If a dice rolls in a way that deviates from a null model, I would question the fairness of the dicee and the person rolling the dice.
I come to this conclusion not through statistics, but from the laws of probability. Statistics and probability are not the same thing. Stats can show what’s probable, but they do not determine probability unless they are statistics drawn from controlled experiments.
What?!? I wasn’t talking about you, I was talking about Bricker. You implied that I said you had accused Bricker of some things, but I never did. All I said was that it was only after you challenged the assumptions that he was making in that other thread, that he decided to come out with all these splendid disclaimers about “codependent variables”.
God, I thought your last failure to read for comprehension was just a one time thing. But you are clearly weak in this area.
Now that’s ironic.
…comprised of random events.
Which rape is not.
Why do I have a feeling that rape is the subject that you with the face has trouble with and not stats or probability questions?
In the Girls Gone Wild thread, Bricker mearly said that if what the woman told the reporter was true, then she was indeed raped but Mr. GGW denies the claims and then asked why people automatically believed the woman.
Then ywtf jumped in to hijack the thread and then later created this one. (btw, I don’t think Polish is a race)
Erm, what do you mean by “dicee”? (this is an honest question.) I’m trying to describe a dice that, rolled by an honest person (let’s say … you) rolls mostly fives. We would be stupid to presume that a given roll has equiprobable outcomes, given a large data corpus with 5 outcomes.
Similarly, if blacks make rape allegations against whites at a significantly higher rate than those rapes actually occur, then we would indeed be justified in assigning a lower prior probability of truth to such claims. I strongly doubt that this is the case, and I believe that studying such a statistic is highly impeded by the observability of key figures, but I don’t believe it to be an intrinsically stupid thing to study. I don’t see the point of your supposed distinction between “stats” and probability; of course we are assuming unbiased figures are available - what would be the point of debating in the context of unreliable numbers? Our job is to determine what the available numbers actually represent (which is why I have consistently distinguished between conviction rates and occurrence rates, for example), and how those relate to the probability figures we would like to assess. I have consistently and painstakingly acknowledged (hell, I’ve been hammering on the point) that the numbers we can observe are mere approximations to the numbers we would like to obtain, but my basic point (to people like you with the face at least) is that the basic probabilistic effort is not intrinsically ludicrous.
you with the face: dear Lord, this is tortuous. No, I did not say you said I accused Bricker of anything - I objected to my posts being used as support for you accusing him of something. Like I said: leave me out of your godawfully banal blatherings; I have no desire for my posts to be associated in any way with your inanity. I hope this is sufficiently clear. I say “I hope”, of course; but it’s sort of the forlorn hope you get when trying to explain the internet to an elderly uncle, or when trying to teach calculus to your cat. Springs eternal, and all that.
Going of just those facts the accusation againt CEO White McWhiterson of course, but once again that has no bearing on the innocence or guilt of the CEO Black McBlackerson. That in itself is my very arguement, I’m not sure what your trying to prove with that.
At no time did I get the impression from anyone participating in the thread that race was somehow a determining factor, or indeed the factor that should be judged at all in a court of law. The only ones who seems to have allowed that thought to enter their mind is yourself or monstro.
I am not aware of anyone advocating that black women are statistically shown to be liars and are supposed to be believed far less often because they are raped by white men in smaller amounts when compared to white women. The prudent judgement, as you said monstro would be to wait until you could get as much evidence as possible, which is what I would expect the court to do in situations such as this. However if race (or height, weight, ect.) is all we have to wager on, I’m going to bet on whatever the statistics show every time. Read again: if that is ALL we have, meaning nothing else, that is the smart play to make. We are not a jury here, nor are we judging a criminal case, in a purely academical sense I fail to see how you cannot grasp this if looking at the facts without prejudice.
Once other fragments of the evidence come to light, the issue of race (height, weight, ect.) becomes rather unimportant in the grand scheme of things, but that is once other evidence comes to light. Sure the white woman can say that her rapist flew down on say a young or very young green dragon (whichever your preference) prior to raping her, which would cast an enormous shadow on her stories credibility. After digesting that information the black woman’s accusation would be far more credible barring some erroneous portion of her story.
