A Multivariate Approach to the Method of Correlated Vectors?

(Unedited post.)

A while back, I deduced that either shared environment produces Jensen Effects or that the race-genes-IQ debate should be over. And then I reasoned that since the debate is ongoing, shared environment probably produces Jensen Effects. To quote myself:

a. either the gap is causal cultural or causal biological.
if the gap is causal biological, it is either causal environmental-biological or causal genetic-biological.
the gap can not be causal environmental-biological (for reasons discussed before).
ergo, the gap must be either causal cultural or causal genetic-biological.
b. either the B/W gap is shared environmental or additive genetic (this follows from the logic of differential regression to the mean, discussed before)
c. either shared environmental influences can produce g(+) gaps or they can not.
the gap is decidedly g(+).
ergo, either the gap is not shared environmental or the gap is shared environmental + shared environmental influences can produce g(+) differences.

this reduces down to…
d. either the gap is additive genetic or the gap is shared environmental and shared environmental influences can produce g(+) gaps
now…
e. either the debate over cause (wholly environmental versus some genetic) is over or it is not
it’s not over.
f. the Jensen effect on the MZA correlation either necessitates an anti-Jensen effect on the URT correlation or not.
if it does, since URT is a direct measure of the influence of shared environment, shared environmental influence must induce an anti-Jensen effect.
ergo, by d, the B/W would obviously have to be genetic.
h. the B/W gap is not obviously genetic. the inductive evidence of this is that the debate is still going on. And that I’m not compelled.

I believe that Jason Malloy and MH were the only ones who grasped my point. Whatever the case, I did and I decided to investigate. As expected, I found a meta-analytic Jensen Effect on Shared Environment. (This analysis is ongoing and the issue is not settled — but if I had to bet, based on n twins = 3000 K=10, I would say that the r (g x c^2) will come out to be non trivially positive and of a magnitude approaching r (g x h^2) — a conclusion that I don’t particularly appreciate). For example, here were the results from a large Austrian Twin Study based on the multidimensional aptitude battery n (tested twin pairs) ~ 1000.

Brisbane Adolescent Twin Study
(Brisbane Adolescent Twin Study: Outline of study methods and research projects)

MCV gives a r (g x c^2) of close to unity. This is even larger than the r (g x c^2) found in Samuelsson et al. (2007) based on achievement tests, n twin pairs = ~ 800. The upshot is that while James Flynn was completely wrong with regards to his suggestion that ethnic differences could be due to unshared environment (an idea on which I wasted an extraordinary amount of time) and with regards to his argument that the Flynn Effect was a Jensen Effect, Rushton, Jensen, Levin, Murray, Gottfredson, Woodely, Meisenberg, Hunt and many others seem to have been off in their arguments, suggestions, and imputations that a Jensen Effect implies biological or genetic etiology. Here, for example, is a typical example of the discussion on this topic:

If population group differences are greater on the more g- loaded and more heritable subtests, it implies they have agenetic origin (Jensen, 1973, 1998).

…A Jensen Effect for heritability has also been found, with the g loadings from various subtests correlating with the heritabilities of these same subtests (Jensen, 1998). A Jensen Effect for heritability provides biological evidence for a true genetic g, as opposed to the mere statistical reality of g .It makes problematic theories of intelligence that do not include a general factor as an underlying biological variable, but only explain the positive manifold, such as the model proposed by Dickens and Flynn (2001), and the mutualism model by van der Maas, Dolan, Grasman, Wicherts, Huizenga, and Raijmakers (2006)
(The rise and fall of the Flynn Effect as a reason to expect a narrowing of the Black–White IQ gap).

[T]he existence of the Jensen effect would appear to have significant ramifications. Firstly it effectively falsifies Thomsonite models of the development of intelligence (e.g. Bartholomew, Deary, & Lawn, 2009; Thomson, 1916; van der Maas et al., 2006), which argue that g arises chiefly from random sampling or mutualistic reinforcement amongst distinct ‘neural elements’ or bonds, rather than from the existence of a special quality or ‘mental energy’, as was first posited by Spearman (1927). This is because the Jensen effect reveals the existence of an apparent nexus amongst diverse biological variables and g , which suggests that the factor corresponds to something very fundamental to brain neurophysiology and genetics, rather than being a mere statistical regularity (Eysenck, 1987; Rushton & Jensen, 2010).

