Mathematical Analsysis Of James Woods IQ Drama

[PancakeBoy]

This is a post RE:
http://forumserver.twoplustwo.com/showfl...0&fpart=all (http://forumserver.twoplustwo.com/showflat.php?Cat=0&Board=tv&Number=5667561&page=0& fpart=all)



[ QUOTE ]
According to this (http://www.geocities.com/rnseitz/Definition_of_IQ.html), about one in every 3,500,000 people have an IQ of 180 or higher. So with 6.5 billion people on Earth, our entire planet should expect to have about 1,857 people with an IQ of 180+.

I'm going to go out on a limb and say that James Woods is probably not one of them.

Then again, you could come to that conclusion sooner if you realize that he read the script to "Be Cool" and still agreed to act in it.

And screaming "I AM A POKER GOD" during Celebrity Poker Showdown doesn't help, either.

[/ QUOTE ]

Okay here is a hugely nerdy analysis of why I think James Woods probably doesn't have an IQ of 184 or greater.


The probablity that any person's IQ is > 180 is 1/3.5 million as stated in the above article. So if we knew nothing about James Woods we would say P(JWIQ > 180) = 1/3.5 million

Now we know James Wood scored 184 on his IQ test So we want to calculate P(JWIQ > 180 | T), that is, the probability that James Wood's IQ is greater than 180 given he scored so high on the test.

So we apply Bayes' Rule:

P(JWIQ > 180| T) =

P(T|JWIQ > 184)*P(JWIQ > 184)
----------------------------------
(P(T|JWIQ >184)*P(JWIQ > 184) + P(T|JWIQ < 184)*P(JWIQ < 184)

Now, we know all but one variable, let me fill them in


P(JWIQ > 180| T) =

1*1/3.5m
----------------------------------
1*1/3.5m + P(T|JWIQ < 184)*3499999/3.5m

SO, we have to make an assumption, what is the probability that he would score that high on the test given his IQ wasn't that high, i.e. that the IQ test was inaccurate, or that James Woods just got lucky on his test (it's multiple choice afterall). Let's give a very very generous number, say, the IQ test is wrong or James Woods gets lucky only one in ten thousand times:



P(JWIQ > 180| T) =

1*1/3.5m
----------------------------------
1*1/3.5m + .00001*3499999/3.5m

= .03~

So assuming James Woods gets lucky on his IQ test or that the IQ test is inaccurate 1 in 10000 times, i have 3% certainty that his IQ is indeed > 184. That is to say, he probably doesn't.

To make the probability of JW's IQ being > 50%, we would have to assume that the accuracy of the test and the probability James doesn't get lucky has to be: 2.86*10^-7.

Is the probabilty that the IQ test is inaccurate AND that he didn't get lucky (knew all the answers) THIS LOW? Probably not; more than likely he got a couple of the real hard questions right because he made some lucky guesses OR the IQ test could be an inaccurate measure of his intelligence.



CONCLUSION: James Woods IQ is probably not > 184.

IN YOUR FACE JAMES WOODS NUTHUGGERS.


/images/graemlins/grin.gif

[HedonismBot]

His IQ is 184

[madnak]

You can't apply probabilistic math against outcomes that have already happened. Also, I don't think any IQ test that goes that high is multiple-choice.

[MadTiger]

He didn't apply "probabilistic math" so much as used probability and statistical analysis to come his conclusion.

[madnak]

Fair enough, I didn't actually look at the math before responding. My bad. Still. The 3.5 million figure is improperly applied here. It's not a question of whether the test "succeeded" or "failed." And this is an inappropriate use of Bayes' theorem. It doesn't apply to every damn thing. In a strict sense, it's highly limited.

But regardless. Even assuming the math is completely valid, as well as the linked data (which is [censored] in at least one regard) it doesn't approach the full implications of the test result. For one thing, it fails to consider the probability that James Woods has an IQ > 160. Even if there is a good chance that his IQ is < 180, it's virtually certain that he's a super-smart guy. And at that level the whole scale falls apart anyhow - nobody can agree on what the basis is. Also it's highly unlikely that intelligence beyond a certain range actually represents a normal distribution in any meaningful sense, so the attempts of the IQ scale to function according to such a distribution make it even more meaningless.

[oneeye13]

[ QUOTE ]
According to this (http://www.geocities.com/rnseitz/Definition_of_IQ.html), about one in every 3,500,000 people have an IQ of 180 or higher. So with 6.5 billion people on Earth, our entire planet should expect to have about 1,857 people with an IQ of 180+.

[/ QUOTE ]

funny that we have so many of them on this board

[BigSoonerFan]

[ QUOTE ]
[ QUOTE ]
According to this (http://www.geocities.com/rnseitz/Definition_of_IQ.html), about one in every 3,500,000 people have an IQ of 180 or higher. So with 6.5 billion people on Earth, our entire planet should expect to have about 1,857 people with an IQ of 180+.

[/ QUOTE ]

funny that we have so many of them on this board

[/ QUOTE ]

I'm pretty sure mine is at least 200. It just can't be measure adequately!

[Peter666]

[ QUOTE ]
This is a post RE:
http://forumserver.twoplustwo.com/showfl...0&fpart=all (http://forumserver.twoplustwo.com/showflat.php?Cat=0&Board=tv&Number=5667561&page=0& fpart=all)



[ QUOTE ]
According to this (http://www.geocities.com/rnseitz/Definition_of_IQ.html), about one in every 3,500,000 people have an IQ of 180 or higher. So with 6.5 billion people on Earth, our entire planet should expect to have about 1,857 people with an IQ of 180+.

