you’re doing the reverse. yes, the cops should at least consider the parents, but should not decide that this stranger is less likely because parents (in general) are more likely. and that’s the use that’s being suggested in this case. that because there exists some data point somewhere that says whites rarely rape blacks, that fact should be used to evaluate the specifics of this case. there exists quite a bit of evidence that suggests the guys are innocent (in case some other ijiot again suggests that my stance includes believing the accuser in this particular case, despite the evidence). but the data on white on black rape isn’t part of it.
Be careful John. You’re trying to shift too many paradigms at a time. The boys were really happy about the whole “vanishingly rare” thing and you just burst their bubble.
Well, the issue in this case is that virtually all the evidence seems to support the statistical likelihood. If all the evidence pointed to the parents, I wouldn’t have a problem tossing in “and in cases like this, statistics show us that parents are the most likely culprit as well.”
So in my hypothetical, tossing “parents usually do it” into the mix would be wrong?
it would elicit a 'so what;?
seriously, try mentioning that data point in the next pit thread about some monster stranger who killed a kid.
Yeah. Why is the stranger accused in the first place? Witnesses maybe? There is no reason to place the parents on the differential list if there is a good reason to believe the stranger may be linked to the crime.
Which corresponds in no way to the situation here.
Here . . . several weeks ago . . . when less public information was available . . . I made an en passant reference to statistics of the government that sent several people into a batshit coma. I typed thousands of characters, in doing so, that said: “in the absence of other, dispositive proof.”
Every dissenter here ignored those qualifications. I regret the moments of my life that I spent in typing them.
As I have (also) wasted thousands of characters posting: the DoJ data is now almost overwhelmed by other information making the accuser’s accusations seem dubious as all Hell.
Or . . . are you suggesing that in this case, “all the evidence pointed” initially to the accuser’s story being true, as opposed to being, um, untrue?
My detractors have utterly ignored these points, and have backtracked as though to claim they did not originally believe her. But, I remain convinced, they did – in the face of “all the evidence” to the contrary. I find it significant that they are still arguing over the DoJ statistics (which, for the point of argument, I begged them to pretend did not ever exist so as to confound their very worldview, and which, however, they continue to freak out about, calling me more or less a Nazi for citing government tables, in a different thread).
None of the Kool-Aid crowd has responded to my challenge:
Did you initially believe or give some credence to the Duke accuser – in your heart of hearts?
Why did/did you not believe her?
How has that affected your approach to this thread?
That tells me volumes (not that I would have expected honest responses from this lot on these issues).
Right, so you’re a lot happier appealing to the stats when they support your particular worldview. I’m glad that the presence of the Atlantic means I’m unlikely to be affected by any epidemic you’re ever working one, and I gladly yield you my place on the short bus (assuming I’m allowed to say “bus” to a black American woman without it being considered hate speech).
Construe:
Exercise 1: I am an extremely truthful individual - I have never been caught telling an untruth. I arrive late for work one morning and I apologise for my tardiness with the excuse that I had to call the funeral director after my father was struck and killed by a meteorite. Do you assume that my statement is consistent with my prior reputation for truthfulness, or do you weigh it against the statistic that there have been exactly zero deaths from direct meteorite strikes in recorded history? :dubious:
Exercise 2: A key witness in a hit-and-run motoring case identifies the offending vehicle as maroon, a very unusual colour. Registration data reveals that, of cars that are either red or maroon, there are 49 red cars for every maroon one. The witness is tested under conditions comparable to the hit-and-run incident and shows up with the impressive ability to tell a red car from a maroon one 95% of the time. What is the probability that the witness is right? (As in the wonderful mathematical world of flat frictionless horizontal planes and rigid inextensible infinitely thin rods, assume that no other colours need be considered.)
It breaks down like this. Of every 1000 cars that are either maroon or red:
[ul]
[li]980 are red, of which 49 are mistakenly identified as maroon and 931 are correctly identified as red[/li][li]20 are maroon, of which 1 is mistakenly identified as red and 19 correctly identified as maroon[/li][/ul]
Therefore nearly 3 cars out of 4 identified as maroon will be mistaken identities - the extreme rarity of maroon cars outweighs the great reliability of the witness - and the odds are nearly 3-1 against the witness being right.
but , as you admit here, there was additonal info availalbe, which makes your introduction of irrelevant info more than a trifle odd. that has been my only interest in that thread, the case itself held little interest for me. but will answer the rest
as one of your detractors, I am very tired of you keeping on with this claim that some how or another my objection to your introduction of irrelevant data into the mix somehow equated to belief that the accuser is telling the truth, despite repeated explicit statements I’ve made that I don’t. repeatedly you assert, as you do here that arguing against you is the same as support for the accusor. leaves us with two choices, doesn’t it?
by the time I heard about the story, some evidence was already out that the woman was lying.
but on the face:
- I have no problem believing that some drunken frat/sports guys hired a stripper.
