[QUOTE=brazil84]
Did I do that somewhere?
[/QUOTE]
Sorry, not you. I didn’t mean to imply that. If you ever get around to reading the thread (only if you’re really bored) you’ll find that several people did. The argument was: 1 in 30K people have IQs of 160 or over. There are 5K Dopers. Therefore, it is very unlikely that there is even one, let alone 2, people with 160 IQ or over here.
[QUOTE=ArizonaTeach]
Sure, but how big’s your penis?
[/QUOTE]
DAMN those typos.
[QUOTE=Voyager]
I have no idea what the distribution of IQ is here, but using the distribution of the population as a whole to make predictions about what you’ll find here is fallacious, ..
[/QUOTE]
So … one can use the distribution of a characteristic in a sample to make predictions about the population, but never the other way around?
That can’t be right. Are you just referring to this being a self-selected sample?
[QUOTE=Sleeps With Butterflies]
This comment seems like the spittle flying, keyboard pounding, nasty intentioned rantings of a hateful lunatic…
You took a reasonable comment from another member and went to Defcon 1,…
[/QUOTE]
I feel obliged to thank you for your defense, and for understanding the intended tone of my post. (And for your opening line; I will honored to inspire a spirited critique from you some day.)
In askeptic’s defense I must point out that I am fat.
[QUOTE=j666]
So … one can use the distribution of a characteristic in a sample to make predictions about the population, but never the other way around?
That can’t be right. Are you just referring to this being a self-selected sample?
[/QUOTE]
If the sample is not randomly selected, you can’t draw any inferences either way.
[QUOTE=ultrafilter]
If the sample is not randomly selected, you can’t draw any inferences either way.
[/QUOTE]
Yes; but can’t one make predictions about a random sample from the characteristics of the population? And Voyager is referring to the Board being a non-random sample?
[QUOTE=j666]
So … one can use the distribution of a characteristic in a sample to make predictions about the population, but never the other way around?
That can’t be right. Are you just referring to this being a self-selected sample?
[/QUOTE]
Absolutely. That’s what polling and other examples of sampling theory are all about.
Yes, the Dope is a self-selected sample, which is why you can’t estimate characteristics in the population from it. My example on the last page showed that if we used the distribution of Jeopardy contestants in the Dope to estimate the population of Jeopardy contestants in the country as a whole, we get an absurd result.
The most famous example of this (and I apologize if you’ve heard of it) was the Literary Digest presidential poll of 1936. It showed Alf Landon winning handily, while FDR won in a landslide. The problem was that it was a telephone poll, and back then telephones were far from common, and those who could afford them skewed Republican.
Presidential poll sample today try to take from all population subgroups, and also try to get likely voters. I think one of the reasons they’ve messed up so much in the primaries is that lots of unlikely voters voted.
[QUOTE=j666]
I feel obliged to thank you for your defense, and for understanding the intended tone of my post. (And for your opening line; I will honored to inspire a spirited critique from you some day.)
In askeptic’s defense I must point out that I am fat.
[/QUOTE]
No thanks are necessary 
Eh, I don’t care if you’re 500 pounds, he’s akin to a 10 year old to pull the fat jokes out. “You’re a fat loser, wah wah wah.” Insults are easier to sling than actually composing a valid response to a well thought out comment. Take a look at his posting history and you’ll see he’s a fan of the low road. People like that have opinions and thoughts which are easy to dismiss.