I'm going to go out on a limb and say that James Woods is probably not one of them.

Then again, you could come to that conclusion sooner if you realize that he read the script to "Be Cool" and still agreed to act in it.

And screaming "I AM A POKER GOD" during Celebrity Poker Showdown doesn't help, either.

[/ QUOTE ]

Okay here is a hugely nerdy analysis of why I think James Woods probably doesn't have an IQ of 184 or greater.


The probablity that any person's IQ is > 180 is 1/3.5 million as stated in the above article. So if we knew nothing about James Woods we would say P(JWIQ > 180) = 1/3.5 million

Now we know James Wood scored 184 on his IQ test So we want to calculate P(JWIQ > 180 | T), that is, the probability that James Wood's IQ is greater than 180 given he scored so high on the test.

So we apply Bayes' Rule:

P(JWIQ > 180| T) =

P(T|JWIQ > 184)*P(JWIQ > 184)
----------------------------------
(P(T|JWIQ >184)*P(JWIQ > 184) + P(T|JWIQ < 184)*P(JWIQ < 184)

Now, we know all but one variable, let me fill them in


P(JWIQ > 180| T) =

1*1/3.5m
----------------------------------
1*1/3.5m + P(T|JWIQ < 184)*3499999/3.5m

SO, we have to make an assumption, what is the probability that he would score that high on the test given his IQ wasn't that high, i.e. that the IQ test was inaccurate, or that James Woods just got lucky on his test (it's multiple choice afterall). Let's give a very very generous number, say, the IQ test is wrong or James Woods gets lucky only one in ten thousand times:



P(JWIQ > 180| T) =

1*1/3.5m
----------------------------------
1*1/3.5m + .00001*3499999/3.5m

= .03~

So assuming James Woods gets lucky on his IQ test or that the IQ test is inaccurate 1 in 10000 times, i have 3% certainty that his IQ is indeed > 184. That is to say, he probably doesn't.

To make the probability of JW's IQ being > 50%, we would have to assume that the accuracy of the test and the probability James doesn't get lucky has to be: 2.86*10^-7.

Is the probabilty that the IQ test is inaccurate AND that he didn't get lucky (knew all the answers) THIS LOW? Probably not; more than likely he got a couple of the real hard questions right because he made some lucky guesses OR the IQ test could be an inaccurate measure of his intelligence.



CONCLUSION: James Woods IQ is probably not > 184.

IN YOUR FACE JAMES WOODS NUTHUGGERS.


/images/graemlins/grin.gif

[/ QUOTE ]

The only problem with this analysis is that it doesn't account for the fact that his IQ is already 184. No need to make "assumptions". Perhaps "Be Cool" will be considered the greatest film ever 100 years from now. Only our Lord and Saviour James Woods recognizes that because of his 184 IQ.

[Copernicus]

There are two fatal problems with this analysis. First,

it assumes that JW comes from a random population and therefore P(IQ>184) is 1/3.5 million.

In fact he comes from some narrower population that can be used to increase that probability substantially. Some of the criteria to narrow that population include SAT scores, success in school, success in life, success in tearing RK a new ahole etc.

The second and bigger problem is that the calculation is extremely sensitive to the assumption that he could get lucky 1/10000 times. As formulated that 1/10000 chance is the probabilty that an average person could achieve a score of 184 on luck alone. From what Ive seen of IQ tests recently, if there are 40 questions, mean intelligence of 100 is achieved at much lower than 20 questions answered correctly..as low as 10 possibly. Ie the sensitivity of the test is focused on the high side, and to achieve a 184 score might require as many as 20 or 25 additional correct answers over the mean of 10. Also the tests Ive looked at dont have just 5 possible answers, but include many questions with two 3x3 matrices and the answer involves one element from each matrix, so a pure guess has only a 1/81 probability of being correct.

But lets be truly generous (as opposed to the 1/10000 assumption which is far from generous). Assume it only takes 20 extra right answers out of 30 to get to 184, and that each question only has 5 answers. (that would mean that 10 more questions are used to judge IQs between 184 and the maximum which I believe is 200, so I think I am being generous) 20 out of 30 correct or more is 11+ standard deviations from random...a probability far lower than the 1 in 4 million needed for there to be a >50% chance that his IQ is >=184.

There are tradeoffs between the two errors...the higher the probability that his IQ is high (> than say 130), the easier it is for him to get lucky (going from 130 to 184 instead of 100 to 184).

I dont know what the right numbers are, but I do know they are a helluva lot better than 3%.

[mindflayer]

standard tests do not accurately measure IQ beyond 150-160.

My son was tested for giftedness and spent two 3 hour sessions answering "questions" and doing tests with zero multiple choices.

The tests measured many different abilities..

one of the "tests" involved repeating increasingly long series of letters and numbers and then being able to manipulate them.
the tester would start with 5g and ask the subject to repeat the them.. then go on to 1B6 , then L4GF, then 15Fy7, then 1k8j8px4 etc.
Then he was asked to repeat the letters and numbers in order but numbers first then letters etc.. so for 5g7k4l9p4
his answer should be 574194gkp
then he was asked to do them backwards.. etc.

BTW I took a multiple question IQ test online the other day.. and then because I wanted to make sure, I took it again the next day, and lo and behold I knew they were wrong the first time, the next day i got 20/20 and it says i am a genuis!!
/images/graemlins/wink.gif