- I have no problem believing that in such circumstances some groping behavior could occur. consensual? perhaps. but I also have no problem believing that drunks in a group can act in manners inconsistent with proper behavior. seems to be some data on that. race of the participants and accuser would not factor into it for me. her occupation would factor into it for me to the extent that I think that her occupation would translate into consent for some level of contact in the minds of some folks (doesn’t mean that it’s ok, doesn’t mean that it happened in this case, just that to some folks, if you work as a stripper, obviously a grope or two would also be 'ok[)
- but, when other info came out, the 911 calls, the photos, the DNA, the time line etc, it seems clear to me that the specific horrific accusations are not at all likely.
now can you stop with the bullshit about ‘your detractors’ ? thanks, ever so.
Please stop projecting emotions on to me because you’re only making yourself look bad. It’s clear you guys were simply ecstatic about the idea that white-on-black rape was “vanishingly rare” because otherwise you wouldn’t have attributed disappointment to someone who never even bothered to dispute your “vanishingly rare” intrepretation of the data.
Now, when confronted with data inconsistent with your worldview, you continue to project, revealing simultaneously your own emotional, unobjective responses to the statistics.
Come on, Malacandra. That was too lame to be insulting. I would think you’d pick your last ditch barbs better than this.
Actually, if I can attest to your credibility, I would not take the default position that you were lying or even “probably” lying. This is not the same thing as believing you whole cloth, though.
If you were to say you were late because your alarm clock failed to go off, the same principle would be at work. Your good credibility means that I couldn’t, at a minimum, suspect you are lying without having evidence. But that doesn’t mean I blindly accept your accuse as the God’s honest truth. It’s called “suspending judgement until you have evidence”.
The witness’s reliability, I presume, has been established by this test. This means that the witness is pretty credible as far as recognizing car color accurately. Which means that regardless of how common maroon cars are, the witnesses testimony carries weight.
I fail to see how either one of your scenarios relate to the Duke case, but it was fun playing.
I despoiled your spoiler because if you are trying to eradicate ignorance, it makes no sense to hide your reasoning.
Please explain how you deduce that 75% of cars identified as maroon will be wrong. It’s not clear to me how you reached this answer. If a person is able to correctly ID a maroon car 95% of the time under test conditions, why should that change all of sudden in the real world?
I thought you were supposed to be the statistics expert. I’m not anything of the sort, yet this makes perfect sense to me. One more piece of evidence that you must be a very poor scientist indeed.
Then you should have problems explaining it then, right? Right?
crickets
No problems, I mean.
Well, the witness identifies 49+19=68 cars as maroon. Only 20 cars are actually maroon. 20/68=.291, so 70.59% of cars identified as maroon weren’t maroon.
I’m wondering if this holds, however: is being able to accurately identify 95% of maroon cars as maroon (19 out of 20) the same as inaccuratly picking 5% of red cars as maroon (49 out of 980)? I’m not a stats expert, someone help me out.
Please stop projecting emotions on to me because you’re only making yourself look bad. I personally was not ecstatic about the statistics - I personally found them astonishing - and I challenge you to find anything I posted that implied I was. I do think, however, that you can’t just toss them out when you don’t like them, and then embrace them when they look more likeable.
No, I observe dispassionately that you’re a lot happier with the statistics now they appear to be more supportive of your position. Your worldview was that whites just had to be raping blacks at a much higher rate than stated.
Did you? That’s interesting. Personally I don’t know what I’ve done (other than disagree with you) to deserve being consigned to the “short bus”… I’m not sure what the idiom means, but from context I doubt it’s intended as a compliment. Anyway there’s nothing “last ditch” about it; I’m not all that invested in this argument, but neither am I anywhere near my personal last ditch, especially not when I’m up against someone with reasoning skills as poor as you’re displaying. Of which, more shortly.
But there isn’t any middle ground. Either I am telling the truth about my father being struck by a meteorite or I am not. And as to my credibility, it’s sweet that you would be willing to personally trust me to be telling the truth about something that has never happened before in recorded history (as opposed to, say, previously truthful people departing from the truth once in a while, which happens somewhat more often), but it displays something other than a Vulcan-like preoccupation with logic.
Indeed, given my reputation for truthfulness, you’d be right to trust me. Alarm clocks do malfunction once in a while (more usually thru user error than anything else, I guess), whereas no-one ever gets killed by a meteorite - at least, no-one has ever been known to.
