Comment by katz | Hacker News Reader

Comment by katz | Hacker News Reader

Parent Full thread

> I would be more inclined to believe any differences in intelligence, if there really are any, are due to the environment in which the child was raised, rather than genetic advantages and disadvantages.

Intelligence measured by the general intelligence factor is highly heritable. The general intelligence factor is a factor analysis of intelligence and is a measure of “innate” intelligence and attempts at “boosting” this fails. The test that measures this the best is Raven's progressive matrices which is culture and language negative. The general intelligence factor is highly correlated with IQ and highly correlated with future success. The performance of a person in these tests has to do with the physical make-up of the brain (there are a lot of online articles by (dr/prof) Haider that is fairly readable that sheds light on this subject.)

> Here is a list I have compiled. IQ scores may be disputed, but the intelligence of these people generally are not.

A note about your cited IQs – it is probably not correct. The popular press like to “guesstimate” the IQ of historic persons – but this is of little value. Rigorous intelligence tests (such as IQ tests) is a 20th century phenomenon. Leibniz lived in the 17th century – well before these tests. So to try and determine any IQ or measure of intelligence from him is a waste of time.

Also – you collected only a point sample - I bet someone could get the same example with selecting a different group of famous persons. You also only selected the top 1% and not the general populations – so to draw assumptions of this sample to the general population would be wrong.

I do agree with you that the study is somewhat worthless. Here are several conflating factors:

People involved in menial labour (i.e. without tertiary qualifications) generally get children at a younger age. A lot of older financially secure men would generally marry younger less intelligent women (trade-up/Russian bride phenomenon). The age of child bearing of different social classes differs significantly.

Also - a three point IQ difference with all these conflating factors would probably fall under the heading "not statistically significant".

The test that measures this the best is Raven's progressive matrices which is culture and language negative.

I've attended a lecture by John Raven, the publisher of that test, and he doesn't make that claim.

The long, careful examination of what the Raven tests show, in conjunction with other evidence, can be found in James R. Flynn's excellent book What Is Intelligence?, which is about to come out in a new, expanded edition.

http://www.amazon.com/What-Intelligence-Expanded-PB-Beyond/d...

Also very good for reconsidering the importance of IQ tests is Keith Stanovich's What Intelligence Tests Miss.

http://yalepress.yale.edu/yupbooks/book.asp?isbn=97803001238...

http://www.amazon.com/What-Intelligence-Tests-Miss-Psycholog...

> I've attended a lecture by John Raven, the publisher of that test, and he doesn't make that claim.

I never said that he made that claim. Several other people compared RAPM and other intelligence tests to determine the g-loading. It was found by other people (not Dr. Raven) that RAPM has a high g loading. Testing for the general intelligence factor and the use of RAPM is widely accepted. Here is an example of a recent article: “Just one g: consistent results from three test batteries”, (Wendy Johnson, Thomas Bouchard, etc...) 2004. Using Google scholar you can find several other results. Even if you reject RAPM as a measure of g, there are less g-loaded tests (such as IQ) that can be used that predicts more or less the same thing (heritability of intelligence, etc...).

> The long, careful examination of what the Raven tests show, in conjunction with other evidence, can be found in James R. Flynn's excellent book What Is Intelligence?, which is about to come out in a new, expanded edition.

My problem with most of Flynn's books is that he is a political scientist and a passionate left-wing campaigner. He is basically a Gould-light – distorting science to fit his own world view.

I usually prefer to read papers or books on intelligence matters from people with a PhD in areas such as educational psychology or neurology when it concerns brain structure (e.g. prof. Haier).

My problem with most of Flynn's books is that he is a political scientist and a passionate left-wing campaigner. He is basically a Gould-light--distorting science to fit his own world view.

