Charts

Showing the full self-selected sample (14–60). Switch to population-raked estimates (Appendix W has exact numbers).

Big Kink Survey · pornography use · analysis June 2026

What a million kinky people
say about porn

Every porn-related finding in the Big Kink Survey, split by biological sex, age-controlled where it matters — with special attention to the four items added in May 2026 (shame, healing, sex-life impact, porn-vs-sex).

N = 1,068,333 (403k M / 665k F) live pull 2026-06-09 ages 14–60 · median 21 unweighted, self-selected 95% CIs throughout

Never use porn/erotica

4.6% / 8.3%

weighted: 6.2 / 11.4

Use weekly+

85% / 62%

weighted: 82 / 56

Use daily+

44% / 19%

weighted: 39 / 17

Median onset age

13 / 13

mean 12.5 / 13.0

Net sex-life effect

−7 / +13

% improved − % damaged

Throughout: ■ Male · ■ Female. "Weighted" = population-raked (sex×age×cis/trans×BMI×politics×ethnicity), ages 14–34, Western countries.

How many people are using pornography?

Nearly everyone. 95% of men and 92% of women in BKS use porn or erotica at least sometimes; the modal man uses it multiple times per week, and 18% of men report multiple times a day. Population-weighting moves these only a few points (full weighted versions of every number: Appendix W).

Age-confound check on the headline sex gap: the base rates above pool ages 14–60, and men in the sample are ~2 years older than women (median 22 vs 20) while use rises with age — so part of the M/F gap could in principle be age composition. It isn't: age-standardizing the male rates to the female age distribution moves daily+ only from 44.0% → 43.3% and never from 4.6% → 5.2% (vs women's 19.2% / 8.3%). The ~25-point sex gap survives essentially untouched, and the weighted numbers are age-raked by construction. Per-age-bin rates are in the table under Fig 2 and Appendix T1b.

Frequency distribution by sex — Fig 1. Full frequency distribution. BKS asks "watch **or read** pornographic/erotic content… for arousal" — including erotica is why female numbers run higher than video-only surveys.

Key finding "Use peaks in adolescence" is wrong for frequency. Male daily+ use peaks at ages 26–34 (47%), not in the teens (41%), and the share of never-users falls with age in both sexes. Only initiation is adolescent (§2).

Daily+ use by age — Fig 2. % using daily+ by current age. Female use is nearly flat from 14 to 60.

Confounders: politics matters a little, current devoutness matters a lot

Very-conservative men use less than very-liberal men — but over three-quarters of very-conservative men still use weekly+. The religion story depends entirely on which religion question you use: the religion someone was raised in barely predicts use, but current devoutness predicts substantially less — devout men 63% weekly+ vs non-religious 84%, and devout never-use jumps to 20–24% (vs 6–10%).

Use by politics — Fig 3. Weekly+ use by politics, ages 18–34. Gradient real but modest.

Use by current religiosity — Fig 4. Weekly+ use by **current** religiosity ("Are you currently religious?"), ages 18–34. Clean downward gradient. NB: childhood-religion *denomination* shows near-identical bars — most raised-religious respondents are now secular, which is why the raised-in measure looks null.

Use by relationship status — Fig 5. Singles use most; married people use no less than people in serious relationships. Same weekly+ metric as Figs 3–4 for comparability; the daily+ version of the gap: single men 49% vs partnered 38%, single women 22% vs 14%.

Onset: the real adolescent peak — and a generational convergence

Initiation peaks at 11–14 for both sexes: 78% of male users and 68% of female users started by 14. Meanwhile masturbation onset is essentially identical by sex (mean 11.7 vs 11.8) — the sex difference is in porn adoption, not sexual development.

Onset age distribution — Fig 6. Age began using semiregularly (users only). Women show a longer late tail (7% started at 19+, vs 2.8% of men).

Key finding — the convergence chart Women aged 45–60 mostly started porn as adults (30% by age 14); girls today start at 11, same as boys. The female onset curve converged to the male one within about twenty years.

Cohort onset convergence — Fig 7. % of users who started by 14, by current age (= the generation they grew up in). Caveats: recall bias; youngest bins right-censored.

Onset age vs current heavy use — Fig 8. Earlier start → heavier adult use, perfectly monotonic in both sexes (men starting ≤8: ~62% now daily+, vs 34% for adult starters). Not causal: higher-drive people plausibly start earlier *and* use more.

What they consume

The biggest modality split: 26% of women consume mostly/entirely written erotica vs 5% of men. Any survey that asks only about "watching porn" structurally undercounts women — a definitions point directly relevant to inconsistent measurement of porn prevalence.

Written vs visual — Fig 9. Written ↔ visual preference among users.

Key finding — counterintuitive Women consume more violent porn than men. "Most or all of it violent": 14% of women vs 6% of men. Among daily+ women, half say a moderate amount or more of their porn depicts nonconsent, struggling, or aggression. Pairs with BKS's broader finding that nonconsent fantasy skews female.

Violent porn distribution — Fig 10. "How much of the porn or erotica you watch is violent?" (users).

Violent porn by frequency — Fig 11. Violent content scales with use frequency — fastest for women.

