Questions on _g_ and intelligence

[Nereid]

I don’t know --- 12 more pages of questions??

Mon ami Tigers2B1, pardonnez-moi, tu et si mignon, mais semblez si stupide! Have you not been reading the posts here? Can you not count the number of times that hitssquad and Mandrake have given conflicting answers, to the same simple questions??

 

[Tigers2B1]

No, I have a part-time job stomping out crop circles for The National Enquirer you know. Plus, my cranium so big it has to be wheeled around on a tricycle --- so just give me time.

 

[BlackVision]

Mandrake, you ever watch TV Land? On the TV Land channel there’s a program call Columbo, which features a private detective who catches his bad guy by simply using long lists of seemingly innocent questions – Columbo always seems confused --- in need of help. Then, near the end of the show, Columbo has this Got-Ya Moment when the ‘bad guy’ slips up on an answer to one of his voluminous questions.

Well, our Nereid is forever like Columbo – except he never has that Got Ya Moment.

For those unfamiliar with Columbo (but know Nereid) – I’ve linked this site ---

http://www.tvtome.com/tvtome/servlet/ShowMainServlet/showid-1011/Columbo/

This is ALL Nereid ever does for something that she has the inability to disprove. The infamous watermelon tactic as shown here:

https://www.physicsforums.com/showthread.php?p=207833&highlight=watermelon#post207833

 

[Nereid]

Intelligence appears to have at least some negative correlation with parity.

How much (range)? How extensively has this been tested?

I don't know. It is something that I have seen mentioned in various books and papers as part of the discussion of within family variance. Unfortunately, Jensen's books are rather poorly indexed (unlike The Bell Curve) so it is always difficult to find material in them.

Miller says that the levels of testosterone that children are exposed to may be related to parity. Presumably he meant the intrauterine environment.

Searching my files didn't work. The problems involve multiple meanings of the term "parity" and the difficulty in removing "disparity" from the search.

This is another of my questions that hasn't yet been answered (so I'm repeating it).This is another of my questions that hasn't yet been answered (so I'm repeating it).

Have you attempted to answer the question with the help of a search engine? It may be difficult, for the reasons I gave in my prior message. The topic is discussed in the literature, but is tedius to locate by searching. It was addressed in Storfer's book Intelligence and Giftedness, but does not appear in the index. The same is true with the other books I have. Most are not throughly indexed.

Let's see why this might be important.

Suppose there are two population groups, one in which all the children are only childs, the other in which every family has exactly two children. Suppose the 'parity effect' is very simple, the second child has a _g_ which is 1$$\sigma$$ below that of the first child (and first children in both groups have exactly the same _g_, on average).

If nothsuR, a well-known, if somewhat controversial, psychometrican, measures the average _g_ of children in each group, controlling for all the environmental variables she knows about (but not parity), what will she find? That the second group has an average _g_ 0.5$$\sigma$$ lower than that of the first group.

Now, suppose this 'parity effect' was only discovered after the MN study was completed (and Burt long since dead); how should honest, good scientist psychometricians respond?

So, let me repeat my question, and amplify: what is the size and nature of the 'parity effect'? In which countries has it been studied? When was the first case of it being characterised (close to the contemporary consensus view)? To what extent have historical data been reanalysed to account for this effect?

 

[Nereid]

This is ALL Nereid ever does for something that she has the inability to disprove. The infamous watermelon tactic as shown here:

https://www.physicsforums.com/showthread.php?p=207833&highlight=watermelon#post207833

Thank you for your insightful and perceptive contribution to the nature of the scientific method BV.

Perhaps you know the answer to some (one?) of my questions? How about this one:

In which countries did Cattell do his research?

 

[BlackVision]

In which countries did Cattell do his research?

You see this? See this is the difference between you and real scientists. Actual scientists do not expect their opponents to do their research. If you wanted to find out where Cattell did his research, why would you ask your opponent instead of finding out yourself? Now I trust you've read many scientific statements by scientists that attempt to refute one another. Do you ever see them doing this? Or do you see them try to present their own evidence and attempt to counter with statements. Yes statements not questions. I know this is not something you're used to but this is the common way in one carries a debate. Presenting evidence. In which it is something you rarely do. You can not win a debate this way.

 

[quantumdude]

You see this? See this is the difference between you and real scientists.

Knock it off right now. If you don't want to answer the question, then don't answer it.

 

[Mandrake]

So, let me repeat my question, and amplify: what is the size and nature of the 'parity effect'? In which countries has it been studied? When was the first case of it being characterised (close to the contemporary consensus view)? To what extent have historical data been reanalysed to account for this effect?

The entire contribution from all environmental causes, plus all errors, has a variance in the range of 20% to 30% (depending on age). The parity effect is quite small. If you are interested in finding the numerical value, I gave you a good reference. Please read Storfer's book. It will not take you any longer to find it that it would take me to find it.

If the other questions are interested enough to you that you actually want to know the answers, I trust that you can find them via the same mechanisms that I would use. If you take the time to look them up, I hope you will share your findings with the rest of us.

 

[Nereid]

The entire contribution from all environmental causes, plus all errors, has a variance in the range of 20% to 30% (depending on age). The parity effect is quite small.

Thanks Mandrake, this is helpful.

Just so that I'm quite clear though:
- this 20-30% 'environmental contribution + all errors' refers only to the US; per Jensen, insufficient work has been done to characterise these contributions - size and nature - in other countries
- 'environmental contribution' does not include factors such as illness, (head/brain) injury, 'drug taking', etc (more later)
- 'environmental contribution' does not include age (or, perhaps more accurately, _g_ is adjusted to include the observed, average age-related decline in general mental capabilities ... at least in the US); specifically, possibly irreversible declines (or advances) in _g_ due to 'lifestyle choices' are not included in the 'environmental causes'
- the 20% to 30% does include (an unquantifiable?) average 'practice effect'.

 

[Nereid]

If the other questions are interested enough to you that you actually want to know the answers, I trust that you can find them via the same mechanisms that I would use. If you take the time to look them up, I hope you will share your findings with the rest of us.

Thanks for the comment.

What I was hoping for was access to an expert, who can quickly point a novice to online material which presents good, neutral summaries of the current state of the findings of researchers, much as many PF members do for those who enquire with physics, astronomy, etc questions.

 

[Nereid]

Summary, (some) results so far

One of the things I am trying to get a better understanding of is the 'systematic errors' (to use the language of physics) that need to be accounted for - or at least identified - when reporting estimates of _g_ (or IQ) from a test given to a subject (or estimates from a meta-analysis of the reported results from many studies).

So far, thanks to Mandrake and hitssquad, I have learned that it's quite difficult to get a clear understanding of these systematic effects. Partly this seems to be because much of the research is reported in terms of correlations, and accurately translating these results into (say) 'IQ points', or 'delta _g_, measured in terms of some population standard deviation' is nigh impossible, and partly because the researchers have tried to control the test environment so as to avoid as many systematic effects as they could think of (so they don't test people who are obviously drunk, or ill, for example).

While I've not yet followed up on all the material which Mandrake and hitssquad have posted (and if it's not available online, I won't be able to), only two systematic effects appear to have been somewhat characterised: age and practice. The former seems it could be up to 30 IQ points (e.g. age 70 person vs age 20 person), the latter maybe 1 SD.

 

[Nereid]

Intention, and its effect on an estimate of an individual's _g_

or, a person is not buckyball!

While I don't think I've seen, or taken, any of the highly g-loaded tests (e.g. matrix relations), I imagine they take the form of a time-limited multiple-choice questionnaire.

