Discussion in 'Off-Topic' started by Bootstoots, Aug 1, 2018.
so does having people like you
This is definitely wrong and countless sources have been posted already that prove exactly the opposite. One little anecdote is definitely not enough to disprove that. Your anecdote, by the way, is easily explained. I will refer you to my post regarding Arthur Jensen and his idea of test taking ability. In fact, people will actually do better on repeated tests, people will actually receive higher scores as they get better, more efficient at doing tests. The reason why you didn't improve is because we have found that both IQ and test taking ability level out once someone reaches their 20s, I suppose you were over 20 already when you took that undergrad test.
I do realize you are a psychologist/psychiatrist ? and all, and I respect you, but at least read the thread.. Or the last page.. If you are going to "correct" someone.
I don't think you are right. Many brilliant ideas were ridiculed in their time simply because the person propagating it wasn't liked for some reason. Many uncomfortable truths have been ignored. Ted kaczynski felt himself forced to kill someone to have his ideas heard by anyone.
Many of the concept/ideas we are now dealing with as "established" are precisely that only because they were shilled and repeated endlessly. Very rarely does an idea get accepted on merit alone, it needs to be spread, promoted, by the right people.
Point taken. Though I happen to disagree with some of the published research.
I take the Planckian view, which is that science advances one funeral at a time.
First of all, thanks for the detailed post with lots of interesting studies.
What I'm seeing from the first study is that there's a fairly weak negative correlation (estimated by the authors at -.38 after lots of statistical work) between g-loading and Flynn effect IQ gains over the period up to the 1990s, after which the Flynn effect probably ended in most developed countries.
That doesn't seem to show that g itself failed to increase substantially up to the end of the 20th century, it just suggests that the increase was somewhat lower than the increases in scores on various IQ tests would show. The article also notes the strong positive correlation between the g-loading of various tests and the racial IQ gap of each test, but doesn't do anything to link them.
I couldn't seem to track down Flynn's book chapter that ycj had linked to. Flynn's paper you linked does acknowledge that the gap in IQ can't be explained away by the Flynn effect alone, but the point is mostly that environmental changes can have effects that are of the same rough size as the black-white IQ gap. His overall conclusions seem to be that it is quite plausible that environmental differences that exist could explain the gap, but the Flynn effect itself doesn't explain them; it just demonstrates that an improved environment can increase IQ.
The gap does appear to have narrowed somewhat over the last few decades, although it is nowhere near vanishing. Dickens and Flynn found in 2006 that the gap had declined by 4-7 points from 1972 to 2002; Jensen responded by saying that they had improperly excluded some studies and made a dubious projection, and the real change was 0-3.44 points. I don't know enough to adjudicate that.
(As an aside, diving into this subject - I've read or significantly skimmed about a dozen papers in the past couple of days - reminds me of why I get frustrated with social science. Everything is too complex to really get a firm handle on, so everyone all the way up to the top researchers in the field cites the studies that back up their beliefs and discounts those that don't. There's still value in finding out where a field stands generally, like what things are generally accepted and what is controversial, but it's a pain to get there.)
So IQs are usually reported as the result of a single test, with or without subtests (and then normed to mean=100 and SD=15). That's how I thought it worked at first.
But I still don't get how exactly they determine g-loading. I get what g-loading is, but not how it's measured. I'm aware g was introduced by Spearman in the early 1900s to explain the correlations between results in different IQ tests, but since it is unmeasurable in and of itself, how do they determine which tests are most and least g-loaded?
Here I'm mostly trying to explain the difference between the really low (~70) results found by Lynn in many developing countries, and the higher results we find among developed country minorities. I don't believe that micronutrient deficiencies are a major cause of racial IQ inequality in the developed world, but they're really common in developing countries.
Most of the studies I've found have been messy in some way or another, with at least one thing that stood out to me as questionable, although I usually don't have the expertise to know which things are truly red flags and which aren't. A big gender gap like that would be unusual though.
