Sat in a philosophy class today.

[Splorf22]

Originally Posted by Alawen [You must be logged in to view images. Log in or Register.]

I already stated my one-sentence summary: he is asserting that exploratory factor analysis cannot prove the existence of g.

Yes, but what does this mean?

[Alawen]

Originally Posted by Splorf22 [You must be logged in to view images. Log in or Register.]

Yes, but what does this mean?

I'm not sure what you're asking. In this context, the practical meaning is that Charles Spearman's attempt to derive a unitary factor through factor analysis of correlation was flawed. I think he even mentions that some relevant statistical method hadn't been invented yet.

The simplified Gf-Gc model which evolved into the current Cattell-Horn-Carroll (CHC) theory was derived empirically and not from factor analysis. Like I said, dead straw man is dead.

[Splorf22]

This is not what I am saying at all.

I googled "example IQ test questions" and got the following: http://www.stevezervos.com/iq-test.htm

So I go to the first question: Which letter does not belong in this series? H D R Y L

I have seen similar questions before, so I start running through my list of possibilities: Is it a mixture of letters and vowels? no. Are the letters almost in order? no. What do the letters look like in standard book code? 8 4 18 25 12. Now I start running through my list of number sequence possiblities. Are the numbers in order? no. Are they primes? no. Are they a mix of odd and even? Yes. So my answer is Y is the odd one out. Is this right? I have no idea.

I go to the second question: Which alpha-numeric is the odd one out? 9D 8F 3E 6C

I again begin to enumerate. The obvious answer is that 3E contains 3, an odd number. The second possibility is that these numbers are in hex. In that case both 8F = 143 and 9D = 157 are odd, while 3E = 62 and 6C = 108 are even, so that's out. Another possibility is that the numbers are in hex and one is prime. In this case 62 (31x2), 108 (9x12), and 143 (11x13, I checked a list of prime numbers [You must be logged in to view images. Log in or Register.]) are not prime (11 x 13, I googled) but 157 is. Both solutions match so I select the hexadecimal one a-priori because prime numbers are cool.

The point I am trying to make is that my thought process here is entirely specific to IQ tests. I know several ways of looking at sequences of numbers for patterns. I apply them. If the test uses one of the ones I know, I get the problem right. If it doesn't, I don't. So I check the answers: (1) Y does not belong.. Every other letter is the third letter following a vowel. (2) 8f. The word for the number 9 contains 4 letters (nine) so the fourth letter of the alphabet (D) is paired with it. This rule applies to the other pairs as well, except 8F.

I have not seen either of those types of patterns before, so I never considered them. I definitely prefer my first solution to his as it is far simpler. I also prefer my second solution as "hexidecimal primes" is simpler than "pairing the book code of a number's english length with the number".

The point I am trying to make by this is that there is nothing GENERAL or FACTORED to what I am doing. It is 100% specific to IQ tests. I don't believe in Spearman's 1 factor or Gardner's 10 factors or whatever.

[Splorf22]

Originally Posted by Alawen [You must be logged in to view images. Log in or Register.]

I'm not sure what you're asking. In this context, the practical meaning is that Charles Spearman's attempt to derive a unitary factor through factor analysis of correlation was flawed. I think he even mentions that some relevant statistical method hadn't been invented yet.

The simplified Gf-Gc model which evolved into the current Cattell-Horn-Carroll (CHC) theory was derived empirically and not from factor analysis. Like I said, dead straw man is dead.

I am saying that you could make that statement by reading the section headings. Pretend I am an idiot (maybe this is not hard, I don't know [You must be logged in to view images. Log in or Register.]) and explain it to me.

[Cecily]

So the big take away from this is that IQ tests are just a front to find out who can save American by breaking Nazi codes.

[Alawen]

Originally Posted by Splorf22 [You must be logged in to view images. Log in or Register.]

I am saying that you could make that statement by reading the section headings. Pretend I am an idiot (maybe this is not hard, I don't know [You must be logged in to view images. Log in or Register.]) and explain it to me.

Yeah, this is a real asshole game that you're playing with me.

In 1904, there was this guy Charles Spearman. He had a bunch of data from giving different mental tests to a bunch of people. I don't know how many tests or how many people. He noticed that there seemed to be similarity between the scores for different tests by the same person. He used math to try to isolate the cause of that similarity. Then he developed a model of how intelligence works based on his math.

In 2007, an obnoxious physicist named Cosma Rohilla Shalizi used over 10,000 words and a lot of unnecessary commas to say that the kind of math Charles Spearman used can't identify the cause that Spearman was looking for. He says that this kind of math can never identify an underlying cause, because lots of different causes could produce the same kind of data that Spearman started with.

I'm pretty irritated with your relentless implication that I didn't understand the paper when in fact it's you who don't know what the fuck you're talking about.

[Alawen]

Shalizi, by the way, is a really fucking horrible writer who intentional obfuscates a pretty simple idea in an apparent attempt to impress people like you.

[Splorf22]

Originally Posted by Alawen [You must be logged in to view images. Log in or Register.]

I'm pretty irritated with your relentless implication that I didn't understand the paper when in fact it's you who don't know what the fuck you're talking about.

Actually if anyone should be irritated it's me: the first thing you said is that I hadn't even read the paper. But I never really get irritated over the Internet.

Anyway if I am such an idiot then how about you explain to me what I'm missing?

[Alawen]

Shalizi is basically backing up Glymour responding to Herrnstein and Murray.

Debunking the century-old incarnation of Spearman's theory of generalized intelligence is a straw man. You don't appear to understand the context of Shalizi's paper, you missed important details that make it irrelevant to anything I've been talking about, and you seem to have willful ignorance of contemporary intelligence theories. You apparently dismiss all social sciences as pointless, so why are we even discussing this?

I mentioned at the beginning of this discussion of IQ that I think there is a disconnect between IQ and intelligence; I am not a strong proponent of IQ as a measure of anything. I just think psychometrics are interesting and I know a few things about them. Neuroscience will probably answer many of these questions in the next few decades. I'm going to bed.

[Hasbinbad]

Alawen what happens if you eat a big meal of fish 12 hours before you take an IQ test?