Factor Analysis of Kink Preferences

In an old tweet, Aella presented a factor analysis of her kink survey, identifying eight factors:

Power/Sadomasochism/Bondage
Big Boobies/Bimbos/Skirts/Pigtails
Romance/Kissing/Gentleness
Orgies/Freeuse/Anonymous Sex/Exhibitionism
Furries/Bestiality/Monsters
Body Fluids/Urine/Disgust/Scat
Genderplay/Futa/Gender Transforms
Ageplay/Incest/Age Gaps

Given biological sex strongly influences sexual preferences, we might expect these factors to differ between men and women. For example, a BDSM factor might emphasize dominance for men and submission for women. Additionally, novel factors may emerge from analyzing men and women separately. To test these hypotheses, I performed separate factor analyses for cis men and cis women. (Trans respondents were excluded due to insufficient sample sizes for reliable analysis.)

Factor Analysis Results

Cis Men (9 Factors)

Cis men shared the same 8 factors as Aella’s original analysis, plus 1 additional factor:

BDSM (sadomasochism, bondage, dominance)
Hyperfemininity (big boobs/butts, bimbos, high heels, pigtails)
Sensuality/Romance (sensuality, makeouts, romance, teasing)
Conventional Kink (orgies, voyeurism, exhibitionism, sex toys)
Monsters/Animals (monsters, furries, bestiality)
Disgust (farts/belches, poop, necrophilia, feedism)
Genderplay (genderplay, sissification, futa, pegging)
Incest/Ageplay (incest, ageplay, nonconsent)
Gay (men)

Cis Women (14 Factors)

Cis women shared the same 8 factors as Aella’s original analysis, plus 6 additional factors:

BDSM (sadomasochism, bondage, submission)
Hyperfemininity (lesbians, big boobs/buts, skirts)
Sensuality/Romance (sensuality, makeouts, romance, teasing)
Conventional Kink (orgies, voyeurism, free use, anonymous sex)
Monsters/Animals (monsters, bestiality, furries)
Disgust (farts/belches, poop, incest, creepy crawlies)
Genderplay (genderplay, sissification, futa, pegging)
Incest/Ageplay (incest, ageplay)
Cum (semen, creampies, blowjobs, bukkake)
Toys (painful toys, electric toys, sounding)
Sensory/Roleplay (latex/leather, specific smells/textures, roleplay)
Ethnicities (white people, black people, Asian people)
Nonconsent (rapeplay, mindbreak, bad ends)
Brutality (violence, body horror, blood, executions)

Higher-Order Factor Correlations

Since the factors are themselves correlated, I performed a second factor analysis—which has poor statistical fit yet is valuable for exploratory purposes—to determine how strongly each factor loads onto a hypothetical “general kink” factor.

Cis Men

Hyperfemininity (0.75)
Conventional Kink (0.73)
BDSM (0.60)
Monsters/Animals (0.50)
Genderplay (0.48)
Disgust (0.41)
Sensuality/Romance (0.33)
Incest/Ageplay (0.30)
Gay (0.15)

Cis Women

Sensory/Roleplay (0.76)
Toys (0.74)
Genderplay (0.67)
Monsters/Animals (0.61)
BDSM (0.59)
Nonconsent (0.57)
Ethnicities (0.54)
Cum (0.52)
Conventional Kink (0.49)
Brutality (0.48)
Hyperfemininity (0.44)
Disgust (0.43)
Sensuality/Romance (0.28)
Incest/Ageplay (0.22)

The loading patterns align with intuition: factors loading strongly on the general kink factor (e.g. Gay, Sensuality/Romance) are less likely to be labeled as “kinks” than factors loading weakly (e.g. BDSM, Sensory/Roleplay).

Visualizations

The following plots are what I call “Tailcalled diagrams”. Basically, we create a standardized (mean=0, SD=1) latent variable summarizing the displayed items in the bottom of the plot. The rows below the histogram show the median response to each question for a given level of the latent variable. For these plots, while the exact loadings are inflated, the rank orderings are roughly accurate.