Cranial Size and Intelligence Link

[Nereid]

Head size is a relatively-easy-to-measure proxy for total brain size. Total brain size, in turn, is a proxy for the sizes of the specific parts of the brain most associated with g. As we zero in on the parts of the brain most closely associated with g correlations between variance in volume and variance in g grow larger and larger, approaching unity.

• In frontal brain regions, a regionally specific linkage has
  previously been found39 between g and metabolic activity measured
  by positron emission tomography (PET), suggesting that
  general cognitive ability may in part derive from a specific
  frontal system important in controlling diverse forms of behavior.
  Frontal regions also show task-dependent activity in tests
  involving working (short-term) memory, divided and sustained
  attention, and response selection40. Genetic factors may therefore
  contribute to structural differences in the brain that are
  statistically linked with cognitive differences. This is especially
  noteworthy, as cognitive performance seems
  to be linked with brain structure in the very
  regions where structure is under greatest genetic
  control (Figs. 2 and 3). This emphasizes the
  pronounced contribution of genetic factors to
  structural and functional differences across individuals,
  as detected here in frontal brain regions.

(http://www.loni.ucla.edu/~thompson/thompson.html

 

[Nereid]

The human population was bottlenecked at one point no? I recall that from somewhere. If anyone has more info on this please share among us. I could be wrong though.

Genetic evidence suggests that all humans today are descendants of a relatively small group of early Homo sap. individuals. However, it certainly doesn't follow that Homo sap. almost became extinct at one point! For example, there may have been many, many other individuals at the time ... they appear to have no descendants among us, but we have no way of knowing why (e.g. maybe they were all killed 10,000 years later, by our ancestors?)

 

[hitssquad]

More correlates of g

Some of the stamps we'd put in our 'dluoG stamp album' (physical or biological characteristics NOT apparently correlated with _g_) might be:
- (spoken) language ability

This might be classifiable as verbal ability. The verbal scale of the WISC (Wechsler Intelligence Scale for Children) (the Wechsler tests consist of two main scales, each composed of several subtests, and the other scale being the performance scale) might serve as a suitable proxy for language ability:
b

• the items of the Verbal and the Performance scales of the Wechsler Intelligence Scale for Children (WISC) correspond rather closely to Gc and Gf types of tests, respectively. In the national standardization sample, a general factor extracted from just the Verbal scale subtests is correlated .80 with the general factor extracted from just the Performance scale subtests. The average g loading of the Verbal and Performance scales is therefore √.80 = .89. This is almost as high as the reliability of IQ on the WISC, and correction for attenuation would bring this average g very close to unity.

(Arthur Jensen. The g Factor. p126.)


And part of the verbal scale is a vocabulary subtest:

• Lynn, Wilson, & Gault, 1989. A study by Deary (1994b) obtained correlations between the nontimed Seashore pitch discrimination test (PD), auditory inspection time (AIT), the Mill Hill Vocabulary test (MHV), and Raven's Standard Progressive Matrices (SPM) on 108 thirteen-year-old students. (The score on AIT was a factor-score composite of two different methods of measuring AIT.) The g loadings (here represented by the first unrotated principal component) of these four variables are as follows: PD = .36, AIT = .72, MHV = .77, SPM = .78.

(Arthur Jensen. The g Factor. pp263-264.)

• The Vocabulary + Block Design composite of the WISC-R has the highest correlation with the WISC-R Full Scale IQ of any other pair of subtests, and both Vocabulary and Block Design are highly g loaded.

(Arthur Jensen. The g Factor. p500.)

- personality

All personality correlates that I have seen correlation numbers for show non-zero correlations with g. A fanous example from the Big 5 is C, trait conscientiousness. It has been found to have a consistent and moderate negative correlation with g:

• http://dx.doi.org/10.1016/j.paid.2003.11.010
  
  Conscientiousness (Control) was significantly negatively correlated with abstract reasoning (fluid intelligence), but not with verbal reasoning (crystallized intelligence). This was interpreted as indicating that the negative relationship between intelligence and Conscientiousness is due to fluid intelligence affecting the development of Conscientiousness, in an educated and need-achieving population.

I'll run through the rest of the list from memory and mostly without citations (Plus or minus sign mean correlates with IQ or g positively or negatively, respectively):

• body mass (not stature) . . . (Body mass index (BMI)) - (more muscle and less fat = higher IQ)
  
• eye colour . . . (lighter eyes, especially blue) +
  
• hair colour . . . (lighter hair, especially blonde) +
  
• skin colour . . . (lighter) +
  
• baldness . . . +
  
• blindness . . . + (though I need to do more research on this) (blindness may correlate positively with IQ because blind persons produce more melatonin, a powerful antioxidant with activity in the brain)
  
• deafness . . . (I don't know yet; there seems to be a lot of research on this question, though, and Nathan Brody has a recent research article on it)
  
• thickness of nails on right big toe . . . (thickness of nails should be positive, since this would fall under health indices)
  
• sleep patterns . . . +
  • Author
    Busby, Keith A; Pivik, RT.
    Title
    Sleep patterns in children of superior intelligence.
    Source
    Journal of Child Psychology & Psychiatry & Allied Disciplines. Vol 24(4) Oct 1983, 587-600.
    Blackwell Publishing, United Kingdom
    Abstract
    Examined the relationship between superior intellectual functioning and physiological patterns and events during sleep in 8-12 yr old males. Six males with superior IQs (WISC-R) and 5 males with average IQs were recorded for 5 consecutive nights of sleep using standard EEG measures. Compared to normal controls, superior IQ Ss had greater amounts of total sleep time, Stage 2, Stage 3, total NREM sleep, a longer average NREM cycle length, and significantly less average REM density. In addition, significant negative relationships were obtained between Full Scale IQ and REM density and between Verbal IQ and REM density. Results suggest that patterns and amounts of sleep stages in superior-IQ children do not differ in any dramatic fashion from those of children with average IQ. However, the negative correlations between IQ measures and eye movement density during REM sleep are consonant with previous notions relating eye movement density to waking information processing strategies and suggest a carry-over of such strategies from wakefulness to sleep.
• (clinical) depression . . . (proneness to depression) +
  
• epilepsy . . . -
  • Author
    Joinson, C; O'Callaghan, FJ; Osborne, JP; Martyn, C; Harris, T; Bolton, PF.
    Title
    Learning disability and epilepsy in an epidemiological sample of individuals with tuberous sclerosis complex. [References].
    Source
    Psychological Medicine. Vol 33(2) Feb 2003, 335-344.
    Cambridge Univ Press, US
    Abstract
    Intellectual impairments are a recognized feature of tuberous sclerosis complex (TSC), but the frequency and degree of intellectual impairments has not been systematically studied in large epidemiological samples using standardized measures. As such, the form of the IQ distribution (uni- or bi-modal) has not been established and the relationship between IQ and other features (e.g. epilepsy history) is poorly delineated. To address these shortcomings, we assessed the intellectual abilities of a large epidemiological sample of individuals with TSC, drawn from the 'Wessex' area of SW England and compared them with the abilities of their unaffected siblings. Standardized tests were used to estimate the abilities of 108 (56 males, 52 females, median age = 25, range = 4-75) individuals with TSC and 29 unaffected siblings (14 males, 15 females, median age = 18, range = 6-55). Seizure history was obtained from informants and medical records. Estimated IQ was bi-modally distributed: 55.5% had an IQ in the normal range; 14% had mild to severe impairments; and 30.5% had profound disability (IQ < 21). Forty-four per cent of the individuals with TSC had an IQ < 70. In the subset of normally intelligent individuals with TSC, IQ was normally distributed with a mean of 93.6...
• schizophrenia . . . -

• [Biological correlates prove] that g is not just an artifact of the way psychometric tests are constructed, nor is g a mere figment of the arcane mathematical machinations of factor analysis.