Fact	Men	Women
Ever paid for porn	36%	12%
Pay regularly	2.8%	0.9%
Mostly/entirely animated content	18%	22%
Find opposite-gender pairing erotic (two women for men / two men for women)	74%	57%

The new items: shame, healing, sex life, porn vs sex

Added May 2026 — only n ≈ 15–18k each, but plenty for stable estimates. These are the freshest numbers in the room.

4a · Shame

About 1 in 3 users feels moderate-to-large shame about their use — and men and women are statistically indistinguishable. The surprise is the direction of the frequency gradient:

Shame distribution — Fig 12. "Do you feel shame about your pornography/erotica use?"

Shame by frequency — Fig 13. **Shame is highest among the LIGHTEST users** (48% of rare-user men) and lowest among daily users (~30%). The non-user bar is split: **never-exposed** people have little shame (14–18%), while **"quit"** (exposed, zero current use) men run 35% — the apparent never-user shame belongs to people with porn histories. See §4e.

Key finding — moral incongruence Measured with current religiosity, this is a clean dose-response: devout men 66% moderate+ shame vs non-religious men 27% (devout women 53% vs 28%), rising monotonically across "a little → moderately → very devout." Same-behavior-different-meaning: among daily+ users alone, shame still scales with devoutness. This strongly confirms the moral-incongruence account rather than qualifying it.

Shame by current religiosity — Fig 14. Porn shame by **current** religiosity (users 18–34). **Methodological note:** the religion someone was *raised in* washes this out almost completely (Christian 33% vs nonreligious 27%) — pick the wrong religion variable and you'd wrongly conclude moral incongruence is weak. Which measure you use changes the answer.

Shame by politics — Fig 15. Conservative men ~39% mod+ shame vs liberal men 27%.

4b · Healing / therapeutic

52% of male users and 57% of female users say their porn use has been at least a little healing or therapeutic. It rises with frequency, and — the notable part — with mental-illness burden:

Therapeutic by frequency — Fig 16. Daily+ women: 31% "moderately/largely healing" (vs 6% of rare users).

Therapeutic by mental illness count — Fig 17. People with 3+ diagnoses are ~2× as likely as the diagnosis-free to call their use therapeutic. Directly relevant to "coping can be adaptive" and "where's the positive-effects research?"

Shame and healing coexist: their correlation is only −0.09, and 27% of people reporting moderate+ therapeutic benefit also report moderate+ shame. They are not opposite ends of one axis.

Porn shame and therapeutic feelings by age, use level and sex — Fig 17b. **Both items across age, split by how much you use.** Porn **shame falls steeply with age** in every use-group and both sexes (women using weekly: 38% moderate+ in the teens → 9% by 40+; daily-using men: 37% → 9%) — the same inverse pattern as Fig 13, now shown to ease off over the lifespan. The **therapeutic/healing** read is flatter, and for heavier-using women it actually *rises* with age. Heavier users sit higher on **both** at most ages — consistent with the coexistence point above. Metric: % "moderate or larger"; shaded 95% CI; n ≈ 18k (shame) / 16k (healing). Use-degree excludes non-users; toggle for the weighted twin (ages 14–34).

4c · Effect on sex life — the money item

Headline finding Women are net-positive about porn's effect on their sex life (+13); men are net-negative (−7) — and the male damage is concentrated in young men. Male net bottoms out at −13 (ages 22–25) and flips positive at 35+, reaching +25 by 45–60. Women are positive at every age, rising to +39.

Sex life impact distribution — Fig 18. "Exposure to erotic content has ___ your sex life." Half of everyone says no impact.

Sex life net effect by age — Fig 19. Net (% improved − % damaged) by age. The "porn damaged my sex life" population is specifically **young men**.

The same split by bodycount sharpens it further — and it's one of the cleaner sex-divergence stories in the deck:

Key finding — the damage is concentrated in inexperienced men Men flip from net-negative to net-positive as partner count rises — virgins −13, crossing zero around 10–19 partners, reaching +17 at 50+. Women are net-positive at every level except as virgins (near-neutral +2), jumping to +16 with even one partner and +31 by 10+. So the "porn hurt my sex life" report is specifically a low-experience-men phenomenon; for anyone with a moderate sexual history — and for women almost regardless — porn reads as a net positive.

Sex-life effect of porn by bodycount and sex — Fig 35. Net effect by lifetime partner count (men 18–60 n≈5,952 / women n≈6,119). Plausible readings: porn fills in for absent partnered sex (and disappoints by comparison) when experience is low, but complements an active sex life once it exists. Caveat: bodycount correlates with age, and this isn't age-controlled — older, more-experienced people also skew positive (Fig 19).

Fine print: by frequency, male damage is flat-ish across use levels — and worst among non-users with porn histories, the "quit" group examined in §4e. Women get more positive with more use. Conservative men are most negative (−16 to −19) vs significantly-liberal men at −1; women are positive across the whole political spectrum.

4d · "Porn is usually more satisfying than sex with a real person"

21% of women agree vs 16% of men — women agree more, which pairs neatly with the written-erotica and violent-content findings: porn delivers content and control that partnered reality often doesn't. A majority of both sexes disagree.