…It appears that dysgenic fertility is indeed a Jensen effect. This finding using a large and representative sample of the US population, along with subpopulations and a well validated measure of intelligence, therefore allows us to place dysgenic fertility into the genetic nexus of the Jensen effect. (A Jensen effect on dysgenic fertility: An analysis involving the National Longitudinal Survey of Youth.)

A non trivial positive r (g x c^2) makes the above reasonings untenable. What is suggested by this is that there is a strong anti-Jensen Effect on unshared environment, rather than a strong Jensen Effect on genes. In terms of path diagrams, while the effect of unshared environment is smaller on g than non g, the effect of both genes and shared environment is larger on g than non g. If you were just to pit genes versus shared environment — something I have yet to do — you might find no difference in terms of the correlation with g-loadings. Following Jensen & Rushton’s reasoning this would imply a shared environmental origin to g. Or, better, one might just say a non-shared environmental non-origin.

So I think that this is a problem for Jensen inferences. And this leads me to the question: Can MCV be salvaged?

Recently, I proposed to Meng Hu a multivariate approach to MCV. In instances where you have a vector of difference, a vector of g-loading, and a vector of cause e.g., heritability, biologicality, shared environmentality, etc. simply run a regression. You would want to look for mediation. Oddly, no one seems to have tried this.

To given an example, Kan (2011) found a Jensen Effect on heritability and on cultural loadings. This led him to devise a complex Cov(GE)-g model. To super simplify: people seek out, in accordance with their genetic dispositions, environments which increase g — hence: r g-loadings-“culture loading” and r g-loadings-h^2. With a path running, presumably: genes–>cultural–>g. But I noted that cultural loaded tests are more cognitively complex than cultural less loaded tests and so likely to be more g-loaded. In counter, Kees-Jan argued that, regardless, culturally loaded tests should be less heritable, because they are more culturally loaded. Presumably, he and his Co. make the same argument in “Kan, Wicherts, Dolan, and van der Maas. On the Nature and Nurture of Intelligence and Specific Cognitive Abilities: The More Heritable, the More Culture-Dependent? Psychological Science”. But this was just an argument from the “heritability paradox”. On this paradox, Jensen, 2006 pg. 133) noted:

This is the name given to the frequent and seemingly surprising finding that h2 increases as a function of task complexity and also as a function of the degree to which tasks call for prior learned knowledge and skills. The notion that this is paradoxical results from the expectation of some theorists that the more elemental or basic components of performance in the causal hierarchy of cognition, going from stimulus (the problem) to response (the answer), are far less removed from the underlying brain mechanisms than are the final outcomes of complex problem solving. The tests of greater complexity are considered to be more differentially influenced by environmental conditions, such as prior cultural–educational opportunities for having acquired the specific knowledge or skills called for by the test. Therefore, it is expected that individual differences in tests that are more dependent on past-acquired knowledge and skills, such as typical tests of IQ and scholastic achievement, should reflect genetic factors to a lesser degree than RT on relatively simple ECTs, which are assumed to depend almost exclusively on the most basic or elemental cognitive processes. Presumably, individual differences in RT on such relatively simple ECTs scarcely involve differences in environmental factors influencing specific cultural or scholastic knowledge and skills, and therefore ECTs should have higher h2 than the more experientially loaded psychometric tests. It turns out, however, that the empirical finding described as the “heritability paradox” is not really paradoxical at all. It is actually an example of a well-known effect in psychometrics — the aggregation of causal effects, whereby the sum or average of a number of correlated factors has greater reliability and generality than the average of the reliability coefficients of each of the aggregate’s separate components. The factor common to a number of somewhat different but correlated ECTs, therefore, should be expected to have a higher phenotype–genotype correlation (and thus higher h2) than its separate elements.