(Arthur Jensen. The g Factor. p138.)

I don't know if the 'prove' is Jensen or hitssquad

The paraphrase biological correlates prove was a rewording of it also proves. Here is the full paragraph:

• First, psychometric tests were never intended or devised to measure anything other than purely behavioral variables. Constructors of IQ tests, in fact, have tried to eliminate any source of test item variance that might reflect individual differences in physical attributes such as muscular strength and sensory acuity. Certainly there has never been the least intent that mental tests should reflect any strictly anatomical or physiological variables, which are directly measurable by other methods. It would therefore be most surprising and remarkable if IQ tests were significantly correlated with physical variables. Yet they are. IQ – especially the g factor of IQ tests – is correlated with a variety of physical variables. What does this mean? For the time being, about all one can say with certainty is that whatever is measured by IQ tests –mostly g – is somehow enmeshed in a host of organismic variables and therefore involves something beyond the purely psychological or behavioral. It also proves that g is not just an artifact of the way psychometric tests are constructed, nor is g a mere figment of the arcane mathematical machinations of factor analysis. Obviously, a correlation between psychometric g and a physical variable means that g is somehow connected with underlying biological systems.

(Arthur Jensen. The g Factor. p138.)

 

[BlackVision]

Genetic evidence suggests that all humans today are descendants of a relatively small group of early Homo sap. individuals. However, it certainly doesn't follow that Homo sap. almost became extinct at one point! For example, there may have been many, many other individuals at the time ... they appear to have no descendants among us, but we have no way of knowing why (e.g. maybe they were all killed 10,000 years later, by our ancestors?)

Humans Came Close to Extinction

A new hypothesis about recent human evolution suggests that we came very close to extinction because of a "volcanic winter" that occurred 71,000 years ago.

Some scientists estimate that there may have been as few as 15,000 humans alive at one time.

The volcanic winter lasted about six years. It was followed by 1,000 years of the coldest Ice Age on record.

It brought widespread famine and death to human populations around the world. It also affected subsequent human evolution.

This was because of a so-called bottleneck effect. The rapid decrease, in our ancestors' populations, in turn, brought about the rapid differentiation - or genetic divergence - of the surviving populations.

...entire article here:

http://news.bbc.co.uk/1/hi/sci/tech/166869.stm

 

[Mandrake]

part 1

Indeed, all parts of the brain would seem to be related to biology, which is partly why I'm quite puzzled that there is, AFAIS, no biological theory of intelligence.
Mandrake: I don't understand your observation. What do you think would constitute a biological theory of intelligence? How does that differ from existing models?
Nereid: If you're referring to the nice stamp collection of 'biological correlates', then that hardly constitutes a 'biological theory.'

If you think a stamp collection is a fair analog to the information known in the field of psychometrics, you might want to reconsider. Stamps in a collection do not tell us anything about other stamps. They do not fit together to show an overall picture. Psychometrics is a science and the knowledge base that it has produced over 100 years is extensive. Just as in any other science, the big picture continues to clarify as investigational tools improve. Consider what has been found in astronomy in the past 50 years. Things progressed slowly until deep space probes and Hubble suddenly provided mountains of new information. Psychometrics is now benefiting from sophisticated laboratory instrumentation and techniques. A better (but still low quality) analogy would be that the individual pieces of information in any science are somewhat like the pieces of a jigsaw puzzle. The shape of one gives us a hint about the shape of another. When joined, the image from several pieces is additive.

Perhaps this succinct statement is closer? "Jensen (1993), as well as others, synthesized these facts and conjectured that "the most obvious hypothesis is that speed of information processing is the essential basis if _g_, and one possible neurological basis of speed of processing is the speed of transmission through nerve pathways" (p. 54).

I happen to be familiar with the source of this comment, but it does not appear on page 54 of the paper that I know in which he presented his hypothetical model. What is the source you have in mind? Why did you give the year, but nothing else? Jensen published several things in 1993. Will you please give the full reference?

The speed of information transmission can be reasonably well measured or extrapolated from reaction time scores. Therefore, if an individual has faster neural processing speed, then he or she have a better reaction time. In turn, given that reaction time is highly correlated with IQ, then those individuals with faster neural processing speeds have higher IQ's. Consequently, neural processing speed determines the level of intelligence of the individual; this intelligence is the one general intelligence, _g_." If so, then where are the studies which show that estimates of _g_ correlate with well known factors affecting reaction time (e.g. drowsiness, illness, drunkeness)?

What? Do you think that scientific research in psychometrics is conducted by idiots? Do you really think that they include subjects who are impaired by obvious factors? I am amazed by this comment. What is your reason for this? Have you ever read papers on chronometric measurements? I am absolutely in disbelief that you cannot separate mental impairment with the foregoing comments.

There may be some special interest in RT as it pertains to impaired people, but most examples I can think of would include MT.

Similarly, one would expect that those who are good at pingpong, or fencing, would have elevated _g_ (reaction time is a key factor in excellence in these sports); and that southpaws smarter than normal folk.

You are badly confused.

RT measurements are typically based on a simple laboratory apparatus, which consists of a home button and multiple response buttons.[1] The device used by Jensen consists of a home button and 8 response buttons, arranged at equal distanced from the home button in an arc of 180 degrees.[2] Each response button has a light directly above it (or, in later devices incorporated lighted buttons). The person being tested holds the home button down and then must press the button closest to the light that is illuminated as the “external stimulus.” The measurement device records the time the stimulus lights goes on; the time the home button is released; and the time the response button is pressed:

T1 ………. T2 ………. T3

The time interval from T1 to T2 is the reaction time, RT. The time from T2 to T3 is the movement time (MT).[3] This simple test is an example of an Elementary Cognitive Task (ECT) and can be completed by any adult (even with IQs as low as 15 to 20), usually in less than 1 second. Individuals with IQs below 40 require some acclimation to the test. Jensen’s tests covered individuals from IQ 15 to 150. There is essentially no correlation between MT and intelligence.

[1] Multiple response buttons are used for reaction time testing known as choice reaction time. There is an IQ correlation to even the most elementary RT testing, known as simple reaction time. Simple reaction time is measured by having the test subject release a button when he is presented with an external stimulus. Discrimination reaction time is a variant of simple reaction time which also uses one button, but requires the test subject to release the button only when the stimulus matches a predefined condition.