Porn more satisfying by frequency — Fig 20. Agreement rises with frequency (daily+ women: 35%) and singleness (single women 27% vs partnered 17%).

4e · Who stops? The "quit" group — and a survivorship warning

There is no direct quitting item in BKS — but the sex-life question offers "I haven't been exposed" and isn't gated on use. So respondents reporting zero current use who still rated porn's effect are, by their own account, exposed non-users. They look substantially like former real users rather than incidental browsers: 52% say porn induced fetishes in them (vs 26% of the never-exposed), 8% have paid for porn (vs 1%), and 60% answered "has your porn use been therapeutic" rather than "I haven't viewed any." Read them as a blend of true quitters and light past exposure.

Quit group vs current users on four outcomes — Fig 36. The quit group vs never-exposed and current users, four outcomes (18–34). Quitters report the **worst sex-life damage** (men −24, women −11 — while every current-user female group is net *positive*) yet the **best dating and depression numbers** among the porn-experienced.

Key finding — survivorship bias, measured The people reporting the most porn damage are precisely the ones no longer in the user pool. Whoever felt harmed appears to have selected out — which means every cross-sectional study of current users (including this one) underestimates harm and overestimates benefit. The women's numbers make it vivid: ex-exposed women are net −11 on sex-life impact while every current-user group runs +12 to +20. This is a general limitation of cross-sectional designs, and it speaks to "why do people stop?": the data pattern is consistent with stopping because it felt harmful — and with quitting being feasible for the more-functional (quitters have lower depression and better dating than daily+ users).

Honest labeling "Quit" is an inference from gating, not a survey answer; the group can't distinguish a former daily user from someone who saw porn a handful of times. n is small (≈290 men / 390 women at 18–34). A direct item — "Did you previously use porn/erotica regularly? Why did you stop?" — would settle it, and at current intake (~15–18k/month on new items) would yield a real quitting dataset within weeks.

Full-sample analogs (n > 1M) The older arousal-pattern items replicate at scale: 66–67% of both sexes agree "I am ashamed/embarrassed about at least some of what arouses me"; 60–62% agree engaging with what arouses them "feels therapeutic or healing." Contrast worth noting: shame about what arouses you rises with use frequency; shame about use itself (Fig 13) falls.

Porn × dating, drive & relationships

How does porn use line up with real-world sexual and romantic life — current arousal, the drive to seek out real sex, dating difficulty, casual-sex experiences? Three of these items (sexmotivated, badatdating, plus the May-2026 satisfaction item) are recent additions; horniness and hookup quality come from the large older items. Read everything here as correlational — the master confound is libido itself: high-drive people both watch more porn and report more of nearly everything below.

5a · Current horniness

Heavier users are simply hornier in the moment: 51% of daily+ men and 50% of daily+ women say they're "quite" or "very" horny right now, vs 34% / 26% of never-users. The jump is concentrated at the daily+ level.

Horniness by porn frequency — Fig 26. The libido confound made visible. This does *not* show porn "causing" arousal — it's the shared-cause pattern you'd expect if drive drives both. Worth stating up front so the dating/drive results below are read correctly.

5b · Drive to seek real sex

Key finding — counters the substitution narrative The motivation to seek out real-life sexual encounters rises with porn use, not falls — daily+ men 48% vs never-users 40%; daily+ women 43% vs 34% — and it holds within both single and partnered groups. So at the population level porn looks like a complement to partner-seeking, not a replacement for it. The "porn kills young men's drive to go find a partner" story isn't the average case.

Sex drive by porn frequency — Fig 27. "I am very motivated to seek out real-life sexual encounters." Note this cuts against a naïve reading of the sex-life and satisfaction items — most heavy users still want real sex *more*, not less.

5c · Dating difficulty — the strongest real-world correlate

This is the steepest gradient in the whole dating/drive cluster. Self-reported difficulty with dating climbs sharply with porn frequency, especially for men: 61% of daily+ men say "I am bad at dating" vs 40% of male never-users (women 53% vs 35%).

Dating difficulty by porn frequency — Fig 28. "I am bad at dating (asking people out, going on dates)." Direction is ambiguous — porn as a refuge from frustrating dating is at least as plausible as porn eroding dating skill — and social anxiety plausibly drives both. But it's a robust association.

And critically, it survives controlling for relationship status — it's not just that single people both date worse and watch more. Within singles and within people in serious relationships, the gradient persists:

Dating difficulty by porn frequency and relationship status — Fig 29. Single daily+ men 72% vs single never-users 58%; partnered daily+ men 43% vs 29%. The gradient is real inside each stratum.

The sharper control is lifetime partner count (sexcount) — a direct proxy for real-world sexual success, and the obvious worry behind "maybe it's just inexperienced people watching porn." It does two things at once:

Key finding — survives the partner-count control Partner count is the bigger driver (men using porn non-weekly: 59% bad-at-dating at 0 partners → 21% at 10+). But the porn gradient persists inside every partner-count band: even among men with 10+ partners, daily+ users report dating difficulty at 37% vs 21% for non-weekly users (+13–20pp slope in every male stratum). So it is not just "virgins watch porn and can't date" — the association holds among the sexually experienced too. And heavy users don't have fewer partners (men daily+ mean 6.9 vs never 4.3), so it isn't raw partner-substitution either.