“Cultural loaded” — really: “informationally loaded” — tests can index a broader range of functioning which gives a more reliable index of psychometric g and by way of psychometric g of heritable g. “Cultural loadedness” will only reduce heritability if the background information is heterogeneous with respect to the sample — in this case between twin pairs. This trivially obvious point seemed to have missed my interlocutor.

Whatever the case, this is speculation either way. Culture-loadings correlate with heritability and with g. And heritability correlates with g. Does g mediate the cultural-heritability correlation as I would predict or does culture mediate the heritability-g correlation? For example, here was some data from Kan (2011):

kancultureg1

And here were the bivariate correlations:

kancultureg3

Well, to make a shorter explanation even shorter, here were the multivariate results:

kancultureg2

Controlling for test battery, the association between heritability and cultural loading was fully mediated by g, but the association between heritability and g was only partially mediated by cultural loading. So, based on this analysis, it would look as if my position was more correct. But, of course, one would want to conduct several Multivariate Method of Correlated Vector (MMCV) analyses before deciding anything.

My point here is simply to (re?)-introduce a less simplistic version of MCV, a version which might be of use since, if I am right about C^2 and g, the Jensen Effect network is more complex than previously thought.

Now, what can be explored? Well, for starters, Kan (2011) chapter 4 presents g-loadings, cultural loadings, and B/W difference data. My prediction: Same as above — the association between cultural loading and b/w d is mediated in full by g. Also, Spitz (1988) presents data on g-loadings, mental retardation scores, and heritability estimates. I guess the immediate issue of interest is whether various r (vector1 x h^2) is mediated by g. For MR, that can be checked with the Spitz results.

(Let me know what you get MH.)

Advertisements
This entry was posted in Uncategorized. Bookmark the permalink.

32 Responses to A Multivariate Approach to the Method of Correlated Vectors?

  1. Marlo says:

    All things being equal, the tails of a normal intelligence distribution will be disproportionately comprised of blacks. Greater genetic variance=more outliers. That should be obvious to a “numbers guy”.

  2. KJK says:

    1) It is true that “ethnic differences could be due to unshared environment”, even if g would be a unitary, highly heritable, biological, and when Jensen-effects for heritability and for mean IQ group differences exist.

    See: http://i161.photobucket.com/albums/t222/Kaziman/MCV-Groupdif.gif

    The reasoning behind the MCV is invalid in the discussion; the vector correlations have no bearing on the origin of the group differences.

    2) The following is irrelevant with respect to the heritability paradox: “[T]the empirical finding described as the “heritability paradox” is not really paradoxical at all. It is actually an example of a well-known effect in psychometrics — the aggregation of causal effects, whereby the sum or average of a number of correlated factors has greater reliability and generality than the average of the reliability coefficients of each of the aggregate’s separate components. The factor common to a number of somewhat different but correlated ECTs, therefore, should be expected to have a higher phenotype–genotype correlation (and thus higher h2) than its separate elements.”

    After taking into account differences in reliability, the relations between ECTs and g and aggregated effects of ECTS and g are empirical as well as theoretical questions, and open to debate.

    http://s161.photobucket.com/user/Kaziman/media/HeritabilityECTs.gif.html

    Moreover, IF it would be true, then mutualistic models would automatically account for a Jensen effect for heritability.

    A last remark, if g is supposed to account for correlations between the ECTS, the ECTS do provide the account for g. (And if the ECTS account for g, then the relations between the ECTS are not accounted for by g).

    3) You can test whether g mediates the relation between genetic influences and cultural influences or whether cultural influences mediate the relationship between genetic influences and cultural influences, which is not the model of Kan et al. by the way. The model in Kan et al. is not a one way genetic influences -> cultural influences -> g model. Seen cross-sectionally, it is a genetic influences -> cognitive abilities cultural influences model. Regression techniques will not help you, as the reciprocal model you need to fit then is not identified. You can factoranalyse the variables though, so you end up with a statistical g factor (that’s why it makes sense to speak of ‘cognitive abilities -> g’ ). Seen developmentally, the model of Kan et al is a genetic influences -> cognitive abilities -> cultural influences -> cognitive abilities …. model. This model testable in principle.

    Testing whether ‘correlations among correlations’ mediate other ‘correlations among correlations’ makes no sense.