[2] Other researchers have used the same measurement devices. The topic of RT testing is discussed at length in numerous places. Bias in Mental Testing is now nearly a quarter of a century old, but has a good discussion. For a more recent reference, see The g Factor.

[3] Some psychometricians define RT as the total time from the start of the test to the end and divide that time into two components, designated “decision time”(DT) and MT. In this case, RT = DT + MT. Jensen designates RT as the first component and does not bother to discuss the sum.

The thing you are describing is MT.

However, a biological theory of intelligence should do more than just conjecture;

A theory resides (among scientists) between a hypothesis and a natural law. An example is the Theory of Relativity. An example of a law is Ohm's Law. It is, in my opinion, scientifically incorrect to imply that a hypothesis is a theory. At present, we have good understandings of various components of intelligence, such as the fact that the variance in the speed of information intake is fundamental to the variance in intelligence. Your quote ignores that we also know that the standard deviation of RT correlates negatively to _g_ and that this correlation is independent of the mean RT correlation. I have never seen a single explanation as to why both happen. It is apparent that there are two factors at work (working memory efficiency and neural noise).

If you are interested in another model of information processing (besides the one Jensen reported in 1993), please see Figure 3 in Chapter 4 of Brand, C. (1996). The _g_ Factor: General Intelligence and Its Implications. Chichester, England: Wiley.

 

[Mandrake]

part 2

Nereid: On this basis, one could naively say that there is no biological theory of intelligence, because "speed of information processing" doesn't seem to have much to do with head size (for example), but _g_ does.

"Naively," is an appropriate comment. There are many other factors that are reasonably well understood and which relate to intelligence. The speed of information processing is one factor, but I believe that Brand's wording "information intake speed" is much more in line with what is being measured. I do agree that there is no present model of how the brain works that accounts for everything that is known about the variance in intelligence. I don't think we are even close to having a complete model.

On alternative theories, I've already come across "Fluid and Crystallized Intelligences (Cattell, 1971)", the "Structure of the Intellect model (Guilford, 1967)", "PASS Theory", "emotional intelligence", "implicit theories of intelligence", and a cryptic reference to "alternative perspectives (Rea, 2001; Ritchhart, 2001)".

Do you personally think that these are real theories, or are they more appropriately designated as hypotheses?

Originally Posted by Nereid
(e.g. how many 'personality psychometricians' quote, or do concurrent research in, 'intelligence'?).

Mandrake: You have mentioned these people on several occasions. I am unaware of their names. Can you tell us who you have in mind? Where do they publish? What are the standard textbooks that they reference? What else can you tell us about them? Thank you.

Nereid: I'm still looking; so far all I've found is a very few results on the relationship between personality and intelligence (none). I'm still flabbergasted that two groups, both psychologists, both calling themselves psychometricians, apparently don't use tools developed by the other group, ignore each other's work, even though the seat of everything they study - the human brain - is the same! It's almost as if condensed matter physicists were ignorant of the work of plasma physicists.

I have not previously seen any reference to psychometricians who work in the area you claim. The only ones I have encountered study intelligence and the relationships between intelligence and other factors.

**************************

Originally Posted by Mandrake
Is it true that skull size at birth is proportional to brain volume in adulthood? Since the difference in brain volumes that accounts for the variance is relatively small in adults, why would you extrapolate that this would cause so much increase in size of the head in a baby that it would result in more difficult child birth? Can you give us some verifiable numbers to support your suggestion?

Nereid: You raise an interesting point Mandrake ... the research that shows a correlation between brain volume and _g_, was that done with young adults? young children?

If you have been following this thread, you know that I have already posted a reference that directly answers the volume versus age question: J. P. Rushton, 1997, Intelligence 25, P. 15. He shows that the brain volume for the three primary racial groups are 315, 332, and 335 cm^3 at birth. He also gives the values for 4 months, 1 year, 7 years, and adults.

To what extent did the researchers tap into other research, on changes in brain volume of Homo sap. from birth to adulthood? To what extent did these brain volume studies attempt to measure the volumes of the different brain structures?

Please read the Rushton reference. Modern studies of brain volumes are based on MIR and fMIR. I have previously posted a reference to the study of localized parts of the brain by Richard Haier.

While we're at it, what is the nature of the change in brain volume as a person grows (from birth to adulthood), and decays (from young adulthood to extreme old age), does anyone know? If there are changes, to what extent are they uniform across all brain structures?

This paper is has the answer:
INTELLIGENCE 21, 109-119 (1995)
Aging, Brain Size, and IQ
ERIN D. BIGLER
STERLING C. JOHNSON
CARLOS JACKSON
DUANE D. BLATTER

It shows that brain size begins to decline by age 26 and continues monotonically and in almost a straight line. Likewise, the decline in performance scale scores begins at the same point and is almost a perfect straight line. Verbal scores increase in males after age 35 into the 70s. Verbal scores for females decline monotonically, but slightly,after age 50.

 

[Mandrake]

Thus my questions about the correlation coefficient (which do not seem robust to me). Plus, what confidence interval are we talking about? (95%? 50%?)

I have posted this before. Here it is yet again:

Murray and Herrnstein: "A crucial point to keep in mind about correlation coefficients, now and throughout the rest of the book, is that correlations in the social sciences are seldom much higher than .5 (or lower than -.5) and often much weaker -- because social events are imprecisely measured and are usually affected by variables besides the ones that happened to be included in any particular body of data. A correlation of .2 can nevertheless be "big" for many social science topics. In terms of social phenomena, modest correlations can produce large aggregate effects. Witness the prosperity of casinos despite the statistically modest edge they hold over their customers."[2]

It is possible to measure IQ as well with chronometric tests as it is with traditional IQ tests. That is, the chronometric results correlate as well with IQ tests as IQ tests correlate with other IQ tests. This happens, even though individual response time (RT)[3] measures correlate from -.2 to -.4, sometimes less. But, when a battery of these tests are given, the end result is a correlation of up to 0.745.[4] The point is that even relatively small correlations produce variances which, in some situations, are additive. If you make enough measurements of additive components, the net measurement can be significant.

Another example of small, but meaningful and robust correlations is inbreeding depression. This phenomenon is observed to affect numerous traits, including IQ and is consistently mentioned in psychometric texts as one of the most indisputable proofs of the strong genetic component of intelligence. When inbreeding is very close (siblings or parent-child), the effect is quite large; but most studies are based on first cousins, where effects on physical traits are typically .05 sigma to .10 sigma.[5]

[2] The Bell Curve, page 67

[3] RT is defined as the time required by the brain to respond to an external stimulus.

[4] Jensen. The g Factor, p229

[5] Jensen reports a number of studies pertaining to inbreeding depression in his book The g Factor. See the chapter titled “The Heritability of g.”

