Poll

Brett Kavanaguh: Yay or Nay?

Yay!
Nay!
Who cares? The SCOTUS doesn't matter anyways.

Author Topic: Brett Kavanaguh: Yay or Nay?  (Read 197742 times)

Cache_Stash

  • Bristles
  • ***
  • Posts: 314
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1000 on: October 02, 2018, 04:00:18 PM »
I wont bother to invest my time in looking up actual bits of testimony

You are not alone among Kavanaugh supporters in refusing to look at facts, or listen to testimony, to determine if he is fit for the court.  Even republican senators have joined you in this complete abdication of responsibility to the truth.

And I'm not asking you to refute anything, I was just laying out some of the times we know that he lied under oath about his youthful indiscretions.  How can you say you disagree with these facts if you won't even go look at the evidence?  Is your party loyalty so strong that you will deliberately preserve your ignorance of these facts long enough to maintain plausible deniability through the end of the confirmation process?


I understand that there are lots of Americans who like Brett Kavanaugh's stated position on total executive autonomy and unchecked power, and on abortion.  I have no qualms with those views, though they differ from mine, because you are allowed to hold them and support candidates who share them.  But there are lots of other candidates who share those views who are not sniveling liars, who don't voluntarily perjure themselves to conceal their personal failings, and who have never committed a sexual assault.  Just nominate one of those.

Even if the nomination had to wait a few months, the Senate is approximately 75% likely to stay in republican hands after the midterms.  I think the only possible way for the republicans to lose the Senate is to needlessly push through the Kavanaugh confirmation.  I think that's the one thing that might actually piss off enough voters to turn the senate red. 

Have you seen the polls that suggest 34% of Americans want Kavanaugh confirmed?  Millions of conservative republicans and independents, many of them women and survivors of sexual assault, are watching the republican party push through a truly toxic man to the highest court in the land.  They may vote their displeasure if Mitch goes through with this.

To the bolded:

My time has value. I am old and have had enough experience to know that you and i can listen to the same tv broad ast ast and draw different conclusions. That is likely what would happene abouit each incident of “lying” you point out. That,s why I wont spend any more time on it.

As it is,
I spent the time noted above and came to conclusions noted above.

So you care enough to peruse multiple discussion boards and add your opinion to the discussion here, but not enough to view the material in question?

There is at least one point that is a provable lie, he stated that three witnesses have "refuted" he ever attacked Ford. That's not true, they said they were not aware of the attack. Doesn't that mean it's a lie?

And before you bring it up, Kavanaugh is a judge. He clearly knows the difference between failing to confirm something and refuting it. Even if he misspoke he could have clarified, he did not.

This is the problem with each side.  Refuting or failing to confirm end up with the same result.  She said the witnesses were there and each one denied it.  Yet, you are holding a class in semantics and trying to draw a conclusion of lying based on such.

Ridiculous.

None of the witnesses said they were not there.  They said they did not recall, or were unaware of such an occasion.

I can refute with absolute certainty that I was not in Alaska on December 2, 1986.  I could not tell you that I wasn't in HEB in Austin Texas on that date; I could only say I don't recall being there.

One is refuting, the other is failing to confirming.
Legally, this is a BIG difference, not semantics.

To the bolded:

Witnesses did not corraborate Dr. Ford’s testimony, that is the main point.

What Kavanaugh said about what the witnesses said and how they said it, and how HE said it—are of far less interest to me than to you and others here.

This, I agree with.

Kris

  • Walrus Stache
  • *******
  • Posts: 7306
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1001 on: October 02, 2018, 04:00:50 PM »
This is the problem with each side.  Refuting or failing to confirm end up with the same result.  She said the witnesses were there and each one denied it.  Yet, you are holding a class in semantics and trying to draw a conclusion of lying based on such.

Ridiculous.

What do you mean by this?

You can't refute something that happened if you weren't present.  But, you can say you don't recall something ever happening because you weren't present.  If the witnesses all say they don't recall something happening, yet she says they were there, then either all three are lying or she is lying.

No. I'm the same age as Ford. There are plenty of things that happened in my life that I don't remember now, because they were unremarkable events to me. That doesn't mean I wasn't there. 

Eg: one of my students telling me about a conversation we had that profoundly impacted them, but I don't remember the conversation. That doesn't mean either the student is lying or I'm lying. It means that the conversation was remarkable and memorable to her/him, but not to me.


sol

  • Walrus Stache
  • *******
  • Posts: 8433
  • Age: 47
  • Location: Pacific Northwest
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1002 on: October 02, 2018, 04:09:18 PM »
You can't refute something that happened if you weren't present.  But, you can say you don't recall something ever happening because you weren't present.  If the witnesses all say they don't recall something happening, yet she says they were there, then either all three are lying or she is lying.

You're not following along, Cache.

Some of the witnesses say they don't remember this specific party, but they all agree that they did attend such parties in this time frame, with these people.  Christine Ford certainly remembers it, and Brett even wrote it on his calendar!  This supports Ford's version of events, not Kavanaugh's.  You have mischaracterized this situation, just like Brett has, by saying that someone who says "I don't remember this night" is really saying "this never happened".  Lawyers (and judges) definitely know the difference.

Only three of those people were present in the bedroom during the alleged assault.  One is the victim, she says it happened.  One is the supreme court nominee, he says it didn't.  The third is a famous addict who said he couldn't remember, just like he can't remember other large portions of his life as a result of his lifelong alcohol addiction.  That's not a refutation.  A refutation would be "I clearly remember the night in question, and Dr. Ford is lying."  Instead, he seemingly said "I am not a reliable witness to anything, because I often drank to the point of blacking out."

Then he was interviewed by the FBI, and we don't yet know what he told them.  Hold your horses until Friday!  Would you change your mind if the FBI report says "Mark Judge confirms that he and Brett Kavanaugh pushed Christine Ford into a bedroom while drunk, wrestled with her, and then she escaped."?  Would that do it for you?  If two of the three people present offered the same story, in direct opposition to what Brett is claiming?
« Last Edit: October 02, 2018, 04:25:19 PM by sol »

Cache_Stash

  • Bristles
  • ***
  • Posts: 314
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1003 on: October 02, 2018, 04:16:16 PM »
You can't refute something that happened if you weren't present.  But, you can say you don't recall something ever happening because you weren't present.  If the witnesses all say they don't recall something happening, yet she says they were there, then either all three are lying or she is lying.

You're not following along, Cache.

The three witnesses say they don't remember this specific party, but they all agree that they did attend such parties in this time frame, with these people.  Brett even wrote it on his calendar!  This supports Ford's version of events, not Kavanaugh's.  You have mischaracterized this situation, just like Brett has, by saying that "I don't remember this night" is the same thing as "this never happened".  Lawyers (and judges) definitely know the difference.

Only three of those people were present in the bedroom during the alleged assault.  One is the victim, she says it happened.  One is the supreme court nominee, he says it didn't.  The third is a famous addict who said he couldn't remember, just like he can't remember other large portions of his life as a result of his lifelong alcohol addiction.  That's not a refutation.  A refutation would be "I clearly remember the night in question, and Dr. Ford is lying."

Then he was interviewed by the FBI, and we don't yet know what he told them.  Hold your horses until Friday!  Would you change your mind if the FBI report says "Mark Judge confirms that he and Brett Kavanaugh pushed Christine Ford into a bedroom while drunk, wrestled with her, and then she escaped."?  Would that do it for you?  If two of the three people present offered the same story, in direct opposition to what Brett is claiming?

That would do it.  Corrborating evidence.

shenlong55

  • Pencil Stache
  • ****
  • Posts: 528
  • Age: 41
  • Location: Kentucky
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1004 on: October 02, 2018, 04:23:25 PM »
This is the problem with each side.  Refuting or failing to confirm end up with the same result.  She said the witnesses were there and each one denied it.  Yet, you are holding a class in semantics and trying to draw a conclusion of lying based on such.

Ridiculous.

What do you mean by this?

You can't refute something that happened if you weren't present.  But, you can say you don't recall something ever happening because you weren't present.  If the witnesses all say they don't recall something happening, yet she says they were there, then either all three are lying or she is lying.

ETA: I realized I misunderstood one of your points and my other point was already covered better by other posters.

« Last Edit: October 02, 2018, 04:34:07 PM by shenlong55 »

sol

  • Walrus Stache
  • *******
  • Posts: 8433
  • Age: 47
  • Location: Pacific Northwest
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1005 on: October 02, 2018, 04:29:47 PM »
Would you change your mind if the FBI report says "Mark Judge confirms that he and Brett Kavanaugh pushed Christine Ford into a bedroom while drunk, wrestled with her, and then she escaped."?  Would that do it for you?  If two of the three people present offered the same story, in direct opposition to what Brett is claiming?