There are a lot of Ph.D. psychologists who disagree with that statement. Here is what Arthur Jensen said about Flynn quite a few years ago, and he hasn't retracted this statment: "Now and then I am asked . . . who, in my opinion, are the most respectable critics of my position on the race-IQ issue? The name James R. Flynn is by far the first that comes to mind." Modgil, Sohan & Modgil, Celia (Eds.) (1987) Arthur Jensen: Concensus and Controversy New York: Falmer. Here's what Charles Murray (all right, not a psychologist nor a geneticist) says in his back cover blurb for Flynn's book I mentioned in this thread: "This book is a gold mine of pointers to interesting work, much of which was new to me. All of us who wrestle with the extraordinarily difficult questions about intelligence that Flynn discusses are in his debt." As Mackintosh (1998, p. 104) writes about the data Flynn found: "the data are surprising, demolish some long-cherished beliefs, and raise a number of other interesting issues along the way." Flynn has earned the respect and praise of any honest researcher who takes time to read the primary source literature. Robert Sternberg, Ian Deary, Stephen Pinker, Stephen Ceci, Sir Michael Rutter, and plenty of other eminent psychologists recommend Flynn's research.

As for what we can conclude about the Raven Progressive Matrices test, one thing we certainly can't conclude is that it is not influenced by the cultural background of the test-taker. The test publisher says culture matters in Raven scores.

http://www.eyeonsociety.co.uk/resources/RPMChangeAndStabilit...

> are the most respectable critics of my position on the race-IQ issue?

The race-IQ issue is separate from the issue of general intelligence and neuroscience of the brain. Even if he is a successful political scientist, it would not make him an educational psychologist or neuroscientist. Much of the popularity of Flynn is his opposition to certain views of Intelligence - not his research.

> one thing we certainly can't conclude is that it is not influenced by the cultural background of the test-taker

With language and cultural negativeness I meant that there is no cultural specific references. One of the main complaints of the IQ test was that it contained western concepts and examples and language (i.e. it was culturally loaded). For example – the pictures in RPM are the same and contains the same meaning from whatever culture you are. Here is a quote from “How Biased are Culture-Loaded Tests?” (Jensen, A):

> Probably the best known and most widely used of such tests s Raven's Progressive Matrices (citations). Such nonverbal tests are expressly designed to reduce item dependence on acquired knowledge and to keep cultural and scholastic content to a minimum while getting at basic processes of intellectual ability.

Also note that that study did not just use RPM but included a verbal section.

The study you gave me showed differences in RPM between what they labelled “cultures” - other authors used other “labels” to try and explain the difference between groups. Here is a quote from the study you gave me:

> On the one hand, the absence of cross-cultural differences in RPM scores between cultures which do differ on a variable

Thanks for the reply. My friendly response to the suggestion that I should consider mostly what subject a person has earned a Ph.D. in as the basis for whom to believe on what subject is that sometimes good ideas for one subject come from people who learned another subject. Sometimes that's the best way to make progress in a subject that has become stuck in outmoded theories. I observe that Ph.D. psychologists

1) were taken by surprise by Flynn's research on IQ test score trends over time, having initially predicted precisely the opposite trend,

2) proposed very stringent standards for acceptable data before believing the trend was real (and Flynn credits Jensen, who has just the sort of credentials you have recommended, as being especially tough in demanding certain kinds of evidence),

3) reviewed his publications before allowing them to appear in the leading journals in the field of psychology,

4) invited Flynn to attend major conferences written up in specialized monographs with acknowledged experts in psychology,

http://www.amazon.com/Nature-Intelligence-No-233-Novartis-Fo...

5) praised his work in standard textbooks on human intelligence,

6) devoted whole monographs specifically to his findings,

http://books.apa.org/books.cfm?id=431712A

and generally accorded Flynn greater honor as a researcher on psychological phenomena than most psychologists gain in a lifetime.

http://www.cambridge.org/us/catalogue/catalogue.asp?isbn=978...

Anyway, scientific issues aren't decided by credentials, but by data. Flynn found and analyzed some very important data that caused a lot of psychologists to reexamine their assumptions. John Raven takes care to write, "The results surprise many psychologists. Eductive ability has turned out to be more easily influenced by appropriate educational and developmental experience than reproductive ability." In other words, what Jensen and others took as a "culture fair," "g-loaded" IQ test turns out to reflect modifiable educational experiences that children have or do not have.