 

[Nereid]

If you think a stamp collection is a fair analog to the information known in the field of psychometrics, you might want to reconsider. Stamps in a collection do not tell us anything about other stamps. They do not fit together to show an overall picture. Psychometrics is a science and the knowledge base that it has produced over 100 years is extensive. Just as in any other science, the big picture continues to clarify as investigational tools improve. Consider what has been found in astronomy in the past 50 years. Things progressed slowly until deep space probes and Hubble suddenly provided mountains of new information. Psychometrics is now benefiting from sophisticated laboratory instrumentation and techniques. A better (but still low quality) analogy would be that the individual pieces of information in any science are somewhat like the pieces of a jigsaw puzzle. The shape of one gives us a hint about the shape of another. When joined, the image from several pieces is additive.

I'll be coming back to this, never fear.

I happen to be familiar with the source of this comment, but it does not appear on page 54 of the paper that I know in which he presented his hypothetical model. What is the source you have in mind? Why did you give the year, but nothing else? Jensen published several things in 1993. Will you please give the full reference?

Yep, Paik, August 1998 (I thought I'd put a source link in my post, but clearly I hadn't).

What? Do you think that scientific research in psychometrics is conducted by idiots? Do you really think that they include subjects who are impaired by obvious factors? I am amazed by this comment. What is your reason for this? Have you ever read papers on chronometric measurements? I am absolutely in disbelief that you cannot separate mental impairment with the foregoing comments.

I answered a similar question in your thread in SS; basically, if you're working primarily from correlations, you need to study - in some detail - how known impairment factors influence your data collection; the impairments will very likely generate systematic errors, and you can't control for them if you don't know how they influence your measurements. To take an analogy, AFAIK a modest consumption of alcohol contributes to keeping heart attacks at bay (I'm sure many PF readers could give us much more accurate statements), zero intake and your risk goes up; too much and it also goes up. However, the relationship is not linear. Now, wrt _g_, it may be that being the tiniest bit drunk helps (your measured _g_ goes up), and there's a plateau over a considerable range of intake. Or it could be a slightly elevated temperature at the onset of a cold really helps, or maybe even 0.1^o change in body temp produces a quite marked change in measured _g_, an effect no doubt too small for most researchers to worry about.

A more important reason for wanting to study how impairment affects your desired variable is to understand how _g_ works. If you're happy with correlations, no worries; if you want to develop a biological theory of _g_, then one way to get a handle on it is see how factors such as illness impact it.

RT measurements are typically based on a simple laboratory apparatus, which consists of a home button and multiple response buttons.[1] The device used by Jensen consists of a home button and 8 response buttons, arranged at equal distanced from the home button in an arc of 180 degrees.[2] Each response button has a light directly above it (or, in later devices incorporated lighted buttons). The person being tested holds the home button down and then must press the button closest to the light that is illuminated as the “external stimulus.” The measurement device records the time the stimulus lights goes on; the time the home button is released; and the time the response button is pressed:

T1 ………. T2 ………. T3

The time interval from T1 to T2 is the reaction time, RT. The time from T2 to T3 is the movement time (MT).[3] This simple test is an example of an Elementary Cognitive Task (ECT) and can be completed by any adult (even with IQs as low as 15 to 20), usually in less than 1 second. Individuals with IQs below 40 require some acclimation to the test. Jensen’s tests covered individuals from IQ 15 to 150. There is essentially no correlation between MT and intelligence.

[1] Multiple response buttons are used for reaction time testing known as choice reaction time. There is an IQ correlation to even the most elementary RT testing, known as simple reaction time. Simple reaction time is measured by having the test subject release a button when he is presented with an external stimulus. Discrimination reaction time is a variant of simple reaction time which also uses one button, but requires the test subject to release the button only when the stimulus matches a predefined condition.

[2] Other researchers have used the same measurement devices. The topic of RT testing is discussed at length in numerous places. Bias in Mental Testing is now nearly a quarter of a century old, but has a good discussion. For a more recent reference, see The g Factor.

[3] Some psychometricians define RT as the total time from the start of the test to the end and divide that time into two components, designated “decision time”(DT) and MT. In this case, RT = DT + MT. Jensen designates RT as the first component and does not bother to discuss the sum.

The thing you are describing is MT.

No. To be a good pingpong player, or fencer, you need to do more than just be quick on MT; you also, in your paradigm, need to be very sharp on RT. If you get a chance, watch a slow motion video of a good pingpong game sometime, both effects are clearly observable.

A theory resides (among scientists) between a hypothesis and a natural law. An example is the Theory of Relativity. An example of a law is Ohm's Law. It is, in my opinion, scientifically incorrect to imply that a hypothesis is a theory. At present, we have good understandings of various components of intelligence, such as the fact that the variance in the speed of information intake is fundamental to the variance in intelligence. Your quote ignores that we also know that the standard deviation of RT correlates negatively to _g_ and that this correlation is independent of the mean RT correlation. I have never seen a single explanation as to why both happen. It is apparent that there are two factors at work (working memory efficiency and neural noise).

So, in my stamp collecting analogy, if I find three completely unrelated variables which have statistical relationships similar to those between RT, _g_, etc I can make a psychometric theory from them?

 

[Nereid]

Do you personally think that these are real theories, or are they more appropriately designated as hypotheses?

Personally? I continue to be amazed at what seems to have been done in psychology; some of it is downright spooky, some looks like stamp collecting, and some I simply can't make head nor tail of. Wrt the ones in my post, somewhere between speculation and idea - want to ask me what I personally think of Jensen's _g_?

But, as I said to BV, it doesn't matter what I personally believe; science rolls onward with or without Nereid.

I have not previously seen any reference to psychometricians who work in the area you claim. The only ones I have encountered study intelligence and the relationships between intelligence and other factors.

Moonbeam started a thread on this topic, in SS; she found over 29,000 webpages with just a simple google search.

 

[Phobos]

A correlation of .2 can nevertheless be "big" for many social science topics.

Yep, that's what I was asking about earlier. Thanks, M.
It's shockingly low compared to other areas of science, but I can see how social science studies would be tough (hard to control/account for the variables).

 

[humanino]

I would like to report that we found a new and very interesting result : we need further investigations, but maybe a correlation between intelligence and using glasses. Sounds weird.

 

[hitssquad]

Myopia and IQ

we need further investigations, but maybe a correlation between intelligence and using glasses.

