"This is where my concept of ‘correlations’ breaks down. I think this effect we’re seeing is real and meaningful ... . But the correlations here are negligible. "
It's a base rates issue. If A is rare and B is common, then A can't be strongly correlated with B even if A strictly implies B; most cases of B will have nothing to do with A.
For instance, imagine an adventurer's guild. Each member is either a fighter or a wizard, and either a human or an elf. Half the humans are fighters, and half the humans are wizards. All the elves are wizards. So conditional on being an elf, the likelihood of being a wizard is doubled. Intuitively that's a strong relationship. But what's the correlation between species and job? It depends on the ratio of elves to humans.
With 10 humans and 10 elves you get 10 elf wizards, 5 human wizards, and 5 human fighters. The job/species correlations is R = 0.58.
With 90 humans and 10 elves you get 10 elf wizards, 45 human wizards, and 45 human fighters. Now the job/species correlation is only R=0.3. With 990 humans and 10 elves, the correlation drops to R=0.1.
Basically, even though knowing someone is an elf tells you a lot about their job, if elves are very rare then knowing who the elves are doesn't help you much if you're trying to predict everyone's jobs.
To put it another way, a correlation of 1.0 between a life history and a kink would mean both that everyone with the history has the kink, and everyone with the kink has the history. So the correlation is reduced for each person who has the history but not the kink, but it's also reduced by each person who has the kink without the history.
Also remember a sample size as huge as Aella's is going to pick up a lot of subtle correlations.
What the correlation coefficient actually measures is the *square root* of the percentage of variation explained by the model (that's why you often see R-squared quoted in more complex models, easier to understand its relationship to the data). So these things only explain 1-2% of the total variation in interest in the population (0.1 or 0.14 squared), even though they're definitely real.
Predictably, many people in the comments are interpreting "X is correlated with Y" as "X causes Y", not realizing that it could equally well be Y causing X, or some third thing causing both.
May be worth editing the title to "Are people with rape fetishes more likely to report being sexual assault survivors?" to try to ward off those people jumping to conclusions.
Is it possible that the cause of the correlation you're seeing is the other way around? Perhaps people who report a rapeplay fetish are more likely to fantastically "remember" or claim to be victims of sexual assault?
Correlations aren't directional. If people with X are more likely to have Y, people with Y are also more likely to have X.
You're probably thinking of causation; that raping someone causes them to develop a fetish for rapeplay. And indeed it could also be the other way around; that having such a fetish makes people more likely to report borderline actions as sexual assault. Or they could both stem from a third cause, like growing up in an abusive household.
Chiming in with an n of 1 from a throwaway to offer some causal theorizing (assuming the correlation holds).
I am male, 27, sexually assaulted at 18 by a former partner (probably categorizable as mild), would report cnc fantasties (receiving only) probably starting a year or two after my assault, would not prior. No childhood abuse, and that assault is the only time I have been on either end of a non consensual sexual experience.
My fantasies *feel* like they developed in a way that was heavily tied to my assault (like, my earliest memories of them were remembered my assault and noticing some uncomfortable arousal). However, I want to raise some other issues:
- I may have had compatible fantasies before. While I can’t know how my fetishes would have developed without being assaulted, I’d put the odds of my developing some kind of cnc fantasy anyway at around 30%, and the odds of my developing similar-but-less-extreme fantasies (e.g. being tied up) at 90%+
- I was aware of the theory that sexual assault leads to interest in cnc. This might mean that I subconsciously pushed myself in that direction through some social-norms-mechanism that I don’t have a clear theory of. It also might mean that, combining with the point above, being assaulted gave me “permission” to develop a fantasy I was already prone to
- simple association could be doing a lot of work here. It’s natural to think about sex when remembering my assault, and it’s natural to remember my assault when thinking about sex. Over time, it doesn’t seem like a stretch to imagine that leading to some kind of mild kink
It’s not uncommon for people who have survived sexual assault to come to climax during it despite feeling utterly violated. This can happen once or repeatedly and can instill a huge association with rape and arousal.
ESPECIALLY given the fact that sexual assault can often be violent or scary - this physiological and cognitive arousal state of fear is intense and by misattribution of arousal, it can cross-wire their brains to associate rape with intense arousal (fear-based adrenaline and excitement-based adrenaline are the same in many ways).
Another factor is that some individuals often process their trauma through rapeplay. I recall recently reading a story where a girl found resolution in masturbating to her sexual assault memories from her Uncle. After she finally let herself enjoy the feeling while thinking of the memory, she felt resolution.
