I do believe you’re committing the same error, however, at a later stage. You’re claiming that knowing the prior probability P(rapedbywhite|black) gives you some intuition as to the likelihood P(blackrapedbywhite|claimmade). This is in fact not the case unless you know something about claim rates too. The only way you could (casually) infer a lower reliability for an arbitrary black person claiming rape by a white would be if black people claimed rape against whites at a rate disproportionate to the actual incidence of those rapes. This is why you are reaching a false conclusion in your linked post; you take two isolated claims, and reach an inference about their reliability, despite not having any information whatsoever about the prior reliability of such claims. If you had 100 rape claims against whites from each ethnicity for a given jurisdiction, and a significantly lower white-on-black rape rate, then you could start saying that the prior probability of a white-on-black rape claim being true is lower, but you have only two claims, and can make no such inference.
I say “casually” above because even to make that inference would be to assume constant probability of conviction across ethnicity groups, approximation of conviction rates to rape rates, and several other shortcuts of varying dubiousness. In reality, you know neither of the above figures with any certainty, since you have no direct access to the real rape figures, viewing them only through the filter of the criminal justice system, which, given the low and wildly varying conviction rates for rapes, is probably just as compromising a factor as your initial error.
I’m very definitely not accusing you of racism or anything (analysing numbers is a neutral game), nor do I really wish to resurrect what looked like a disastrous trainwreck, but I do think your application of statistics in this instance is quite heavily off-base.
Having seen some of face’s ability to comprehend a simple mathematical problem, and back herself into a corner where she was desperately changing the hypothetical every time she touched the keyboard in an absurd attempt to prove she was right all along, I’ve formed my own opinion on her intelligence and professionalism, and I agree someone’s too far gone in his bias here, but it ain’t necessarily Bricker.
But you articulated very clearly for someone with his tongue wedged firmly between face’s ass-cheeks.
Instead of responding to me, respond to the two posts above yours which clearly detail exactly why you’re a blithering fuckwit. I assume you are also accusing those posters of being incapable of comprehending “a simple mathematical problem”?
I haven’t posed these posters a simple mathematical problem and watched them floor it the way face did, for a start. And since I haven’t said anything about black women lying about being raped, perhaps you can tell me where either of those posts detail exactly why I’m a blithering fuckwit.
Now get back to licking face’s fanny-crack, buttmunch.
Let’s not go putting words into people’s mouths, hmm? I disagree strongly with your points re: population versus the individual, and believe that you are becoming massively over-offended at the very idea of prior probability. One can come up with all sorts of discriminating factors to try and predict people’s behaviour, and in many instances you’ll get some interesting and valuable results. Where one has to be careful is in focusing on a weakly discriminating factor when other possibilities are manifold and often stronger. Nonetheless, these sorts of calculations occur constantly; every time you’ve applied for credit, you’ve been subjected to just the sort of heinous calculation against which you appear to be railing. The contention that one cannot apply a prior probability to an individual event is insupportable; people do not exist in a vacuum, and exterior factors affect their action in ways of varying predictability. No, it’s nothing like as precise as studying 100 throws of a dice, since the common factors are so much looser and ill-defined, but that doesn’t mean that the effort is impossible.
Even if one did analyse the data, and conclude in a statistically valid manner that black people are more likely than whites to make a false rape claim against whites, this would still not be the same as stating that their ethnicity causes such a statistic. In such a hugely inter-dependent model, ethnicity as a discriminating factor would be an agglomeration of a huge number of other, less observable factors such as economic status, locality etc. This does not mean that the statistic itself can not be observed, however; merely that it should be treated with more care. No, I do not believe that ethnicity is a causative factor in human behaviour, but it may sometimes be an observable proxy for other factors.
Incidentally, your claim that race has nothing to do with this is rather roundly contradicted by this quote:
…the prevalance of the type of attack has some slight, tiny, non-zero influence on our judgement of credibility of a single report of such an attack.
Specific evidence about the specific attack is overwhelmingly MORE probative. But in the ABSENCE OF OTHER EVIDENCE, with only the past statistics and nothing else on the table, the past statistics are a non-zero influence on our judgement. Tiny. Miniscule. But NOT ZERO.
No, he can not. He thinks that makes perfect sense. And because he took a class that touched on statistics back in the 80’s some time or another, that means he’s right. Who are we to question him?
You would think a lawyer would know this–and not only know it but embrace it–but that’s a bad assumption apparently. Silly us.
Please try to follow the argument. I know it’s complicated and has lots of big words. But try to focus your meager reasoning power on this sequence of posts:
Answer: NOTHING!