Try PubMed; a search for the terms <IQ myopia> returns 13 hits, including the following:

• CONCLUSIONS: Nonverbal IQ may be an independent risk factor of myopia, and this relationship may not be explained merely by increased reading (books per week) among myopes. An interesting observation is that nonverbal IQ may be a stronger risk factor for myopia compared with books read per week. The complexity of the relationships between nonverbal IQ, reading, and myopia warrant additional studies to clarify any cause-effect relationship.
  PMID: 15326105
  
  
• The objective of the investigation was to confirm the authors' experience and data from the literature that refractory myopia is as a rule associated with higher intelligence. In the first part one of the authors evaluated a group of 14-year-old myopic children with hypermetropic children of the same age. For evaluation she used the intelligence quotient (IQ) and assessment by teachers before the children left elementary school. The results are clearly in favour of the myopic children. In the second part the authors evaluated the intelligence of 15-18-year-old myopic secondary school pupils as compared with their classmates. Assuming that myopia is usually associated with higher intelligence and makes thus more profound education possible, they compared the students with apprentices trained to be cooks and waiters. The myopic students won again. Among students there were 36.8% myopic, among apprentices only 8%. The myopic students have in general better progress and better marks in mathematics than their classmates without refractory defects (1.86:2.07 and 2.14:2.39). Based on the results of the investigation the authors assume that they were able to confirm data in the literature that myopic students are on average more intelligent that their peers of the same age.
  PMID: 7586044
  
  
• With a paired study, 23 factors were investigated in 204 school children aged 9-14 in Taiyuan. The results of analysis of logistic regression showed that there was a close relationship between juvenile myopia and TV-watching distance, myopia in parents, Zn: Cu ratio in hair, N score of EPQ, Pscore of EPQ, and verbal IQ.
  PMID: 1782828
  
  
• At age 11 both those with myopia and with pre-myopia had increased verbal and performance IQ, while those with hypermetropia had slightly reduced verbal and performance IQ, in comparison with the children without refractive errors.
  PMID: 3234604
  
  
• The results of a postal questionnaire distributed to British members of Mensa failed to confirm an association of superior intelligence with torsion dystonia, retinoblastoma, or phenylketonuria, but were consistent with real associations between high IQ and infantile autism, gout, and myopia. Further confirmation of these findings in other populations might well indicate that genes producing these disorders have more or less direct effects on cerebral development and function.
  PMID: 7334499
  
  
• A school population has been screened to locate same-sexed twins with myopia and also to compare intelligence test performance of myopic and nonmyopic individuals. Augmentation of the twin data by a survey of the world literature has led to the identification of a total of 106 MZ twin pairs, 100 of them concordant for myopia, as well as 41 DZ pairs, 12 concordant. Myopic students score eight points higher on IQ tests than nonmyopes, the entire bell shaped distribution curve being shifted to a higher range. The intellectual gain precedes in time the development of nearsightedness.
  PMID: 1036378

In addition, Arthur Jensen reports (in his 1998 book, The g Factor):

• Myopia and IQ (Intrinsic). It has long been known that myopia, or nearsightedness, is related to high IQ. The evidence, reviewed elsewhere,[16] is based on many studies and huge samples. In terms of correlation the r is about +.20 to +.25. Myopia is highly heritable and a single gene that controls the shape of the eyeball has been identified as mainly responsible. Myopia in adolescents and adults can be predicted by ocular examination in infants as young as one year of age.
  
  The "near-work" hypothesis that myopia is solely caused by excessive use of the eyes for "near-work" such as reading, sewing, and the like has been largely discredited by modern researchers. Major chromosomal anomalies, such as trisomy 21 (Down's syndrome), which override the effects of the normal polygenic causes of individual differences in mental ability and result in severe mental retardation, militate against reading and most other forms of "nearwork." Yet the incidence of myopia among persons with these conditions is the same as in the general population. Also, myopia has high heritability. As myopia is a continuous trait, it appears that an interaction between a genetic predisposition and at least some slight degree of engagement in "near-work," such as most schoolwork, during childhood are necessary to produce a degree of myopia in adolescence or adulthood that calls for corrective eyeglasses.
  
  Individual differences in degree of myopia and in IQ are positively correlated in the general population. Children in classes for the intellectually gifted (IQ > 130), for example, show an incidence of myopia three to five times greater than the incidence among pupils in regular classes.
  
  The question arises of whether the relation of myopia to IQ is an intrinsic or extrinsic correlation (as defined on pages 139-40). The correlation could well be extrinsic due to population heterogeneity in both myopia and IQ, because various racial groups differ in the incidence of myopia and also differ, on average, in IQ. To find the answer to this question, the degree of myopia was measured as a continuous variable (refraction error) by means of optical techniques in a group of sixty adolescents selected only for high IQs (Raven matrices) and their less gifted full siblings, who averaged fourteen IQ points lower, a difference equivalent to 0.92σ. The high-IQ subjects differed significantly from their lower-IQ siblings in myopia by an average of 0.39σ on the measure of refraction error.[16] In other words, since there is a within-families correlation between myopia and IQ, the relationship is intrinsic. However, it is hard to think of any directly functional relationship between myopia and IQ. The data are most consistent with there being a pleiotropic relationship. The causal pathway through which the genetic factor that causes myopia also to some extent elevates g (or vice versa) is unknown. Because the within-family relationship of myopia and IQ was found with Raven's matrices, which in factor analyses is found to have nearly all of its common factor variance on g,17 it leaves virtually no doubt that the IQ score in this case represents g almost exclusively.
  
  
• 16. Cohn, Cohn, & Jensen, 1988.
  
  17. The Raven's g loading when factor analyzed among a large and diverse battery of mental tests is typically about .80. If it were only the Raven's specificity that was involved in its correlation with myopia, one would not expect myopia to be correlated with any other mental tests in which matrix items are absent. In fact, however, myopia is correlated with a wide variety of g-loaded tests and various g-loaded achievements (Jensen & Sinha, 1993, pp. 212-217).
  
  
• Cohn S. J., Cohn C. M. G. & Jensen A. R. (1988). "Myopia and intelligence: A pleiotropic relationship?" Human Genetics, 80, 53-58.
  
  Jensen A. R. & Sinha S. N. (1993). "Physical correlates of human intelligence". In P. A. Vernon (Ed.), Biological approaches to the study of human intelligence (pp. 139-242). Norwood, NJ: Ablex.

(Arthur Jensen. The g Factor. pp149-150, 167, 602, 615)

 

[humanino]

Hitssquad, your competences encompass domains even beyond my humor. Apparently science has no boundary
Is there a correlation between intelligence and fast speech ?

 

[hitssquad]

Correlations between talking speed and IQ

Is there a correlation between intelligence and fast speech?

In Mogdil & Mogdil's (Eds.) 1987 book Arthur Jensen: Consensus and Controversy, Chris Brand (pp253-255) reported "a list of g's numerous wider correlates" which included the item talking speed as a positive correlate. His reference was:

• Rabbitt, P. M. A. (1984) 'Decision times and motor skills in childhood and old age', Bulletin of The British Psychological Society, 37, A19.

See Brand's entire list of "wider" g correlates here.

 

[humanino]

Among which :
Variables correlated with g:
Eye color, lightness of
Hair color, lightness of
Height
Symmetry of facial features

Variables inversely correlated with g:
Racial prejudice

No comment. You know what you're doing.

Also, facsinating idea this red highlighting.

 

[humanino]

Again about China :
they find it very useful to have cheap manpower.

 

[Mandrake]

Mandrake: I happen to be familiar with the source of this comment, but it does not appear on page 54 of the paper that I know in which he presented his hypothetical model. What is the source you have in mind? Why did you give the year, but nothing else? Jensen published several things in 1993. Will you please give the full reference?