On the other point, though, let's assume the race-IQ gap were entirely genetic, that whites have mean IQ=100 and blacks 85. The army goes through and conscripts a large number of people, but rejects the ones with IQ<85. The distributions, and thus the means, of the accepted recruits would still differ by race because the entire left half of the bell curve was cut off for blacks but only the left 16% of the white curve.
I don't know how to calculate the new means analytically, so I had Excel do it for me using about 2000 normally-distributed random numbers from two distributions, one with mean 85 and the other with mean 100, then rejected the ones from both "races" with IQ<80 (assuming that was the military cutoff) and took the means of the accepted samples. The mean of the accepted "whites" was about 102.5, that of "blacks" was about 94. So a gap of about 8.5 points would remain. At a cutoff of 85, the gap narrows slightly to 7.5.
The effect of removing the low IQ members of two distributions with different means therefore narrows the gap somewhat, but still leaves a little over half of the difference intact. Further, I'd imagine that we would expect some regression to the mean as well.
So the fact that they found no difference at all is a little bit of evidence against that gap being genetic. Although now that I look at their numbers, I doubt they could have detected a 4-point gap with their sample sizes (~8 points from the original black-white military gap, divided by 2 to account for the German mother). Maybe it really was just totally inconclusive.
(This sort of thing always happens when I look at a social science or medical study in depth. It takes an hour or two to really understand it and work through the details, only to find the evidence is too weak to really draw any firm conclusions one way or the other. Gah!)
I'm just throwing this in here because I can't find the random thoughts thread, but it's really jarring seeing all these people with very Slavic names who are like, literal Nazis. I can't help wondering, you know the actual Nazis would have viewed you as a subhuman and put you right in the gas chamber, right?
Moderator Action: stop the trolling - ori
Please read the forum rules: http://forums.civfanatics.com/showthread.php?t=422889
Sure, but they like the idea of a world where they're the ones in control of that gas chamber.
I suppose to be fair the Nazis did use plenty of Slavic collaborators. Unless I'm quite mistaken though I believe the long-term plan was for the physical extermination of the Slavic nations. The Nazis had already gotten a pretty good start on the Poles by the time the Red Army showed up.
They wanted to spare a few millions in Europe according to plan Ost, IIRC.
In general, yes, we've got our share of idiots. Some people just say we should have surrendered to them and join united and enlightened Europe. They would develop our economy and we'd drink German beer and eat sausages by now.
I find it interesting that those who are quite happily proclaiming incontrovertible proof that there are IQ differences between races do not (it seems to me) belong to the disadvantaged groups in this respect. I wonder whether racial identity is a significant factor in the propensity to believe this supposed fact.
Scientists have already found 538 genes that are linked to intelligence. All we'd need to do is check to see if they're evenly distributed.
So what does IQ measure? Test taking ability, or real intelligence? The answer is both. It is possible to train for IQ tests, but the gains are not g-loaded, and they fade. I'm not entirely sure what you think Lemon Merchant is wrong about, but it seems to me that the anecdote is pretty much in line with existing research on IQ.
It seems to me that Uppi was merely suggesting that raw natural talent can, to some extent, compensate for likeability or contacts. Your philosophizing about the nature of world-changing ideas is interesting, but I don't think that it is ultimately at odds with what Uppi said.
They found a small negative correlation. Meaning that as one goes up, the other goes down. There have been some studies suggesting that genotypic IQ has actually been going down. Of course, technically, that study only looked at genes linked to education, rather than IQ, but I would assume that they are related.
Ah, the narrowing of the gap. This is a source of constant debate. I've looked at a fair bit of studies (I'm thinking of a certain VOX article specifically). From what I've seen, adult samples show zero narrowing of the gap, whereas samples from children are subject to the Wilson effect (to varying degrees). I believe this might why you see such inconsistent results.
Welcome to the mudpit. My solution to this problem has been to debate the issue. I get all the relevant (and many irrelevant) papers thrown at me, and if I get something wrong, I get my stuff pushed in. Usually I learn something new in every debate.