Dating difficulty by partner count and porn use — Fig 32. Lines slope down (partner-count effect) but stay separated (porn effect within stratum). One nuance: virgin women (0 partners) show almost no porn gradient (+6pp) — they're uniformly high; the porn gradient is a feature of women *with* experience.

Dating difficulty across the full bodycount range, men, by porn use — Fig 33. Same idea across the full partner-count range (men 18–60, n = 11,419), with three clean frequency bands. The "never / ≤1×year" line **stops at 10–19 partners** — men who barely use porn but have 50+ partners are too rare to estimate, which is itself a data point. The 50–99 dip and 100+ uptick are small-n noise (wide CIs).

5d · Casual-sex experience (women)

The casual-hookup-quality item is women-only. Counter to a "porn ruins women's real sex" expectation, women who use more porn report better casual sex: among daily+ women, 46% say hookups have been good and 28% bad, vs 33% good / 37% bad among never-users. Consistent with the §4c finding that women are net-positive on porn's effect on their sex lives.

Women's hookup quality by porn frequency — Fig 30. "In general, casual sexual hookups have been a ___ experience for me" (women only, n≈82k in this 18–34 cut).

5e · The substitution subgroup is real — but a minority

All of the above is the average. There's a distinct minority for whom porn genuinely looks substitutive: people who agree "porn is more satisfying than real sex" (~16% of men, 21% of women) are meaningfully more dating-averse, lower-drive, and more often single. So both stories are true at once — net complement, minority substitute.

Fig 31. Porn-preferrers vs porn-dispreferrers on three real-life markers. The gap is consistent but the group is small — useful for "are there populations for whom porn is uniquely harmful?" without overclaiming it's the norm.

5f · What confounds the porn ↔ dating link? (the capstone)

The dating-difficulty result survived controlling for relationship status (Fig 29) and partner count (Figs 32–33) — so what does explain it? Mapping every candidate variable by how it correlates with both porn use and dating difficulty isolates the real confounders: variables sitting in the shaded "loads-on-both" quadrants.

Confounder map: correlation with porn vs with dating difficulty — Fig 34. Each dot is a survey variable (men 18–34). **Red = correlates with both in the same direction** → a genuine confounder. **Self-rated attractiveness** is the strongest, with a cluster of temperament/mental-health traits (shyness, autism, neuroticism, depression, mental-illness load, social anxiety). Note the contrasts: **partner count** sits on the y-axis (drives dating, *not* porn — which is why Fig 32 survived it); **sex motivation** and **horniness** sit in opposite quadrants (they're suppressors, correlating with porn the other way).

Key finding — the link is mostly a shared cause Partial correlation of porn × dating-difficulty, before and after controlling these confounders:

	raw ρ	+ attractiveness	+ everything
Men	+0.136	+0.091	+0.058
Women	+0.137	+0.107	+0.033

Controlling attractiveness + the temperament/mental-health cluster absorbs ~57% (men) to ~76% (women) of the association. A weak residual survives (men ρ≈0.06), and the porn gradient still shows up within attractiveness strata (+6 to +13pp), but the headline is: most of "heavy porn users are bad at dating" is a third-variable story — less-attractive, more socially-anxious people both date with more difficulty and use more porn. This argues against a causal reading in which porn causes dating difficulty.

Caveat on the top confounder hotterthanothers is self-rated attractiveness, itself entangled with self-esteem and mental health — so it likely absorbs some variance that isn't "looks" per se. Treat it as "self-perceived desirability," not an objective beauty measure.

Interpretation In short: porn frequency tracks higher drive and (for women) better real sex, but also more dating difficulty — and a minority treat it as a substitute. Every line here is cross-sectional and libido-confounded; none of it licenses a causal "porn makes men bad at dating" claim. The dating-difficulty result is best read through a "refuge vs cause vs shared social-anxiety" lens rather than as evidence of a causal effect.

Mental health

Within sex and age band, depression rises with frequency — but only by ~10 points from never to daily+, and nobody's mental health is better among non-users. The much steeper correlate is onset age.

Fig 21. Depression by frequency, ages 18–34 (gradient survives the 18–24 / 25–34 split). Anxiety is flatter for men (~4pp).

Depression by onset age — Fig 22. Women who started at ≤6: 62% depression and mean 4.7 diagnoses, vs 45% / 2.6 for late starters.

Causality caveat — say this out loud Cross-sectional. Very-early porn exposure is plausibly a marker of chaotic or abusive childhood, not a cause of adult mental illness. But it usefully reframes "porn and mental health" as an exposure-timing question rather than a dose question. Also remember this is a high-burden sample (57% of women 18–34 report anxiety); weighting drops levels (any-MH 76→69%) without changing shapes.

