Composite Faces of Study Of Us

The average face of 13,823 people, split by how they answered the survey. October 2026.

Each face below is a composite: hundreds or thousands of real Study Of Us members' photos, lined up by their features and averaged into one face. The people in each pair answered a survey question in opposite ways. On the left are the people at one end of the question, on the right the people at the other.

Within each pair, both sides were weighted to the same mix of ages and ethnicities, so a difference isn't just one side being older or less white. Under each figure is the question exactly as it was asked, and which answers went on each side. A figure marked * shows a difference that held up when we split the data in half at random and compared. † marks a weaker one, and anything unmarked could be chance. Press blink to flip between the two faces in place, which makes small differences easier to catch.

A composite shows the average, not what any one person looks like. Some differences come from how people take photos (angle, lighting, makeup, glasses) as much as from their faces.

Childhood and background
Sex and dating
Mind, politics and belief
Gender identity
Body and life
Personality, drive and mood

Exaggeration: · · · The page opens at ×5; pick actual to see the real averages. Exaggeration pushes each pair apart from its midpoint, so whatever differs is easier to see. It magnifies chance differences just as much as real ones.

Plate I. Childhood and background

Composite of 6,176 people: No — **Fig. 1.** Abused as a child. Left, no; right, yes. Men, N = 7,568; women, N = 2,392. Asked: “Were you (at least semiregularly) abused as a child?”

Composite of 1,392 people: Yes — **Fig. 1.** Abused as a child. Left, no; right, yes. Men, N = 7,568; women, N = 2,392. Asked: “Were you (at least semiregularly) abused as a child?”

Composite of 2,766 people: Lower-middle class or poorer — **Fig. 2.** Family class growing up. Left, lower-middle class or poorer; right, upper-middle class or richer. Men, N = 5,232; women, N = 1,718. Asked: “Growing up, your family was”

Composite of 2,466 people: Upper-middle class or richer — **Fig. 2.** Family class growing up. Left, lower-middle class or poorer; right, upper-middle class or richer. Men, N = 5,232; women, N = 1,718. Asked: “Growing up, your family was”

* The difference shows up again in random halves of the data (beats all 1,000 shuffles). † Weaker: beats 99% of shuffles. Unmarked rows could be chance.

Plate II. Sex and dating

Composite of 5,269 people: 0 to 4 — **Fig. 3.** Number of sex partners. Left, 0 to 4; right, 10 or more. Men, N = 8,400; women, N = 2,640. Asked: “How many people have you had sex with?”

Composite of 3,131 people: 10 or more — **Fig. 3.** Number of sex partners. Left, 0 to 4; right, 10 or more. Men, N = 8,400; women, N = 2,640. Asked: “How many people have you had sex with?”

Composite of 5,603 people: Vanilla or a little kinky — **Fig. 4.** Kinkiness. Left, vanilla or a little kinky; right, pretty or extremely kinky. Men, N = 9,349; women, N = 2,804. Asked: “How kinky would you say you are?”

Composite of 3,746 people: Pretty or extremely kinky — **Fig. 4.** Kinkiness. Left, vanilla or a little kinky; right, pretty or extremely kinky. Men, N = 9,349; women, N = 2,804. Asked: “How kinky would you say you are?”

Composite of 6,745 people: Straight — **Fig. 5.** Sexual orientation. Left, straight; right, leaning straight, bi or gay. Men, N = 8,308; women, N = 2,505. Asked: “You are:”

Composite of 1,563 people: Leaning straight, bi or gay — **Fig. 5.** Sexual orientation. Left, straight; right, leaning straight, bi or gay. Men, N = 8,308; women, N = 2,505. Asked: “You are:”

Composite of 1,132 people: Once a month or less — **Fig. 6.** Porn habit. Left, once a month or less; right, several times a week or more. Men, N = 8,032; women, N = 2,104. Asked: “How often do you watch or read pornographic/erotic content?”

Composite of 6,900 people: Several times a week or more — **Fig. 6.** Porn habit. Left, once a month or less; right, several times a week or more. Men, N = 8,032; women, N = 2,104. Asked: “How often do you watch or read pornographic/erotic content?”

Composite of 7,159 people: No — **Fig. 7.** Has done sex work. Left, no; right, yes. Men, N = 7,511; women, N = 2,378. Asked: “Have you done sex work (as a performer/provider/etc.?)”

Composite of 352 people: Yes — **Fig. 7.** Has done sex work. Left, no; right, yes. Men, N = 7,511; women, N = 2,378. Asked: “Have you done sex work (as a performer/provider/etc.?)”