Yep, Paik, August 1998 (I thought I'd put a source link in my post, but clearly I hadn't).

FWIW, the information was also presented in Psychometric _g_ and achievement. In B. R. Gifford (Ed.), Policy perspectives on educational testing (pp. 117-227).

Mandrake:
What? Do you think that scientific research in psychometrics is conducted by idiots? Do you really think that they include subjects who are impaired by obvious factors? I am amazed by this comment. What is your reason for this? Have you ever read papers on chronometric measurements? I am absolutely in disbelief that you cannot separate mental impairment with the foregoing comments.

I answered a similar question in your thread in SS; basically, if you're working primarily from correlations,

Have you established that point? I get the impression that you mentally delete the large amount of discussion about laboratory measurements.

To take an analogy, AFAIK a modest consumption of alcohol contributes to keeping heart attacks at bay (I'm sure many PF readers could give us much more accurate statements), zero intake and your risk goes up; too much and it also goes up. However, the relationship is not linear. Now, wrt _g_, it may be that being the tiniest bit drunk helps (your measured _g_ goes up),

And there may be a teapot in orbit around Pluto. Why are you speculating? It would be easier to discuss what is known, so why not simply tell us what you have determined as a known?

Mandrake
The thing you are describing is MT.
No. To be a good pingpong player, or fencer, you need to do more than just be quick on MT; you also, in your paradigm, need to be very sharp on RT. If you get a chance, watch a slow motion video of a good pingpong game sometime, both effects are clearly observable.

What time values have you observed? You may be correct, but it is difficult for me to see that it is so on the basis of an assertion. If you are right, then there should be a very real correlation between _g_ and ping pong ability and between _g_ and fencing ability. Are those sports more intelligence dependent than boxing? It seems to me that boxing requires very fast movements, but is not a sport that is dominated by highly intelligent people.

So, in my stamp collecting analogy, if I find three completely unrelated variables which have statistical relationships similar to those between RT, _g_, etc I can make a psychometric theory from them?

You tell me. As I have stated, I think your stamp collection analogy is misleading and designed to disparage. Perhaps you simply don't have a good grasp of the range and relationships of what is known in psychometrics? Is that true or not? If it is true, how can you hope to fabricate worthwhile analogies?

 

[Mandrake]

Yep, that's what I was asking about earlier. Thanks, M.
It's shockingly low compared to other areas of science, but I can see how social science studies would be tough (hard to control/account for the variables).

One of the most extreme examples of the value of a low correlation is that associated with a single test item in an IQ test. The correlation is low and the variance is, obviously, very low. But the variances combine to produce a single value that has much more net meaning.

I am sure we could all come up with examples of how very tiny effects are very important. I once listened to Val Fitch as he explained his Nobel Prize which was about The Discovery of Charge – Conjugation Parity. It is my understanding that the parity violation here is quite small, yet it has had a huge role in the development of matter in the cosmos.

The usual example of a small probability is the statistical edge held by the house in gambling games. The house makes money. ;-)

I am not going to look it up right now, but the time required for a 1% survival advantage to overtake 99% of a breeding group is not all that many generations. The numbers are given in The _g_ Factor.

 

[Nereid]

I would like to report that we found a new and very interesting result : we need further investigations, but maybe a correlation between intelligence and using glasses. Sounds weird.

You may have overlooked one of hitssquad's correlations ... if you get a tan (go to the beach in summer, go to a tanning studio in winter), you run the risk of lowering your _g_

 

[Nereid]

_g_ apparently correlates with every possible (human) physical variable?!

hitssquad quoted Jensen (here):
"First, psychometric tests were never intended or devised to measure anything other than purely behavioral variables. Constructors of IQ tests, in fact, have tried to eliminate any source of test item variance that might reflect individual differences in physical attributes such as muscular strength and sensory acuity. Certainly there has never been the least intent that mental tests should reflect any strictly anatomical or physiological variables, which are directly measurable by other methods. It would therefore be most surprising and remarkable if IQ tests were significantly correlated with physical variables. Yet they are. IQ – especially the g factor of IQ tests – is correlated with a variety of physical variables. What does this mean? For the time being, about all one can say with certainty is that whatever is measured by IQ tests –mostly g – is somehow enmeshed in a host of organismic variables and therefore involves something beyond the purely psychological or behavioral. It also proves that g is not just an artifact of the way psychometric tests are constructed, nor is g a mere figment of the arcane mathematical machinations of factor analysis. Obviously, a correlation between psychometric g and a physical variable means that g is somehow connected with underlying biological systems."

hitssquad also gave a list of correlations (well, in many cases just the sign, not the values), and (for some) lists of sources. From this long list, I selected all those which seem to be + (and -) and made these composites:

people who are tall, muscular, bald, blind, blond-haired, albino, blue-eyed, with a large head and brain, fast alpha brain waves, a high rate of brain glucose metabolism, good general health (but are myopic and prone to depression); fast-talkers, butterflies.

people who are short, fat, hairy, sighted, black-haired, brown-eyed, heavily tanned, with a small head and brain, slow alpha brain waves, a low rate of brain glucose metabolism, poor general health (esp epilepsy and schizophrenia; but not myopic, nor prone to depression); laconic, conscientious.

{There's also a whole lot of personality correlations, apparently; when I find these, I'll be able to add them to the composites. I wasn't sure how to deal with 'latency and amplitude of evoked brain potentials'}

Since all these are based on correlations (and no well-tested biological hypotheses have yet emerged to account for them), there is no a priori reason to think any of these physical variables (other than _g_?) have any correlation, so when the datasets are analysed appropriately, how do the correlations relate? E.g. are they additive? Also, if all these physical attributes are included, what does the _g_ correlation rise to?

 

[Nereid]

Yep, that's what I was asking about earlier. Thanks, M.
It's shockingly low compared to other areas of science, but I can see how social science studies would be tough (hard to control/account for the variables).

[nitpick]There is a category confusion here; it would seem that intelligence psychometrics can be lumped into 'social sciences' (e.g. with economics and anthropology), but that some of what some psychometricians claim is biological. AFAIK, correlations of 0.5 or lower in biology aren't thought very interesting, and certainly would need to be followed up with further studies to find what (if anything) is giving rise to the correlations. In this sense we could perhaps consider this sub-discipline to have found some interesting results, but to still lack the basics of anything that could be considered part of mainstream biology.[/nitpick]

 

[Moonbear]

[nitpick]There is a category confusion here; it would seem that intelligence psychometrics can be lumped into 'social sciences' (e.g. with economics and anthropology), but that some of what some psychometricians claim is biological. AFAIK, correlations of 0.5 or lower in biology aren't thought very interesting, and certainly would need to be followed up with further studies to find what (if anything) is giving rise to the correlations. In this sense we could perhaps consider this sub-discipline to have found some interesting results, but to still lack the basics of anything that could be considered part of mainstream biology.[/nitpick]

From everything presented on psychometrics in the various threads on this topic, and from my own digging around into the literature, it seems to be quite firmly planted as a subdiscipline of pscyhology. Most psychology departments have sort of a split personality. There are psychologists who focus more on the social science side of things and really aren't interested in the biological mechanisms underlying what they study. The other camp in psychology are the psychobiologists. They focus far more on the biological basis for behavior (behavioral neuroscience has grown out of this field). Psychobiologists are not content with correlations below 0.5 (and would even be cautious about interpreting correlations between 0.5 and 0.75), and interpret such low correlation to mean one of two things: 1) there isn't any real relationship between the two things being studied, it was entirely due to chance, or 2) the study wasn't properly conducted to control for all the variables.

