@lanodan @Terry @p @Dave @fluffy It really isn't. First, there are 7 of these studies, with a combined sample of 387. But the actual validity of the result you get depends on the values you get. These aren't weak values, so the chances of them being completely wrong is low, but the authors calculate the p value anyway, in Firestone 2000 it's <0.05

@lanodan @Terry @p @Dave @fluffy Well, that's about the correlation they're studying there, not the non-offender specifically, but you get my point

@applejack @Terry @p @Dave @fluffy
> combined sample of 387

Lol, as if you could just add different statistics together like that, in programming this would be a fucking Type Error. Don't be JavaScript.

@lanodan @Terry @p @Dave @fluffy This is a valid thing people do. They take multiple studies and get a value by weighing them by their samples

I remember it here as an example from a really good paper
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4739500/pdf/emss-66004.pdf
As an example, a review of the world’s literature on intelligence, which included 10,000
pairs of twins, showed that identical twins are significantly more similar than fraternal twins,
with twin correlations of about 0.85 and 0.60, respectively, with corroborating results from
family and adoption studies, implying significant genetic influence (Bouchard & McGue,
1981, as modified by Loehlin, 1989)

@lanodan @Terry @p @Dave @fluffy When they actually do it they account for any differences, exclude based on criteria, correct for X, Y, and Z, etc, etc, but the point still stands that if a series of studies get numbers significantly higher than people generally think (1-2%), it's very very unlikely it's random chance

@lanodan @Terry @p @Dave @fluffy Like, 50 sounds like a small number, but think about it in context

You roll random numbers, there is a 1/100 chance for it to be true, what are the chances you get true 25x if you roll 50x? Pretty low

Okay, now what are the chances you get similar numbers again 7x?

@lanodan @Terry @p @Dave @fluffy *25% not 25x
*6x more times, can't believe I did a OBOE

@applejack @lanodan @Terry @Dave @fluffy

> This is a valid thing people do. They take multiple studies and get a value by weighing them by their samples

Literally nobody. Unless the methodology is the same, you can't do that.

> https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4739500/pdf/emss-66004.pdf

This is just a meta-analysis. They are not adding random numbers together.

@p @Terry @lanodan @Dave @fluffy Hint: they're not quoting a single study with 10,000 people, it's 111 of them

>Unless the methodology is the same
It's not, if you looked at it you'd see them talk about correcting for that
https://www.gwern.net/docs/iq/1981-bouchard.pdf

This is really the lengths we have to go to, bickering about how you can cite a study, what studies you're allowed to post, what sources it's allowed to be from, you need one single study instead of a bunch, all decided by you

You're really disingenuous and boring

@applejack @Terry @lanodan @Dave @fluffy

> It's not, if you looked at it you'd see them talk about correcting for that

A random meta-analysis of some genetic study has nothing to do with this. Methodological variables are much easier to control for, but "Here's a photo, did you get a boner?" and "We played them an audio tape with a cuff hooked up to their dicks" are so radically different that it is actually impossible to control for the variables in the methodology. Not just that, but the essay you're citing that cites the studies is not a meta-analysis to begin with: it's an essay written by a pro-pederasty activist. What you are suggesting is stupid.

> https://www.gwern.net/docs/iq/1981-bouchard.pdf

This is a meta-analysis of IQ. This has nothing to do with whether you can just add numbers from other studies (where you have not even attempted the impossible task of correcting for methodological differences)

> bickering about how you can cite a study

That's you wanting to bicker about that. I said you haven't read these studies. (You haven't.) I said that you would be able to come up with at least the titles of the papers that you're citing if you had read them. (You linked to an essay that cited them. You haven't read them, you just cite them.)

> what studies you're allowed to post

You have posted only one of the studies you cited. I said if you're going to cite them, you should have at least read them. An essay is not a study.

> what sources it's allowed to be from

I did not bring this up, but I did say an essay written by a pro-pederasty activist is not a study. I said this because an essay written by a pro-pederasty activist is not a study.

> all decided by you

You are making shit up about pedophilia and also making shit up about the conversation.

> You're really disingenuous and boring

This is in the debate script, right after "Try to get them to call you names and then tell them they're triggered." It says you should be incredulous that I didn't accept your "evidence" (an essay written by a pro-pederasty activist is not a study), and then declare that you are unwilling to engage in such a long conversation and bail. So this means you're giving up and bailing, right?

@applejack @lanodan @Terry @Dave @fluffy

> First, there are 7 of these studies, with a combined sample of 387.

That is such horseshit. The methodologies aren't comparable, so you can't just combine the sample size. Meta-analyses are a completely different class of study and this isn't even a meta-analysis, it's just "herp derp i added the numbers".

@p @Terry @lanodan @Dave @fluffy I already covered this

@applejack @Terry @lanodan @Dave @fluffy You did not. You blew smoke.

@lanodan @applejack @Dave @Terry @fluffy They're actually omitting a lot of data in those charts. The reason the sample sizes were small is that this chart was made by taking this studies that had 200 rapists and 40 guys that got $50 to be the control group. So this chart, they take the control group information and present it on its own, it's an intentionally deceptive chart.

@p @Terry @lanodan @Dave @fluffy They included a control group and the control group had a high arousal, it's stating 100% what it's stating

@p @Terry @lanodan @Dave @fluffy And it applies even in other studies, you just obsess over that one
https://www.sciencedirect.com/science/article/pii/S0005789405800395
Self-reported and physiological sexual arousal to adult and pedophilic stimuli were examined among 80 men drawn from a community sample of volunteers. Over ¼ of the current subjects self-reported pedophilic interest or exhibited penile arousal to pedophilic stimuli that equalled or exceeded arousal to adult stimuli

@applejack @Terry @lanodan @Dave @fluffy

> And it applies even in other studies, you just obsess over that one

I can't have been obsessing over that one, because I picked another one and downloaded it and read it over right before I wrote that reply. (I picked "Fedora" because it sounded like it wasn't real: 60 guys in the control group, plus 227 rapists.) So far, almost all of these studies are basically that. You'd probably know that if you had read them.

> blah blah hall95.pdf

I've read that one. That's the one the fake chart cites. The chart looked like bullshit and turns out to be bullshit. I wrote about this nearly a year ago.

Speaking of your "noooooooo these are representative samples", Hall wrote "Samples of normal men have been small and unrepresentative in previous studies."

> Over ¼ of the current subjects self-reported pedophilic interest or exhibited penile arousal to pedophilic stimuli

Congrats, you can read an abstract. Three of those 16 guys turned out to just be child sex offenders, and they weren't discarded from the study. That brings the number down to 13/77. The study was designed to measure whether they could inhibit responses by effort. "The most arousing slide for 72 [of 80] subjects was one of the adult female slides." and "The hypothesis that sexual arousal to pedophilic stimuli is a function of general sexual arousability factors was supported [...]". Not just that, but "it is also possible that self-reporting sexual arousal created a performance demand that may partially account for the self-report/physiological discrepancy." In other words, they had a cuff hooked up to their dick and they were asked a bunch of questions about what gave them boners. As many of them popped wood for the contrast slides as for the nude images.