rate race-related and degradation content more taboo; men rate male-receptive acts slightly more taboo):

139 sexual interests ranked by mean taboo rating

09Methods & limitations

Surveys: porn-and-incentives v1 (n=2,862), v2 (n=2,779), v3 (n=2,812), recruited mostly via Twitter (~75–90%) and FetLife; raters 89% male-identified; each rated a random subset → 12,871 act-ratings total, ~110–390 per act×frame×person cell (95% CIs throughout). Taboo scores: taboo-ratings survey (n=6,426, ~2,100 raters/item), item-level means used as fixed act properties; male-only taboo means give the same ordering. Third-vs-first comparisons use men only, matched frames. BKS comparison uses male daily-vs-never prevalence ratios with approximate category matches (rapeplay↔nonconsent, spanking↔sadomasochism, pee/scat↔secretions, heels↔clothing, choking↔brutality).

Beliefs, not effects. Nothing here measures what porn actually does — except by contrast with the BKS correlations, which are themselves not causal.
Vignette wording differs between third person ("thumb drive") and first person ("forced to watch 100 hours, while aroused") — direction of bias likely favors first-person effects, making the third-person gap conservative, but it's a confound.
Sparse cells. Some act-cells have n≈100; CIs are drawn everywhere and the headline patterns (r=0.71, the CSA gap) are far outside them.
Self-selected, very-online sample, skewed male and kink-tolerant; a general population might show even larger third-person effects (this sample is porn-friendly).
CSA items: interpret with care — they measure beliefs about a hypothetical man, and self-reports about oneself on a maximally stigmatized topic are exactly where socially desirable responding is strongest.