It may be difficult to do 'better' in such a test simply by consciously wanting to do as well as one could, but it's certainly easy to do worse through conscious intent! For example, one could simply choose an answer at random. I'm told there's a whole global industry called 'sports psychology', which (among other things) seeks to give the sportsperson the best possible 'frame of mind' for their competition, event, or whatever. Hitssquad has told us that there is a well-known, significant 'practice effect' wrt _g_; I wonder if there is an equivalent to the sporting 'psyched to win'?

In sports you can certainly be 'psyched to lose', and you may be very conscious of some aspects of this (e.g. fear); maybe you could also be quite unconscious of an attitude or 'mental state' (sports psych speak) that predisposes you to lose (or do less than your best). Is there some equivalent wrt taking IQ tests? To what extent has this been studied?

Now, a reader may think "what a stupid question! why on Earth would anyone *want* to do less than their best in an IQ test?!" To some extent this objection is irrelevant; the point is that when researchers give tests to subjects, they assume the subjects will be trying their best; from another angle, it's misplaced - for example, if your personality predisposes you to poorly in 'IQ test environments' (e.g. fear), you may not be aware that (at some level) you are 'wanting' to do less than your best.

 

[hitssquad]

One-stop intellectual resources for differential psychology students

quickly point a novice to online material which presents good, neutral

The request as stated cannot be honored. To my knowledge, techniques have not yet been developed to objectively quantitatively measure in academic material the traits good and neutral.

summaries of the current state of the findings of researchers

The most thorough review of the most broadly professionally-accepted findings regarding the g factor is Arthur Jensen's 1998 book The g Factor. It is available online for a subscription fee (monthly, quarterly, and yearly subscriptions are available; the cost for the yearly subscription at the current discount rate works out to 30 cents USD per day).

The most thorough review of the most broadly professionally-accepted findings regarding the presence within the United States of bias in mental testing is Arthur Jensen's 1980 book Bias in Mental Testing. It is not available online.

 

[Nereid]

The request as stated cannot be honored. To my knowledge, techniques have not yet been developed to objectively quantitatively measure in academic material the traits good and neutral.

Thanks hitssquad

So, how about 'good' as in 'well-written, easily readable by an non-specialist', and 'neutral' as in 'presents both mainstream views of the field as well as alternative views and critiques from related fields'. Perhaps an example might help: within astronomy and cosmology, such a paper would cover not only the concordance models (including dark matter and dark energy/quintessence/cosmological constant), but also MOND, and even plasma cosmologies.

Within the sub-discipline of intelligence psychometrics, is there a tradition of review papers?

 

[hitssquad]

A note to the reader of The g Factor

quickly point a novice to online material which presents good, neutral

To my knowledge, techniques have not yet been developed to objectively quantitatively measure in academic material the traits good and neutral.

So, how about 'good' as in 'well-written, easily readable by an non-specialist'

Those are not quantities.

Jensen describes The g Factor in its Preface as readable by "typical ... college graduates":

• A NOTE TO THE READER
  
  Although much of the material in this book is admittedly, though unavoidably, at a fairly difficult conceptual level, I have tried to present it in such a way that it can be understood not only by specialized readers with a background in psychology, psychometrics, statistics, or behavioral genetics, but by any interested persons of whatever educational background whose reading comprehension is up to the level of what I presume is typical of college graduates. I had thought of providing a glossary of the more specialized terms, but discovered that nearly all of the entries I would have included are given quite adequate definitions in the Random House Unabridged Dictionary (Second Edition, 1993).

If you would like to judge for yourself whether the book suits your standards, Matt Nuenke has published http://home.comcast.net/~neoeugenics/jen12.htm in its entirety on his website.

and 'neutral' as in 'presents both mainstream views of the field as well as alternative views and critiques from related fields'.

Those also are not quantities.

I do not recall ever having read any academic publication by Jensen that did not include critiques and alternate views. Chapter 5 of The g Factor is devoted to "challenges to g," and Bias in Mental Testing is itself composed largely of critiques and responses to those critiques.

 

[Mandrake]

Just so that I'm quite clear though:
- this 20-30% 'environmental contribution + all errors' refers only to the US; per Jensen, insufficient work has been done to characterise these contributions - size and nature - in other countries

Since this same question appears in many of your messages, I previously noted that I will not attempt to research it each time it is asked. The value of h^2 has been researched and reported on by psychometricians from various countries. To the best of my knowledge, there are data sets from a number of countries that have been evaluated and reported. The comments in The _g_ Factor have not been restricted to findings reported in the United States, nor to research conducted by US scientists.

- 'environmental contribution' does not include factors such as illness, (head/brain) injury, 'drug taking', etc (more later)

The above statement is incorrect. These are micro environmental factors. Macro environmental factors involve social interactions; they have a near zero presence after age 17.

- 'environmental contribution' does not include age (or, perhaps more accurately, _g_ is adjusted to include the observed, average age-related decline in general mental capabilities ... at least in the US); specifically, possibly irreversible declines (or advances) in _g_ due to 'lifestyle choices' are not included in the 'environmental causes'

Psychometric data is usually restricted in such a way as to eliminate variations that would skew the data of interest. Age is a factor in many psychometric studies and is almost always discussed in the papers. It would be silly to pretend that age related factors have no relevance to psychometric studies.

- the 20% to 30% does include (an unquantifiable?) average 'practice effect'.

Most serious psychometric research is now reported in terms of _g_. Practice can change the _g_ loading of a conventional paper and pencil test, so that variable has to be accounted for. Passive measures of _g_ by RT and electroencephalography do not show a practice effect, but these studies usually include a familiarization phase, in which the test subject becomes familiar with the test procedure. All of this is explained in detail in The _g_ Factor.

 

[Mandrake]

While I don't think I've seen, or taken, any of the highly g-loaded tests (e.g. matrix relations), I imagine they take the form of a time-limited multiple-choice questionnaire.

The most highly _g_ loaded standard IQ test is the Raven's (there are three tests for different levels). The appropriate method of administering this test is with no time limit. Sometimes time limits are used and do not degrade the test, if the time is generous. When time limited tests are used, the _g_ loading necessarily decreases as the time limit is shortened. This is because _g_ loading is a thinking relation. As time becomes more and more important, the _s_ loading increases. People who already have knowledge of the answer can call it up faster than those who have to think about it.

Arguably the most central of all IQ tests is the WISC, which is not a group test. It must be administered by a psychologist; some parts are timed and some are not.

For example, one could simply choose an answer at random.

That could be done with some tests and not with others. Tests administered to one person by one psychologist are not subject to such guessing. The test administrator would quickly determine that the test subject was not being cooperative and would stop the test. Given the high dollar cost of such tests, someone would be very upset with this outcome.

In sports you can certainly be 'psyched to lose', and you may be very conscious of some aspects of this (e.g. fear); maybe you could also be quite unconscious of an attitude or 'mental state' (sports psych speak) that predisposes you to lose (or do less than your best). Is there some equivalent wrt taking IQ tests? To what extent has this been studied?

There has been a good bit of study of testing errors and factors that may influence scores. Much of Bias in Mental Testing is devoted to this topic. It was necessary to evaluate this subject in order to demonstrate that it was not the cause of blacks scoring lower than any other population group. Of course the same concern exists with respect to the differences between all other population groups, but political pressure applies only to blacks. Sophisticated analysis of the tests can show such important factors as rank order difficulty. If someone takes a test and answers correctly below a given level of difficulty and then begins to answer incorrectly when the difficulty is higher, the test is probably doing its job. If he misses easy questions with an equal frequency to mid-level or hard ones, he is obviously trying to skew the results downward.

