No study ranks the “craziest” videos — the evidence measures categories, coded samples, and associations, each with hard limits

Eight complete peer-reviewed results published 2018–2026, spanning one meta-analysis of 59 studies, five association surveys (50 to 4,024 respondents), and two coded content samples (50 videos; 390 forum posts). None supports a ranking of individual videos, and no association result establishes causation.

association evidence — cross-sectional, cannot establish causation coded content — describes a narrow sample only meta-analytic synthesis — pooled correlations

Circle area — sample size (participants, coded items, or pooled studies; units differ by study). Hover any mark for the finding and its stated limitation. Year for the hijab content analysis is taken from its published citation (2022); the source row leaves the year field blank.

59 studiesThe 2020 meta-analysis found no association between nonviolent pornography and sexual aggression, and only a weak correlation for violent material that could not distinguish selection from socialization; publication bias pushed the effect below the interpretive threshold. christopherjferguson.com
R² = 0.71In 677 Dutch university students, the violent-pornography link to sexual-aggression risk in men was conditional — amplified when the material was perceived as realistic and peer rape-myth acceptance was high — not a uniform effect of exposure. link.springer.com
51.6%Among 4,024 Spanish adolescents aged 14–17, just over half reported intentional past-year access and 10.6% daily use — a prevalence estimate, not evidence about content extremity or harm. link.springer.com
757 adultsReported distress about use is not the same as use itself: scrupulosity was the strongest factor across every dimension of problematic pornography use for both genders, so moral incongruence must be separated from frequency and content (doi 10.1080/10720162.2019.1567410).

Ethics and scope note

This report does not name, rank, link to, thumbnail, embed, or make playable any pornographic video or pornography website. Embedding source material would create consent risks (performer consent cannot be reliably inferred from a finished video), age-verification and copyright exposure, privacy harms to identifiable performers, and research-ethics violations — and it is unnecessary for the scientific question, which concerns categories and associations, not individual clips. Explicit act-level codes from the content analyses are summarized here only in neutral terms: aggression, coercion, objectification, power imbalance, and performer agency.

The report is intended for adult academic and public-health discussion, not clinical diagnosis. Every design shown except the meta-analysis is cross-sectional or descriptive: association is not causation, self-report and selection effects apply throughout, coding definitions vary between studies, and niche samples must not be generalized to all pornography.

The eight complete study rows

StudyYearDesignSampleKey finding (neutral summary)Stated limitation
Intentional pornography access among Spanish adolescents2026Cross-sectional survey
association
4,024 adolescents, 14–17, Spain51.6% intentional past-year access, 10.6% daily; access associated with lower moral emotions and greater risk-behavior involvement.Authors acknowledge several limitations; a first quantitative description, not a causal test.
Compensatory and dissociative coping, maltreatment, and problematic use2026Cross-sectional latent profile analysis
association
608 young adults, 18–27, IsraelDistinct compensatory and dissociative coping profiles show different patterns of trauma exposure and pornography-related behavior; coping style moderates how trauma relates to behavior.Limitation field not extracted; design is cross-sectional, so no causal inference.
Experiences with AI-generated pornography (Reddit)2025Manual quantitative content analysis
content coding
390 public English-language Reddit postsPosts covered production (59.5%), content (60.8%), effects (37.2%), ethical-legal issues (35.1%), use (12.8%); reactions ranged from concern over harms to curiosity and economic benefit.Self-reported posts may diverge from actual behavior through bias, memory distortion, framing, and self-presentation.
Moderators of the violent pornography–sexual aggression link2025Online survey
association
677 Dutch university studentsFor men, violent-pornography use predicted sexual-violence perpetration risk (Nagelkerke R² = 0.71), amplified by perceived realism and peer rape-myth acceptance; for women, the link appeared when peer rape-myth acceptance was high.Perceived realism may matter most in developmental periods before real-life sexual experience.
Hijab pornography content analysis2022Content analysis
content coding
50 coded videos from a niche categoryIn all coded videos aggression was directed only at women; pervasive objectification, male control of the encounter in nearly all videos, coercion coded in about a third, and power imbalance in half.Niche genre not representative of mainstream material; performer intent and consent could not be coded; English-language requirement excluded much of the genre.
Hentai use, rape myths, and sexually aggressive strategies2022Quantitative analysis
association
906 college students who used this genreFrequency and intensity of hentai use predicted sexually aggressive strategies; rape-myth endorsement was an important factor in both relationships.Limitation field not extracted; self-selected genre users, so selection effects apply.
Pornography and sexual aggression: can meta-analysis find a link?2020Meta-analysis
synthesis
59 correlational, experimental, and population studiesNo association for nonviolent pornography; violent pornography weakly correlated with sexual aggression, with selection versus socialization effects indistinguishable.Identified publication bias reduced the effect size below the level used for interpretation.
Is scrupulosity behind problematic viewing, depression, anxiety, and stress? (doi 10.1080/10720162.2019.1567410)2018Multi-group structural equation modeling
association
757 adults (507 women, 250 men), online US sampleScrupulosity was the strongest factor across all dimensions of problematic use; models explained 12–27% of variance depending on dimension and gender.Cross-sectional data preclude causal inferences.

Method: evidence extracted from 10 rows of one academic result set; 2 near-blank duplicate rows were dropped in favor of their complete versions, leaving 8 study results published 2018–2026. Each mark and table row is one published result; circle area encodes the study’s own sample size in its own units (participants, coded videos, forum posts, or pooled studies), which are not directly comparable across studies. Statistics are reported as extracted; no effect sizes were recomputed. Explicit act-level content codes were condensed into neutral categories (aggression, coercion, objectification, power imbalance, performer agency) for space and to keep the report non-graphic. One study’s publication year (2022) comes from its citation string because the extracted year field was blank; one 2018 study’s source is available only as a DOI, shown as text.

Keenable · made with SELECT · 3,972 pages in 6m 30s · Open the conversationShare: XLinkedInReddit