Porn-induced fetishes

~85% of fetish-havers say porn induced new fetishes in them; a quarter of men say "new and totally different" from anything pre-existing. Dose-responsive in both frequency (93% among daily+ vs 68% among rare users) and onset (early starters report more "totally different"). Self-reported induction also tracks actual fetish counts: mean 6.2 fetishes for "no" vs 9.9 for "totally different."

Special populations

Porn performers

Among sex workers (n≈20k), 8–12% have studio porn-acting experience. Of those, 69% of men and 60% of women rate the experience positively — but the negative minority is real (40% of women), and performers who left the industry traumatized are plausibly underrepresented in a kink survey.

Respondents reporting pedophilic arousal (handle with extreme care, if at all)

BKS gates a question to respondents reporting any arousal to prepubescent children (n≈15k): would consuming erotic content about children increase or reduce their own likelihood of acting? 48% say reduce, 38% no effect, 14% increase — sexes nearly identical.

Content-as-risk belief — Fig 25. This is the population's **self-theory**, not outcome data; self-serving bias is obvious. But it's a rare direct measurement bearing on catharsis-vs-escalation, and on "populations for which porn is uniquely helpful or harmful." These are fantasy-reporters, not offenders.

Methodology notes

1 · Definitions drive prevalence. "Watch or read erotic content for arousal" yields 92% female use; surveys saying "pornography" (read: video) lose the 26% of women who are mostly-written consumers. Part of every male–female "gap" in the literature is a measurement artifact.

2 · Self-selection, measured. BKS is enormous, but its self-selected sample is sex-positive, very liberal, and extremely online. Raking to population demographics moves prevalence a few points (men daily+ 44→39%) and changes no qualitative conclusion — a useful empirical answer to "does your weird sample matter?" Selection plausibly matters less for correlations than for levels. The deeper selection problem is survivorship within the user pool (§4e): those who felt harmed appear to quit, so current-user samples — everyone's, not just BKS's — skew toward people porn works for.

3 · Which measure you pick changes the conclusion. The sharpest example in this dataset: childhood religion barely moves either use or shame, while current religiosity strongly predicts both (less use, far more shame). Use the raised-in variable and you'd report a null on moral incongruence; use the current-devoutness variable and you confirm it. Same caution applies to self-report quirks worth flagging — never-users reporting shame about use; "therapeutic" and "shameful" co-endorsed by the same people; the rare-user shame spike — single items hide motive structure.

4 · "Adolescent peak" conflates onset with use. Onset peaks at 11–14 and is converging across sexes by cohort; frequency peaks at 26–34 for men and is flat for women. The pleasure→coping framing looks shaky too: therapeutic-framing rises with frequency and mental-illness load while use-shame falls — heavy adult users look more like comfortable copers than escalating addicts.

Common questions about pornography use

"How many people are using pornography?"

Nearly all: 95% of men, 92% of women (and that's with erotica counted — definitions matter). §1

"Why do people begin, continue, and stop using pornography?"

Begin: at 11–14, alongside masturbation onset. Continue: no direct motive item, but over half call it at least a little healing; the mentally burdened say so most; coping looks adaptive. Stop: the exposed-non-user ("quit") group reports the worst sex-life damage but the best dating/depression — consistent with stopping because it felt harmful. §4, §4e

"Do motives matter?"

The same behavior carries opposite valence by person: identical frequency → shame for conservatives/religious, "therapeutic" for the mentally burdened. Context, not dose. §4

"Can porn be good for mental health, or must it be neutral/negative?"

Self-report says yes-for-some: women net-positive on sex-life impact at every age; 1 in 3 daily-user women calls it moderately+ healing. The negative population is specific: young men. §4, §6

"Does porn replace the drive to find a partner / cause dating problems?"

Not on average: real-sex motivation rises with porn use, women's casual sex gets better. But dating difficulty rises steeply (daily+ men 61% vs 40%), surviving relationship-status controls — and a ~16–21% minority treat porn as a substitute. Complement on average, substitute for a minority. §5

"Why is use framed as peaking in adolescence?"

Because onset does. Frequency doesn't — it peaks at 26–34 in men. §2

"Are there populations for which porn is uniquely helpful/harmful?"

Helpful: women (sex-life net +13, better hookups), the mentally burdened (therapeutic ×2). Harmful-feeling: young men (net −13 at 22–25), conservative/religious men (shame, damage reports), the dating-averse substitution minority. Plus the gated MAP self-theory data. §8

Appendix · Weighted results

Frame: ages 14–34, Western countries, n=481,560 (Kish effective n≈160k). Raking: sex×age×cis/trans×BMI×politics×ethnicity. The unweighted columns are recomputed on the same frame (weights off), so differences show the pure effect of weighting; they can differ slightly from main-report numbers computed on the full 14–60 sample. Weighted CIs use Kish effective-n. Because raking includes sex×age, weighted numbers are inherently age-adjusted.

Not weightable from this extract: diagnosis count, self-rated attractiveness, autism/shyness (§5f confounder analysis), performer ratings, fetish counts, and anything 35+ (the male sex-life age-flip, the 45–60 onset cohort tail).