That would do it.  Corrborating evidence.

What if he told the FBI something like "I certainly acted in regrettable ways when I was young, but due to my heavy drinking there are many nights I cannot remember, and this may be one of those nights."

What if he told the FBI something like "I remember attending many drunken parties with Brett Kavanaugh, and we sometimes engaged in rough horseplay with girls, but I always assumed it was consensual.  I regret to learn that Christine didn't think so."

Cache_Stash

  • Bristles
  • ***
  • Posts: 314
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1006 on: October 02, 2018, 05:08:09 PM »
Would you change your mind if the FBI report says "Mark Judge confirms that he and Brett Kavanaugh pushed Christine Ford into a bedroom while drunk, wrestled with her, and then she escaped."?  Would that do it for you?  If two of the three people present offered the same story, in direct opposition to what Brett is claiming?

That would do it.  Corrborating evidence.

What if he told the FBI something like "I certainly acted in regrettable ways when I was young, but due to my heavy drinking there are many nights I cannot remember, and this may be one of those nights."

What if he told the FBI something like "I remember attending many drunken parties with Brett Kavanaugh, and we sometimes engaged in rough horseplay with girls, but I always assumed it was consensual.  I regret to learn that Christine didn't think so."



If he can't defend himself or his behavior, then yes.  In your second supposition, I would say there are too many variables that haven't reared their head to address.e

anisotropy

  • Pencil Stache
  • ****
  • Posts: 681
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1007 on: October 02, 2018, 05:24:35 PM »
I voted nay, not because of the allegations but because how emotionally stirred he was and how partisan he seemed.

Contrast that with calm demeanor Gina Haspel displayed when she went through her hearings, I found him too temperamental for my own comfort. I understand being accused of a sexual assault perpetrator (whether true or false) definitely played a part in his reactions, but still, perhaps I am asking for too much.

EDIT: WHAT I DID WAS NOT FACTUAL. DELETED.

« Last Edit: October 06, 2018, 05:39:59 PM by anisotropy »

Glenstache

  • Magnum Stache
  • ******
  • Posts: 3493
  • Age: 94
  • Location: Upper left corner
  • FI(lean) working on the "RE"
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1008 on: October 02, 2018, 05:41:26 PM »
Ugh math. So, the false accusation rate is 5% (just taking your numbers). If there are 100 accusations, then there will be 5 who were wrongfully accused. Your math is using multiplication of probabilities out of order.

Cache_Stash

  • Bristles
  • ***
  • Posts: 314
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1009 on: October 02, 2018, 05:44:23 PM »
I voted nay, not because of the allegations but because how emotionally stirred he was and how partisan he seemed.

Contrast that with calm demeanor Gina Haspel displayed when she went through her hearings, I found him too temperamental for my own comfort. I understand being accused of a sexual assault perpetrator (whether true or false) definitely played a part in his reactions, but still, perhaps I am asking for too much.

Regarding the allegations, aside from waiting for the FBI investigation results, this is something I whipped up and I would like to run it by some of the more mathematically attuned posters:

Suppose you suspect that you are allergic to something, which 1% of the population does, you went to the clinic and did a test. The test turned out to be positive.

The test has a false positive rate of 10% and a false negative rate of 20%. What's the chance of you actually being allergic?

The standard 2x2 table would look like this

                              1% have it                      test yes                          test no
Have allergy                10                                 8                                      2
Don't have                 990                                99                                   891
                                1000                              107                                  893

so, 107 people are test positive but only 8 people are actually allergic; even with a positive test, your chances of being actually allergic is only 7%, despite the test having only 20% false negative rate.

Now, I am going to do the same math but using stats related to male on female rape. I am assuming 5% of the male are rapists. The false accusation rate is known to be 2-10%, so I will take the avg 5% as false positive rate. Only about 1/3 rape are reported, so I am using 2/3 as false negative rate. You can substitute numbers you find reasonable.

The 2x2 table would look like this

                              5%                             accused                          not accused
Rapists                   50                                  17                                   33
Not rapists              950                                48                                  902
                             1000                               65                                  935

so, 65 people are accused but only 17 people are actually rapists; that's 28% chance of a person being actual guilty when accused of rape, despite the false accusation rate being only 2-10%. Obviously with multiple accusers this number would go up, classic Bayesian results. This is why I always refrain from making judgements without hard forensic evidences. Hopefully the FBI will clear things up.

Kavanaugh is most likely going through the grieving process.  He has a sense of loss from what his family has gone through, his honor being trashed, etc...  The first stage of the process is anger.  Let that soak in.

“If you make the same claim to me today,” he said, “it would be scorched-earth. I don’t care if it would cost me $10 million in court for 10 years, you are not taking my name from me, you are not taking my name and reputation from me, I’ve worked too hard for it, I’ve earned it, you can’t just blow me up like that.” - Matt Damon more than a year ago.

Then he goes and spoofs Kavanaugh on SNL.  That is some truly sad hypocritical BS. 


anisotropy

  • Pencil Stache
  • ****
  • Posts: 681
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1010 on: October 02, 2018, 05:50:03 PM »
Not being rude or anything, but I think you are mistaken Glen. As far as I remember, this is how to describe the probabilities of an event, based on our prior knowledge of conditions. As long as there is a false positive rate and a false negative rate, this is how we would calculate the actual probabilities and not just the "intuitive" 5 people wrongly accused out of 100 because of the 5% false accusation rate.

https://www.mathsisfun.com/data/probability-false-negatives-positives.html

Statistics is often counter intuitive (to me at least), recall the monty hall problem.

sol

  • Walrus Stache
  • *******
  • Posts: 8433
  • Age: 47
  • Location: Pacific Northwest
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1011 on: October 02, 2018, 05:52:35 PM »
Why are we talking about arbitrary probabilities?  In addition to not knowing what the correct numbers to use might be, this is an individual case.  It's either a yes or a no.  Probabilities do not apply to individuals.

And this individual wasn't able to credibly deny the allegations against him.  Why worry about math, when you already know the guy is a liar?

anisotropy

  • Pencil Stache
  • ****
  • Posts: 681
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1012 on: October 02, 2018, 05:56:01 PM »
Why are we talking about arbitrary probabilities?  In addition to not knowing what the correct numbers to use might be, this is an individual case.  It's either a yes or a no.  Probabilities do not apply to individuals.

And this individual wasn't able to credibly deny the allegations against him.  Why worry about math, when you already know the guy is a liar?

How do you (or any of us) know what really happened without hard forensic evidence. Statistics is important when we have no information on the case whatsoever other than what he said and what she said and what they said.

To me, Kavanaguh is just a random guy who is being accused, I dont know anything else concrete about the case. Then yes, the probabilities definitely apply, and there is 28% of chance that he is actually guilty based on a single allegation.

Glenstache

  • Magnum Stache
  • ******
  • Posts: 3493
  • Age: 94
  • Location: Upper left corner
  • FI(lean) working on the "RE"
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1013 on: October 02, 2018, 05:58:59 PM »
Kavanaugh is most likely going through the grieving process.  He has a sense of loss from what his family has gone through, his honor being trashed, etc...  The first stage of the process is anger.  Let that soak in.

“If you make the same claim to me today,” he said, “it would be scorched-earth. I don’t care if it would cost me $10 million in court for 10 years, you are not taking my name from me, you are not taking my name and reputation from me, I’ve worked too hard for it, I’ve earned it, you can’t just blow me up like that.” - Matt Damon more than a year ago.

Then he goes and spoofs Kavanaugh on SNL.  That is some truly sad hypocritical BS.

... you may want to put that quote in context to realize what he was actually saying.
https://abcnews.go.com/Entertainment/matt-damon-opens-harvey-weinstein-sexual-harassment-confidentiality/story?id=51792548

What he is saying was taken out of context in a lot of ways, and he could have said things better. But, he also was a bit prescient where he said effectively what Sol has been saying: people will shift to total denial regardless of the truth because those who admit are pilloried regardless of level of contrition and level of offense. He was also pretty clear that none of it was acceptable. The specific quote above was in the context of confidentiality agreements and that he would rather go court than pay someone off to avoid a public fight.

Glenstache

  • Magnum Stache
  • ******
  • Posts: 3493
  • Age: 94
  • Location: Upper left corner
  • FI(lean) working on the "RE"
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1014 on: October 02, 2018, 06:06:35 PM »
Not being rude or anything, but I think you are mistaken Glen. As far as I remember, this is how to describe the probabilities of an event, based on our prior knowledge of conditions. As long as there is a false positive rate and a false negative rate, this is how we would calculate the actual probabilities and not just the "intuitive" 5 people wrongly accused out of 100 because of the 5% false accusation rate.

https://www.mathsisfun.com/data/probability-false-negatives-positives.html

Statistics is often counter intuitive (to me at least), recall the monty hall problem.

in your math, you have a total of 65 accusations of which 48 are not rapists, eg false accusations. That is a 73% false accusation rate. The false accusation rate (which you apply at 5%) is the number of accusations that are in the category "not rapists".Yes. Math is fun.