    • Chuck says:

      You said: “It is true that “ethnic differences could be due to unshared environment”, even if g would be a unitary, highly heritable, biological, and when Jensen-effects for heritability and for mean IQ group differences exist.”

      You missed the whole point behind the argument from MCV. You made the same mistake in your thesis. Sure, if you start with an unadulterated group difference in g, then group differences with correlate will g regardless of the cause (a, c, e). But the whole argument is that a random sample of unshared environmental effect with not cause an unadulterated group difference in g, because random samples of unshared environment (and environment in general) will also load on non-g factors too — and they will load on non-g more. This will cause a negative correlation between e^2 (and e^2+c^2) and g-loading.

      I discussed this here:
      http://occidentalascent.wordpress.com/2013/05/09/spearmans-hypothesis-tested-on-the-admixture-data-in-scarr-et-al/

      Generally, you’re being silly. Regardless, MCV arguments are finished because I found a meta-analytic correlation between c^2 and g no less than that between h^2 and g, K=23, N~8000. But my point would be that were the r ( c^2 x g) negative or zero as some have suggested (e.g., Jensen, 1997, te Nijenhuis, variously), then this would have constituted an argument. It’s annoying that you keep missing this obvious point.

      You said: “2) The following is irrelevant with respect to the heritability paradox”

      Maybe I was misinterpreting “and generality” in:

      “whereby the sum or average of a number of correlated factors has greater reliability and generality”

      I assumed that Jensen meant that a broader measure would be more g-loaded simply because it was a broader measure (irrespective of test-test reliability).

      Whatever the case, the high correlation between g, h^2, and informational/cultural load is not a problem for biological g.

      a. First, here was the h^2 x g correlation for tests that were informationally loaded (e..g, vocabulary, information. etc.)
      http://lesacreduprintemps19.wordpress.com/?attachment_id=1140 Controlling for cultural/informational load does not abolish the relation.

      b. Second, cultural wouldn’t be expected to reduce heritability unless the population was differentially exposed to the relevant information. Jensen 1973 made this same point.

      c. Third, cultural load would correlate with g/h^2 if cultural loaded tests better indexed g by assessing a broader range of capacity. This is plausible since e.g., vocab indexes one’s brain’s ability to absorb info over an extensive span of time. Controlling for reliability doesn’t address this.

      Generally, I see no a prior reason why crystal intelligence would not exhibit higher g-loadings than fluid g given a biological g model. But I am open to being convinced otherwise.

      You said: “Testing whether ‘correlations among correlations’ mediate other ‘correlations among correlations’ makes no sense.”

      Well, then, I misinterpreted what you said — probably because I am stuck in my cross sectional frame! The point is that you see a h^2 x g correlation if you control for dichotomous cultural load.

      My larger point was that you could use multivariate MCV to see if x,y,z mediates a correlation. You can also use this MMCV to partially address the specificity critique of Dolan et al. For example, you can control for influence on non-g factors. So, for example, when I control for broad factor influences for the Deaf-Hearing difference (which is driven by a verbal deficient) I abolish the Jensen Effect. Doing the same for the B/W data increases it — because the B/W difference is a g difference unlike the Deaf-Hearing difference. Generally, you should be able to explore specificity within the MCV framework — at least somewhat.

      But since there is a Jensen Effect on c^2, this doesn’t matter to me. But not for the reason you said. Or for the reason Dolan said in email reply. Jensen Effects are causally informative in principle.

      BTW, what would your critique of the following study be?

      Rowe and Cleveland (1996) “Academic Achievement in Blacks and Whites: Are the Developmental Processes Similar?

      Can this issue not be resolved with Biometric Modeling? Would it be a waste to try more of these studies. I ask because I have two data sets on which I can use biometric modeling — and one is longitudinal. But if doing so is worthless, I won’t bother (i.e., I wont pay someone to do this for me).

  3. KJK says:

    The negative relation between e2 and g-loading is because of standardization. This is exactly why the MCV is not useful in the discussion of the genetic and environmental sources of group differences (it can be useful in other respects, I’m not against the MCV itself or something). Group differences can be due to numerous nonshared influences (see the plot in my previous post), without the MCV revealing this. I cannot understand why you do do not want to see this (and that’s why I feel I I need to quit further discussion). BTW, who says the nonshared influences should be randomly sampled? As I understand, Flynn referred to systematic differences (such as educational chances).