 

[Nereid]

'parity', or 'family size and birth order'

Have you attempted to answer the question with the help of a search engine? It may be difficult, for the reasons I gave in my prior message. The topic is discussed in the literature, but is tedius to locate by searching. It was addressed in Storfer's book Intelligence and Giftedness, but does not appear in the index. The same is true with the other books I have. Most are not throughly indexed.

I spent a fair bit of time searching the internet for results of good research into the size of the 'family size' and 'birth order' effects on _g_ (or intelligence).

I found that:
a) research seems to have been done only in countries whose economies are service-based (e.g. US, western Europe)
b) findings have been confusing and apparently contraditory - some research found a clear effect, some found none
c) the research protocol has a big impact on what the results are; e.g. across a population (at one time) vs across time (following a clearly defined group)
d) none of the sources I found online gave quantatitive results (e.g. x IQ points, with an experimental error of +/- y); plenty talked about 'the larger the family size, the lower the average IQ', or 'the more older siblings, the lower the average IQ'.

 

[Nereid]

languages, language fluency, and _g_

Another area I researched was the relationship between language fluency and _g_ (or 'intelligence', or 'IQ'). I was interested in this because one's ability to develop a 'mother tongue' fluency in a (spoken) language becomes dramatically different for all but a fortunate few once one is past puberty. This suggests that there is some deep biological process at work in the brain, perhaps related to the processes that give rise to _g_?

I couldn't find anything! Apparently Jensen, Rushton, even the founders (Binet, Spearman) haven't researched this at all!

Now, I'll be the first to admit that my research has been purely online, and that there could well be a veritable cornucopia of research results that I didn't find.

hitssquad, Mandrake, BlackVision, nuenke - can you help please?

What I'm looking for is any differences in the estimated _g_ of those who are at least bilingual from their earliest childhood vs those who are monolingual, or who acquired a second language essentially post-puberty. I'm referring entirely to oral capability; literacy is a different dimension which I will explore in a later post.

 

[Nereid]

_g_ and literacy

The extent to which an individual's estimated _g_ is related to that individual's literacy (ability to read and write in their 'mother tongue') is a third area I researched (online).

The first thing I found was that such research would be difficult, if not impossible, in service-based economies; they all have 'universal education', and those who do not achieve functional literacy by adolescence (or puberty) are quite likely to have already known low _g_ (e.g. Down's syndrome, physical injury, severe drug addiction, mental illness).

Next, it seems that what little research has been done among non-literate adults, in agriculture-based economies for example, suffers from heavy biases, e.g. Rushton and Lynn's work.

Finally, there seems to have been no work done to estimate the relative _g_ impact of different writing systems (e.g. logographic vs alphabetic); indeed, the conclusion from studies which appear to show that the populations in Japan, Korea, China, Hong Kong, Singapore, Taiwan etc have a higher average IQ seems to have been 'they are genetically superior wrt _g_' rather than 'literacy based on a logographic writing system results in higher estimated _g_'

 

[Mandrake]

The extent to which an individual's estimated _g_ is related to that individual's literacy (ability to read and write in their 'mother tongue') is a third area I researched (online).

Literacy is presumably intended to mean "education." I have not found anything that suggests that education can raise or lower _g_. The _s_ loading of some tests is probably increased by education and may cause some measurable increase in IQ.

The first thing I found was that such research would be difficult, if not impossible, in service-based economies; they all have 'universal education', and those who do not achieve functional literacy by adolescence (or puberty) are quite likely to have already known low _g_ (e.g. Down's syndrome, physical injury, severe drug addiction, mental illness).

Within any society, there are people who get more or less education as a matter of circumstance. Not every high _g_ person is highly educated. We have discussed the cases of the people who were featured by Esquire as being among the smartest in the US. Two of the 4 hold doctorates. Two (to the best of my knowledge) hold no college level degrees. At least two are living at a level that may be fairly categorized as impoverished. Among those not listed, I know a good number (in Hi-Q societies -- way above Mensa) who have not earned college degrees. My point is that intelligence is not caused by education.

Next, it seems that what little research has been done among non-literate adults, in agriculture-based economies for example, suffers from heavy biases, e.g. Rushton and Lynn's work.

If one is studying sub-Saharan Africans, the cultures will be have little literacy and will be non-industrial. What are the "heavy biases" that you have found in their work?

Finally, there seems to have been no work done to estimate the relative _g_ impact of different writing systems (e.g. logographic vs alphabetic); indeed, the conclusion from studies which appear to show that the populations in Japan, Korea, China, Hong Kong, Singapore, Taiwan etc have a higher average IQ seems to have been 'they are genetically superior wrt _g_' rather than 'literacy based on a logographic writing system results in higher estimated _g_'

I don't know how much work has been done with respect to language. It is something to study and may have been studied. Brand has written a number of comments about deaf people, which I think may be related. For example:

... deaf children have entirely normal levels of performance on gf tests despite having missed much of the supposedly enriching and stimulating world of language and verbal communication; but, especially in childhood, they do have lower scores on gc tests requiring knowledge of language (Braden, 1994). In both cases, the tests are valid; but one type, gc, requiring normal verbal skills, registers - quite properly, and indeed quite fairly - a real handicap.(18)

... deaf children, despite their gross cultural deprivation, have no special problems with non-verbal tests that are well known as good measures of IQ.

Braden (op.cit.) especially considers the idea that minority children are handicapped in access to the ways of the 'dominant culture' - e.g. because their parents do not know it, do not like it, or anyhow cannot communicate it to their children; and thus that minority children will be deficient in the knowledge which is sometimes thought to be especially tapped by IQ-type tests. By such criteria, deaf children clearly have a massive handicap in accessing the 'dominant culture'; yet they have entirely normal levels of gf.
[Brand, C. (1996). The _g_ Factor: General Intelligence and Its Implications. Chichester, England: Wiley]

 

[Nereid]

Thanks Mandrake.

Since this same question appears in many of your messages, I previously noted that I will not attempt to research it each time it is asked. The value of h^2 has been researched and reported on by psychometricians from various countries. To the best of my knowledge, there are data sets from a number of countries that have been evaluated and reported. The comments in The _g_ Factor have not been restricted to findings reported in the United States, nor to research conducted by US scientists.

This conflicts with what hitssquad posted earlier, at least as far as Jensen is concerned. Also, in my (limited) online research, I've found that the well-controlled, lab-based, large-scale work seems to have been done almost exclusively in just a few countries, and that none come remotely close to the US in terms of quantity. Further, there are few, if any, pieces of in-depth research done in countries which are predominantly agricultural, or even industrial.

Of course, generalisation to all countries, all of Homo sap. is pervasive (again, with the notable exception of Jensen); detailed explanations for why such generalisations are justified scant.

'environmental contribution' does not include factors such as illness, (head/brain) injury, 'drug taking', etc

The above statement is incorrect. These are micro environmental factors.

I looked; no one seems to have attempted to study how estimates of _g_ would vary if the subjects were (for example) drunk (and quantify the effects by time and quantity of grog consumed).

Macro environmental factors involve social interactions; they have a near zero presence after age 17.

This statement seems to reflect research done in the US; there doesn't seem to be anything on these effects for people who reach adolescence at significantly different ages than US kids, nor who remain living with family (or an institution, e.g. priests, nuns) beyond adolescence.

Psychometric data is usually restricted in such a way as to eliminate variations that would skew the data of interest. Age is a factor in many psychometric studies and is almost always discussed in the papers. It would be silly to pretend that age related factors have no relevance to psychometric studies.