Lack of reporting is not a false negative rate. Your analogy to allergy testing is incorrectly formulated.

And yes, as Sol pointed out the statistics fall away once you get to an individual boolean case.

Glenstache

  • Magnum Stache
  • ******
  • Posts: 3493
  • Age: 94
  • Location: Upper left corner
  • FI(lean) working on the "RE"
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1015 on: October 02, 2018, 06:12:11 PM »
And in the word of Bart O'Kavanaugh himself, he certainly didn't seem like a choir boy who had no intention of fooling around with girls. To the contrary, hooking up appears to have been a much wanted "wish" and that the group was an "obnoxious" group of drunks.
https://www.nytimes.com/2018/10/02/us/brett-kavanaugh-georgetown-prep.html?smid=fb-nytimes&smtyp=cur



DarkandStormy

  • Handlebar Stache
  • *****
  • Posts: 1498
  • Age: 34
  • Location: Midwest, USA
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1016 on: October 02, 2018, 06:20:33 PM »
He told Senator Leahy he wasn't Bart O'Kavanaugh.

Glenstache

  • Magnum Stache
  • ******
  • Posts: 3493
  • Age: 94
  • Location: Upper left corner
  • FI(lean) working on the "RE"
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1017 on: October 02, 2018, 06:27:27 PM »
He told Senator Leahy he wasn't Bart O'Kavanaugh.
Actually, he just purposefully dodges the question by throwing up chaff.

Quote
LEAHY: Judge Kavanaugh, I’ve heard your — your line (ph) and you state it over and over again. And I have that well in mind. But let me ask you this. He authored a book titled, “Wasted: Tales of a Genx Drunk.” He references a Barthold (ph) Kavanaugh vomiting on someone’s car during Beach Week and then passing out. Is that you that he’s talking about?

KAVANAUGH: Senator, Mark Judge was…

LEAHY: To your knowledge, is that you that he’s talking about?

KAVANAUGH: I’ll explain it if you let me.

LEAHY: Proceed, please.

KAVANAUGH: Mark Judge was a friend of ours in high school who developed a very serious drinking problem, an addiction problem that lasted decades and was very difficult for him to escape from.

And he nearly died. And then developed — then he had leukemia as well, on top of it.

Now, as part of his therapy — or part of his coming to grips with sobriety, he wrote a book that is a fictionalized book and an account….

(CROSSTALK)

KAVANAUGH: I think he picked out names of friends of ours to throw them in as kind of close to what — for characters in the book. So, you know, we can sit here…

LEAHY: So you don’t know — you don’t know whether that’s you or not?

KAVANAUGH: … we can sit here and you (ph) like (ph), make — make fun of some guy who has an addiction.

LEAHY: I’m not making…

(CROSSTALK)

KAVANAUGH: I don’t think that really makes — is really good…

LEAHY: … Judge Kavanaugh, I’m trying to get a straight answer from you under oath. Are you Bart (ph) Kavanaugh that he’s referring to, yes or no? That’s it (ph)…

KAVANAUGH: You’d have to ask him.

MasterStache

  • Magnum Stache
  • ******
  • Posts: 2907
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1018 on: October 02, 2018, 06:50:22 PM »
Kavanaugh is most likely going through the grieving process.  He has a sense of loss from what his family has gone through, his honor being trashed, etc...  The first stage of the process is anger.  Let that soak in.

Actually the first step of the grieving process is denial. Perhaps you are grieving? (Feel free to fact check that). 

sol

  • Walrus Stache
  • *******
  • Posts: 8433
  • Age: 47
  • Location: Pacific Northwest
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1019 on: October 02, 2018, 06:52:15 PM »
How do you (or any of us) know what really happened without hard forensic evidence. Statistics is important when we have no information on the case whatsoever other than what he said and what she said and what they said.

But we do!  We have SO much more. 

Quote
To me, Kavanaguh is just a random guy who is being accused, I dont know anything else concrete about the case. Then yes, the probabilities definitely apply, and there is 28% of chance that he is actually guilty based on a single allegation.

How about if there are three allegations?

Also, I feel it only right to point that your math is deliberately misleading.  You can't put an a priori 1% infection rate under a test with a 10% false positive rate and then claim you "only have a 7% chance" of being infected.  A single positive test gives you a 700% increased chance from the random 1% chance of being infected.  The 7% is artificially low because the false positive rate is artificially high relative to the specified infection rate. 

But I'm presuming you already knew that, if you're the kind of person who would pose such an example.  Which leads me to believe it was done with bad intentions. 

In such instances, with high false positive rates, what would a doctor do?  He'd order a repeat test.  Kavanaugh now has three accusers.  Go ahead, run the numbers on three positive tests on the same patient, and then report back with the likelihood of infection.  I'll wait.

Here's an inverse test for you that I think is more relevant.  If you have a list of 25 nominees for the supreme court and only one of them stands credibly accused of sexual assault, what are the odds that congressional republicans go the mattresses for that one specific guy?

It still won't matter.  I think Brett will be confirmed on Friday, no matter what the FBI investigation finds.  Remember boys and girls: sexual assault isn't a bug, it's a feature.

anisotropy

  • Pencil Stache
  • ****
  • Posts: 681
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1020 on: October 02, 2018, 06:54:17 PM »

in your math, you have a total of 65 accusations of which 48 are not rapists, eg false accusations. That is a 73% false accusation rate. The false accusation rate (which you apply at 5%) is the number of accusations that are in the category "not rapists".


Exactly, the false accusation rate only applies to the "non rapists". Thus, when we combine both real rapists and non rapists who are accused, we arrive at a grand total of 73% false accusation rate. This back of envelope calculation shows the much publicized 2-10% false accusation rate is very misleading when we actually take a look deeper.


Lack of reporting is not a false negative rate. Your analogy to allergy testing is incorrectly formulated.

And yes, as Sol pointed out the statistics fall away once you get to an individual boolean case.



You are more than welcome to suggest an alternative false negative rate with good reasoning. I used what I thought was suitable (rapists not being accused).

Statistics are meaningless when we get to the individual case ONLY if we have additional information that changes the prior conditions. See the allergy testing analogy, it is as individual as it gets (you being allergic) yet the calculation still applies. So no, I disagree on this completely.

partgypsy

  • Walrus Stache
  • *******
  • Posts: 5196
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1021 on: October 02, 2018, 06:59:06 PM »
He didn't say he wasn't Bart OKavanagh, he just refused to answer the question.

I think something needs to be said, is that false rape claims fall into certain patterns of who reports them and for what reason. Ford's accusation doesn't fall into any of those patterns.
In turn Kavanaugh displayed a number of risk factors for men in the community that self-report non consensual sex.
So, it's not just a numbers game; it's using what information we have to determine what is a credible versus non-credible account. 

https://www.vox.com/first-person/2018/9/18/17874504/kavanaugh-assault-allegation-christine-blasey-ford


If you don't think that Kavanaugh could have possibly done it, men who commit sexual assault are commonly out there in population and often do not consider themselves as doing anything wrong. Some do it only 1 or two times, and stop. Others become repeat offenders. The vast majority, never accused or prosecuted.

https://www.nytimes.com/2017/10/30/health/men-rape-sexual-assault.html

"...Dosage of what? Certain factors — researchers call them “risk factors” while acknowledging that these men are nonetheless responsible for their actions — have an outsize presence among those who commit sexual assaults.

Heavy drinking, perceived pressure to have sex, a belief in “rape myths” — such as the idea that no means yes — are all risk factors among men who have committed sexual assault. A peer group that uses hostile language to describe women is another one."

« Last Edit: October 02, 2018, 07:20:31 PM by partgypsy »

shenlong55

  • Pencil Stache
  • ****
  • Posts: 528
  • Age: 41
  • Location: Kentucky
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1022 on: October 02, 2018, 07:06:40 PM »

in your math, you have a total of 65 accusations of which 48 are not rapists, eg false accusations. That is a 73% false accusation rate. The false accusation rate (which you apply at 5%) is the number of accusations that are in the category "not rapists".


Exactly, the false accusation rate only applies to the "non rapists". Thus, when we combine both real rapists and non rapists who are accused, we arrive at a grand total of 73% false accusation rate. This back of envelope calculation shows the much publicized 2-10% false accusation rate is very misleading when we actually take a look deeper.

