VisorCentral.com - One Year On

Pages (24): « First ... « 5 6 7 8 9 [10] 11 12 13 14 15 » ... Last »
Show 20 posts from this thread on one page

VisorCentral.com (http://discussion.visorcentral.com/vcforum/index.php)
- Off Topic (http://discussion.visorcentral.com/vcforum/forumdisplay.php?forumid=6)
-- One Year On (http://discussion.visorcentral.com/vcforum/showthread.php?threadid=26965)

Posted by Toby on 10-17-2002 07:35 PM:

quote:
Originally posted by KRamsauer
Guess not. Didn't you say you were a statistician? That's one.....

No, I never said I was a statistician, and I hope you're joking, because this was exactly the flaw that started it.

Posted by K. Cannon on 10-17-2002 07:35 PM:

quote:
Originally posted by KRamsauer
I was using it as a lower limit of drowning depth to show that any given point of water is most likely not going to kill someone. It seemed like a reasonable estimate given the height of people and such.

Ah, but there you go estimating and assuming again!!

Don't most people drown in their bathtub? (I either heard that somewhere or it's a Kelley-ism, I'm not sure.)

Let's bring this back to where we started. Basically, Yardie said that Joe Lieberman would be in favor of the war against Iraq. I compared that to folks who assumed JFK would be in favor of the Pope's wishes over USA interests. KRamsauer came in and basically said it was okay to assume that a Jewish person would be for the war on Iraq. I quote: " To assume someone's political view because of their religion is not to assume anything about their worth as an individual."

Which is not the point I was trying to make--I wanted to say "Can you really assume someone's political view b/c of their religion?"

Then followed a discussion of songs from films and KRamsauer offered to buy me a beer (which my husband later vetoed

)

And I said:
"I'm just saying that every time you make assumptions about someone's beliefs based on one of those factors, you should also prepare for the possibility that (1) are are wrong about what that person believes and (2) that you have offended that person by assuming they believe a certain way."

Then Jonathan, who is Jewish, said He was not in favor of the war on Iraq.

Therefore, KRamsauer's theory (that you can figgur out a person's political leanings ahead of time, based upon their religion) was disproved.

NOW LET ME SAY THIS: KRamsauer, a long time ago and way up higher in this thread, I said that people use race, religion, gender, intelligence, etc, etc, to make value judgments every day. It is called generalization. And while it is not always wrong, it sometimes is and (to me) the one time I risk offending someone (even if, as you previously insisted, you are making no judgment of the worth of that person) for no reason is enough (to me) to try not to generalize.

Whew.
Kelley

Posted by KRamsauer on 10-17-2002 07:37 PM:

Re: Re: Re: Re: Re: Re: Simplification

quote:
Originally posted by Toby
Because your premise is flawed. Not all Jews consider the current incarnation of Israel their country. Not all Jews consider there to be a threat. I doubt that even all Jews in _Israel_ consider there to be an eminent threat (given history and Santanaya's good track record).

Look closely. I never said all. Sorry if I implied that, but I thought I was being clear.

Posted by Toby on 10-17-2002 07:37 PM:

quote:
Originally posted by KRamsauer
I got the joke, I just thought it was a poor way to discredit statistics, that's all.

If you thought the point of the joke was to discredit statistics, you didn't get the joke.

quote:
You're right that it was foolish to insist he would absolutely live. In my book he's not even a statistician because any competent mind would interpret the notion of average correctly (this recently deceased statistician probably expects to see one third of a kid when walking into an "average" household). As for my statement, analyzing the situation I determined at most, one third of all spots in a body of water would cause someone to drown. So selecting single spots and dumping statisticians in randomly will result in fewer than 50% dying. That's all. Of course your one statistician may drown, but odds are the next person to try (at random, at a random spot on the body of water) won't.

You ever heard a quote about foolish consistencies? Please don't start down such a path.

Posted by KRamsauer on 10-17-2002 07:39 PM:

quote:
Originally posted by Toby
I generally don't put much stock in opinion polls or their scientific basis. If they actually published the survey and the methodology, I might, but they almost never do.

That's a good point. There have been many studies done that were just flat-out crazy. I believe the Hoover v Roosevelt election was predicted using a phone poll. DUMB! That resulted in huge selection bias in favor of Hoover. You're right in that methods are very important. I find it hard to believe people who respond to polls aren't systematically different from those who don't in every instance, but I'm comfortable in most instances dismissing such selection error as unsystematic.

Posted by KRamsauer on 10-17-2002 07:45 PM:

Great summary! I have only a few comments (you knew I would!).

quote:
Originally posted by K. Cannon

Then Jonathan, who is Jewish, said He was not in favor of the war on Iraq.

Therefore, KRamsauer's theory (that you can figgur out a person's political leanings ahead of time, based upon their religion) was disproved.