Quite. However, to repeat, my interest at present is systematic effects on estimates of an individual's _g_ obtained from a single test (or a series of tests done in one sitting). In this sense, your comment, while it may be accurate, is irrelevant.

Most serious psychometric research is now reported in terms of _g_. Practice can change the _g_ loading of a conventional paper and pencil test, so that variable has to be accounted for. Passive measures of _g_ by RT and electroencephalography do not show a practice effect, but these studies usually include a familiarization phase, in which the test subject becomes familiar with the test procedure. All of this is explained in detail in The _g_ Factor.

Jensen's? or Brand's?

I want to return to these 'passive measures', esp wrt the Flynn effect, but for now I merely note that few if any such detailed studies appear to have been done with folk other than young, literate urban adults in the US and a few other service-based economies. I also note that details of the research protocols - esp the extent to which pre-defined, double-blind ones are used, and non-manual measurement, are scanty (at least, I've not been able to find much online).

 

[Mandrake]

Originally Posted by Mandrake
Since this same question appears in many of your messages, I previously noted that I will not attempt to research it each time it is asked. The value of h^2 has been researched and reported on by psychometricians from various countries. To the best of my knowledge, there are data sets from a number of countries that have been evaluated and reported. The comments in The _g_ Factor have not been restricted to findings reported in the United States, nor to research conducted by US scientists.

Nereid: This conflicts with what hitssquad posted earlier, at least as far as Jensen is concerned.

Jensen is a US scientist and has done most of his work in California, presently under the rule of an Austrian.

Also, in my (limited) online research, I've found that the well-controlled, lab-based, large-scale work seems to have been done almost exclusively in just a few countries, and that none come remotely close to the US in terms of quantity.

The United States leads all countries in scientific research. Look at the Nobel Prizes in the sciences before and after WW2. There was a sudden shift from Germany to the US. I can't think of any reason to expect another country to do more research in psychometrics than would be the case in astronomy or particle physics.

Further, there are few, if any, pieces of in-depth research done in countries which are predominantly agricultural, or even industrial.

Huh? What is the point of that comment? It applies quite well to magnetic data storage, DNA research, space travel, and lasers. I don't get it. I think you have made similar comments before. Why?

I looked; no one seems to have attempted to study how estimates of _g_ would vary if the subjects were (for example) drunk (and quantify the effects by time and quantity of grog consumed).

Is there some reason why psychometricians should be looking at temporary phenomena? If someone goes to sleep, I would bet that his score on an IQ test would not be high, but that it would not be meaningful either. My prior comment was in reference to permanent environmental factors that relate to the micro environment. I could not believe that you were really asking about temporary impairment. Do you think that there is scientific research concerning human running ability? I assume there is. If so, would it be your concern that they should study runners with broken legs, drunk runners, and runners who have not eaten for six days?

Mandrake: Most serious psychometric research is now reported in terms of _g_. Practice can change the _g_ loading of a conventional paper and pencil test, so that variable has to be accounted for. Passive measures of _g_ by RT and electroencephalography do not show a practice effect, but these studies usually include a familiarization phase, in which the test subject becomes familiar with the test procedure. All of this is explained in detail in The _g_ Factor.

Jensen's? or Brand's?

As a convention, I assume that the reader understands that Jensen, A. R. (1998). The g factor: The science of mental ability. Westport, CT: Praeger, is the standard textbook in the field of psychometrics. When I wish to reference Brand's book, I do it by writing The _g_ Factor: General Intelligence and Its Implications.

I want to return to these 'passive measures', esp wrt the Flynn effect, but for now I merely note that few if any such detailed studies appear to have been done with folk other than young, literate urban adults in the US and a few other service-based economies.

This comment is not correct. Passive testing has been done in the US, Australia, New Zealand, England, and probably Germany and other countries. The IQ ranges of subjects reported by Jensen starts at around 40. Since retardation starts at 70, I do not agree that all of these subjects were literate. The service based economy thing is spurious. If you must keep including it, please explain your reason each time. Science is not dependent on whether the researchers live in a service based economy or not. I should add that you have not established that your claim is even true. I doubt that some of the contributing countries have service based economies. Would you please give us a complete list of the countries that have service based economies and some sound and verifiable information that shows your list is correct?

I also note that details of the research protocols - esp the extent to which pre-defined, double-blind ones are used, and non-manual measurement, are scanty (at least, I've not been able to find much online).

I have no idea what you are reading. The Intelligence 32-4 and 32-5 arrived in my mailbox today. A quick look through them (and any past issues) shows that your observation is false. Since you think that the US is alone in this area of science, you might be interested in the article in 32-5 by a large team of Chilean scientists who studied brain volumes, IQ, and SES. The paper includes a detailed description of the methodology and a massive presentation of the data collected, along with brain images. The subjects included both high and low SES people of both sexes. "Independently of sex, brain volume was the only brain parameter that contributed to explain IQ variance."

 

[Nereid]

Literacy is presumably intended to mean "education." I have not found anything that suggests that education can raise or lower _g_. The _s_ loading of some tests is probably increased by education and may cause some measurable increase in IQ.

Sorry I wasn't clear; I meant basic literacy, as opposed to 'illiteracy'; I specifically do not mean 'education'. IOW, basic literacy means being able to read simple signs (e.g. "STOP", "Walk"), numbers, and in the western tradition is what most kids reach by kindergarten. AFAIK, there are hundreds of millions, possibly billions, of adults who are illiterate; for example, despite years of propoganda, it seems that a significant proportion of the adult population in China's 'countryside' are illiterate or barely literate.

If one is studying sub-Saharan Africans, the cultures will be have little literacy and will be non-industrial. What are the "heavy biases" that you have found in their work?

It's not only there; the majority of such folk are in Asia, particularly in China and India. In fact, it's probably less than 50 years - not even two generations - since the vast majority of Homo sap. were living in economies that were primarily agricultural (of course, it's much longer for the US and western Europe, but they comprise <15% of the world's people).

I am getting to the heavy biases; patience please.

 

[Nereid]

Jensen is a US scientist and has done most of his work in California, presently under the rule of an Austrian.


The United States leads all countries in scientific research. Look at the Nobel Prizes in the sciences before and after WW2. There was a sudden shift from Germany to the US. I can't think of any reason to expect another country to do more research in psychometrics than would be the case in astronomy or particle physics.


Huh? What is the point of that comment? It applies quite well to magnetic data storage, DNA research, space travel, and lasers. I don't get it. I think you have made similar comments before. Why?

If all that you have is stamps (correlations), and the only stamps you have collected are from a few countries, how do you know that the rest of the world has stamps, or that they aren't made of diamond and not paper? How can you be confident that the biological correlates are the same for people whose life experiences are significantly different from those who you have studied?

Is there some reason why psychometricians should be looking at temporary phenomena? If someone goes to sleep, I would bet that his score on an IQ test would not be high, but that it would not be meaningful either. My prior comment was in reference to permanent environmental factors that relate to the micro environment. I could not believe that you were really asking about temporary impairment. Do you think that there is scientific research concerning human running ability? I assume there is. If so, would it be your concern that they should study runners with broken legs, drunk runners, and runners who have not eaten for six days?

To reiterate the point about correlations, if you have no idea how 'temporary phenomena' affect the arbitrary constructs you are studying, how can you tell that there aren't 'temporary phenomena' which are systematically biassing your results? Further, if you hypothesise that _g_ is something to do with brain processes, don't studies of impairment tell you a great deal? AFAIK, this is just how much progress has been made into the neurophysiology (?) of language.