No, I'm pretty sure Glenstache is right.  The false accusation rate is the percentage of accusations where an investigation establishes that no crime was committed or attempted.  Not the percentage of "not rapists" that are accused as you seem to be using.  I don't think your analogy works.

anisotropy

  • Pencil Stache
  • ****
  • Posts: 681
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1023 on: October 02, 2018, 07:15:32 PM »

How about if there are three allegations?

Also, I feel it only right to point that your math is deliberately misleading.  You can't put an a priori 1% infection rate under a test with a 10% false positive rate and then claim you "only have a 7% chance" of being infected.  A single positive test gives you a 700% increased chance from the random 1% chance of being infected.  The 7% is artificially low because the false positive rate is artificially high relative to the specified infection rate. 

But I'm presuming you already knew that, if you're the kind of person who would pose such an example.  Which leads me to believe it was done with bad intentions. 

In such instances, with high false positive rates, what would a doctor do?  He'd order a repeat test.  Kavanaugh now has three accusers.  Go ahead, run the numbers on three positive tests on the same patient, and then report back with the likelihood of infection.  I'll wait.


Hi what? bad intentions? Did you not take any statistics in college? That's how the stat questions are usually framed? You can see the link I provided earlier to Glen.

How else would you frame the question: Given 1% of population are allergic, test has 10% false positive and 20% false negative, what's the chance you are actually allergic with a positive result. Stat is stat, don't bring ideology (bad intentions) into everything please.

And yes, I did mention multiple allegations would bring the numbers up quickly. We need to keep in mind the nature of exposing oneself (to Ramirez) is different from rape and should be removed from this calculation. If we include Swetnick's accusation and run the calculation twice we would be around 70%.

I think the FBI investigation result matters, even if he gets confirmed. The result may provide grounds for impeachment, he could be the second one!

sol

  • Walrus Stache
  • *******
  • Posts: 8433
  • Age: 47
  • Location: Pacific Northwest
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1024 on: October 02, 2018, 07:17:27 PM »
You are more than welcome to suggest an alternative false negative rate with good reasoning. I used what I thought was suitable (rapists not being accused).

In the normal parlance of sexual assault allegations, a "false allegation" is one in which the accused is proven innocent.  Don't conflate that with a "false positive" in your medical exam example.  They are very different.

In your example, a better "false negative" would be someone who was evaluated for a history of sexual assault and found to be free of all credible charges, and yet was still an abuser.  Jerry Sandusky was a false positive for years, repeatedly exonerated of molesting children.  Whoops!

Quote
See the allergy testing analogy, it is as individual as it gets (you being allergic) yet the calculation still applies. So no, I disagree on this completely.

As I've already tried to point out, it is deliberately misleading.  We're not talking about a population of people who all take a test to determine if they are abusers.  Most of the population in this case is untested.  We're only talking about the population of people who stand accused of sexual assault (in your example, people who have already tested positive at least once), and those people are uniformly more likely to have committed sexual assault than the population at large and Glen's formulation is more appropriate.  If you just wanted to pursue this purely mathematical analysis of guilt, then multiple allegations against the same person would increase the likelihood that he's an abuser.  I think we can all see the problems with that idea.

anisotropy

  • Pencil Stache
  • ****
  • Posts: 681
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1025 on: October 02, 2018, 07:23:02 PM »

in your math, you have a total of 65 accusations of which 48 are not rapists, eg false accusations. That is a 73% false accusation rate. The false accusation rate (which you apply at 5%) is the number of accusations that are in the category "not rapists".


Exactly, the false accusation rate only applies to the "non rapists". Thus, when we combine both real rapists and non rapists who are accused, we arrive at a grand total of 73% false accusation rate. This back of envelope calculation shows the much publicized 2-10% false accusation rate is very misleading when we actually take a look deeper.

No, I'm pretty sure Glenstache is right.  The false accusation rate is the percentage of accusations where an investigation establishes that no crime was committed or attempted.  Not the percentage of "not rapists" that are accused as you seem to be using.  I don't think your analogy works.

I see where the confusion is, I believe the 2-10% is the false positive rate and should be used as an input, you guys believe its the final result.

If we assume this 2-10% false accusation rate is the final result and not the mere "false positive" rate, we would arrive at some  troubling result. Either the number of rapists in the general population would need to be incredibly large (approaching 60%), or the false positive rate need to be super tiny, ie, sub 0.001%. I find these assumptions unrealistic.


partgypsy

  • Walrus Stache
  • *******
  • Posts: 5196
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1026 on: October 02, 2018, 07:25:57 PM »
yes, the formulation is assuming that accusations are made randomly. Yet common sense dictates that people who actually committed sexual assault will be accused at a higher rate than the population at large.

anisotropy

  • Pencil Stache
  • ****
  • Posts: 681
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1027 on: October 02, 2018, 07:36:30 PM »
Sol, I disagree on your "false allegation" definition, if you want to use

false negative" would be someone who was evaluated for a history of sexual assault and found to be free of all credible charges, and yet was still an abuser.

then the population breakdown would need to be redefined. Because by your definition then rapists are tried and proven guilty, instead of simply having committed an act with or without being accused. Anyone know whats the % of male rapist in the male populationÉ

its not deliberately misleading, this is how stat scenarios are framed in academics, its just that in daily life we (and most people) are not used to such way of thinking. I have said multiple times yes more allegation would raise the likelihood and I have shown you the result, classic Bayesian, perhaps you missed it somehow.

Another poster: common sense dictates that people who actually committed sexual assault will be accused at a higher rate than the population at large.


Yup, hence the low false positive rate.
« Last Edit: October 02, 2018, 07:38:15 PM by anisotropy »

sol

  • Walrus Stache
  • *******
  • Posts: 8433
  • Age: 47
  • Location: Pacific Northwest
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1028 on: October 02, 2018, 07:43:10 PM »
Hi what? bad intentions? Did you not take any statistics in college?

Lol.  Yes.  Also in graduate school!  Also as part of my job (which I just quit because I got tired of arguing about statistics).

Quote
How else would you frame the question: Given 1% of population are allergic, test has 10% false positive and 20% false negative, what's the chance you are actually allergic with a positive result. Stat is stat, don't bring ideology (bad intentions) into everything please.

The bad intentions aren't in how you state the problem, they're in the problem you've chosen.  You don't get meaningful results from a test with a 10% false positive rate if the true incidence is only 1%.  MOST of the positives will be false in that scenario, as you've chosen to highlight, because you've structured a problem with almost no true sick people, and a really shitty test.  This has no bearing at all on the veracity of sexual assault allegations, unless you a priori believe that the incidence of sexual assault is much lower than the incidence of false accusations of sexual assault.

But the math is a distraction here, anisotropy.  You've already told us you don't think Brett shouldn't be confirmed.  But there are multiple people in this thread who think the whole series of allegations are a Chinese Hoax, and your poorly chosen example lends them credence.  It teaches them to discount the allegations as "probably false" when that is entirely unwarranted in this case.  Your mathematical approach is equally valid for Ford as it is for Kavanaugh, in the absence of other information, so it's not really very helpful.  What is helpful is looking at the details of the two accounts and determining which seems more reliable.  Brett Kavanaugh is clearly not a reliable witness, on a whole laundry list of topics, so his denials don't carry much weight with me.  Why should I believe a man who lies about easily disproven stuff, dodges questions, refuses to answer, changes the subject, opposes an investigation to find the truth, and and then screams about a vast Clinton conspiracy?  Dude is a single engine train to crazy town, and I think anyone who watches his testimony can see it.
« Last Edit: October 02, 2018, 07:49:00 PM by sol »

iris lily

  • Walrus Stache
  • *******
  • Posts: 5657
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1029 on: October 02, 2018, 08:43:21 PM »
Iris-What are the rules to Devil's Triangle?

Whitehouse asked what "Devil's triangle" referred to, and Kavanaugh replied that it was a drinking game with three glasses in a triangle played in a similar fashion to quarters.

Do you think Brett Kavanaugh is lying about a Devil's Triangle or is the rest of the pre-2018 internet lying?  There are no pre-September 2018 references to a Devil's Triangle drinking game.  However, there are several references to a sexual practice.

How about what a Renate Alumnus is?

"That yearbook reference was clumsily intended to show affection, and that she was one of us…It was not related to sex."

If several member of the football team had printed in the yearbook that they were "Iris Lily Alumnus", would you have been flattered at this sign of respect that is not about sex?  How about if several members of the football team had written that about your daughter?  And then, decades later, testified under oath about how affectionate and respectful their gesture was?   Would you have considered that a lie or just the gosh-darn-honest truth?

I dont even know what you are talking about.

anisotropy

  • Pencil Stache
  • ****
  • Posts: 681
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1030 on: October 02, 2018, 08:49:02 PM »
Sol,

Congratz on leaving a job that brought you headaches. I think I can relate to that, I also did stat (for the purpose of reservoir characterization) for a living and it was at times incredibly difficult trying to explain things to non tech (or even non stat) people.