T1 · Headline use rates (ages 14–34)

Weighting rakes sex×age to population, so weighted numbers are inherently age-adjusted.

metric	M unweighted	M weighted	F unweighted	F weighted
% never	5.8	6.3 ±0.2	9.7	11.5 ±0.2
% weekly+	83.7	82.1 ±0.3	60.1	55.7 ±0.4
% daily+	42.2	39.2 ±0.3	18.4	17.1 ±0.3
% multiple/day	17.1	15.3 ±0.2	6.7	6.0 ±0.2

T1b · Base rates by age bin

Within-bin rates remove the M/F age-composition difference entirely.

metric	M unweighted	M weighted	F unweighted	F weighted
% daily+ · age 14-17	41.2	35.6 ±0.7	21.5	19.8 ±0.5
% daily+ · age 18-21	37.4	35.5 ±0.5	16.6	16.2 ±0.3
% daily+ · age 22-25	44.2	40.2 ±0.6	17.6	15.6 ±0.5
% daily+ · age 26-29	46.2	42.5 ±0.8	18.2	16.9 ±0.7
% daily+ · age 30-34	46.9	42.3 ±0.9	18.6	16.9 ±0.9
% never · age 14-17	8.0	10.1 ±0.4	10.9	13.7 ±0.4
% never · age 18-21	8.4	9.2 ±0.3	13.3	14.6 ±0.3
% never · age 22-25	4.2	5.3 ±0.3	7.8	10.5 ±0.4
% never · age 26-29	2.9	3.9 ±0.3	5.9	8.8 ±0.6
% never · age 30-34	2.1	2.9 ±0.3	5.7	9.8 ±0.7

T2 · Onset

Cohort gradient is truncated here (frame caps at 34; the 45-60 cohort lives only in the unweighted full sample).

metric	M unweighted	M weighted	F unweighted	F weighted
% of users started by 12	44.7	43.9 ±0.3	40.1	34.1 ±0.4
% of users started by 14	80.8	79.7 ±0.3	66.3	58.0 ±0.4
mean onset age (binned)	12.23	12.32 ±0.0	13.16	14.05 ±0.0
mean masturbation onset (fapage)	11.72	11.77 ±0.0	11.93	12.38 ±0.0
cohort: % started by 14 · now 14-17	91.5	91.8 ±0.4	88.3	87.2 ±0.4
cohort: % started by 14 · now 18-21	81.7	81.9 ±0.4	70.7	69.8 ±0.5
cohort: % started by 14 · now 22-25	78.1	78.0 ±0.6	58.6	53.4 ±0.7
cohort: % started by 14 · now 26-29	76.7	76.0 ±0.7	49.7	43.7 ±1.0
cohort: % started by 14 · now 30-34	71.0	70.7 ±0.9	42.2	34.4 ±1.2
% NOW daily+ · started ≤10 (18-34)	57.6	54.7 ±1.1	28.8	27.0 ±1.1
% NOW daily+ · started 15-16 (18-34)	37.3	34.9 ±0.9	15.7	16.6 ±0.8

T3 · Content type & payment (ages 14–34)

metric	M unweighted	M weighted	F unweighted	F weighted
% mostly/entirely WRITTEN (users)	4.4	3.0 ±0.1	26.6	25.0 ±0.3
% mostly/entirely animated (users)	16.8	11.2 ±0.2	20.3	17.2 ±0.3
% violent porn moderate+ (users)	22.5	20.2 ±0.3	34.9	33.2 ±0.4
% violent porn most/all (users)	5.6	4.7 ±0.1	13.3	12.2 ±0.3
% violent mod+ among DAILY+ users	28.2	25.7 ±0.5	49.1	47.9 ±0.8
% agree opposite-gender pairing erotic	75.0	80.1 ±0.3	53.1	47.0 ±0.4
% ever paid for porn	33.5	34.2 ±0.3	12.8	10.9 ±0.2

T4 · Shame (NEW item; ages 18–34)

New-item cells are small in the weighted frame (total n≈8.4k split by sex × group) — expect wide CIs.

metric	M unweighted	M weighted	F unweighted	F weighted
% mod+ shame (all respondents)	31.2	30.4 ±2.5	26.8	26.2 ±3.6
% mod+ shame · QUIT group	39.5	38.4 ±9.6	19.4	18.1 ±12.1
% mod+ shame · rare users (1-3)	49.4	40.5 ±15.4	37.9	40.7 ±12.6
% mod+ shame · daily+ users	28.9	28.4 ±4.1	24.2	23.3 ±9.3
% mod+ shame · users, not religious	25.0	23.9 ±3.4	23.9	21.6 ±4.7
% mod+ shame · users, very devout	69.4	66.4 ±10.0	52.8	54.3 ±16.5
% mod+ shame · users, sig-liberal	27.0	24.1 ±3.6	24.8	24.0 ±3.4
% mod+ shame · users, mod/sig-conservative	37.2	34.1 ±5.6	37.0	41.4 ±11.2

T5 · Healing/therapeutic (NEW; 18–34)

Mental-illness-count gradient can't be weighted (TotalMentalIllness not in weighted extract).