That's not a proof. That's a counter example, but in my wording I was careful to avoid saying "all" "every" or "completely." What you're putting forth is akin to saying women don't have babies because your great-aunt passed away childless.

quote:
NOW LET ME SAY THIS: KRamsauer, a long time ago and way up higher in this thread, I said that people use race, religion, gender, intelligence, etc, etc, to make value judgments every day. It is called generalization. And while it is not always wrong, it sometimes is and (to me) the one time I risk offending someone (even if, as you previously insisted, you are making no judgment of the worth of that person) for no reason is enough (to me) to try not to generalize.

I think that's a fair assessment. I must point out though that the beauty of thought is that no one else knows about it. :-) I can walk into a bar thinkign "oh man, that sign outside is really cruddy, this place is going to smell." However unless I scream "you smelly pigs" as I walk in, no one knows and I realize the truth, that it is actually a sewing circle and I'm the wrong place.

Posted by Toby on 10-17-2002 07:45 PM:

Re: Re: Re: Summary

quote:
Originally posted by KRamsauer
Okay, that's true. I tried to lay out in a theoretical framework that once parameters are known predictions can be made.

And you'd still be trying to use them poorly in the case that was at hand.

quote:
And being familiar with statistics yourself you know parameters can be derived from samples not a "census."

AAMOF, the most used Census data is conducted by sample (the long form isn't sent to everyone). It would still be flawed to find someone in a census block, though, and assume they fit the generalized criteria of it.

quote:
We were never enemies? Don't open the package that's getting delivered tomorrow then. ;-)

Considering that I don't open the mail at the only publicly accessible address available for me, I don't think that'd be a problem.

Posted by KRamsauer on 10-17-2002 07:46 PM:

quote:
Originally posted by Toby
If you thought the point of the joke was to discredit statistics, you didn't get the joke.[B]You ever heard a quote about foolish consistencies? Please don't start down such a path.

No, I'm insisting the next person will be safe upon discovering each trial has a 1/3 chance of failure, and that figure is completely independent. I'm not claiming that after flipping 10 heads in a row you'll get a tails at above chance levels. :-) Was the point of the joke to laugh at stupid people? Perhaps those who drown?

Posted by Toby on 10-17-2002 07:48 PM:

Re: Re: Re: Re: Re: Re: Re: Simplification

quote:
Originally posted by KRamsauer
Look closely. I never said all. Sorry if I implied that, but I thought I was being clear.

You implied that your default assumption would be that they would because (in your estimation) the majority would. The whole point of this digression is that such would be a flawed assumption.

Posted by KRamsauer on 10-17-2002 07:50 PM:

Re: Re: Re: Re: Summary

quote:
Originally posted by Toby
And you'd still be trying to use them poorly in the case that was at hand.

How so? I assigned a numerical value to the joke while preserving its intent.

quote:
AAMOF, the most used Census data is conducted by sample (the long form isn't sent to everyone). It would still be flawed to find someone in a census block, though, and assume they fit the generalized criteria of it.

I realize most of the useful data from the census comes from sampling hence my use of quotes to indicate my reference to the initial notion of a census, or complete canvassing.

Posted by KRamsauer on 10-17-2002 07:52 PM:

Re: Re: Re: Re: Re: Re: Re: Re: Simplification

quote:
Originally posted by Toby
You implied that your default assumption would be that they would because (in your estimation) the majority would. The whole point of this digression is that such would be a flawed assumption.

I was trying to say that given knowledge of a population and its traits it's possible to reason behond chance level the characteristics of a given subset. BEYOND CHANCE, not 100%

Posted by Toby on 10-17-2002 07:52 PM:

quote:
Originally posted by KRamsauer
[...] Was the point of the joke to laugh at stupid people? Perhaps those who drown?

The point of the joke is pretty much the same point that we've been trying to get across. Be careful of how you try to apply statistics (no matter how reliable or unreliable) in the real world.

Posted by KRamsauer on 10-17-2002 07:53 PM:

quote:
AAMOF, the most used Census data is conducted by sample (the long form isn't sent to everyone). It would still be flawed to find someone in a census block, though, and assume they fit the generalized criteria of it.

It would be wrong to state with 100% certainty, yes. It wouldn't be wrong to form a prediction though. Of course your prediction may be wrong (and indeed may be wrong in the overwhelming majority of cases) but knowing the characteristics of the group allows you the luxury of above chance predictions.

Posted by KRamsauer on 10-17-2002 07:56 PM:

quote:
Originally posted by Toby
The point of the joke is pretty much the same point that we've been trying to get across. Be careful of how you try to apply statistics (no matter how reliable or unreliable) in the real world.