To give a possibly irrelevant example; several decades ago smoking was widespread, and even then nicotine was understood to have neurological effects. Presumably Jensen et al tried very hard to control for this 'temporary phenomenon', but given the near ubiquity of the social habit, the persistence of the drug (and metabolites) in the brain, and people's imperfect veracity when it comes to reporting such habits, are you confident that the effect of smoking as a temporary phenomenon has been completely eliminated from old data?

This comment is not correct. Passive testing has been done in the US, Australia, New Zealand, England, and probably Germany and other countries.

All of which are service-based economies (see below)

The IQ ranges of subjects reported by Jensen starts at around 40. Since retardation starts at 70, I do not agree that all of these subjects were literate.

That's not my point (again, apologies for not being clear); the subjects have included few, if any, 'average' adults who just happened to be illiterate.

The service based economy thing is spurious.

Au contraire, mon ami! As I said above, extrapolating *correlations* found in one set of circumstances is contraindicated in good science; if all you have is correlations (and our discussion of the state of play wrt neuroscience and theory certainly seems to indicate that we've little else today), you need extremely good reasons to declare validity beyond the domains within which you obtained them.

If you must keep including it, please explain your reason each time. Science is not dependent on whether the researchers live in a service based economy or not.

It's not where the researchers live so much as where the *subjects* live.

I should add that you have not established that your claim is even true. I doubt that some of the contributing countries have service based economies. Would you please give us a complete list of the countries that have service based economies and some sound and verifiable information that shows your list is correct?

How about The World Bank? The link is to a page from which you can get employment by sector ('primary' = agriculture, fishing, animal husbandry, and mining; 'secondary' = industry; 'tertiary' = services), and also gives breakdowns by gender. The panel at the left allows you to navigate through a huge wealth of data.

Employment by sector (%; male/female; primary, secondary, tertiary):
US: 3/1; 32/12; 65/87
Australia: 6/3; 30/10; 64/87
New Zealand: 12/6; 32/12; 56/82
UK: 2/1; 36/11; 62/88
Germany: 3/2; 44/18; 52/80

Thailand: 50/48; 20/17; 30/35

Of course, these data need to be read in combination with the extensive notes; in particular, the agricultural sector for developing countries tends to be larger than the data would otherwise appear to indicate. On these World Bank pages, you will also find sector data, by economic activity (cf employment).

 

[Mandrake]

Mandrake:
Jensen is a US scientist and has done most of his work in California, presently under the rule of an Austrian.

The United States leads all countries in scientific research. Look at the Nobel Prizes in the sciences before and after WW2. There was a sudden shift from Germany to the US. I can't think of any reason to expect another country to do more research in psychometrics than would be the case in astronomy or particle physics.

Huh? What is the point of that comment? It applies quite well to magnetic data storage, DNA research, space travel, and lasers. I don't get it. I think you have made similar comments before. Why?

If all that you have is stamps (correlations), and the only stamps you have collected are from a few countries, how do you know that the rest of the world has stamps, or that they aren't made of diamond and not paper?

First, let me note that you did not address the issue of why it is significant, from your perspective, that the US leads all other nations in psychometric research. It does the same in most fields of science. Now, let's look at your comment above:
If all that you have is stamps (correlations)
This seems to imply that this is all that exists in psychometrics. We have discussed this topic before, but you have apparently forgotten it. Psychometrics is based on correlations AND laboratory measurements. I have previously listed the specific measurements and discussed some of them in detail. Suggesting that these don't exist comes across as very difficult to understand. Do you forget, or do you want to distort facts?
the only stamps you have collected are from a few countries

The stamp analogy is poorly considered and designed to be dismissive and misleading. The above comment is not correct. Psychometric data is collected and analyzed by scientists in a rather large number of nations. I have previously listed some and commented that the list of contributors to the journal Intelligence includes a large number of nations in each issue. The subjects of the studies include locals (from many countries) and citizens of other nations. If you really believe the things you are writing, you are misdirected to a point that your conclusions are highly likely to be incorrect (as is the case).
how do you know that the rest of the world has stamps

Stamp collecting is a canard. We know what research is being conducted in other nations by reading their published papers. I do. You apparently don't. Some of your comments indicate that you are following the discussion and asking good questions. But, then along come questions that have previously been discussed and answered.

How can you be confident that the biological correlates are the same for people whose life experiences are significantly different from those who you have studied?

Scientists study closed groups to eliminate variables. Within group studies were used to establish every point made in the first 12 chapters of The Bell Curve and the group in question was a single population group. When identical results are found between groups, there is a very high probability that the item being observed is not following different mechanistic processes in each group. You seem to be imagining how psychometric data has been collected and analyzed. I would like to suggest that a better approach would be to read what has been done, instead of guessing.

Mandrake:
Is there some reason why psychometricians should be looking at temporary phenomena? If someone goes to sleep, I would bet that his score on an IQ test would not be high, but that it would not be meaningful either. My prior comment was in reference to permanent environmental factors that relate to the micro environment. I could not believe that you were really asking about temporary impairment. Do you think that there is scientific research concerning human running ability? I assume there is. If so, would it be your concern that they should study runners with broken legs, drunk runners, and runners who have not eaten for six days?

To reiterate the point about correlations, if you have no idea how 'temporary phenomena' affect the arbitrary constructs you are studying, how can you tell that there aren't 'temporary phenomena' which are systematically biassing your results?

You are making the presumption that temporary impairment is not understood. You are wrong. I would expect that there are pertinent studies that address some forms of impairment. I have read studies that address stress as it applies to IQ testing. If you read discussions of testing procedures, you will see many discussions of factors that can adversely affect test results. Some of these are small effects and some are large. There is also a matter of common sense. It does not take a complex scientific study to determine that drunk people are not suitable for research on topics other than as they relate to drunkenness. You are imagining a bias that you have not bothered to study. The most detailed text on the subject is Jensen, A.R. (1980). Bias in mental testing. New York: Free Press. I have pointed you to this reference on several occasions.

Further, if you hypothesise that _g_ is something to do with brain processes, don't studies of impairment tell you a great deal?

Yes. The literature is rich with papers that address permanent impairment from disease and injury. Animal studies are important with respect to injury, since they include the intentional destruction of various brain regions, as I described to you in an earlier message. There may be literature on temporary impairment that extrapolates to some useful conclusions, other than that the person soul not be included in a psychometric study.

AFAIK, this is just how much progress has been made into the neurophysiology (?) of language.

What is "this?" You previously told us that neuroscience was "cool" to intelligence research. That is not true. So, I must conclude that your research into this topic should continue.

To give a possibly irrelevant example; several decades ago smoking was widespread, and even then nicotine was understood to have neurological effects. Presumably Jensen et al tried very hard to control for this 'temporary phenomenon', but given the near ubiquity of the social habit, the persistence of the drug (and metabolites) in the brain, and people's imperfect veracity when it comes to reporting such habits, are you confident that the effect of smoking as a temporary phenomenon has been completely eliminated from old data?

This is a good topic for you to research. I have not read any studies about it. What data have you seen that suggests intelligence impairment from smoking?

Mandrake:
This comment is not correct. Passive testing has been done in the US, Australia, New Zealand, England, and probably Germany and other countries.
All of which are service-based economies (see below)

The information you presented didn't list all of the countries where scientists are presently studying psychometrics. There is nothing to suggest that the science of psychometrics is actually a variable that is dependent on the number of people working in one sector or another. Psychometrics has been around for about a century. In that time the economic demographics of most advanced countries have changed. So what?

