A Psychometric Analysis of Tailcalled's Gender Satisfaction Survey

Tailcalled ran a “Gender satisfaction and other things” survey, so I decided to analyze the data. The responses came from 185 cis men recruited on Prolific.

Factor Structure

I was able to extract 9 factors:

• Affective Gender Identity
• BDSM Interest
• Masculine Gender Norm Endorsement
• Masculine Gender Role Insecurity
• Gender Dysphoria
• Homosexuality
• Perceived Female Advantage
• Mental Illness
• Autogynephilia

Factor 1: Affective Gender Identity

This factor contained items about wanting to be a woman. Among the emotions one might experience if they were to magically wake up as a woman, comfort (which loaded positively) and disgust (which loaded negatively) had the strongest loadings. Interestingly, sexual arousal at the idea of sex with tomboyish/masculine women also loaded positively on this factor.

This scale has very high reliability across most levels of affective gender identity, due to many items with high loadings.

Factor 2: BDSM Interest

This factor contained items about sexual interest in BDSM-related activities. Notably, both dominance-themed items and submission-themed items had very strong loadings.

This scale has high reliability for people with average and high interest in BDSM, it has low reliability for those with low interest in BDSM. Part of this is due to the number of respondents who, as far as the scale can tell, had essentially zero interest in BDSM. But, if we compare this scale to the scale formed from Aella’s Kink Survey, we see that another issue is the limited number of items; increasing the number of items would likely improve reliability, even taking into account the low interest in BDSM.

Factor 3: Masculine Gender Norm Endorsement

This factor contained items endorsing masculine gender norms, expressing pride or comfort in being male, and claiming being a man was a core part of one’s identity. Belief in negative stereotypes of women (e.g., stereotypes about women being prone to overthinking or generating unnecessary drama) loaded positively on this factor, along with the belief that men and women have large psychological differences. The belief that one can choose their gender loaded negatively.

This scale has high reliability for most levels of masculine gender norm endorsement, due to many items with high loadings.

Factor 4: Masculine Gender Role Insecurity

This factor contained items about various insecurities related to masculine gender roles. Item loadings were strong.

This scale has high reliability for people with moderately below-average to high levels of masculine gender role insecurity, though reliability starts dropping for those with very low levels of masculine gender role insecurity.

Factor 5: Gender Dysphoria

This factor contained capturing distress about one’s male gender or the desire to be female. Thinking of oneself as transgender had a strong positive loading on this factor.

This scale has high reliability for those with average-to-high levels of gender dysphoria, but low reliability for those with low levels of gender dysphoria. Part of this is due to the number of items, but another contributor is simply that many respondents do not experience gender dysphoria.

Factor 6: Homosexuality

This factor contained items dealing with sexual interest in having sex with men (along with an item about sexual interest in having sex with women, which had a weaker and negative loading.)

This scale had moderate reliability for the upper end of the distribution and atrocious reliability for the lower end. This isn’t surprising given that the scale consists of three items, and also, many respondents simply are not gay.

Factor 7: Perceived Female Advantage

While most of the items load positively on this factor are about various aspects of womanhood being benefits, the highest positively loading item was an aspect of manhood—the expectations of toughness—being a benefit. I’m unsure why this item loads (positively) on this factor. Another interesting item was the belief that there are more than two genders, which loaded negatively on this factor. It’s possible that this belief correlates with social liberalism, which in turn correlates with a general belief that women have things pretty rough.

This scale had moderate reliability across the distribution. Given the number of items, along with the varying difficulties, the only problem seems to be the relatively low loadings of the items. I would suggest adding items that are more direct about asking which gender has it better, but the item that asks that question directly, “Do you feel men or women have it easier in life?”, has a relatively low loading, even for the scale.

Factor 8: Mental Illness

This factor contained all the items dealing with mental conditions, which all had strong loadings on this factor.

This scale has moderate reliability for the high end of the distribution, but low reliability for the low end. Many respondents appear to not have been diagnosed with mental conditions or suspect themselves of having mental conditions. It might be worth including items that ask if respondents have specific symptoms of various mental conditions.

Factor 9: Autogynephilia

Items concerning sexual arousal from behaving in feminine ways or from imagining oneself with female anatomy or bodily functions loaded strongly on this factor. Items about being sexually aroused by sexual interactions with (possible) MtFs or as MtFs loaded (a bit) less strongly, though their loadings were still strong.

This scale had very high reliability for average-to-high levels of autogynephilia, but low reliability for low levels of autogynephilia. As with several previous scales, this likely reflects the low prevalence of autogynephilia in the sample. This factor has similar items to the genderplay factor in Aella’s Kink Survey. It may be worth adding items from that survey, especially the item about being submissive in sexual scenarios, which had higher levels of endorsement than other items of the genderplay factor.

Factor Interactions

Factor Correlation Matrix

In the following correlation matrix, each cell displays the estimated Pearson correlation coefficient in large font, with the 94% credible interval in smaller font. A notable correlation cluster contains affective gender identity, gender dysphoria, and autogynephilia, which show substantial intercorrelations, reflecting their common association with aspects of identifying as female.

Partial Correlation Matrix

The partial correlation between two variables is their correlation after eliminating the influence of a third variable (or set of variables). Inverting a covariance matrix yields its precision matrix, from which partial correlations and partial variances can be extracted. The resulting partial correlations are those between two variables while controlling for the influence of every other variable. The resulting partial variances reflect how much variance remains unexplained by all the other variables.

Compared to the correlation matrix, the partial correlation matrix has much smaller correlations. This suggests that the Markov network describing the interaction between the factors is sparser than it appears when examining the raw correlation matrix.

The BDSM Interest factor has notable partial correlations with:

• Homosexuality (+.23)
• Autogynephilia (+.22)
• Masculine Gender Norm Endorsement (+.19)
• Masculine Gender Role Insecurity (+.18)

Given that the BDSM Interest, Homosexuality, and Autogynephilia factors all consist of questions about sexual arousal, there may be a general sexual arousal (horniness) factor that links them all together. That would also explain the partial correlation between BDSM Interest and Masculine Gender Norm Endorsement. The association between BDSM Interest and Masculine Gender Role Insecurity remains unclear; it may be noise.

Deriving a Markov network from the remaining variables yields the following structure:

Thanks to Tailcalled for sharing the data!