I am sorry if you feel my post is providing ammunition to the group that are pro-Kavanaguh. Yet I believe math is math, ideology should not be part of it, so I choose to present things as they are. The back of envelope calculation is meant to demonstrate (and perhaps counter) that commonly touted "believe all women no matter what because false accusation rate is only 2-10%" is very misleading when we actually work things out and the likelihood of being actually guilty off one single allegation is low. Instead of 90%, its 26%.

Like you and I both noted, however, given multiple allegations (I used 2), the chance of him being guilty is now ~70% based on calculation alone. Note how even with two allegations, its still lower than 90%.

For what it's worth, most of medical conditions (or any low chance occurances) occur less than 1% of the time, and this is what people have to deal with using stats so I think its entirely valid. I merely applied the same method to evaluating the odds of this guy being actually guilty given the known stats, like I said, math is math (objective truth perhaps), it should not be partisan.

At the end of the day, stat is just stat, once the FBI releases the result, we can all make better decisions from there.

Turkey Leg

  • Bristles
  • ***
  • Posts: 258
  • Location: US
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1031 on: October 02, 2018, 08:53:49 PM »
Iris-What are the rules to Devil's Triangle?

Whitehouse asked what "Devil's triangle" referred to, and Kavanaugh replied that it was a drinking game with three glasses in a triangle played in a similar fashion to quarters.

Do you think Brett Kavanaugh is lying about a Devil's Triangle or is the rest of the pre-2018 internet lying?  There are no pre-September 2018 references to a Devil's Triangle drinking game.  However, there are several references to a sexual practice.

How about what a Renate Alumnus is?

"That yearbook reference was clumsily intended to show affection, and that she was one of us…It was not related to sex."

If several member of the football team had printed in the yearbook that they were "Iris Lily Alumnus", would you have been flattered at this sign of respect that is not about sex?  How about if several members of the football team had written that about your daughter?  And then, decades later, testified under oath about how affectionate and respectful their gesture was?   Would you have considered that a lie or just the gosh-darn-honest truth?

I dont even know what you are talking about.
Then your opinion about Kavanaugh is uninformed and should be disregarded.

Davnasty

  • Magnum Stache
  • ******
  • Posts: 2793
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1032 on: October 02, 2018, 09:17:16 PM »

in your math, you have a total of 65 accusations of which 48 are not rapists, eg false accusations. That is a 73% false accusation rate. The false accusation rate (which you apply at 5%) is the number of accusations that are in the category "not rapists".


Exactly, the false accusation rate only applies to the "non rapists". Thus, when we combine both real rapists and non rapists who are accused, we arrive at a grand total of 73% false accusation rate. This back of envelope calculation shows the much publicized 2-10% false accusation rate is very misleading when we actually take a look deeper.

No, I'm pretty sure Glenstache is right.  The false accusation rate is the percentage of accusations where an investigation establishes that no crime was committed or attempted.  Not the percentage of "not rapists" that are accused as you seem to be using.  I don't think your analogy works.

I see where the confusion is, I believe the 2-10% is the false positive rate and should be used as an input, you guys believe its the final result.

If we assume this 2-10% false accusation rate is the final result and not the mere "false positive" rate, we would arrive at some  troubling result. Either the number of rapists in the general population would need to be incredibly large (approaching 60%), or the false positive rate need to be super tiny, ie, sub 0.001%. I find these assumptions unrealistic.

I don't think you do. 2-10% is the final result, as in, 2-10% of accusations are concluded false. The actual % of false accusations is almost certainly higher. For reference, only ~20% of accusations are concluded positive and that's with ~80% guilty pleas.

You could add to the equation lots of other variables too:
-This is a highly politicized figure at a very divisive time. I would suggest this makes the chance of a false accusation much higher than normal.
-The accuser is well educated, mentally stable, and has no potential personal gain (excluding the previous point). This makes a false accusation less likely.
-The numerous lies Kavanaugh has told to protect his character takes away from his credibility. Boofing is not a term for flatulence. C'mon Brett, we have the internet.
-The body language displayed at the hearing. Sure, anyone falsely accused would rightfully be upset, but the way he squirmed when he just couldn't dodge the question any longer suggests he's lying. The way he refused to answer questions or at other times gave more information than was requested. People often do this when they lie. In contrast, Ford was calm and spoke directly.

What I'm getting at here is that applying general statistics to this case is all but meaningless. And you're doing it wrong.

former player

  • Walrus Stache
  • *******
  • Posts: 8724
  • Location: Avalon
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1033 on: October 02, 2018, 09:47:24 PM »
I voted nay, not because of the allegations but because how emotionally stirred he was and how partisan he seemed.

Contrast that with calm demeanor Gina Haspel displayed when she went through her hearings, I found him too temperamental for my own comfort. I understand being accused of a sexual assault perpetrator (whether true or false) definitely played a part in his reactions, but still, perhaps I am asking for too much.

Regarding the allegations, aside from waiting for the FBI investigation results, this is something I whipped up and I would like to run it by some of the more mathematically attuned posters:

Suppose you suspect that you are allergic to something, which 1% of the population does, you went to the clinic and did a test. The test turned out to be positive.

The test has a false positive rate of 10% and a false negative rate of 20%. What's the chance of you actually being allergic?

The standard 2x2 table would look like this

                              1% have it                      test yes                          test no
Have allergy                10                                 8                                      2
Don't have                 990                                99                                   891
                                1000                              107                                  893

so, 107 people are test positive but only 8 people are actually allergic; even with a positive test, your chances of being actually allergic is only 7%, despite the test having only 20% false negative rate.

Now, I am going to do the same math but using stats related to male on female rape. I am assuming 5% of the male are rapists. The false accusation rate is known to be 2-10%, so I will take the avg 5% as false positive rate. Only about 1/3 rape are reported, so I am using 2/3 as false negative rate. You can substitute numbers you find reasonable.

The 2x2 table would look like this

                              5%                             accused                          not accused
Rapists                   50                                  17                                   33
Not rapists              950                                48                                  902
                             1000                               65                                  935

so, 65 people are accused but only 17 people are actually rapists; that's 28% chance of a person being actual guilty when accused of rape, despite the false accusation rate being only 2-10%. Obviously with multiple accusers this number would go up, classic Bayesian results. This is why I always refrain from making judgements without hard forensic evidences. Hopefully the FBI will clear things up.


Your supposed stats suffer from the usual problem of "garbage in, garbage out".  Firstly, why are you using an example of rape when the accusation is attempted rape and sexual assault?  Why are you assuming 5% of males as rapists?  Why are you assuming 5% of false accusations, and why are you basing this on a range of 2% to 8% which is not "known" but is itself an uninformed guess?


And why are you specifying the need for "hard forensic evidence"?  That's a scientists approach, not a law courts approach or a job interview approach, and is both inappropriate and I suspect a misunderstanding of "forensic" as meaning "scientific".  In either a court or a job interview it is entirely appropriate to take into account witness demeanour: this is why witnesses in court are required to give evidence in person - contrast Ford's demeanour with Kavanaugh's).  And there is hard evidence in Kavanaugh's calendar, which proves the existence of a party as mentioned by Ford as the occasion of the assault: location and place (Bethesda, house) people present (Kavanaugh, PJ, Judge, and another), and activity (drinking beer).    Why don't you apply your stats to the chances of Ford imagining that party out of nothing and then getting all those details right?  And there is hard evidence in Kavanaugh's lying, evasions and inconsistencies.

anisotropy

  • Pencil Stache
  • ****
  • Posts: 681
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1034 on: October 02, 2018, 10:05:00 PM »
Dabnasty,

Quote
I don't think you do. 2-10% is the final result, as in, 2-10% of accusations are concluded false. The actual % of false accusations is almost certainly higher. For reference, only ~20% of accusations are concluded positive and that's with ~80% guilty pleas.

Let's use your numbers and work backwards and see what we get in terms of population breakdown. Based on your replies, we will assume the % of false positive is 20%. For simplicity sake we will keep the false negative (rapists not accused) to be the same number I used, ie, 2/3

                                                            accused                          not accused
Rapists                   A                                 1/3A                                   2/3A
Not rapists              B                                0.2B                                    0.8B
                                                           
we need to find values of A and B so that 1/3A / (0.2B+1/3A) = ~95%, so that it satisfy 2-10% being the final result (again, I took the avg here to be 5%)

We get A/B = 11.5, in English, this means in the population more than 90% of the men are rapists. Here's the breakdown for population of 1000.

                            A/B=11.5                             accused                          not accused
Rapists                   920                                     307                                613
Not rapists              80                                        16                                  64
                             1000                                   323                                 679

16/323= 0.05

This is what we get by using 2-10% being the final result, which is clearly unrealistic and wrong, 92% of men are rapists? Like I said, if we assume 2-10% being the final result, we will get unrealistically large numbers of rapists in the population, or super low false-accusation rate.