Recent studies reported in just the last two issues of Intelligence came from:
Spain, England, Chile, Ireland, Denmark, The Netherlands, Canada, Germany, Norway, Australia, France, USA, and Estonia. If you were to take the time to look back further, papers from other nations would be found. Science is science. It is practiced by nations that have enough intellect to pursue it. I doubt that you will find papers from Uganda, Kenya, Haiti, etc.

Mandrake:
The service based economy thing is spurious.

Au contraire, mon ami! As I said above, extrapolating *correlations* found in one set of circumstances is contraindicated in good science;

You have a great imagination. Populations have been studied from all over the world. Interestingly, there are studies of population groups that have lived together for centuries and studies of the same groups (genetically) within other countries (no admixture).

if all you have is correlations (and our discussion of the state of play wrt neuroscience and theory certainly seems to indicate that we've little else today)

You have not understood the discussions. We have studies based on chronometrics, electroencephalography, MRI, PET, fMRI, autopsy, and the measurement of biological parameters (such as brain pH).

Mandrake:
If you must keep including it, please explain your reason each time. Science is not dependent on whether the researchers live in a service based economy or not.
It's not where the researchers live so much as where the *subjects* live.

Yes. As we know, researchers have not confined themselves to their homelands, nor have they confined themselves to any particular type of economy.

 

[Nereid]

First, my thanks to Mandrake, hitssquad, Evo, Moonbear, SelfAdjoint and other posters for stimulating me to read more about this highly controversial* subject!

So, some tentative conclusions (so far). Perhaps in terms of Mandrake's opening post?

1 - Intelligence is best represented by _g_.
A definition, a convention, a shorthand, ... certainly bad science - tries to re-define a social construct (what Joan Public and Joe Sixpack mean when they use the word) with an abstract value obtained from statistical analysis of test results from many individuals.

2 - Virtually all of the external validity of IQ tests comes from their _g_ loading.
Circular; requires independent, objective definitions of 'IQ tests'

3 - What we know about _g_ is that it correlates strongly with various physiological conditions: nerve conduction velocity, pH, brain volume (and more specifically we now can see that particular areas of the brain are the actors and that their volumes correlate strongly with _g_), myelination, and information intake speed.
Hugely overstated; many correlations are only 'strong' wrt some found in social sciences; if examined through the lens of biology, they're weak (at best). Research results re 'particular areas of the brain' (etc) are, at best, early day results.

4 - These factors influence working memory which is now known (seen the most recent issue of the journal Intelligence) is predicted almost perfectly by _g_.
Overstated again? Statement is unclear in any case.

5 - All of the physiological measurements are seen between the population groups that are known to differ in mean IQ scores.
Vastly overstated; extensive physiological measurements, and their relationship to 'mean IQ scores' have been done in only a few countries. In this, and all the above, the results are *only* correlations (no biological hypotheses to test!)

6 - It is possible to measure _g_ by elementary cognitive tests (which are based on response time chronometrics), with a result that correlates as well with standard IQ tests as those tests correlate with each other.
Correlations have indeed been reported; the hypothesis that these correlations are universal ('apply to the mammal Homo sap. in general') has been tested in only a tiny minority.

7 - It is likewise possible to determine _g_ by electroencephalography using several different techniques and with similar accuracy.
Breath-taking extrapolation from very limited data.

8 - Both of these techniques are essentially passive, not subject to practice effects, and are totally blind to all social factors.
The first part may be accurate; the second doubtful; the third a wild assumption.

Generally, I have learned that those in the 'intelligence' field seem to reach for sweeping generalisations from their limited correlation results (Rushton and Lynn are particularly notable examples; Jensen an exception?). When faced with criticism, unlike E. O. Wilson (of sociobiology fame) who is reported to have said something like he clearly needed to do a great deal more work to establish his tentative findings, many in this field seem to have blindly carried on, finding yet more correlations (but doing little biology; there are exceptions, of course).

When compared to other fields which study how the brain works (relevance? I'd be astonished if any intelligence psychometricians claimed _g_ was *not* related to brain function!), the intelligence folk seem to have a curious reluctance to state and then test biological hypotheses. Contrast how the human vision system came to be understood (through studies of the 'visual acuity of nations'?), or how neuroscientists teased out the relationships between the three interacting sets of brain structures and the three core elements of language (crudely, semantics, phonemes, and grammar).

Or look at the extraordinarily limited basis for the sweeping claims of the genetic component. Twin studies? Only within a very narrow range of socio-economic conditions (how many New Guinea hunter gatherers? Mongolian nomadic herdsmen? Indonesian subsistance rice farmers?) which pertain to a tiny minority** of Homo sap. individuals, especially when the whole history of Homo sap. is concerned.

Some, e.g. Linda Gottfredson (Scientific American, 1998), acknowledge these limitations ("The foregoing findings on g's effects have been drawn from studies conducted under a limited range of circumstances - namely, the social, economic, and political conditions prevailing now and in recent decades in developed countries [e.g. almost all those in any list Mandrake has given] that allow considerable personal freedom. It is not clear whether these findings apply to populations around the world, to the extremely advantaged and disadvantaged in the developing world, or, for that matter, to people living under restrictive political regimes."

Reading much of this 'intelligence' literature, I was struck by the gulf between the claims (e.g. correlation does not imply causation) and the research directions (e.g. a race correlation is taken as genetic determinism).

*This is not a comment on the work of any scientist, or racist masquerading as a scientist, just an observation of the apparent public perception of the topic, at least in the US.

**

Recent studies reported in just the last two issues of Intelligence came from:
Spain, England, Chile, Ireland, Denmark, The Netherlands, Canada, Germany, Norway, Australia, France, USA, and Estonia. If you were to take the time to look back further, papers from other nations would be found. Science is science. It is practiced by nations that have enough intellect to pursue it. I doubt that you will find papers from Uganda, Kenya, Haiti, etc.

With the exception of Chile (population ~16 million) and Estonia (population ~1 million), but both of which are highly urban, all these are service-based economies, with near universal literacy, very low numbers of children per adult female, very low infant mortality, highly urbanised living, and so on. These conditions pertain to ~<15% of Homo sap.

 

[Mandrake]

part 1 of 3

So, some tentative conclusions (so far). Perhaps in terms of Mandrake's opening post?

Your comments strike me as further evidence that you are promoting a nihilistic agenda. Your conclusions are at odds with the findings of scientists who have devoted long careers to the understanding of psychometrics. When I see someone who is obviously unfamiliar with a subject attempting to dismiss it, I can only assume that the person in question is attempting to satisfy a personal or social agenda and is not attempting to fairly evaluate or discuss the topic.

I would be interested in learning your personal explanation of how it happens that you have reached conclusions that are opposite to those we find in respected scientific journals? How is it that you managed to do this so quickly? Does this indicate incompetence on the part of the people who have devoted their careers to the subject of psychometrics?

I will not bother to comment on the 8 points you dismissed, but will simply say that my reading of the literature leads me to the conclusion that your dismissive analysis is incorrect.

When compared to other fields which study how the brain works (relevance? I'd be astonished if any intelligence psychometricians claimed _g_ was *not* related to brain function!), the intelligence folk seem to have a curious reluctance to state and then test biological hypotheses.

This seems to imply that you have familiarized yourself with the literature to the point that you know this to be true. When you insert terms, such as "biological hypotheses," you imply that there is some definition of this term that excludes the various biological hypotheses that have appeared in the literature and which have been discussed here. Are you trying to play tricks with words? I have previously discussed Miller's myelination hypothesis. Why do you dismiss it? There are also various hypotheses that pertain to working memory, nerve conduction velocity, T1 and T2 relaxation times, and brain volume.