If it is the case that being assaulted by more than one victimizer is correlated with higher rates of rapeplay fetish, you would see this same "any particular victimizer subgroup is higher than the overall average" effect.
I think this is more likely than the cause being related to honesty.
To give a concrete example, imagine that the 4 groups "father" "mother" "both" and "neither" all have the same number of people and the interest in rapeplay is as follows:
Both: 60%
Only one: 30%
Neither: 20%
Then the average among people assaulted by either parent is (60%+30%+30%)/3 = 40%, while the average among those assaulted by a specific parent is (60%+30%)/2 = 45%.
I don't think they were. Aella doesn't report the number or % of people reporting sexual assault by any given person. The figures are % of people reporting interest in rapeplay, given that they were assauted by X, not the % of people reporting assault by X.
Checking females who reported either ‘moderate’ or ‘severe’ childhood sexual assault (n=21k), here’s the % of the groups who reported interest in rapeplay, binned by who they say sexually assaulted them in childhood.
So over on ACX, someone pointed out there are correlations between rationalism, polyamory, kink, LGBT, neurodivergency (autism, ADHD), and being a huge nerd.
I despaired at ever finding any data on this sort of thing, but...you know who would?
I think most of your followers would find it of interest. What correlates with what? It's no secret that if A correlates with B, and B correlates with C, A doesn't necessarily correlate with C (think of education, income, and politics), but it would be neat to see what correlates with what!
Of course sexual assault survivors are more likely to have rape fantasies, for the same reason that people who were sexually assaulted as children or began masturbating as children are more likely to be pedophiles: orgasms train sexual interest.
"This is where my concept of ‘correlations’ breaks down. I think this effect we’re seeing is real and meaningful ... . But the correlations here are negligible. "
It's a base rates issue. If A is rare and B is common, then A can't be strongly correlated with B even if A strictly implies B; most cases of B will have nothing to do with A.
For instance, imagine an adventurer's guild. Each member is either a fighter or a wizard, and either a human or an elf. Half the humans are fighters, and half the humans are wizards. All the elves are wizards. So conditional on being an elf, the likelihood of being a wizard is doubled. Intuitively that's a strong relationship. But what's the correlation between species and job? It depends on the ratio of elves to humans.
With 10 humans and 10 elves you get 10 elf wizards, 5 human wizards, and 5 human fighters. The job/species correlations is R = 0.58.
With 90 humans and 10 elves you get 10 elf wizards, 45 human wizards, and 45 human fighters. Now the job/species correlation is only R=0.3. With 990 humans and 10 elves, the correlation drops to R=0.1.
Basically, even though knowing someone is an elf tells you a lot about their job, if elves are very rare then knowing who the elves are doesn't help you much if you're trying to predict everyone's jobs.
To put it another way, a correlation of 1.0 between a life history and a kink would mean both that everyone with the history has the kink, and everyone with the kink has the history. So the correlation is reduced for each person who has the history but not the kink, but it's also reduced by each person who has the kink without the history.
Thank you!
Is there another term for this? Like I want to be able to say "It's not strongly correlated, but there is a strong ____"
Hey, I only just now realized that I never properly answered this. I don't know if it's relevant to you now, but I threw some suggestions down in a Google doc: https://docs.google.com/document/d/1mGX2v6usSZYOJEVs0iROH67rbQd2jht16F0kAke7e6Q/edit?usp=sharing
Keywords for further reading would be "odds ratios," "conditional probability", or "relative probability."
Also remember a sample size as huge as Aella's is going to pick up a lot of subtle correlations.
What the correlation coefficient actually measures is the *square root* of the percentage of variation explained by the model (that's why you often see R-squared quoted in more complex models, easier to understand its relationship to the data). So these things only explain 1-2% of the total variation in interest in the population (0.1 or 0.14 squared), even though they're definitely real.
Predictably, many people in the comments are interpreting "X is correlated with Y" as "X causes Y", not realizing that it could equally well be Y causing X, or some third thing causing both.
May be worth editing the title to "Are people with rape fetishes more likely to report being sexual assault survivors?" to try to ward off those people jumping to conclusions.
Is it possible that the cause of the correlation you're seeing is the other way around? Perhaps people who report a rapeplay fetish are more likely to fantastically "remember" or claim to be victims of sexual assault?
Correlations aren't directional. If people with X are more likely to have Y, people with Y are also more likely to have X.
You're probably thinking of causation; that raping someone causes them to develop a fetish for rapeplay. And indeed it could also be the other way around; that having such a fetish makes people more likely to report borderline actions as sexual assault. Or they could both stem from a third cause, like growing up in an abusive household.