metric	M unweighted	M weighted	F unweighted	F weighted
% at least 'a little' healing (viewers)	46.6	47.3 ±2.9	54.0	52.1 ±4.4
% moderate+ healing (viewers)	12.0	11.8 ±1.8	15.5	12.7 ±2.9
% mod+ healing · daily+ users	18.2	17.4 ±3.5	31.9	23.3 ±9.4
% mod+ healing · rare users (1-3)	5.1	8.6 ±9.2	5.2	2.7 ±4.4

T6 · Sex-life impact, net = %improved − %damaged (NEW; 18–34)

The 35+ male flip to positive is OUTSIDE this frame (caps at 34). Bodycount capped at 10+ here (older high-count people excluded by frame).

metric	M unweighted	M weighted	F unweighted	F weighted
net sexlife (all answering)	-13.0	-18.1 ±4.1	+19.2	+18.5 ±5.9
net · rare 1-3	-20.0	-15.8 ±23.0	+15.6	+13.2 ±16.3
net · weekly 6-7	-11.7	-17.9 ±6.3	+23.6	+23.1 ±8.4
net · daily+ 8-9	-10.7	-15.8 ±7.2	+29.0	+32.6 ±16.1
net · QUIT group	-24.6	-29.7 ±12.8	-9.4	-2.9 ±19.7
net · bodycount 0	-16.3	-22.5 ±7.0	+4.6	+4.5 ±9.1
net · bodycount 1	-12.9	-17.4 ±8.0	+24.5	+18.0 ±14.5
net · bodycount 2-3	-11.0	-14.4 ±9.1	+21.5	+19.1 ±12.7
net · bodycount 4-9	-15.9	-24.3 ±9.8	+21.8	+18.9 ±13.1
net · bodycount 10+	-7.3	-13.1 ±11.4	+26.0	+27.2 ±13.0

T7 · 'Porn more satisfying than real sex' (NEW; 18–34)

metric	M unweighted	M weighted	F unweighted	F weighted
% agree porn > real sex	14.3	12.2 ±1.7	21.0	20.1 ±3.0
% agree · daily+ users	21.2	18.3 ±3.3	36.9	32.3 ±9.5
% agree · single	15.0	12.7 ±2.9	27.4	21.8 ±5.4
% agree · serious relationship	12.9	11.9 ±2.4	16.6	18.4 ±4.2

T8 · Dating, drive, horniness (18–34)

Hookup rows: women-only item, value shown in the M columns. Attractiveness/autism/shyness confounder analysis can't be weighted (cols not in extract).

metric	M unweighted	M weighted	F unweighted	F weighted
% quite/very horny now · never-users	36.5	40.9 ±2.3	29.1	31.8 ±2.3
% quite/very horny now · daily+	53.4	56.8 ±1.2	53.8	57.9 ±2.6
% high real-sex motivation · never-users	41.7	45.9 ±7.0	34.9	33.8 ±7.9
% high real-sex motivation · daily+	50.6	57.8 ±3.5	45.2	42.6 ±8.3
% bad at dating · never	40.5	41.7 ±6.9	33.2	35.2 ±8.0
% bad at dating · weekly	48.2	42.5 ±3.4	43.8	38.6 ±5.0
% bad at dating · daily+	57.6	51.5 ±3.5	51.6	41.9 ±8.3
% bad at dating · SINGLE daily+	71.7	69.6 ±4.4	65.1	50.4 ±11.3
% bad at dating · SINGLE never/rare-mo (0-5)	55.7	54.1 ±8.5	53.5	49.1 ±8.3
% bad at dating · PARTNERED daily+	39.7	33.9 ±6.1	35.4	30.7 ±14.1
% bad at dating · PARTNERED never/rare-mo	31.0	25.9 ±5.5	27.9	23.2 ±5.8
% bad at dating · bodycount 0, not-weekly (0-5)	53.9	55.9 ±8.9	55.5	58.2 ±8.1
% bad at dating · bodycount 0, daily+	79.0	78.2 ±5.0	63.1	52.4 ±11.3
% bad at dating · bodycount 10+, not-weekly (0-5)	23.4	15.4 ±10.0	26.9	21.7 ±9.3
% bad at dating · bodycount 10+, daily+	30.7	23.8 ±7.4	42.2	33.5 ±17.6
women: % hookups GOOD · never-users	32.2	30.7 ±2.8	—	—
women: % hookups GOOD · daily+	45.4	44.3 ±2.5	—	—

T9 · Substitution cluster (18–34)

metric	M unweighted	M weighted	F unweighted	F weighted
% bad at dating · agree porn>sex	48.5	44.3 ±6.7	44.9	41.9 ±7.9
% bad at dating · disagree	39.5	35.1 ±2.8	32.1	28.4 ±4.1
% low sex-drive · agree porn>sex	25.8	17.6 ±5.1	32.4	28.6 ±7.2
% low sex-drive · disagree	20.3	16.3 ±2.1	25.9	29.5 ±4.2
% single · agree porn>sex	36.1	32.5 ±6.3	43.7	32.6 ±7.5
% single · disagree	33.3	29.4 ±2.6	29.4	28.4 ±4.1