Point taken. Another one I like: A statistician and a musician are getting out of the car in a dark lot. They parked in a dark corner. The musician, who drove there realizes he dropped his keys when he got out of the car. Of course it's dark over there. So the musician goes to the lit part of the lot and starts to look. Upon being asked why by the statistician, he replies "Because the light is better over here."
No disrespect meant to musicians, btw.
This is an example of not using knowledge of the population of potential key locations to one's advantage.

Posted by Toby on 10-17-2002 07:58 PM:

quote:
Originally posted by KRamsauer
It would be wrong to state with 100% certainty, yes.

Have you ever met a 60% person?

quote:
It wouldn't be wrong to form a prediction though.

I give up.

Posted by KRamsauer on 10-17-2002 08:06 PM:

quote:
Originally posted by Toby
Have you ever met a 60% person?[B]I give up.

I'm not sure what you mean by a 60% person. As per the census question, all I'm stating is that knowing 90% of the people in a sample (or a population) meet a given criteria, it would be wise upon selecting one person at random to predict they meet that criteria as opposed to not meeting it. Of course it's best to ask them, but as a matter of predictive reliability, it's better to say they meet it than they don't.

Posted by KRamsauer on 10-17-2002 08:10 PM:

To move this into better surroundings. Let's play "Let's Make a Deal." Behind one of three doors is a prize. You get to pick one door. Then I will show you a door that doesn't have the prize behind it (of course I won't reveal your door). Do you switch afterward, when given the chance?

According to your logic it doesn't matter if you switch because the knowledge of knowing one of the wrong doors doesn't allow you to make a 100% sure prediction. I say it definitely does matter if you switch, and you should switch every time. I won't be right 100% of the time, but I'll be right more often than someone who pursues any other strategy. This is an example of using knowledge of the population (there is one winner) to my advantage instead of ignoring it.

Posted by Toby on 10-17-2002 08:23 PM:

quote:
Originally posted by KRamsauer
I'm not sure what you mean by a 60% person. As per the census question, all I'm stating is that knowing 90% of the people in a sample (or a population) meet a given criteria, it would be wise upon selecting one person at random to predict they meet that criteria as opposed to not meeting it. Of course it's best to ask them, but as a matter of predictive reliability, it's better to say they meet it than they don't.

For argument's sake, let's say that 99% of the population is not severely allergic to $MEDICATION. You are wheeled unconsciously into the doctor's office with $DISEASE which is treated mainly by $MEDICATION. You just happen to be in the 1% who is severely allergic. You're dead. He was just using predictive statistics. Extreme example, yes, but it's the point all along.

Posted by KRamsauer on 10-17-2002 08:31 PM:

quote:
Originally posted by Toby
For argument's sake, let's say that 99% of the population is not severely allergic to $MEDICATION. You are wheeled unconsciously into the doctor's office with $DISEASE which is treated mainly by $MEDICATION. You just happen to be in the 1% who is severely allergic. You're dead. He was just using predictive statistics. Extreme example, yes, but it's the point all along.

Now what if this disease kills people within 10 seconds of some sign (say a seizure) which I am experiencing if $Medication isn't injected and the test to determine my allergy takes half an hour. Presume there is $Medication2 which cures the 1% allergic to $Medication but kills the 99% who aren't. What do you do? Clearly you use statistics! That is all I'm saying here. You're going to be wrong often, but you will save more people (like me!) than if you were randomly administer (in 50% - 50% ratios) the two or administer none whatsoever. You should use your knowledge of the poplution to cure me.

Posted by Toby on 10-17-2002 08:31 PM:

quote:
Originally posted by KRamsauer
To move this into better surroundings. Let's play "Let's Make a Deal." Behind one of three doors is a prize. You get to pick one door. Then I will show you a door that doesn't have the prize behind it (of course I won't reveal your door). Do you switch afterward, when given the chance?

I'm more of a Jeopardy person.

quote:
According to your logic it doesn't matter if you switch because the knowledge of knowing one of the wrong doors doesn't allow you to make a 100% sure prediction.

No, it doesn't. What on earth would make you think that knowing one of the wrong doors means anything? Giving you a _perceived_ 50/50 shot at it hasn't increased the original odds of you winning.

quote:
I say it definitely does matter if you switch, and you should switch every time. I won't be right 100% of the time, but I'll be right more often than someone who pursues any other strategy.

Let's see the math. If they always show a wrong choice that you didn't pick, your odds have not improved at all, and there's no incentive to switch, and AAMOF, the odds of pursuing a strategy of always _not_ switching are identical.

quote:
This is an example of using knowledge of the population (there is one winner) to my advantage instead of ignoring it.

If there is only one winner, what makes you think it isn't behind the door you already picked?

All times are GMT. The time now is 04:46 AM.

Pages (24): « First ... « 5 6 7 8 9 [10] 11 12 13 14 15 » ... Last »
Show 20 posts from this thread on one page