So no, my calculation was right based on the numbers I used. The 2-10% is the false positive rate and should be treated as an input.  Given a single accuser, the likelihood of the guy being guilty is low, its only with multiple accusations the likelihood increases as I have demonstrated earlier, with 2, its at 70%.

sol

  • Walrus Stache
  • *******
  • Posts: 8433
  • Age: 47
  • Location: Pacific Northwest
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1035 on: October 02, 2018, 10:14:15 PM »
This is what we get by using 2-10% being the final result, which is clearly unrealistic and wrong, 92% of men are rapists? Like I said, if we assume 2-10% being the final result, we will get unrealistically large numbers of rapists in the population, or super low false-accusation rate.

You're still assuming that every man has been accused. 

The fundamental problem with your example is that you have a high false positive rate compared to the rate of actual offenses.  If you make those assumptions, then yes of course the majority of accusations will be false.  You've designed the problem that way.  That doesn't make it instructive.

And as many people have pointed out now, it's totally irrelevant in this case because we literally have testimony from multiple people, calendars, text messages, books, etc that dispute one version of events and not the other.  We are no longer considering probabilities of an arbitrary candidate with a single accusation who exists in a vacuum of no other information, we're considering Brett Kavanaugh, who is apparently a really bad liar with a history of getting blackout drunk and behaving in sexually inappropriate ways.  As determined by real evidence about this real person.  Not probabilistically.

anisotropy

  • Pencil Stache
  • ****
  • Posts: 681
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1036 on: October 02, 2018, 10:22:01 PM »
I voted nay, not because of the allegations but because how emotionally stirred he was and how partisan he seemed.

Contrast that with calm demeanor Gina Haspel displayed when she went through her hearings, I found him too temperamental for my own comfort. I understand being accused of a sexual assault perpetrator (whether true or false) definitely played a part in his reactions, but still, perhaps I am asking for too much.

Regarding the allegations, aside from waiting for the FBI investigation results, this is something I whipped up and I would like to run it by some of the more mathematically attuned posters:

Suppose you suspect that you are allergic to something, which 1% of the population does, you went to the clinic and did a test. The test turned out to be positive.

The test has a false positive rate of 10% and a false negative rate of 20%. What's the chance of you actually being allergic?

The standard 2x2 table would look like this

                              1% have it                      test yes                          test no
Have allergy                10                                 8                                      2
Don't have                 990                                99                                   891
                                1000                              107                                  893

so, 107 people are test positive but only 8 people are actually allergic; even with a positive test, your chances of being actually allergic is only 7%, despite the test having only 20% false negative rate.

Now, I am going to do the same math but using stats related to male on female rape. I am assuming 5% of the male are rapists. The false accusation rate is known to be 2-10%, so I will take the avg 5% as false positive rate. Only about 1/3 rape are reported, so I am using 2/3 as false negative rate. You can substitute numbers you find reasonable.

The 2x2 table would look like this

                              5%                             accused                          not accused
Rapists                   50                                  17                                   33
Not rapists              950                                48                                  902
                             1000                               65                                  935

so, 65 people are accused but only 17 people are actually rapists; that's 28% chance of a person being actual guilty when accused of rape, despite the false accusation rate being only 2-10%. Obviously with multiple accusers this number would go up, classic Bayesian results. This is why I always refrain from making judgements without hard forensic evidences. Hopefully the FBI will clear things up.

Your supposed stats suffer from the usual problem of "garbage in, garbage out".  Firstly, why are you using an example of rape when the accusation is attempted rape and sexual assault?  Why are you assuming 5% of males as rapists?  Why are you assuming 5% of false accusations, and why are you basing this on a range of 2% to 8% which is not "known" but is itself an uninformed guess?


Are you suggesting I should treat Dr. Ford's accusation separately from swetnicks' because it's "not" rape?

I am usually pretty polite and constructive in a serious conversation, it is only when some random person shows up wholly uninformed I get annoyed. You must have never had a disagreement with me here, otherwise you would know I usually use pretty well-sourced data.

Assuming 5% are rapists: "6.4% of men openly admitted of committing the strictest possible definition of rape", I rounded it so it's easier to run calculations.
https://davidlisak.com/wp-content/uploads/pdf/RepeatRapeinUndetectedRapists.pdf

Assuming 5% false accusation: Most experts agree it's between 2-10%. "A multi-site study of eight U.S. communities including 2,059 cases of sexual assault found a 7.1% of false reports (Lonsway, Archambault, & Lisak, 2009)." Again, I rounded it so it's easier to run calc.
https://www.nsvrc.org/sites/default/files/Publications_NSVRC_Overview_False-Reporting.pdf
http://www.scirp.org/(S(i43dyn45teexjx455qlt3d2q))/reference/ReferencesPapers.aspx?ReferenceID=1238871

You are more than welcome to provide your own numbers provided they are backed by research.

anisotropy

  • Pencil Stache
  • ****
  • Posts: 681
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1037 on: October 02, 2018, 10:42:06 PM »
This is what we get by using 2-10% being the final result, which is clearly unrealistic and wrong, 92% of men are rapists? Like I said, if we assume 2-10% being the final result, we will get unrealistically large numbers of rapists in the population, or super low false-accusation rate.

You're still assuming that every man has been accused. 

The fundamental problem with your example is that you have a high false positive rate compared to the rate of actual offenses.  If you make those assumptions, then yes of course the majority of accusations will be false.  You've designed the problem that way.  That doesn't make it instructive.

And as many people have pointed out now, it's totally irrelevant in this case because we literally have testimony from multiple people, calendars, text messages, books, etc that dispute one version of events and not the other.  We are no longer considering probabilities of an arbitrary candidate with a single accusation who exists in a vacuum of no other information, we're considering Brett Kavanaugh, who is apparently a really bad liar with a history of getting blackout drunk and behaving in sexually inappropriate ways.  As determined by real evidence about this real person.  Not probabilistically.

I used the false positive rate suggested by Dabnasty, you can use something low for FP like 1% and the population breakdown would still look horrendous.

Sol, once again I believe stats has a place here because we don't know anything other than what they said and he's just a random guy as far as I know. I was in the audience when Henry Lee (forensic scientist) gave a vivid description of what beyond reasonable doubt means, and so far we don't have enough "evidence".

It is very common for there to be inconsistencies in people's stories. In fact, it's so common that from my perspective being a scientist/engineer, I find most people (myself included) are unreliable narrators under pressure, and I prefer verifiable physical evidence. Perhaps you are good at spotting lies and judging people's characters in general, most people, including myself are not so fortunate.

https://fivethirtyeight.com/features/why-humans-are-bad-at-spotting-lies/?ex_cid=story-twitter

My view is that I am not willing to pass judgement regarding the allegations (chance of being guilt now 70% given 2 allegations) until the FBI investigation concludes. I don't think he is a good candidate for other reasons I mentioned, but I can not condemn him for something he's yet to be proven guilty of.

ice_beard

  • Bristles
  • ***
  • Posts: 251
  • Location: East Bay, CA
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1038 on: October 02, 2018, 11:07:32 PM »
His behavior last week was not judicial.  He has already shown that he is politically biased.  Oh and there is the sexual assault and alcohol abuse stuff too.   
I keep wondering, why him?  Why not one of the other dozens of judges the Heritage Foundation has approved??  Oh yeah, because he has this singular view that sitting presidents cannot be charged with crimes.  Who would be interested in such a whack ass not based on existing standards of law position?? 

This administration doesn't even try to hide their sliminess, it's shameful. 

laserlady

  • 5 O'Clock Shadow
  • *
  • Posts: 60
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1039 on: October 02, 2018, 11:12:02 PM »
You can't apply the false accusation rate as a false negative because the false negative needs to be based on the population as a whole, while the false accusation rate is based on a subset of the population.

In your allergy analogy, 107 people test positive for allergies, but only 8 have allergies.  Thus, if having allergies were a crime and people could be convicted of having allergies based on an allergy test that is given to everyone in the population, then ~93% of allergy convictions would be wrongful convictions, since only 7% of people who test positive for allergies would turn out to actually have allergies. 

In the rape context, on the other hand, we don't need to solve for the false conviction/accusation number because we already have an estimate.  You estimate that approximately 5% of accusations are false.  Thus, in a population of 107 accused rapists, we can estimate that approximately 102 would be actual rapists -- 95% of people who are accused of rape turn out to actually be rapists. 