If you want to learn about factor analysis, then I recommend you look somewhere else. But essentially, it's about which subtests are good at predicting overall performance. For example, if all the people who do well on the test also do well in Raven's progressive matrices, then RPM probably is a good indicator of overall intelligence. If the people who do well on the test have a huge amount of variance in their verbal subtest scores, then verbal subtest probably isn't a good indicator of overall intelligence.
Or, you could do what I do, and simply take the numbers as given. I'm certainly not going to redo all the maths.
Oh, in that case, of course. Usually, I'm happy to concede whichever number people want to use. Usually they prefer Wicherts et al. But yes, if Lynn's numbers are to be believed, then there definitely is a strong environmental effect going on as well.
Usually, I attribute big discrepancies to small-ish sample size. There is a way of calculating confidence intervals for sample sizes, but if the authors themselves don't do it, then I won't either.
I once did some back of the envelope maths myself. The cutoff point is 85. So assuming that the average black IQ back then was 85, that would mean that the lower half was cut off entirely. I calculated the median IQ for the remaining population to be around 92 or 93 (it was a long time ago, I don't remember the exact number). Also, I should have calculated the average rather than the median (I imagine that the average would be a little higher). So given this, the parents' average IQ would be (92,5+100)/2= 96,25. 96 is very well in line with what they got (that is, 97 IQ). Of course, perhaps it should be lower given the regression to the mean (my unofficial, unsourced rule of thumb is that IQ regresses about halfway between the parents' average and the racial average, in this case, it regresses towards 92,5). That would give us a figure of 94,375 (but then again, I used median IQ rather than the average). All in all, this kind of maths is riddled with all kinds of assumptions, and if Eyferth had 8 point discrepancies in his data, I'm not sure what difference a few IQ points make. It's also good to note that the mean age for the kids was 10, which would mean that the results are subject to the Wilson effect.
EDIT: Meant to say median instead of mean
Do you think we're all Asian?
Interesting, though somewhat predictable, that the fact that the in-group is not at the top of the pyramid is so readily trotted out to nullify the possibility of racial bias.
Also, as far as Kyriakos' hypothesis about sports ability goes, I suppose it is possible that there is some kind of correlation between IQ and athletic performance. But I would imagine that the correlation is small, and that it would be far out-weighted by athletic performance. Are there differences in athletic performance? Do we want to go there? Do we need to go there? I think not.
Well, if this all made up racial bias, then why not put the in-group at the top of the pyramid while making stuff up?
An issue is that there are differences in all phenomena which appear to be "the same". Eg there is no reason to think that your mind forms the notion of "1" (or any other) in a manner which is in complete tautology with how it is formed in any other person. This issue was known - and very heavily discussed - already in platonic philosophy, and apparently also by a few presocratics.
We simply do not identify such phenomena other than by their very external manifestation, so two people may even have varying degrees of complexity in how they form the same notion. And notions are just used as building blocks for other things, so the complexity is always increasing, along with the difference from individual to individual.
Imo it would make more sense to examine which person has a more complicated consciousness (which is not the same as trying to note who is more intelligent), rather than to claim that a very small bit of type of intelligence (eg for solving problems with math forms in iq tests) will provide one with a serious insight on who is more intelligent in a group.
Oh boy, you're diving into such deep levels of philosophy. I think I'll stick to IQ, rather than delve into this complicated consciousness. For all of it's flaws, IQ has a lot of research on it, and it is associated with many real world outcomes. It is, in a sense, more grounded.
It doesn't nullify it, but surely you would expect the effect of people "agreeing with a theory because it makes them look good" to be much smaller if these people aren't actually the "winners" of that theory.
woohoo, Ireland comes in tied for 49th
I feel so superior
Interesting East-Asian cluster. Japan and Mongolia are quite different in living conditions, culture and education, but probably close genetically.
Sierra Leone comes in at 91 but Liberia is at 67? I'm going to take that with a pretty big grain of salt.
The site says that this is based on Richard Lynn's work. He came out with substantially lower results for sub-Saharan Africa than other researchers (~70 compared to ~82). Also, he's on the board of the rather openly racist Pioneer Fund and is their primary contact - this suggests (but does not prove) that his research may be biased towards exaggerating the IQ gap.
Separate names with a comma.