Well statistics would show that white people are more likely to commit white collar crimes, as you with the face has helped to point out. But if we were to assume that your statistics are proven correct and the ones on race were not existant, and he was indeed wearing red underwear while most of those types of crimes are commited by white underwear folk, than I would be more apt to say it’s less likely he commited the crime compared to a suspect wearing white underwear. That does not mean Mr. Red Underwear is innocent, nor does it mean Mr. White Underwear is guilty.
I know I’m repeating myself here, but it bears repeating, no one is saying anyone should be judged in a court of law based on race. Rape is a very serious crime, and every single one reported should be investigated thoroughly regardless of the race (height, weight, ect.) of the suspect or victim. Some of you seem to be so sensitive to the subject of race that you would believe that I think otherwise.
Additionally, I believe some of you are struggling with the word credible and it’s usage here. Try replacing it with likely or plausible and see if you that changes the way your looking at the situation a little.
I’m saying you have no basis to say that the accusation against Mr. White is more credible than the other. I was hoping to get you to see that the pertinent factor (being a CEO) is the more important consideration for assigning relative likelihood. Both guys are CEO’s, therefore both are in a position that favors this type of crime, therefore you have no rational reason to conclude that the charge against Mr. Black is less likely if all you have to go by is that he is black. My point ultimately is this: So what if he’s black? He’s a fucking CEO.
When someone uses racial crime data to suggest that a specific allegation is suspect, they are essentially saying that race is a determining factor for that crime. As monstro pointed out, using race like that makes as much sense as using underwear color. People glommed onto a particular characteristic, assumed that it is a relevant factor, and insisted upon its use in assigning a estimate of credibility.
You obviously don’t see a problem with that, and that’s fine are far as this thread is concerned. But I’m not arguing against a strawman.
Actually it doesn’t.
Just in the hope that answering this will help:
My answer is exactly the same as it is for the case of rape allegations: we can not tell, unless we also know the rate of allegations against blacks and whites for fraud. If 97% of allegations are levelled against white men, then we would say that the allegations are equally plausible, since they match the actual crime rates. If 97% of allegations are levelled against black men, then we would say that allegations against black men are massively less plausible, since so few of them result in conviction. From the scenario you present, we have only two allegations, so can make no statistically valid inference as to the rate of allegations against any given race, and should therefore consider each allegation to be equally plausible. This is not the same as declaring race to be irrelevant; it is declaring it do be of inobservable importance, given the posited data set. This is an important distinction.
In no instance should we let our assessment of the prior probability colour our examination of the facts of an individual case.
It was a typo. I meant “dice”.
And I said that if you tell me you’ve rolled the same number 100 consecutive times, I will doubt that you’re rolling randomly. I will look askance at both the dice and the person rolling the dice.
But I’m failing to see how this is at all relevant to the discussion. I’m not being snarky; I just don’t understand why you would assume I don’t get this basic principal based on what I’ve said.
You are using probability in a weird sense. Probability is based on randomness, not on credibility. If a black woman claims to have been raped by a white person, probability-wise, her claim is quite likely. Whites outnumber blacks considerably. Black women are very likely to come in contact with a white person. A null model of black-white encounters would NOT justify skepticism about the accuser’s claim. So no, we don’t lower the “probability”. The probability of a black woman getting raped by a white man is actually higher than it is for a white woman by a black man, in a world that works strictly stochastically.
What if your hypothetical statistic is true, but we also have stats showing that males are fifty times more likely to lie in general than women? Or rapists are fifty times more likely to lie about having sex when they are intoxicated? It might be justifiable to use your stats about race and false accusations, but why stop there? Why should an individual give so much weight to race–a factor that has not been casually linked to rape–to the neglect of more relevant factors, such as gender or intoxication?
Isn’t it easier just to wait and see what the investigation turns up?
Especially not in this thread, where he’s actually been called “disingenous” three times.
I think it means “not a genius”? I’m not sure.
Race is irrelevant if only for the reason that its relevance hasn’t been demonstrated. Sure, it could be relevant. But so could underwear color. Pig’s could fly, too. If they had wings and stuff.
There is no data in evidence which shows that one’s race provides any meaningful information about the likelihood that a particular claim which has already allegedly happened did in fact happen. Therefore, it makes perfect sense to say that on the basis of race alone, you can not conclude whether the two claims that I put forth are more or less credible relative to one another.