    “Controlling for cultural/informational load does not abolish the relation.”

    The point is not whether the one relation is stronger than the other or something. The point is that the relations coexist. This coexistence is what needs explanation. g theories do not provide any account for It (nor any other theory). I do not say that g theories will be unable to explain the coexistence. What I say is that they need more hypotheses to do so. With more I mean more than the complexity and investment theory alone.

    • Chuck says:

      You said: “The negative relation between e2 and g-loading is because of standardization.”

      1- rMZT is a direct measure of e^2. And this shows a large anti-Jensen Effect. How is this just “because of standardization”? Try comparing: rMZA, rURT, and rMZT. There is no a prior reason why these direct measures of genetic, shared environmental, and unshared environmental influence should differ in magnitude and direction with regards to Jensen Effects. The strong anti-JE on e^2 is a fact of the world. You effectively showed this yourself.

      You said: “I cannot understand why you do do not want to see this (and that’s why I feel I I need to quit further discussion). BTW, who says the nonshared influences should be randomly sampled?

      Ooops! You were wrong — and you won’t admit it! :0)

      As for your question:

      Quote: “[…]Strong inference is possible: (1) genetic theory predicts a positive association between heritability and group differences; (2) culture theory predicts a positive association between environmentality and group differences. (Jensen, 2012. Rushton’s contributions to the study of mental ability).

      The claim, which you –not I — challenged was that “environmentality” would induce a negative correlation with g. Does not “environmentality” refer to environmentality in general as opposed “some specific environmental effects”? But I agree with what you are saying concerning particulars. Hence I said: “obviously one can do this; one just has to propose that the group differences are caused by those specific environmental effects that cause g-loaded differences e.g., fetal alcohols syndrome — but the burden is on environmentalists to specify the causal factors because a generic environmental account is insufficient.)”

      So, we are back to: there is something that needs to be explained. What’s your proposal? But really I am looking for the environmental cause myself. And I am using MCV to do this. First, I verify that 1-h^2 does indeed cause an anti- Jensen Effect. Then I move onto 1-h^2-c^2 and 1-h^2-e^2. Since there is a large anti-Jensen effect on e^2 r (e^2 x g) = 0.6 uncorrected, r (e^2 x g) = 0.95 corrected, but not an anti-Jensen effect at all on c^2, I assume that the latter is a more promising region of exploration. Then I move onto specific proposed c^2 causes: schooling, cultural rearing i.e., adoption, etc. But, of course, this is a silly approach because some effects classified as e^2 could in principle cause the consistent r (groups difference x g). Ah, the triumph of “environmentalism in principle”!

      You said: “As I understand, Flynn referred to systematic differences (such as educational chances).”

      Flynn says a lot of rubbish. But my point is that causal explanations can be tested, to a minimal extent, with MCV. The set of proposed factors needs to be able to produce JE+. Do you disagree?

      You said: “The point is not whether the one relation is stronger than the other or something. The point is that the relations coexist. This coexistence is what needs explanation. g theories do not provide any account for It (nor any other theory).”

      I agree that this is something that needs to be explained. And that complexity + investment theory, as you formulated it, can’t do this. I did note before, though, that “biological g theory” shouldn’t be equated with complexity + investment theory as you formulated it. And I quoted Jensen (2011) to clarify this point.

      Whatever the case, to my mind, you left uncertain whether g mediated the cultural-loading h^2 correlation for the Minnesota sample and the cultural-loading BW difference correlation for the Jensen (1985) sample. Maybe you figured that this was obvious. I don’t know. One way that I was planning to explore this further was to see if cultural-load correlated with c^2 using URT. But perhaps you would consider that to be a waste of time?

      You said: “This shows two things: 1) In a homogeneous population heritability of crystallized abilities will approach the heritability of the underlying fluid abilities, not that heritability of crystallized abilities is HIGHER than heritabilty of fluid abilities.”