You're trying to solve for X in an equation when we already know X, and that's making your numbers wrong.  Using your assumptions, we'll say that 5% of the male population are rapists, that only 33% of rapes are reported, and that 5% of rape reports are false.  That should lead to these figures:


                              5%                             accused                          not accused
Rapists                   50                                  17                                   33
Not rapists              950                                1                                    949
                             1000                               18                                  982


Of our 18 men accused of rape, 1 is a false report, which is consistent with your estimate of a 5% false report rate for those who are accused of rape.

anisotropy

  • Pencil Stache
  • ****
  • Posts: 681
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1040 on: October 03, 2018, 12:13:08 AM »
You can't apply the false accusation rate as a false negative because the false negative needs to be based on the population as a whole, while the false accusation rate is based on a subset of the population.


I didn't apply the false accusation rate as a false negative, I used non-reported rape as false-negative, which is 2/3. You must have meant false positive.

Like you said, the false accusation (false positive) rate is based on a subset of the population, ie, the "not rapists", not the final result.

Let's review what false accusation means,: you are accused but you are not a rapist, hence, the 2-10% applies to 950, not the entire population, the 2x2 table should make that apparent. If we go by your table, the false accusation rate is then 1/950, not the 2-10% range.

I will try to paraphrase it so it's easier to understand/wrap our heads around:
Given 2-10% false accusation rate, the likelihood someone actually is a rapist based on one accusation is 26%.

In your scenario/table,
Given 0.1% false accusation rate, the likelihood someone actually is a rapist based on one accusation is 95%.
« Last Edit: October 03, 2018, 12:34:00 AM by anisotropy »

runbikerun

  • Pencil Stache
  • ****
  • Posts: 539
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1041 on: October 03, 2018, 01:21:15 AM »

in your math, you have a total of 65 accusations of which 48 are not rapists, eg false accusations. That is a 73% false accusation rate. The false accusation rate (which you apply at 5%) is the number of accusations that are in the category "not rapists".


Exactly, the false accusation rate only applies to the "non rapists". Thus, when we combine both real rapists and non rapists who are accused, we arrive at a grand total of 73% false accusation rate. This back of envelope calculation shows the much publicized 2-10% false accusation rate is very misleading when we actually take a look deeper.

No, I'm pretty sure Glenstache is right.  The false accusation rate is the percentage of accusations where an investigation establishes that no crime was committed or attempted.  Not the percentage of "not rapists" that are accused as you seem to be using.  I don't think your analogy works.

I see where the confusion is, I believe the 2-10% is the false positive rate and should be used as an input, you guys believe its the final result.

If we assume this 2-10% false accusation rate is the final result and not the mere "false positive" rate, we would arrive at some  troubling result. Either the number of rapists in the general population would need to be incredibly large (approaching 60%), or the false positive rate need to be super tiny, ie, sub 0.001%. I find these assumptions unrealistic.

Or (and I find it deeply unrealistic that this didn't occur to you, because it's been the core of one of the most visible public debates in modern history for the last year), the number of rapists in the general population is uncomfortably large, and the number of false accusations is very low. Didn't someone link a study upthread that indicated that an appalling percentage of men admitted to rape when it was phrased as something other than rape?

I also take exception to your maths: you have a calculation that takes a 5% false accusation rate, and somehow turns that into 72% of accusations being false. Without a very clear explanation of how that works, I can only conclude that there's either an error or some sleight of hand going on. A 5% false accusation rate is not compatible with 72% of accusations being false. And given that I've already seen you on another thread substituting the ECHR's judgement on Turkish politics for EU policy on Muslim immigrants, I'm not inclined to give you the benefit of the doubt and assume there's a remarkable quirk of statistics I'm just not getting.

runbikerun

  • Pencil Stache
  • ****
  • Posts: 539
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1042 on: October 03, 2018, 01:39:39 AM »
I'm going to apply a test I like to use in cases like these: I change something that doesn't impact in the maths, then reread the changed outcomes and see if it looks even vaguely sensible.

Let's replace "false accusation" with "wrongful conviction", and leave all the numbers as they are. The vast majority of the population has never been wrongfully convicted, but now almost three-quarters of our prison populations quite literally never did anything illegal. If anisotropy's math holds up, and if we accept a wrongful conviction rate of 5% (not a mad figure), then the numbers lead us directly to the conclusion that the prison system is designed to punish completely innocent people.

I hold no great love for the American prison system, but even I don't think 3/4 of prisoners are as pure as the driven snow. And yet, if anisotropy's math is correct, that's the case. There are two possible conclusions: the math is fatally flawed, or Western society is a brutalised hellscape. Bad as things are, I really don't think the second is true.

former player

  • Walrus Stache
  • *******
  • Posts: 8724
  • Location: Avalon
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1043 on: October 03, 2018, 01:40:32 AM »
I voted nay, not because of the allegations but because how emotionally stirred he was and how partisan he seemed.

Contrast that with calm demeanor Gina Haspel displayed when she went through her hearings, I found him too temperamental for my own comfort. I understand being accused of a sexual assault perpetrator (whether true or false) definitely played a part in his reactions, but still, perhaps I am asking for too much.

Regarding the allegations, aside from waiting for the FBI investigation results, this is something I whipped up and I would like to run it by some of the more mathematically attuned posters:

Suppose you suspect that you are allergic to something, which 1% of the population does, you went to the clinic and did a test. The test turned out to be positive.

The test has a false positive rate of 10% and a false negative rate of 20%. What's the chance of you actually being allergic?

The standard 2x2 table would look like this

                              1% have it                      test yes                          test no
Have allergy                10                                 8                                      2
Don't have                 990                                99                                   891
                                1000                              107                                  893

so, 107 people are test positive but only 8 people are actually allergic; even with a positive test, your chances of being actually allergic is only 7%, despite the test having only 20% false negative rate.

Now, I am going to do the same math but using stats related to male on female rape. I am assuming 5% of the male are rapists. The false accusation rate is known to be 2-10%, so I will take the avg 5% as false positive rate. Only about 1/3 rape are reported, so I am using 2/3 as false negative rate. You can substitute numbers you find reasonable.

The 2x2 table would look like this

                              5%                             accused                          not accused
Rapists                   50                                  17                                   33
Not rapists              950                                48                                  902
                             1000                               65                                  935

so, 65 people are accused but only 17 people are actually rapists; that's 28% chance of a person being actual guilty when accused of rape, despite the false accusation rate being only 2-10%. Obviously with multiple accusers this number would go up, classic Bayesian results. This is why I always refrain from making judgements without hard forensic evidences. Hopefully the FBI will clear things up.

Your supposed stats suffer from the usual problem of "garbage in, garbage out".  Firstly, why are you using an example of rape when the accusation is attempted rape and sexual assault?  Why are you assuming 5% of males as rapists?  Why are you assuming 5% of false accusations, and why are you basing this on a range of 2% to 8% which is not "known" but is itself an uninformed guess?


Are you suggesting I should treat Dr. Ford's accusation separately from swetnicks' because it's "not" rape?

I am usually pretty polite and constructive in a serious conversation, it is only when some random person shows up wholly uninformed I get annoyed. You must have never had a disagreement with me here, otherwise you would know I usually use pretty well-sourced data.

Assuming 5% are rapists: "6.4% of men openly admitted of committing the strictest possible definition of rape", I rounded it so it's easier to run calculations.
https://davidlisak.com/wp-content/uploads/pdf/RepeatRapeinUndetectedRapists.pdf

Assuming 5% false accusation: Most experts agree it's between 2-10%. "A multi-site study of eight U.S. communities including 2,059 cases of sexual assault found a 7.1% of false reports (Lonsway, Archambault, & Lisak, 2009)." Again, I rounded it so it's easier to run calc.
https://www.nsvrc.org/sites/default/files/Publications_NSVRC_Overview_False-Reporting.pdf
http://www.scirp.org/(S(i43dyn45teexjx455qlt3d2q))/reference/ReferencesPapers.aspx?ReferenceID=1238871

You are more than welcome to provide your own numbers provided they are backed by research.

But what is the correlation between the admitted 5% rapists and the 5% false accusations?  There is a very significant underreporting of rape and sexual assault.  The 5% false accusations is 5% of all accusations, right?  So you need to take account of the underreporting rate too.  I see laserlady has made this point

And as sol says, you are ignoring that 1) this is an individual case not one based on probabilities, and 2) as numerous people including me have pointed out, there is strong evidence supporting Dr Ford's testimony and disputing Kavanaugh's.


anisotropy

  • Pencil Stache
  • ****
  • Posts: 681
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1044 on: October 03, 2018, 02:08:09 AM »

Or (and I find it deeply unrealistic that this didn't occur to you, because it's been the core of one of the most visible public debates in modern history for the last year), the number of rapists in the general population is uncomfortably large, and the number of false accusations is very low. Didn't someone link a study upthread that indicated that an appalling percentage of men admitted to rape when it was phrased as something other than rape?