But it has everything to do with the discussion we were having and the point you posed, which I was answering and you evidently forgot immediately after posting it. Being imprisoned at Riker’s Island is NOT a neutral statement; it DOES imply something about character or behavior, and it IS correct to say that a random pick of a Riker’s Island inmate will be black.
See, how it works is, you say something, and I answer you.
IN THE ABSENCE OF ANY OTHER EVIDENCE. IN THE ABSENCE OF ANY OTHER EVIDENCE. IN THE ABSENCE OF ANY OTHER EVIDENCE. IN THE ABSENCE OF ANY OTHER EVIDENCE. IN THE ABSENCE OF ANY OTHER EVIDENCE. IN THE ABSENCE OF ANY OTHER EVIDENCE. IN THE ABSENCE OF ANY OTHER EVIDENCE. IN THE ABSENCE OF ANY OTHER EVIDENCE. IN THE ABSENCE OF ANY OTHER EVIDENCE.
This post was brought to you by the In The Absence Of Any Other Evidence Singers.
Assigning even probability to a specific claim on the basis of irrelevant data such as:
– race/ethnicity
– gender
– systolic heart pressure
– hips to waist ratio
– age of menarche
– shoe size
– religion
– or any other quality that has no direct bearing on the factor under consideration (such as credibility)
…is worthy of derision.
See, it’s not about race. It’s about your stupidity.
Yet there are more blacks in Rikers Island, percentage-wise, than there are in the
general population in New York City.
Race has no direct bearing on being a criminal. But if I have to make a guess about the race of a random Riker’s Island inmate, I would be foolish not to guess “black.”
If it were to develop that systolic heart pressure had such a correlation, then it, too, would be of some probative value.
On the one hand, I wasn’t going to jump in and defend Bricker, as he is completly able to do that himself.
But on the other hand, You with the face, why are bringing back that other thread into this one? Bricker didn’t mention race in this thread. You did that. Your personal axe to grind with Bricker has completly derailed this over due pitting of the Girls Gone Wild chap, just so you can bring up a months old arguement that you have lost and I can only guess the bitter taste of bile comes up in the back of your mouth when ever you see Bricker’s name. I know that feeling as well, but I got over it.
If you want to argue with Bricker about what he said in a different thread, do it in that thread, or start a new one.
I’m not “massively overoffended” at anything but Bricker’s continued stupidity in misunderstanding this point. Nothing you say is anything I’ve argued with at all, let alone “railed against”. There are circumstances and factors which are valid for use as data points in determining likely qualities of individuals, but this is not one of them. You’re putting words into MY mouth, not the other way around.
Race has nothing to do with the basic argument. It has a lot to do with the adjacent reaction, of course it does. But “race” can easily be replaced by a number of other factors. I’m surprised you don’t understand this.
Not without data regarding claim rates. Indeed, if whites raped whites 9 times more often than they rape blacks, but only 1 percent of rape allegations against whites were from black people, then we might then infer that in fact black people’s rape claims against whites have a higher prior probability of being true. Even then that would contain the dubious assumption that blacks and whites are equally likely to report such a rape if it genuinely occurs, along with the other simplifications I mentioned earlier. The best approximation we can get to a prior probability of a claim’s validity is to assess the percentage of claims that result in a conviction. Obviously this is a poor approximation, but equally obviously it intrinsically factors out the prevalence of the crime.
A rare crime merely means that one’s sample size for estimating the reliability of a class of claims is smaller, so in this instance the only effect that the prevalence of white-on-black crime would have is making our reliability estimate less accurate.
To continue arguing from the middle: you with the face, do you not understand that analysing the probabilities is a matter of discovering which factors are relevant? The whole point is that one can examine various contributing factors, and discover either that they have no statistically significant effect (as one would in the case of shoe size, for example), or one might find that they do, and investigate further to see if this is a causative factor, or is merely co-dependent on a hidden, unexamined factor.
To take your shoe size example, I agree that one might well expect that shoe size is irrelevant, and that to study it is a waste of time. Indeed, if we did study it, I would be amazed to find that it had any effect, but that does not make the act of studying it stupid. What Bricker is saying is that IF one studies it and IF one discovers that shoe size does have a statistically significant bearing on a particular outcome, THEN you can indeed use that as a predictive factor. Clearly it’s a faintly ludicrous factor, and one would want to study why it has an effect, and what the underlying cause is, but it seems to me that you are wilfully ignoring the fact that Bricker is answering his question on the condition that the data shows a particular outcome. Yes, he’s making erroneous inferences at a later stage, but you appear to be mocking the very idea that someone might statistically analyse factors which you deem silly. Which is, itself, silly.
I incidentally find it hilarious (although not suprising) that you failed to understand that Dead Badger was criticizing YOUR position above; your demand that I respond to him is just freakin’ funny.
And very consistent with the rest of your comprehension skills.