Composite of 2,328 people: Disagree or strongly disagree — **Fig. 8.** Bad at dating. Left, disagree or strongly disagree; right, agree or strongly agree. Men, N = 6,217; women, N = 1,890. Asked: “I am bad at dating (asking people out, going on dates)”

Composite of 3,889 people: Agree or strongly agree — **Fig. 8.** Bad at dating. Left, disagree or strongly disagree; right, agree or strongly agree. Men, N = 6,217; women, N = 1,890. Asked: “I am bad at dating (asking people out, going on dates)”

* The difference shows up again in random halves of the data (beats all 1,000 shuffles). † Weaker: beats 99% of shuffles. Unmarked rows could be chance.

Plate III. Mind, politics and belief

Composite of 3,952 people: Centre or right (average) — **Fig. 9.** Politics. Left, centre or right (average); right, moderately left or more (average). Men, N = 7,644; women, N = 2,491. Asked: “Socially / economically speaking, you tend to be more: (average of the two)”

Composite of 3,692 people: Moderately left or more (average) — **Fig. 9.** Politics. Left, centre or right (average); right, moderately left or more (average). Men, N = 7,644; women, N = 2,491. Asked: “Socially / economically speaking, you tend to be more: (average of the two)”

Composite of 4,050 people: Somewhat to totally disagree — **Fig. 10.** Calls themself a narcissist. Left, somewhat to totally disagree; right, somewhat to totally agree. Men, N = 6,729; women, N = 2,158. Asked: “To what extent do you agree with this statement: I am a narcissist.”

Composite of 2,679 people: Somewhat to totally agree — **Fig. 10.** Calls themself a narcissist. Left, somewhat to totally disagree; right, somewhat to totally agree. Men, N = 6,729; women, N = 2,158. Asked: “To what extent do you agree with this statement: I am a narcissist.”

Composite of 1,817 people: Disagrees with both — **Fig. 11.** Depressed and anxious. Left, disagrees with both; right, agrees with both. Men, N = 4,699; women, N = 1,638. Asked: “I tend to suffer from depression / I tend to suffer from anxiety”

Composite of 2,882 people: Agrees with both — **Fig. 11.** Depressed and anxious. Left, disagrees with both; right, agrees with both. Men, N = 4,699; women, N = 1,638. Asked: “I tend to suffer from depression / I tend to suffer from anxiety”

Composite of 2,169 people: Didn't tick Borderline — **Fig. 12.** Borderline personality disorder. Women: left, didn't tick Borderline; right, ticked Borderline. Women, N = 2,426. Shown although it didn't pass the split-half test, so it could be chance. The women's row also includes nonbinary people assigned female at birth. Asked: “Do you have any of the following”

Composite of 257 people: Ticked Borderline — **Fig. 12.** Borderline personality disorder. Women: left, didn't tick Borderline; right, ticked Borderline. Women, N = 2,426. Shown although it didn't pass the split-half test, so it could be chance. The women's row also includes nonbinary people assigned female at birth. Asked: “Do you have any of the following”

Composite of 5,059 people: Slightly to strongly disagree — **Fig. 13.** Enjoys astrology. Left, slightly to strongly disagree; right, slightly to strongly agree. Men, N = 6,466; women, N = 1,999. Asked: “I enjoy astrology”

Composite of 1,407 people: Slightly to strongly agree — **Fig. 13.** Enjoys astrology. Left, slightly to strongly disagree; right, slightly to strongly agree. Men, N = 6,466; women, N = 1,999. Asked: “I enjoy astrology”

Composite of 1,846 people: Less attractive than average — **Fig. 14.** Rates own looks. Left, less attractive than average; right, more attractive than average. Men, N = 5,152; women, N = 1,445. Both sides were also matched on BMI. Asked: “Compared to other people of your same gender and age range, you are”

Composite of 3,306 people: More attractive than average — **Fig. 14.** Rates own looks. Left, less attractive than average; right, more attractive than average. Men, N = 5,152; women, N = 1,445. Both sides were also matched on BMI. Asked: “Compared to other people of your same gender and age range, you are”

* The difference shows up again in random halves of the data (beats all 1,000 shuffles). † Weaker: beats 99% of shuffles. Unmarked rows could be chance.

Plate IV. Gender identity

Composite of 9,944 people: Man (cis) — **Fig. 15.** Cis or nonbinary. Top, cis men vs nonbinary people assigned male at birth; bottom, cis women vs nonbinary people assigned female at birth. Assigned male, N = 10,287; assigned female, N = 3,607. Asked: “What's your gender?”