      I gave you a 2 part reply:
      (a) cultural influences need not reduce h^2
      (b) tests that tap into cultural influences arguably can index a broader range of ability than tests that don’t.

      The latter wasn’t just my musing:

      “These two research groups concluded that the high g loadings of crystallized subtests in batteries such as the Multidimentional Aptitude Battery (MAB) and Wechsler Adult Intelligence Scale-III (WAIS-III), are not due so much to the overrepresentation of verbal subtests, as the fact that fluid ability subtests are more narrowly-defined tasks with more unique variance, leaving crystallized subtests to represent most of the common variance attributed to g”

      Can you code subtests by “task broadness”? Can we see if: Cultural-load –> “task breadth” –> g –> heritability, where we operationalize “task breadth” to more than just cognitive complexity = number of mental manipulations? Generally, I think your basic analysis can be built on by adding more indexes of subests kindness i.e., complexity, breadth, c^2, informationality, etc. So, yes, “the relations coexist” — and this is interesting — but it would be nice to see if the relations is statistically explainable by some other variables.

  4. KJK says:

    “Cultural wouldn’t be expected to reduce heritability unless the population was differentially exposed to the relevant information. Jensen 1973 made this same point.”

    This shows two things: 1) In a homogeneous population heritability of crystallized abilities will approach the heritability of the underlying fluid abilities, not that heritability of crystallized abilities is HIGHER than heritabilty of fluid abilities. 2) While within two homogenous populations h2 may be high (even 100%), mean differences between can be due to mean differences in the exposure to the relevant information (e.g., different chances to higher educational systems)

    let’s put this in mathematical terms:

    Fluid ability = A + E
    Crystallized ability = Fluid ability + E_cultural = A + E + E_education

    In a completely homogeneous sample with respect to E_education var(E_education) is 0 , so that individual differences in Crystallized ability reflect individual differences in Fluid ability, hence individual differences in A and E. Suppose population A is homogeneous with respect to E_education, with mean A=100, mean E = 100, and E_education = 100. Suppose population B is homogeneous with respect to E_education, with the same mean of A (100) and the same mean of E (100), the same variances of A and E. but with mean E_education = 200. Next, ask yourself the question, what are the sources of the individual differences within population A? (answer: A and E [not E_education]). What are the sources of the individual differences within population B? (answer: A and E [not E_education])? What is the source of the group difference? (answer: E_education [not A, not E])

  5. KJK says:

    I do not have much critique on Rowe and Cleveland. Multi-Group Analysis with Structured Means is exactly the way I advise researchers to do their analyses. You see, no mediating biological g is necessary. In fact, a mediating g is unlikely, because the pattern of loadings on the genetic factor are different from pattern of the loadings on the environmental factor (that’s why the model is identified; a model in which g mediates is not).

    Just a pity the researchers did not compare their model with other models, e.g. a model in which the test scores also directly load on the 1. In that way they could have tested whether their model is preffered over a model in which the pattern of the mean is different from the pattern of the covariances. Now we only know that the model fits reasonably and fits with Spearman’s hypothesis, but whether Spearman’s hypothesis is preferred over other hypotheses.

    • Chuck says:

      You said: “because the pattern of loadings on the genetic factor are different from pattern of the loadings on the environmental factor (that’s why the model is identified; a model in which g mediates is not).”

      I imaged as much. Thanks for the clarification as to how this works.

      You said: “In that way they could have tested whether their model is preffered over a model in which the pattern of the mean is different from the pattern of the covariances.”

      I don’t well understand this. But I thought that Rowe discussed this issue here:

      Rowe et al. 1994. No More Than Skin Deep: Ethnic and Racial Similarity in Developmental Process
      https://lesacreduprintemps19.files.wordpress.com/2012/05/no-more-than-skin-deep-ethnic-and-racial-similarity-in-developmental-process.pdf

      But let’s say I paid someone to do a biometric analysis of the recent longitudinal ECLS — and they found similar results — and they also found that the pattern of means didn’t differ from that of the covariances. Would this get me anywhere in terms of understanding the cause of these groups difference? I mean Flynn and Dickens just dismiss this stuff. For example:

      “The more the pattern of black-white differences across different tests resembles the pattern of genetic influence on different tests, the more the statistical procedure will attribute the blackwhite differences to genetic differences. Using this method, David Rowe and Jensen have independently estimated that from one half to two-thirds of the black-white gap is genetic in origin…. […] Those cognitive abilities for which multiplier processes are most important will be the ones that show the largest heritability, because of the environmental augmentation of the genetic differences. But they will also be the ones on which a persistent change in environment will have the biggest influence. Thus we might expect that persistent environmental differences between blacks and whites, as well as between generations, could cause a positive correlation between test score heritabilities and test differences. (Dickens, 2005. Genetic Differences and School Readiness)”

      So, I just assumed that this was a dead end.

  6. KJK says:

    I don’t think there is a dead end. Rowe and Cleveland for example used a model which contains the assumption that genetic and environmental influences are uncorrelated. This assumption is highly unrealistic. The problems are 1) that if GE covariance is present, part of environmental variance will contributed to genetic variance 2) if GE covariance is absent, g theory does not explain the relation between cultural load, g-loading, and heritability.

    The Dickens & Flynn model can account for the interrelations between cultural load, g-loading, and heritability, under certain assumptions, e.g. when weak pleiotropy is assumed (general general are not necessary). If these ‘certain assumptions’ are not made the model accounts for the relation between cultural load and g-loading and the relation between cultural load and heritability, but not for the relation between g-loading and heritability. In the discussion section of my thesis, I point out that g theory can also explain the relation between cultural load, g-loading, and heritability, but also under additional assumptions, e.g., if it includes – among other additional ssumptions – that there is GE covariance (e.g., education depends on g, which would be very reasonable), but then you encounter problem 1. If the additional assumptions are not met, g theory does explain the relation between g loading and heritability, but does not provide an explanation of between g-loading and cultural load and between cultural load and heritability. Again, this is not to say that the theory is UNABLE to explain it, merely that the theory LACKS an explanation.

    In any case, all theories of intelligence need to be more specific and the sources of group differens in intelligence are still open to debate.

  7. KJK says:

    It still holds that results from the MCV are not informative (multivariate or not)

    • Chuck says:

      “It still holds that results from the MCV are not informative (multivariate or not)”

      Well, I disagree. Whatever the case, we found a meta-analytic correlation between c^2 and g that was higher than the correlation between h^2 and g, (now) K=29, n pairs = ~15,000. Using Wechsler’s tests and some others, I was able to replicate your r (h^2 x cultural loading) finding. Strangely, from my perspective at least, there didn’t seem to be a positive r (c^2 x cultural loading) (at least based on how you defined cultural loading). But I would have to look at this more closely — and I will, I guess, if you want. I had thought that I could explain away your results by showing that g mediated the cultural loading x h^2 correlation — which it does seem to — and that the cultural loading x g correlation could be statistically explained by a r (cultural loading x c^2). But possibly not. What do you make of the positive r ( c^2 x g)? Does this matter? I don’t think that it’s just an artifact of parameter estimate standardization or models used — as Dolan suggested — because we found the same using a direct measure of c^2 i.e., the correlation between unrelated sibs e.g.,

      • KJK says:

        You can disagree, but I’ll prove analytically the results of the MCV are not informative.

        [I already explained this. Your phrasing “not informative” is too strong.]

        For the rest, as I said, the sizes of the cor(gl,h2) and cor(gl,cl) etc do not matter. What matters is that gl, h2 and cl are interrelated. Whether the relations c2 between gl and cl are absent or not is not relevant to these relations. It probably means that genetic and environmental influences have different pathways or other dimensionalities. You can keep juggling with and arguing about vector correlations like those, but it’s better to spend time on statistical methods that address your original questions….

        [I agree with the last sentence. As for r (c^2 x g), the question is whether this relation is of interest in its own right. Here were the results by the way. But ya, I am done with MCV.]

        • KJK says:

          So, how is it too strong?