I agree with you that my comparison of Flynn with Gould was not entirely fair. Flynn found the Flynn effect which is an important part. But a book by Flynn is not peer reviewed and may be influenced by his political views (he is author of “How to Defend Humane Ideals: Substitutes for Objectivity” after all).

> Anyway, scientific issues aren't decided by credentials, but by data.

This is true – but data in peer reviewed journals not books.

> In other words, what Jensen and others took as a "culture fair," "g-loaded" IQ test turns out to reflect modifiable educational experiences that children have or do not have.

The majority of psychologists believe that the g-factor is an innate and immutable property. The g-factor correlates with reaction times, glucose metallisation rate and is highly heritable. This indicates that the g-factor most likely have a neurobiological basis (see articles by Haier for example).

In a recent article the g-factor also appeared in a battery of 42 tests (Johnson,W Bouchard, TJ, “Just one g: consistent results form three test batteries”). Here is a quote from the abstract:

> The concept of a general intelligence factor or g is controversial in psychology. Although the controversy swirls at many levels, one of the most important involves g's identification and measurement in a group of individuals. If g is actually predictive of a range of intellectual performances, the factor identified in one battery of mental ability tests should be closely related to that identified in another dissimilar aggregation of abilities. We addressed the extent to which this prediction was true using three mental ability batteries administered to a heterogeneous sample of 436 adults. Though the particular tasks used in the batteries reflected varying conceptions of the range of human intellectual performance, the g factors identified by the batteries were completely correlated (correlations were .99, .99, and 1.00). This provides further evidence for the existence of a higher-level g factor and suggests that its measurement is not dependent on the use of specific mental ability tasks.

A definitive reply to most of that outmoded thinking about g can be found in Nisbet's new book,

http://www.psychologicalscience.org/convention/program_detai...

or in Stanovich's.

http://yalepress.yale.edu/yupbooks/book.asp?isbn=97803001238...

Stanovich is very clear on the distinction between "intelligence" (he is a mainstream psychologist in calling "intelligence" what I would simply call "IQ") and "rationality," an aspect of human cognition hardly tapped at all by most IQ tests. Stanovich comes at these issues from the framework of cognitive science, which also informs much of current behavioral economics. He previously wrote a textbook on understanding psychological research that has gone through a number of editions.

> A definitive reply to most of that outmoded thinking about g can be found in Nisbet's new book,

If you feel that these views are outmoded – give me peer reviewed research articles (not books). I do not have immediate access to books and books are not peer reviewed. The article that I referenced above was written in 2004 – you can find more recent articles in Google Scholar if you want to. I did not see that the above article (or any of the others) have been refuted. I doubt that you can refute thousands of research articles (a lot of them recent) and a discovery that is probably the hallmark of modern psychometrics with a simple reference to a book (calling anything that you oppose “outmoded”).

> Stanovich is very clear on the distinction between "intelligence" (he is a mainstream psychologist in calling "intelligence" what I would simply call "IQ")

Also I think you are confusing “intelligence” with the general intelligence factor (g) (also intelligence does not equal IQ. IQ is just a standardised intelligence test that is fairly well correlated with the g factor). The general intelligence factor is a factor that is common to almost all different tests on intelligence. While the general intelligence factor is immutable other parts of intelligence can be changed by training/learning (i.e. by the environment).

> "intelligence" (he is a mainstream psychologist in calling "intelligence" what I would simply call "IQ") and "rationality," an aspect of human cognition hardly tapped

The purpose of intelligence tests is to test intelligence – not factors such as rationality, emotion and inter-personal skills such as Stanovich's book suggest. The latter are all factors that is undoubtedly environmental and the value of such tests are doubtful to the intelligence debate.

Why is it people don't understand that all groups have different means given large enough N. Statistical significance has nothing to do with the magnitude of difference. Effect size is what you are looking for which looks at the magnitude of the difference in means taking into account the groups variance.

Also, 3 may well be statistically significant, and perhaps have a large effect size. This does not mean it is clinically significant.

Reply on news.ycombinator.com