Composite of 343 people: Nonbinary / other (AMAB) — **Fig. 15.** Cis or nonbinary. Top, cis men vs nonbinary people assigned male at birth; bottom, cis women vs nonbinary people assigned female at birth. Assigned male, N = 10,287; assigned female, N = 3,607. Asked: “What's your gender?”

Composite of 9,652 people: Man (cis) — **Fig. 16.** Cis or trans. Top, cis men vs trans men; bottom, cis women vs trans women, each in the faces they were rated with. Men, N = 9,881; women, N = 4,031. Asked: “What's your gender?”

Composite of 229 people: Man (trans) — **Fig. 16.** Cis or trans. Top, cis men vs trans men; bottom, cis women vs trans women, each in the faces they were rated with. Men, N = 9,881; women, N = 4,031. Asked: “What's your gender?”

* The difference shows up again in random halves of the data (beats all 1,000 shuffles). † Weaker: beats 99% of shuffles. Unmarked rows could be chance.

Plate V. Body and life

Composite of 871 people: Short — **Fig. 17.** Height. Men: 5'6" or under vs 6'0" or over; Women: 5'2" or under vs 5'6" or over. Men, N = 5,028; women, N = 2,042. Asked: “Your height? (inches)”

Composite of 4,157 people: Tall — **Fig. 17.** Height. Men: 5'6" or under vs 6'0" or over; Women: 5'2" or under vs 5'6" or over. Men, N = 5,028; women, N = 2,042. Asked: “Your height? (inches)”

Composite of 2,631 people: Under 22 — **Fig. 18.** Body mass index. Left, under 22; right, 27 or over. Men, N = 4,885; women, N = 1,993. Asked: “Your height? / Your weight (lbs): (combined into BMI)”

Composite of 2,254 people: 27 or over — **Fig. 18.** Body mass index. Left, under 22; right, 27 or over. Men, N = 4,885; women, N = 1,993. Asked: “Your height? / Your weight (lbs): (combined into BMI)”

Composite of 1,511 people: Not really — **Fig. 19.** Works out. Left, not really; right, moderately or more. Men, N = 6,758; women, N = 1,802. Asked: “Do you work out?”

Composite of 5,247 people: Moderately or more — **Fig. 19.** Works out. Left, not really; right, moderately or more. Men, N = 6,758; women, N = 1,802. Asked: “Do you work out?”

Composite of 1,810 people: High school or less — **Fig. 20.** Education. Left, high school or less; right, bachelors or more. Men, N = 8,069; women, N = 2,308. Asked: “What level of formal education have you completed?”

Composite of 6,259 people: Bachelors or more — **Fig. 20.** Education. Left, high school or less; right, bachelors or more. Men, N = 8,069; women, N = 2,308. Asked: “What level of formal education have you completed?”

Composite of 541 people: Under $50k, net worth $0 or less — **Fig. 21.** Income and net worth. Left, under $50k, net worth $0 or less; right, $100k+, net worth $100k+. Men, N = 2,164; women, N = 478. Asked: “Roughly speaking, what is your yearly income in USD? / Your CURRENT net worth is closest to”

Composite of 1,623 people: $100k+, net worth $100k+ — **Fig. 21.** Income and net worth. Left, under $50k, net worth $0 or less; right, $100k+, net worth $100k+. Men, N = 2,164; women, N = 478. Asked: “Roughly speaking, what is your yearly income in USD? / Your CURRENT net worth is closest to”

* The difference shows up again in random halves of the data (beats all 1,000 shuffles). † Weaker: beats 99% of shuffles. Unmarked rows could be chance.

Plate VI. Personality, drive and mood

Composite of 3,271 people: Bottom third — **Fig. 22.** Extraversion (Big Five). Left, bottom third; right, top third. Men, N = 6,375; women, N = 1,766. Asked: “I am quiet around strangers / I am the life of the party / I talk to a lot of different people at parties”

Composite of 3,104 people: Top third — **Fig. 22.** Extraversion (Big Five). Left, bottom third; right, top third. Men, N = 6,375; women, N = 1,766. Asked: “I am quiet around strangers / I am the life of the party / I talk to a lot of different people at parties”

Composite of 2,447 people: Bottom third — **Fig. 23.** Extraversion (HEXACO). Left, bottom third; right, top third. Men, N = 5,306; women, N = 1,385. Asked: “I'm comfortable being the center of attention. / I often feel awkward about myself in social settings. / I usually feel enthusiastic and full of energy.”