I also take exception to your maths: you have a calculation that takes a 5% false accusation rate, and somehow turns that into 72% of accusations being false. Without a very clear explanation of how that works, I can only conclude that there's either an error or some sleight of hand going on. A 5% false accusation rate is not compatible with 72% of accusations being false. And given that I've already seen you on another thread substituting the ECHR's judgement on Turkish politics for EU policy on Muslim immigrants, I'm not inclined to give you the benefit of the doubt and assume there's a remarkable quirk of statistics I'm just not getting.

First off, if i recall correctly, i listed ECHR's judgment of certain parts of Sharia law being incompatible with democracy. I mistook ECHR as part of EU, that is true and I corrected myself once pointed out. You have to realize this mistake on my part in no way reverses their finding and the judgment stands. You are free to disagree with that judgement if you so desire. But say it out loud explicitly.

Secondly, LOL seriously? Because we disagreed on the topic regarding muslim immigrants, more specifically, that I agreed with ECHR that the Sharia laws are not compatible with democracy, it automatically disqualifies anything I say? How tribal can you get? Don't be ridiculous, how's this behavior any different from the tribal Republicans you seem to enjoy making fun of?

Lastly, I think with the sole exception of Sol (perhaps some others), most of you don't really understand how to perceive, let alone how to calculate, conditional probabilities when it involves false positives and false negatives.

It did not "somehow turn that into 72% of accusations being false", there is no slight of hand, take (or retake) stat 101 in college, you will run into similar questions in class, and hopefully learn to do them for the first (or this) time.

Let's review what false accusation means: you are accused but you are not a rapist, hence, the 2-10% applies to 950, not the entire population, the 2x2 table should make that apparent.

I will try to paraphrase it so it's easier to understand/wrap your head around runbikerun:
Given 2-10% false accusation rate, the likelihood someone actually is a rapist based on one accusation is 26%.

-----------------------------------------------

Quote
I'm going to apply a test I like to use in cases like these: I change something that doesn't impact in the maths, then reread the changed outcomes and see if it looks even vaguely sensible.

Let's replace "false accusation" with "wrongful conviction", and leave all the numbers as they are. The vast majority of the population has never been wrongfully convicted, but now almost three-quarters of our prison populations quite literally never did anything illegal. If anisotropy's math holds up, and if we accept a wrongful conviction rate of 5% (not a mad figure), then the numbers lead us directly to the conclusion that the prison system is designed to punish completely innocent people.

I hold no great love for the American prison system, but even I don't think 3/4 of prisoners are as pure as the driven snow. And yet, if anisotropy's math is correct, that's the case. There are two possible conclusions: the math is fatally flawed, or Western society is a brutalised hellscape. Bad as things are, I really don't think the second is true.
Again, your wrongfully accused

If your premise is wrongful conviction, you must then first # people convicted, and # of people charged. Then you need to identify the false positive rate, which i have no clue what, but let's say we use the same 5%. What would you then use as false negative? crimes unreported?

Until we change the numbers to reflect the new premise, we can not have sensible outcomes. Luckily someone already did some of the hard work for us:

"How many other Glenn Fords are still behind bars? How many will die there? Just how often does our venerated justice system fail? Rarely, at least according to U.S. Supreme Court Justice Antonin Scalia. In a 2006 opinion he cited an approximate error rate of 0.027 percent, based on back-of-the-envelope calculations by an Oregon district attorney in a fiery op-ed for the New York Times. The op-ed was in response to a report by Samuel Gross, a law professor at the University of Michigan, cataloguing 340 exonerations between 1989 and 2003. “Let’s give the professor the benefit of the doubt,” the op-ed read. “Let’s assume that he understated the number of innocents by roughly a factor of 10, that instead of 340 there were 4,000 people in prison who weren’t involved in the crime in any way. During that same 15 years, there were more than 15 million felony convictions across the country. That would make the error rate .027 percent — or, to put it another way, a success rate of 99.973 percent.”

https://www.nationalgeographic.com/science/phenomena/2014/04/28/how-many-people-are-wrongly-convicted-researchers-do-the-math/

The error rate, when it comes to being wrongly convicted, is a much smaller 0.027% leading to a success rate of 99.973%. When you use 0.027% as the false positive rate, and do the calculation as Sam Gross did, you arrive at wrongfully convicted being ~4%. I am afraid based on what you said you dont understand conditional probabilities at all.

« Last Edit: October 03, 2018, 02:26:39 AM by anisotropy »

anisotropy

  • Pencil Stache
  • ****
  • Posts: 681
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1045 on: October 03, 2018, 02:19:37 AM »
formerplayer,

The under reporting is being accounted for in my calculation as "False negatives", namely, crime went unreported, which is 2/3 based on studies.

And one more time, Let's review what 5% false accusation means: you are accused but you are not a rapist, hence, the 5% applies to 950, the subset, not the entire population, the 2x2 table should make that apparent.

I will try to paraphrase it so it's easier to understand:
Given 2-10% (avg 5%) false accusation rate (in the not rapist subset), the likelihood someone actually is a rapist, in the entire population, based on one accusation is 26%.

Again, I believe stats apply here, because as far as I am concerned, this is a random guy in the population being accused. With two accusations the guilty likelihood is ~70% based on stats alone, but without hard physical evidence or a positive FBI investigation result, I must decline to see him as guilty.

former player

  • Walrus Stache
  • *******
  • Posts: 8724
  • Location: Avalon
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1046 on: October 03, 2018, 02:34:18 AM »
formerplayer,

The under reporting is being accounted for in my calculation as "False negatives", namely, crime went unreported, which is 2/3 based on studies.

And one more time, Let's review what 5% false accusation means: you are accused but you are not a rapist, hence, the 5% applies to 950, the subset, not the entire population, the 2x2 table should make that apparent.

I will try to paraphrase it so it's easier to understand:
Given 2-10% (avg 5%) false accusation rate (in the not rapist subset), the likelihood someone actually is a rapist, in the entire population, based on one accusation is 26%.

Again, I believe stats apply here, because as far as I am concerned, this is a random guy in the population being accused. With two accusations the guilty likelihood is ~70% based on stats alone, but without hard physical evidence or a positive FBI investigation result, I must decline to see him as guilty.


Let's put your stats to one side: whether accurate or not they are in fact a distraction from the facts that is a diversion tactic from the question of whether Kavanaugh should be confirmed as Supreme Court justice.  Unless you saying that Kavanaugh was just a random guy picked on by Dr Ford to make an accusation against?  That the sworn statements and appearance before the Senate of a serious and respected professional woman were a complete fiction made up for some unknown reason?  That she didn't know Kavanaugh when she was younger?  That she randomly made up the existence of a party described in Kavanaugh's calendar?  That she invented the details of the house in which she was pushed into a bedroom, sexually assaulted and feared suffocation from an attempt by Kavanaugh to keep her quiet, to the extent that decades later she built an extra escape route into her own house?  You are really discounting all of that?  On what rational grounds?  You seem keen on logic, please try and logic all that away, point by point.

Dollar Slice

  • Walrus Stache
  • *******
  • Posts: 9598
  • Age: 46
  • Location: New York City
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1047 on: October 03, 2018, 02:43:16 AM »
And one more time, Let's review what 5% false accusation means: you are accused but you are not a rapist, hence, the 5% applies to 950, the subset, not the entire population, the 2x2 table should make that apparent.

No. You're getting this wrong and it's not because of the math.

It's 2-10% *of rape allegations* that are false.

It is NOT that 2-10% of innocent men are accused. Go check your source and I'm sure you'll see that it says 2-10% *of allegations*.

That means you do not apply it to the population. You apply it to the number of rape allegations that happened. So if there were 17 real allegations of rape there should be 0-2 false accusations. You take 2-10% of 17, because that is the number of allegations. And your statistic is a percent of allegations, not a percent of a population of people.

This is why they teach word problems in school... :-)

runbikerun

  • Pencil Stache
  • ****
  • Posts: 539
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1048 on: October 03, 2018, 03:11:41 AM »
anisotropy - I don't doubt you because we disagree on Muslim immigration. I doubt you because I think you're intellectually dishonest.

And I can see what you're doing with these numbers. You're taking "5% of accusations are false" and "this test is 95% accurate" to be equal concepts, when they're not. The data you're starting with deals only with accusations, or positive tests, and the 5% number applies here. "This test is 95% accurate" and "95% of positive results on this test are correct" are two very different statements, but you've muddied the waters to produce the results you want.

Unique User

  • Pencil Stache
  • ****
  • Posts: 718
  • Location: NC
Re: Brett Kavanaguh: Yay or Nay?
« Reply #1049 on: October 03, 2018, 05:37:11 AM »
This forum is so strange, 21 pages and the thread has come down to lies, damned lies, and statistics.