It’s no use painting yourself as a Fair Witness at this stage of the debate. And anyone who thinks that a faulty alarm clock is a no more or less likely excuse than a direct meteorite strike needs his head examined.
No it doesn’t. We’ll go over this below, Doctor.
I’m glad you admit that any failure is on your part. I’m only sorry you didn’t play better, Doctor.
The spoiler was so you could have a go at the problem without seeing the answer in advance. In our schools, that’s much how they used to teach mathematics. It’s not as though you couldn’t see the worked solution at five seconds’ notice - hardly “hiding” my reasoning. And it’s plain that even with sight of the answer, you completely failed to understand what was going on. I’ll not pretend to be surprised, Doctor. Here we go:
Taking it slowly, Doctor:
Of our imagined 1000-car sample, the witness will see 980 red cars. Her impressive 95% accuracy rate will mean that she says “I see a red car” 95% of the time. That’s 931 times. The remaining 5% of the time, she will say “I see a maroon car”. That’s 49 times. OK so far? 95% of 980 is 931? 931 + 49 = 980?
The witness will also see 20 maroon cars. She will say “I see a maroon car” 19 times. She will say “I see a red car” just once. Still OK?
Now pay close attention, Doctor:
On a total of 68 occasions, the witness has said “I see a maroon car”. On only 19 of those occasions was the car actually maroon. The other 49 times, it was a red car mistakenly identified as maroon. That’s 49/68 * 100%, or 72.06% to four significant figures. I skimped a little and approximated; 49:19 is close to 48:16, which is 3:1. Note that I said “nearly 3 - 1”; I didn’t claim that as an accurate figure. The student was expected to work this out on his own, after correctly understanding the method.
If you’re still struggling with this, I must (and I urge the audience to) reconsider the level of statistical competence you have been claiming: this is very easy mathematics, Doctor. Bottom line: witness reliability is important, but the unlikeliness of the event must be factored in. If you redo the above exercise but with red cars outnumbering maroon ones by only three to one, you will find that even a witness who is right a mere nine times out of ten is more likely than not to be right.
We can get around to the relevance of this to the case once you’ve finished claiming that your professional qualifications as an epidemiologist mean that you’re entitled to patronise anyone who doesn’t bow to your vast knowledge of statistics, Doctor.
You’re smarter in this area than you think you are because you are hinting at the same problem that I did.
Being able to correctly ID a red car is not the same thing as being as able to correctly ID a maroon car. In epidemiology, we talk about “sensitivity” and “specificity” when it comes to test reliability. Let’s say the witness’s ability to correctly ID a maroon car is the test. If that test has a high sensitivity, then he/she will be able to diagnosis a maroon car very accurately but at the expense of accurately diagnosing red cars. Meaning, when confronted with a maroon car, the witness might be able to correctly spot it 99.9% of the time. But when confronted with a red one, he/she may misidentify it 95% of the time.
If that test has a high specificity, then the opposite will be in affect. The witness may almost never ID a red car as maroon, but sometimes they may “miss” a maroon car by calling it red.
There is a balance between specificity and sensitivity; no test can excell equally in both areas. But you can not predict one when you only have the other. Malacandra only gave us the specificity (95% ability to tell a red from a maroon). The more important consideration is sensitivity, since that determines the probability that this witness can accurately spot a maroon car.
I look forward to hearing Malacandra and Weirddave explain how they reached their answers. Perhaps they can eradicate some more ignorance. Or perhaps not.
That’s not necessarily true. Please read my response to Dragon Ash. Being able to correctly spot a red car is not necessarily the same thing as being able to correctly ID a maroon one.
During the test it’s perfectly possible that the witness correctly IDed 100% of the maroon cars, but messed up a few times and called the red ones maroon. This is hardly is the end of the world. If there are 20 maroon cars in a sample of 1000 and the witness was right on the money each time those 20 cars showed up, that indicates that this person is pretty damn reliable. And if out of those 980 red cars, they misidentifed 5 of them as maroon, that is small potatoes. * Because it means 995 times, they were right.*
I won’t bother responding to the rest of what you wrote because it based on unwarranted assumptions.
Laughable. 
So you don’t know what “suspending judgement until there is evidence” means? I ask this sincerely.
She’s teaching you something, dude. If you can’t at least respect her for that, then she’s wasting time talking to you.
http://en.wikipedia.org/wiki/Short_bus
an insulting perogative that has no place in this discussion and one that likely should be discontinued in terms of this discussion if someone is attempting to take the high ground—at least in my opinion.