Composite of 2,859 people: Top third — **Fig. 23.** Extraversion (HEXACO). Left, bottom third; right, top third. Men, N = 5,306; women, N = 1,385. Asked: “I'm comfortable being the center of attention. / I often feel awkward about myself in social settings. / I usually feel enthusiastic and full of energy.”

Composite of 2,002 people: Somewhat to totally disagree — **Fig. 24.** High powered, driven, successful. Left, somewhat to totally disagree; right, somewhat to totally agree. Men, N = 6,453; women, N = 2,003. Asked: “I am high powered, driven, successful”

Composite of 4,451 people: Somewhat to totally agree — **Fig. 24.** High powered, driven, successful. Left, somewhat to totally disagree; right, somewhat to totally agree. Men, N = 6,453; women, N = 2,003. Asked: “I am high powered, driven, successful”

Composite of 2,830 people: Not horny at all — **Fig. 25.** Horny right now. Left, not horny at all; right, moderately or real horny. Men, N = 4,850; women, N = 1,627. Asked: “How horny are you right now?”

Composite of 2,020 people: Moderately or real horny — **Fig. 25.** Horny right now. Left, not horny at all; right, moderately or real horny. Men, N = 4,850; women, N = 1,627. Asked: “How horny are you right now?”

* The difference shows up again in random halves of the data (beats all 1,000 shuffles). † Weaker: beats 99% of shuffles. Unmarked rows could be chance.

Notes on method

Who. 13,823 Study Of Us members with an approved photo: 10,450 cis men from the men's faces and 3,373 cis women from the women's faces. The gender-identity plate adds 343 nonbinary people assigned male at birth and 330 assigned female at birth, 229 trans men and 707 trans women, each from the faces they were rated with. Everyone was 18 or older in the photo they're represented by.

One photo per person. People who submitted many photos count once. We used each person's clearest photo: facing the camera first, then the sharpest. Anyone whose photos were all turned well away from the camera was left out (women were allowed slightly more turn, since there are fewer of them).

Making a composite. Every photo gets 478 facial landmarks (Google's MediaPipe face model). Faces are lined up on those landmarks, warped to the group's average face shape, then averaged pixel by pixel. Hair and background are faded out.

The two sides. Each figure compares two ranges of answers to one question; the caption says exactly which answers are on each side. Answers in between are left out.

Personality scores. The Big Five and HEXACO figures use the three short items Study Of Us asks for each trait, quoted under each figure. One item in each set is reverse-scored, as the Study Of Us app scores them, and the three are averaged. Each figure compares roughly the bottom third of scores with the top third, set separately for men and women.

Weighting. Everyone with an answer on either side is included, and N counts all of them. People were sorted into cells by age band (18–21, 22–25, 26–29, 30–34, 35–39, 40–49, 50+) and ethnicity (white, Hispanic, South Asian, East Asian, Black, other, not given), then weighted so the left and right sides have the same mix of cells. A cell only one side has is left out. A pair is shown only when each side has at least 60 people.

Does it hold up? For each pair we split both sides in half at random and measured the left-minus-right difference in each half, separately for face shape (landmark positions) and for skin and texture (pixel colours). If the difference is real, the two halves agree. We did this 300 times and compared the agreement with 1,000 runs where the left/right labels were shuffled within each age and ethnicity cell. * means it beat all 1,000 shuffles on shape or texture; † means it beat 99% of them. With 103 rows tested, expect one or two † by luck.

Left out. Questions where neither the men's nor the women's faces showed a difference that held up aren't shown: spanked as a kid, sexual upbringing, birth order, age at first sex, dominant or submissive, monogamous or polyamorous, has hired a sex worker, religion, attachment style, worried about AI killing everyone, has kids, relationship status, openness (Big Five), conscientiousness (Big Five), agreeableness (Big Five), neuroticism (Big Five), honesty-humility (HEXACO), emotionality (HEXACO), agreeableness (HEXACO), conscientiousness (HEXACO), openness (HEXACO), finds the supernatural plausible, needs to feel in control, officially tested IQ, shame about porn, baby fever, mood PMS symptoms, on hormonal birth control and ovulating right now.

Exaggeration. The ×2, ×3 and ×5 settings push each pair's two faces apart from their midpoint, by two, three or five times the real gap, in shape and in colour. That makes a small real difference visible, but it magnifies chance differences just as much, so check the marks first.

Limits. These are people who chose to submit a photo to be rated, mostly selfies, and every answer is self-reported. Nothing here is a single person's photo, and a composite of 60 or more people can't be traced back to anyone in it.

← chaosfactor.xyz