For system models, we have previously discussed the one that appears in at least three of Jensen's publications (including The _g_ Factor) and the one that appears in Brand's The _g_ Factor: General Intelligence and Its Implications. Now that we have a lot of research focused on specific locations of the brain, we have a location-specific model. See Gray, J. R., Chabris, C. F. & Braver, T. S. Neural mechanisms of general fluid intelligence. Nature Neurosci. 6, 316–322 (2003).

 

[Mandrake]

part 2 of 3

Or look at the extraordinarily limited basis for the sweeping claims of the genetic component. Twin studies? Only within a very narrow range of socio-economic conditions (how many New Guinea hunter gatherers? Mongolian nomadic herdsmen? Indonesian subsistance rice farmers?) which pertain to a tiny minority** of Homo sap. individuals, especially when the whole history of Homo sap. is concerned.

The above comment is a good stimulus for a discussion about the total body of evidence. Nereid's comment would lead an unsuspecting, uninformed person to the conclusion that the heritability of intelligence is based on one category of study (twins). If this were so, then her comment might merit attention. But, those who are familiar with the big picture know that the heritability of intelligence has been established and quantified by various independent means. When all vectors are pointing the same way and have the same magnitude and are determined by different categories of study, the combined certainty of the total observation is greater than even the sum of the components.

In the case of heritability, the most overwhelming evidence is from inbreeding depression. There is no other explanation for it than genetic heritability. Inbreeding depression studies not only show the expected effects, but the depression fits the magnitude of h^2 that has been determined by other means. Nereid apparently has forgotten this fact, since it has been mentioned before and is well documented in the literature. If she had bothered to read the full range of what is known about the heritability of h^2, I assume she would have discussed inbreeding depression and noted that it cannot be explained by any environmental factors.

I have mentioned path analysis on several occasions. The method of path analysis has been used to determine h^2 and produces results that are virtually identical to the other methods. Naturally, neither path analysis nor inbreeding depression depends on twins as subjects.

Besides these statistical methods, brain comparisons have been made for MZ and DZ twins using MRI. The previously referenced Gray and Thompson study shows the heritability of brain regions via MRI. ["We were stunned to see that the amount of gray matter in frontal brain regions was strongly inherited, and also predicted an individual's IQ score," said Paul Thompson, Ph.D., the study's chief investigator and an assistant professor of neurology at the UCLA Laboratory of Neuro Imaging. "The brain's language areas were also extremely similar in family members.]

It is also significant (but ignored by Nereid) that MZA studies have been carried out by a variety of researchers, at widely separated times, with different cohorts, and in different countries. The results of the studies have been remarkable for their tight consistency.

Among other factors that support the primary findings are the very high heritability of inspection time performance and the 85% heritability of brain volume. [The brain volume factor is now known to be particularly important at specific locations -- Richard Haier, previously discussed.] These factors can and have been determined independently of intelligence measurements, but are known to correlate with intelligence to such a degree that they have become central to the investigation of cognitive functioning.

Throughout science, the total body of evidence has to be taken into account as an indication of what is and what is not important. In this case, we have broad and consistent evidence from diverse observations, all pointing to the same answer. Yet, somehow, Nereid has elected to dismiss all of this. Why? I cannot understand how someone can imagine that they can find answers without effort and to have confidence that they understand the issues, without first considering the full range of facts.

There is yet another way to determine the genetic component of intelligence. It is to observe the variance that is associated with environmental factors. One way this has been done is via adoption studies. We have discussed this at length and I have pointed out that the "shared environment" component exists only in children and vanishes by late adolescence Bouchard and McGue show that the shared environment component is zero by age 12 and rises to over 80% by age 18). The remaining environmental variance is due to micro environmental factors, which amount to about 15% in adults. The error component, per Brand, is about 10%. The remainder is due to genetics.

Related to the above ...
Genetic and Environmental Influences on Human Psychological Differences
Thomas J. Bouchard, Jr., Matt McGue
Wiley Periodicals, Inc. J Neurobiol 54: 4–45, 2003
We conclude that there is now strong evidence that virtually all individual psychological differences, when reliably measured, are moderately to substantially heritable.

Additional evidence concerning the heritability of intelligence comes from the constant _g_ gap between US blacks and whites. Since intelligence data covers a relatively long period of time, one would expect that if _g_ is not primarily heritable, there would be a narrowing of the black vs. all other groups gap. During the past century blacks have enjoyed a huge gain in standard of living and especially in college education, yet the gap remains. There are but two choices: genetics and environment. The evidence is very strong that the predominant factor in the difference between various population groups is traceable to genetics. Even when adopted by white families, black children mature to have mean intelligence values that match those of their genetic peers who were not adopted. Likewise, Asian children mature to match the intelligence means of their genetic peers (which happens to mean that they have higher mean intelligence than their adoptive families -- opposite to the case of blacks adopted by whites in the US and Europe.

 

[Mandrake]

part 3 of 3

Some, e.g. Linda Gottfredson (Scientific American, 1998), acknowledge these limitations ("The foregoing findings on g's effects have been drawn from studies conducted under a limited range of circumstances - namely, the social, economic, and political conditions prevailing now and in recent decades in developed countries [e.g. almost all those in any list Mandrake has given] that allow considerable personal freedom. It is not clear whether these findings apply to populations around the world, to the extremely advantaged and disadvantaged in the developing world, or, for that matter, to people living under restrictive political regimes."

There is good reason to conclude that the same factors that relate to intelligence in the US or Europe also relate in other nations. Within the US, we have large populations of the major population groups and they have been studied for nearly a century. Most of the research has been within-group. When a mechanism is found in group A and in group B and both are seen to act identically, it is not an unfair extrapolation to conclude that the differences in mean values between the groups is subject to the same factors as the differences that are found within group.

While there may well be environmental factors that account for some of the large difference between sub-Saharan Africans and US blacks, there is little to suggest that these factors can be large enough to erase the difference. As much as you would apparently like to discredit Lynn and Rushton, I see no reason to believe that you have a better understanding of this science than these two people who have devoted their lives to actual measurements.

Reading much of this 'intelligence' literature,

Much? How much do you define as "much?" My impression is that you have not yet read enough to even connect the dots on a single psychometric issue, such as heritability. Have you read "much" of the literature on intelligence?

I was struck by the gulf between the claims (e.g. correlation does not imply causation) and the research directions (e.g. a race correlation is taken as genetic determinism).

I would like to suggest that you use quotes from the literature (you have read "much" of it) instead of making assertions that are actually your thoughts. It is impossible to discuss specific issues when one is offering only vague and unsupported assertions. Just tell us what some respected psychometrician has written, tell us his name, and then let us discuss whether or not that person has been scientifically naïve, or not.

Originally Posted by Mandrake
Recent studies reported in just the last two issues of Intelligence came from:
Spain, England, Chile, Ireland, Denmark, The Netherlands, Canada, Germany, Norway, Australia, France, USA, and Estonia. If you were to take the time to look back further, papers from other nations would be found. Science is science. It is practiced by nations that have enough intellect to pursue it. I doubt that you will find papers from Uganda, Kenya, Haiti, etc.
With the exception of Chile (population ~16 million) and Estonia (population ~1 million), but both of which are highly urban, all these are service-based economies, with near universal literacy, very low numbers of children per adult female, very low infant mortality, highly urbanised living, and so on. These conditions pertain to ~<15% of Homo sap.

There are three major racial groups and a number of identifiable population groups. The major groups have been studied extensively and are well understood with respect to the parameters that consistently appear within and between groups. It is, as you should know, not true to imply that groups in various parts of the world have been studied, but it is correct that they have not been studied as extensively as has been the case for population groups within Europe and the US. If one simply writes off all groups that have not been extensively studied, nothing changes with respect to the knowledge of psychometrics.