 

[Mandrake]

There is a category confusion here; it would seem that intelligence psychometrics can be lumped into 'social sciences' (e.g. with economics and anthropology)...

It appears to me that your comments here and in prior messages are designed to advance a nihilistic perspective with respect to psychometrics. The problem with your case is that you constantly demonstrate to us that you are unaware of the depth and breadth of the science you criticize. In fact, when issues are explained to you, we find later comments that ignore the prior explanations. I cannot understand your purpose, just as you do not understand the science of psychometrics.

The comparison to economics is as inappropriate and misleading as your inept stamp collection. Economics does not involve laboratory measurements, nor is it a science that deals with human biology. Economics is incapable of displaying differential observations between population groups and between family members. Economics is not genetically determined. The comparison you made is no more appropriate than comparing astronomy to architecture.

... but that some of what some psychometricians claim is biological.

Would you please explain the above comment? My reading of it is that you wish to imply that you hold a different position and that your position is that intelligence is not biological. We have previously discussed the strong links between physiology and intelligence. What does your comment mean in the context of those links? If intelligence is not biological, what is it? Spiritual? Ethereal?

AFAIK, correlations of 0.5 or lower in biology aren't thought very interesting

Correlations are applied as a means of detecting the presence of variables that coexist with other variables. Since I am not a biologist, I will not attempt to speak for that science, but your argument strikes me as an attempt to discredit something, but which is based on your perception and is counter to the understanding that is not challenged by the very competent scientists who have an in-depth understanding of psychometrics. You have not demonstrated that degree of understanding, nor even a good recollection for the material that you have attempted to discuss.

and certainly would need to be followed up with further studies to find what (if anything) is giving rise to the correlations.

A .5 correlation is usually interpreted as one that corresponds to a variance that explains 25% of the variance in the other parameter. Do you dismiss a 25% overlap in two variables? If so, can you explain why? There is another way to look at a .5 correlation. It is half of the maximum possible correlation. In light of your comment about biology, I would like for you to give us a few examples of biological variables that correlate at .5, but which "aren't thought very interesting." Thank you.

 

[CloakNight]

Psychobiologists are not content with correlations below 0.5 (and would even be cautious about interpreting correlations between 0.5 and 0.75), and interpret such low correlation to mean one of two things: 1) there isn't any real relationship between the two things being studied, it was entirely due to chance, or 2) the study wasn't properly conducted to control for all the variables.

I do not know of any scientist that would consider 0.5 a low correlation. 0.5 is generally thought to be a considerably high correlation. Researches on homosexuality as printed in the journals of Science, General Psychiatry, and American Journal of Psychiatry show a genetic correlation of 0.5 for homosexuality. Would you agree then that homosexuality is a choice and predominately environmentally based? Would homosexuality then be "curable"?

 

[humanino]

Welcome CloakNight !

I know many in physics who require much higher correlations. We investigate certainty, not vague facts. We require 99.9999% likelihood before the community accept a discovery. This is of course not possible in biology I guess. Or maybe you should try to search the various topics some guys in this discussion have been debating, and try to make your own opinion about what purpose they try to serve here. I made mine. I know the caution required here. Some people are clever, we must be careful.

 

[bd1976]

Researches on homosexuality as printed in the journals of Science, General Psychiatry, and American Journal of Psychiatry show a genetic correlation of 0.5 for homosexuality. Would you agree then that homosexuality is a choice and predominately environmentally based? Would homosexuality then be "curable"?

Erm.. Dude your logic is sadly lacking. Even if homosexuality is entirely caused by enviromental effects it could be that these effects cause the brain to change as it develops and so the effects are not reversible once the brain has reached maturity.

The problem with this is that its a symptum of a larger problem. Science can not explain how the brain works. Science has very little idea how the brain operates. This has given rise to these "soft sciences" or "social sciences" like psychology. Where you propose an intelligent sopunding idea and tyhen go off and "prove" it with statistics. Often your trying to prove your own clever suggestion so you have a vested interest in the resuts. Look at your own quote. It's from the American Journal of Psychiatry don't you think this group has a vested interest in their being a correlation. Who is going to treat the homo's ... hmmm... prehaps its... Psychiatrist's! Wow there's a coincidence!

 

[Moonbear]

I do not know of any scientist that would consider 0.5 a low correlation. 0.5 is generally thought to be a considerably high correlation. Researches on homosexuality as printed in the journals of Science and American Journal of Psychiatry show a genetic correlation of 0.5 for homosexuality. Would you agree then that homosexuality is a choice and predominately environmentally based? Would homosexuality then be "curable"?

We have to differ on this, r=0.5 is not a high correlation. It's borderline, at least in the field of biology, and I think biologists tend to be more generous than those in other sciences. Further, if you look back through the threads here, you will see that there are some who are claiming r=0.2 is a high correlation, and that is the background for my comments.

I've discussed the topic of homosexuality in other threads here. There is evidence that some of the biology of sexual orientation is indeed not genetically based, nor is it learned behavior or a choice. However, I also don't view that to be something that requires a "cure." And, would you please cite the reference for the Science article(s) you refer to that say a correlation of r=0.5 is a high correlation. The original Science article demonstrating a genetic linkage for homosexuality is does not report correlations, indeed, such an analysis wouldn't fit with the way they collected their data.

A Linkage Between DNA Markers on the X Chromosome and Male Sexual Orientation
Dean H. Hamer; Stella Hu; Victoria L. Magnuson; Nan Hu; Angela M. L. Pattatucci
Science, Vol. 261, No. 5119. (Jul. 16, 1993), pp. 321-327.

The only other article in Science that I'm aware of on this topic contradicts those earlier findings, though doesn't exclude the possibility of a genetic linkage.

Male Homosexuality: Absence of Linkage to Microsatellite Markers at Xq28
George Rice, Carol Anderson, Neil Risch, George Ebers
Science, Vol 284, Issue 5414, 665-667, 23 April 1999

Both articles are followed by comments in subsequent issues debating the findings, so it's far from an open and shut case.

However, if you'd like to continue to discuss the topic of homosexuality, I suggest taking it to this thread, which is more appropriate to the topic.
https://www.physicsforums.com/showthread.php?t=39163

 

[CloakNight]

Erm.. Dude your logic is sadly lacking. Even if homosexuality is entirely caused by enviromental effects it could be that these effects cause the brain to change as it develops and so the effects are not reversible once the brain has reached maturity.