Do you see yet that I’m not arguing about prior probability? I’m talking about an event that has already happened. The “probability” that a past event has occured is either 100% or 0%. I have no idea what you think is so egregious about my position when nothing you’ve written disagrees with the points I’ve raised.
Nitpicky, yes; arsehole…not so much. 
It seems to me that Bricker is dead wrong. He said that “Janice’s” story was more credible. If he had said her story is more likely to be credible, it would be more or less accurate in a maddeningly irrelevant sort of way, but he did not say that. In my understanding of the concept of credibility, it has squat-diddly to do with statistics and literally only hinges on the specific situation. Looking at a case as a set of data from a distance, one could possibly say that Janice will be “more likely to be credible” (though I strongly question the usefulness of this), but once one actually zooms in on the case as a unique situation, one can’t apply statistics to credibility.
No, you’d have to be a fool to do anything with that kind of “evidence”. In the real world, it’s just a worthless piece of information. If that’s all you have, drop it 'cause it proves nothing and leads nowhere. If you’ve got more than that, use the rest and drop that, 'cause it adds nothing.
Which is exactly my point as well, once again your proving nothing other than what I’m saying. So you will concede that CEO’s tend to commit crimes in a white collar fashion, moreso than perhaps a construction worker? This is a fact that you stated above, that they are both CEO’s and thefore they are more likely to commit a crime than a different party, a party based entirely on profession.
Should I be decrying the fact that your villanizing CEO’s? Absolutely not, they are more likely to commit that particular crime in that particular manner. Besides, much like how I never said the white woman was automatically telling the truth, you never said that means the CEO is indeed guilty.
Now apply that same line of thought to race, white people tend to commit white collar crimes moreso than black people. Sure I could whine and moan that you never told me if the black CEO has a history of committing that type of crime, or if the white CEO is a lying crackhead, but that’s exactly the reason why you make your decision based on all that you can know. To apply it to our previous argument, the white woman could have said 10 different men have raped her prior to this accusation and gone back on what she said, or the black woman could have her rapist on tape, but that does not apply to the situation we were talking about.
Which would you chose if your life was on the line, and you had to guess the race and job of an individual who committed fraud, and the four combinations you had to chose from were a white CEO, a black CEO, a white construction worker, or a black construction worker? That is the only evidence that you get to make your decision from, nothing else. Be honest now, pretend your life was on the line. I’ll tell you if most people had a gun to their head they would say a white CEO, because they have nothing else to go on but the fact that he’s a white CEO and they know that it’s more common for them to commit such an act in that way.
Now of course, that doesn’t mean the black construction worker did not commit fraud anymore than it means the white CEO did, just as it doesn’t mean the black woman wasn’t raped whereas the white woman was. However, given only minimalistic knowledge of these 2 situations such as race, profession, or gender of the parties involved you can still predict which is more likely to have occured if only one out of two options can be true. Pay close attention to the fact that I said more likely to have occured, not that it did or did not happen based off that knowledge.
On preview: to those saying it has nothing to do with the real world, did you not catch the times I said it was purely in a theoretical sense? I would be one of the first people saying a decision was wrong if a judge came back with a sentence stating that the white suspect charged with raping a black woman was innocent because she was black and therefore could not have been raped by him. Additionally Roadfood, I’m not trying to use that evidence to lead anywhere other than what the evidence itself already is. If the evidence shows it’s more likely a white woman will be raped by a white man when compared to a black woman, then it speaks for itself. Ignoring it does not make it go away, but that does not mean that it is an appropriate statistic to use in an a court of law.
Your only thinking of one way credibility can be used, in this case it is interchangable with likely, possible, or plausible. I had a feeling that some may be disagreeing because of the way it was worded, which is not necessarily the way I would have worded it myself.
While we’re addressing hypotheticals, suppose we compare the following two claims:
“In the 1850s, my great-great-grandfather was scalped by the Comanches”
“In the 1850s, my great-great-grandfather was scalped by the Apaches”
Which is more likely, if we happen to know that Apaches did not practise scalping?
Depends. Some people say that the practice of scalping was brought over by the Europeans. Did the people in the forts pay bounty for scalps of, say, Sioux? Was your grandfather dark haired enough that his scalp could be used for bounty?
Depends. What was wrong with the original question that you had to change the hypothetical?