Meanwhile, there remains one fact that cannot be disputed and that is that population groups having very low mean intelligence scores have performed as a group in a manner that can be projected by IQ scores. We have been through this line before and it is beyond argument. Likewise, population groups that have high mean IQs have shown accomplishment that is in line with high intelligence.

 

[selfAdjoint]

I have read Nereid's post twice, carefully, and I just wnat to make one comment. It is an old story for hard scientists to show contempt for the .3 - .7 correlations considered highly significant in the human sciences - this happens throughout these sciences, not just in the IQ area. But that is what nature gives the human scientists, and when they do their math carefully, the difference between correlations of +.3, 0, and -.3 is perfectly clear, and meaningful. Yes it isn't quantitative at the level that astronomical evidence is, but that doesn't mean it's a hash.

Well, I said only one comment but here's another. Nereid, can you confirm that you really studied the modern fMRI-g experiments? Your dismissal of them didn't seem to describe them as I understood them from reading the papers.

 

[Tigers2B1]

It’s my understanding that g is a construct used to describe the phenomenon witnessed when a person makes similar scores on differently constructed, but complex tests of mental ability. Like the construct ‘energy’ that is used to describe events in physics --- while a physicist may have never seen an ‘energy,’ the idea / construct has value and describes the witnessed phenomenon. So, for example, if a person scores a 140 in a spatial relations subgroup in one test, that person is very unlikely to score an 80 in the numerical reasoning subgroup – even though spatial relations and numerical reasoning may have no other relationship outside of the complexity of the problems. The construct g describes this. It describes the ability to ‘figure out’ complex problems. The same phenomenon was observed inter-test -- in that the same person is very unlikely to score a 140 on the Raven Matrix IQ test and score an 80 on another test, if that test uses mentally complex problems - like complex pattern problems. The harder the solution is to find for the average test taker, the more likely the solution calls on reasoning abilities that can translate to other mental abilities and to other tests of mental ability.

In Bias in Mental Testing, p. 250, Arthur Jensen summarizes the construct this way:

By examining the surface characteristics of a great variety of tests in connection with their g loadings, we may arrive at some descriptive generalizations about the common surface features that characterize tests that have relatively high g loadings as compared with tests that have relatively low g loadings. Today we have much more test material to examine for this purpose than was available to Spearman more than half a century ago. This permits broader generalizations about g than Spearnan could safety draw. Spearman characterized the most g-loaded tests essentially as those requiring the subject to grasp relationships—“the eduction of relations and correlates.” That is all perfectly correct. But now we can go further. The g factor is manifested in tests to the degree that they involve mental manipulation of the input elements (“fundaments” in Spearman’s terminology), choice, decision, invention in contrast to reproduction, reproduction in contrast to selection, meaningful memory in contrast to rote memory, long-term memory in contrast to short-term memory, and distinguishing relevant information from irrelevant information in solving complex problems. Although neither the forward nor backward digit-span test of the Wechsler Intelligence Scale, for example, has much g loading. the slightly greater mental manipulation required by backward than by forward recall of the digits more than doubles the g variance in backward as compared with forward digit span (Jensen & Figueroa, 1975). We have seen many examples in which a slight increase in task complexity is accompanied by an increase in the g loading of the task. This is true even for the most mundane and seemingly nonintellectual tasks. Virtually any task involving mental activity that is complex enough to be recognized at the commonsense level as involving some kind of conscious mental effort is substantially g loaded. It is the task’s complexity rather than its content that is most related to g.

 

[Nereid]

I have read Nereid's post twice, carefully, and I just wnat to make one comment. It is an old story for hard scientists to show contempt for the .3 - .7 correlations considered highly significant in the human sciences - this happens throughout these sciences, not just in the IQ area. But that is what nature gives the human scientists, and when they do their math carefully, the difference between correlations of +.3, 0, and -.3 is perfectly clear, and meaningful. Yes it isn't quantitative at the level that astronomical evidence is, but that doesn't mean it's a hash.

Hmm, I'm not sure what a 'human science' is, but fMRI ... 'race' (biological definition (whatever that is), not a social construct) ... frequency of alpha brain waves ... ECT ... latency and amplitude of evoked brain potentials ... hereditability (genetic, not social or cultural) ... rate of brain glucose metabolism ... brain volume ... are not terms one commonly finds in sociology texts I would guess. PF members Moonbear and Phobos have commented on how 'hard' biology is.

Maybe 'intelligence science' just got a whole lot 'harder'?

 

[selfAdjoint]

fMRI has done nothing to refute, and everything to confirm, the concept of g. And yes I have given you references on that before. Let's not just rehash the whole thread all over again. Frequency of brain waves does not yet bear on the issue AFAIK. Heritability is supported by extensive careful twin studies which show those .3 - .7 correlations you don't like. And I do remember your attempts to undercut those studies, which, I am afraid I have to say, reminded me of nothing so much as the attempts of posters on the Astronomy forums to undercut the Hubble expansion with cherry-picked observations. What if this, what if that, we can't accept the studies till they have answered every question I can think up. Sorry for the rough speech but it comes honestly.

 

[Roberth]

The old story again?

I have come across this debate by chance (my main interest is physics) and in a way I am sad about it. It does strike me, however, and therefore I will comment.

What happened to the old and widely correct saying:

"IQ tests measure how well you do on IQ tests!"

Rather than going into a long essay, I will simply make a few points.

1) There has never been a consensus what 'intelligence' actually means. The one that I can best subscribe to is Piaget's. To paraphrase him: "intelligence is what you use when you don't know what to do".

2) IQ tests measure convergent intelligence and there are only a very few tests around which measure divergent intelligence (the closest to creativity). Creativity is at least as, if not more important, than analytical thinking, verbal abilities and all the other classical IQ test areas.

3) There are no good tests around that measure abstraction skills. This ability is definitely one of the key factors that distinguishes a mere high IQ person from an "intellectually intelligent" person. The same is true for people with average intelligence.

4) The most viable numbers on the nature-nurture debate are 60% to 80%. This can mean a difference between an IQ of 78 and 130!

5) Has anybody ever come up with a test measuring the ability for 'deeper insight'? This is a long-term cognitive ability but related to EQ. Many people would agree that this is a strong indication of 'intelligence'. I am not convinced that many people with high IQs are particularly good in this area. Conversely, low level IQ people can certainly have the quintessential deeper insight, 'wisdom'.

6) The practicing side is widely underestimated in IQ tests. Langauge abilities fall in this category. For example, I do not think that Arnold Schwarzenegger would either in English nor in German (his mother tongue) achieve 85% of the result that he achieved when he was 20 (before he moved to the US). His English skills never reached his mother tongue level and his mother tongue level has deteriorated since he moved to the US, 30 years ago.

7) There have been several racial crossing studies which show that there is no significant difference in IQ between races (eg. Eyferth 1961, Tizard 1972, Scarr and Weinberg 1976)

etc.

The major question of this thread, however, remains an ethical one, particularly when it comes to cross-racial studies. I love science and, after physics, psychology has always been my favorite area (I have some formal education and also used it professionally).

The questions are: What is the purpose of these cross-racial IQ studies? Why are people doing them? What knowledge do these studies add that is of any value to anybody?

I cannot see any potential benefits of these cross-racial studies that would outweigh the damage that they create in society.

Roberth

"Now, winning the Nobel prize is one thing, but winning it with an IQ of only 125, that is really something!" - Richard Feynman