The problem with this is that its a symptum of a larger problem. Science can not explain how the brain works. Science has very little idea how the brain operates. This has given rise to these "soft sciences" or "social sciences" like psychology. Where you propose an intelligent sopunding idea and tyhen go off and "prove" it with statistics. Often your trying to prove your own clever suggestion so you have a vested interest in the resuts. Look at your own quote. It's from the American Journal of Psychiatry don't you think this group has a vested interest in their being a correlation. Who is going to treat the homo's ... hmmm... prehaps its... Psychiatrist's! Wow there's a coincidence!

As the comments above are completely ignorant to the basics of science, brain anatomy, and a complete insult to anyone having a scientific background, I will ignore this post.

 

[CloakNight]

We have to differ on this, r=0.5 is not a high correlation. It's borderline, at least in the field of biology, and I think biologists.

Whether or not 0.5 would be a high correlation or a mild correlation, it surely wouldn't be nothing. It would show there is a reasonable genetic connection.

Further, if you look back through the threads here, you will see that there are some who are claiming r=0.2 is a high correlation, and that is the background for my comments.

For biology, I could understand why 0.2 would be considered a low correlation but for social sciences I can see why it would be considered more viable.

The original Science article demonstrating a genetic linkage for homosexuality is does not report correlations, indeed, such an analysis wouldn't fit with the way they collected their data.

Bailey and Pillard (1991): occurrence of homosexuality among brothers

52% of identical (monozygotic) twins of homosexual men were likewise homosexual
22% of fraternal (dizygotic) twins were likewise homosexual
11% of adoptive brothers of homosexual men were likewise homosexual

J.M. Bailey and R.C. Pillard, “A genetic study of male sexual orientation,” Archives of General Psychiatry, vol. 48:1089-1096, December 1991.


Bailey and Pillard (1993): occurrence of homosexuality among sisters

48% of identical (monozygotic) twins of homosexual women were likewise homosexual (lesbian)
16% of fraternal (dizygotic) twins were likewise homosexual
6% of adoptive sisters of homosexual women were likewise homosexual

Bailey, J. M. and D. S. Benishay (1993), “Familial Aggregation of Female Sexual Orientation,” American Journal of Psychiatry 150(2): 272-277.

http://www.worldpolicy.org/globalrights/sexorient/twins.html

 

[Moonbear]

Whether or not 0.5 would be a high correlation or a mild correlation, it surely wouldn't be nothing. It would show there is a reasonable genetic connection.

As I said before, it would be something to interpret with caution.

J.M. Bailey and R.C. Pillard, “A genetic study of male sexual orientation,” Archives of General Psychiatry, vol. 48:1089-1096, December 1991.
Bailey, J. M. and D. S. Benishay (1993), “Familial Aggregation of Female Sexual Orientation,” American Journal of Psychiatry 150(2): 272-277.

Unfortunately, neither of those references is available online (both pre-date the online versions of those journals), so I could only read the abstracts. With regard to this thread, what r-value do they claim? Again, I don't really see how one can analyze data on homosexuality using a Pearson's correlation, so I'm really curious how this is relevant. So far, all the analyses I've seen are comparing proportions of populations that are homosexual or heterosexual using chi-squared analyses. Since I can't access the original articles tonight (or anytime soon...our science/medical library is undergoing renovation, so all hardcopies have to be requested for delivery, which is a slow process right now, so I probably won't be requesting articles that aren't directly pertinent to my research), I'd be grateful if you could provide that information for me, and a little context of the related methods.

 

[Mandrake]

We have to differ on this, r=0.5 is not a high correlation. It's borderline, at least in the field of biology, and I think biologists tend to be more generous than those in other sciences. Further, if you look back through the threads here, you will see that there are some who are claiming r=0.2 is a high correlation, and that is the background for my comments.

Do you consider medicine to be a biological science? When a drug company seeks approval for a new drug, it must submit various studies to the government for approval. If a drug is found to correlate at r=.2 to a side effect that causes permanent disability or death, will the drug be approved? At what value of r would the license be granted? The same question can be asked with respect to drugs that might be given to pregnant women. If r=.2 for deformed babies or death to the fetus, would the drug be allowed for use with pregnant women?

If you were considering an elective operation and knew that the correlation between the procedure and a debilitating outcome was .2, would you have the operation? If not, what value of r would you consider to be acceptable?

 

[humanino]

This is a ridiculous argument ! There are some sides effects, always. If the drug compagnies had the same requirements as in physics, there would be none.

Of course, there would also be no drug passing so high requirements. This is only due to market considerations. If they had long enough time (say over decades) they would be able to produce such medicines.

 

[Moonbear]

Do you consider medicine to be a biological science? When a drug company seeks approval for a new drug, it must submit various studies to the government for approval. If a drug is found to correlate at r=.2 to a side effect that causes permanent disability or death, will the drug be approved? At what value of r would the license be granted? The same question can be asked with respect to drugs that might be given to pregnant women. If r=.2 for deformed babies or death to the fetus, would the drug be allowed for use with pregnant women?

If you were considering an elective operation and knew that the correlation between the procedure and a debilitating outcome was .2, would you have the operation? If not, what value of r would you consider to be acceptable?

In the examples you cite, the data wouldn't be analyzed that way. A Pearson's correlation (r) is the wrong statistic to use. So, if a company seeking drug approval handed a report to FDA that included such a statistic, no, the drug would not be approved because FDA would tell them their analysis was flawed. Drug trials would not include correlations or even post-hoc analyses; those are very frowned upon by FDA. Severe side effects that required people drop out of the study or led to serious health problems would either be reported as a proportion of the subjects reporting the side effects or using a survival analysis. Using a Pearson's correlation requires comparing two variables with a normal distribution and similar standard deviations. The closer r is to 0, the more scattering there is of the values from a linear relationship.

 

[Mandrake]

This is a ridiculous argument ! There are some sides effects, always. If the drug compagnies had the same requirements as in physics, there would be none.

I didn't make any argument. I asked a few questions. We all know that drugs have side effects and are sometimes licensed with side effects that can be serious. The question is whether a .2 correlation is insignificant with respect to biology. Your answer is apparently a resounding YES. But the question I raised is at what correlation is the risk of a serious complication accepted as small enough to allow. I seriously doubt that a drug with life threatening side effects in the r = .2 range would be licensed. If that is true, we can conclude that this example of biology recognizes small correlations as very important.

 

[Mandrake]

In the examples you cite, the data wouldn't be analyzed that way. A Pearson's correlation (r) is the wrong statistic to use.

If you can produce a scatter diagram with the data, you can determine a correlation coefficient. I am not claiming anything about the licensing prodedures of the FDA. The simple question is whether a drug would be licensed if it had that kind of correlation. I asked a similar question with respect to whether or not readers here would undergo elective surgery, if they believed that the correlation between that operation and severe impairment could be represented by a correlation coefficient of .2.