Chiming in with an n of 1 from a throwaway to offer some causal theorizing (assuming the correlation holds).
I am male, 27, sexually assaulted at 18 by a former partner (probably categorizable as mild), would report cnc fantasties (receiving only) probably starting a year or two after my assault, would not prior. No childhood abuse, and that assault is the only time I have been on either end of a non consensual sexual experience.
My fantasies *feel* like they developed in a way that was heavily tied to my assault (like, my earliest memories of them were remembered my assault and noticing some uncomfortable arousal). However, I want to raise some other issues:
- I may have had compatible fantasies before. While I can’t know how my fetishes would have developed without being assaulted, I’d put the odds of my developing some kind of cnc fantasy anyway at around 30%, and the odds of my developing similar-but-less-extreme fantasies (e.g. being tied up) at 90%+
- I was aware of the theory that sexual assault leads to interest in cnc. This might mean that I subconsciously pushed myself in that direction through some social-norms-mechanism that I don’t have a clear theory of. It also might mean that, combining with the point above, being assaulted gave me “permission” to develop a fantasy I was already prone to
- simple association could be doing a lot of work here. It’s natural to think about sex when remembering my assault, and it’s natural to remember my assault when thinking about sex. Over time, it doesn’t seem like a stretch to imagine that leading to some kind of mild kink
Honestly I’m not surprised by this at all.
It’s not uncommon for people who have survived sexual assault to come to climax during it despite feeling utterly violated. This can happen once or repeatedly and can instill a huge association with rape and arousal.
ESPECIALLY given the fact that sexual assault can often be violent or scary - this physiological and cognitive arousal state of fear is intense and by misattribution of arousal, it can cross-wire their brains to associate rape with intense arousal (fear-based adrenaline and excitement-based adrenaline are the same in many ways).
Another factor is that some individuals often process their trauma through rapeplay. I recall recently reading a story where a girl found resolution in masturbating to her sexual assault memories from her Uncle. After she finally let herself enjoy the feeling while thinking of the memory, she felt resolution.
If it is the case that being assaulted by more than one victimizer is correlated with higher rates of rapeplay fetish, you would see this same "any particular victimizer subgroup is higher than the overall average" effect.
I think this is more likely than the cause being related to honesty.
To give a concrete example, imagine that the 4 groups "father" "mother" "both" and "neither" all have the same number of people and the interest in rapeplay is as follows:
Both: 60%
Only one: 30%
Neither: 20%
Then the average among people assaulted by either parent is (60%+30%+30%)/3 = 40%, while the average among those assaulted by a specific parent is (60%+30%)/2 = 45%.
This is a field of study that would really benefit from random assignment. But, then, that pesky IRB.
Yeah, they'd probably frown on genetically-modifying random people to be more into rapeplay.
Yes, quite. Though the punchline I was looking for was, "we couldn't run the experiment because we got informed consent."
[RETRACTED]: I'm very surprised that as many females were sexually assaulted by their mothers as by their fathers.
I don't think they were. Aella doesn't report the number or % of people reporting sexual assault by any given person. The figures are % of people reporting interest in rapeplay, given that they were assauted by X, not the % of people reporting assault by X.
Quote from the post:
Checking females who reported either ‘moderate’ or ‘severe’ childhood sexual assault (n=21k), here’s the % of the groups who reported interest in rapeplay, binned by who they say sexually assaulted them in childhood.
Mother: 49.8%
Father: 46.5%
Someone I didn’t know well: 43.0%
Sibling: 42.7%
Extended family/family friend: 41.3%
"here’s the % of the groups who reported interest in rapeplay"
Each group in this case is "people sexually assulted by X". So it's showing what % of the [people sexually assaulted by X] are into rapeplay.
Oh! Right! I misread that, and misread it again in Andrew's reply..
Hail, Lady of Spicy Statistics!
So over on ACX, someone pointed out there are correlations between rationalism, polyamory, kink, LGBT, neurodivergency (autism, ADHD), and being a huge nerd.
I despaired at ever finding any data on this sort of thing, but...you know who would?
I think most of your followers would find it of interest. What correlates with what? It's no secret that if A correlates with B, and B correlates with C, A doesn't necessarily correlate with C (think of education, income, and politics), but it would be neat to see what correlates with what!
Spanking doesn't seem like a sexual assault to me. But that's just me. When I spank it's part of foreplay so it's mixed with lots of fondling.
Of course sexual assault survivors are more likely to have rape fantasies, for the same reason that people who were sexually assaulted as children or began masturbating as children are more likely to be pedophiles: orgasms train sexual interest.