T10 · Mental health (self-reported mod-severe)

Diagnosis-count (TotalMentalIllness) not in weighted extract; binary has_* used.

metric	M unweighted	M weighted	F unweighted	F weighted
% depression · never-users (18-24)	23.2	21.6 ±1.3	39.5	38.3 ±1.2
% depression · daily+ (18-24)	33.6	29.2 ±0.7	56.1	52.2 ±1.1
% depression · never-users (25-34)	28.9	24.1 ±2.6	45.4	40.0 ±2.4
% depression · daily+ (25-34)	38.6	30.9 ±0.8	61.1	52.4 ±1.7
% anxiety · never-users (18-34)	29.2	25.9 ±1.3	56.6	54.5 ±1.3
% anxiety · daily+ (18-34)	34.4	28.8 ±0.5	65.2	61.1 ±1.0
% depression · porn onset ≤10 (18-34)	40.7	35.2 ±1.0	59.7	55.6 ±1.2
% depression · porn onset 13-14 (18-34)	29.9	24.5 ±0.6	50.2	46.4 ±0.9
% depression · porn onset 15+ (18-34)	28.3	23.2 ±0.7	47.2	42.6 ±0.7

T11 · Porn-induced fetishes (14–34)

metric	M unweighted	M weighted	F unweighted	F weighted
% any porn-induced fetish	86.2	85.5 ±0.2	82.1	80.0 ±0.3
% 'new & totally different'	24.7	23.5 ±0.3	16.2	15.8 ±0.3
% any induced · daily+ (18-34)	92.7	92.5 ±0.3	93.2	92.5 ±0.6
% any induced · rare 1-3 (18-34)	68.5	69.8 ±2.5	66.0	65.6 ±1.6

T12 · Content-as-risk belief (gated: pedophilic-arousal reporters; 14–34)

metric	M unweighted	M weighted	F unweighted	F weighted
% says REDUCE	49.3	49.5 ±2.2	46.1	47.8 ±5.0
% no effect	36.8	36.6 ±2.1	41.3	37.8 ±4.9
% says increase	13.9	13.9 ±1.5	12.7	14.4 ±3.5

T13 · Quit group vs daily+ (18–34)

Quit cells are ~100-200 raw / fewer effective — treat weighted values as directional only.

metric	M unweighted	M weighted	F unweighted	F weighted
sexlife net · QUIT	-24.6	-29.7 ±12.8	-9.4	-2.9 ±19.7
sexlife net · daily+	-10.7	-15.8 ±7.2	+29.0	+32.6 ±16.1
% depression · QUIT	22.1	22.7 ±8.3	45.0	41.5 ±15.5
% depression · daily+	35.8	30.2 ±0.5	58.1	52.4 ±1.0
% bad at dating · QUIT	40.0	42.5 ±9.8	30.0	32.4 ±14.7
% bad at dating · daily+	57.6	51.5 ±3.5	51.6	41.9 ±8.3

T14 · Age-standardized male base rates (full sample 14–60, unweighted)

Male rates re-weighted to the FEMALE age distribution (men are ~2y older; use rises with age, so crude male rates are slightly inflated).

metric	M crude	M age-std to F dist	F crude
% never	4.6	5.2	8.3	(standard)
% weekly+	85.2	84.4	62.0	(standard)
% daily+	44.0	43.3	19.2	(standard)

How many people are using pornography?

Confounders: politics matters a little, current devoutness matters a lot

Onset: the real adolescent peak — and a generational convergence

What they consume

Other content facts worth having loaded

The new items: shame, healing, sex life, porn vs sex

4a · Shame

4b · Healing / therapeutic

4c · Effect on sex life — the money item

4d · "Porn is usually more satisfying than sex with a real person"

4e · Who stops? The "quit" group — and a survivorship warning

Porn × dating, drive & relationships

5a · Current horniness

5b · Drive to seek real sex

5c · Dating difficulty — the strongest real-world correlate

5d · Casual-sex experience (women)

5e · The substitution subgroup is real — but a minority

5f · What confounds the porn ↔ dating link? (the capstone)

Mental health

Porn-induced fetishes

Special populations

Porn performers

Respondents reporting pedophilic arousal (handle with extreme care, if at all)

Methodology notes

Common questions about pornography use

Appendix · Weighted results

T1 · Headline use rates (ages 14–34)

T1b · Base rates by age bin

T2 · Onset

T3 · Content type & payment (ages 14–34)

T4 · Shame (NEW item; ages 18–34)

T5 · Healing/therapeutic (NEW; 18–34)

T6 · Sex-life impact, net = %improved − %damaged (NEW; 18–34)

T7 · 'Porn more satisfying than real sex' (NEW; 18–34)

T8 · Dating, drive, horniness (18–34)

T9 · Substitution cluster (18–34)

T10 · Mental health (self-reported mod-severe)

T11 · Porn-induced fetishes (14–34)

T12 · Content-as-risk belief (gated: pedophilic-arousal reporters; 14–34)

T13 · Quit group vs daily+ (18–34)

T14 · Age-standardized male base rates (full sample 14–60, unweighted)