          • Chuck says:

            g x parameter estimates are causally informative in that they can say something. For example, r (g x e^2) = -0.50 says that unshared environmental effect on average will not cause a positive Jensen Effect. That is, if group differences were caused by a random sampling of the unshared environmental influence that caused within population differences then the between population difference would exhibit an anti-Jensen Effect. Is this itself informative? You argue no because you don’t see why group differences would necessarily have to be caused by a random sampling of e^2. I agree that they don’t have to — but argue that it is a priori less probable that the e^2 causing the group difference would be a JE causing e^2 — meaning that of all group differences caused by e^2, few will exhibit a strong JE. This allows one to construct a probabilistic argument against e^2 causation. Since I can construct such an argument, g x parameter estimates is not completely causally uninformative. And of course, showing that a specific unshared environmental influence did not tend to induce JE would be informative. As I said, theoretically, one could go through numerous causal hypotheses this way as done by some of your compatriots: http://lesacreduprintemps19.files.wordpress.com/2013/05/metzen-2010.pdf

  8. Steve says:

    Hi, I’m having a little discussion with somebody on youtube about the impressive educational attainment of African immigrants to the US. This person is pointing out that African immigrants have the highest percentage of people with degrees of any immigrant group (including the lauded Chinese) and also US whites. Obviously that doesn’t tell us who has the best grades! I was wondering if you know about this, if you can shed any light on it. Also, how selected are they and i there iq data for them?

    thanks

  9. Chuck says:

    Volar,

    This is a imbecilic reply. Does anti-racism only attract the extremes of the curve? Those capable of ingenious spin and those stupid enough to take it seriously? First, Steve’s question concerned dysgenic fertility in general. This is now a global problem. See, for example, “Differential Fertility, Human Capital, and Development”:

    “Discussions of cross-sectional fertility heterogeneity and its interaction with economic growth typically assume that the poor have more children than the rich. Micro-data from 48 developing countries [e.g., Mexico, Bangladesh, Indonesia] suggest that this phenomenon is very recent. Over the second half of the twentieth century, these countries saw the association of economic status with fertility and the association of the number of siblings with their education flip from generally positive to generally negative”

    So, yes, you are wrong on this account, Whether fertility is going down doesn’t matter because it’s increasingly becoming dysgenic with respect to human capital.

    As for the racialists — this is more stupidity on your part. Genetic engineering will come online within the next century. Look up “biological liberalism” and “neoeugenics”. Expect radical “racial” diversification, perhaps subspeciation– for those wanting it. Think: wrath of Khan. Or do you imagine that a global government will repress the realization of a true transhuman diversity? Generally, I find your idiotic reflexive anti-racism to be offensive. “Ya, Amerika! Lev Davidovich Bronshtein, my hero!”

    Your other comments are just sheer buffoonery. I suppose this is why excessive stupidity is undesirable. I’m just not sure that excessive intelligence isn’t. Just look at the types who do your thinking. Take this statement:

    “Also mixed means way, way, way more combinations of genes, which means more diversity.”

    This is just ridiculous. it means less variegation. So putting chocolate, vanilla, and strawberry ice-cream into a blender is an act of “diversification”. If you are going to act like a moron, don’t comment on my blog.

    “So technically “race” mixing is one of the best things that can possibly happen to the human race in every way.”

    Genetic variety is typically seen as valuable in species. It prevents genetic Cul-de-sacs.

    “In addition to the above arguments, there is a much larger reason to embrace human diversity in all its forms, in our view. Humanity’s genetic diversity — small or large, within or among groups — is a resource for, rather than a detri- ment to, creating a more fulfilling and prosperous society. Just as people have come over time to cherish cultural diversity, so we hope that attitudes will warm towards genetic diversity. In the natural world, genetic diversity is a source of evolutionary resilience and adaptability. It buffers against changing environments and allows species to occupy broader and more fluid ecological niches. Even for a single individual, differences between its two copies of
    the genome can often lead to higher fitness. Indeed, sexual reproduction is thought to have evolved as a way for species to take advantage of genetic diversity. Consequently, the loss of a species’ diversity often threatens its long-term survival. The susceptibility of agro-monoculture to sudden disease outbreaks or climate changes is just one example.” (Let’s Celebrate Human Genetic Diversity.)”

    But none of this matters. Because genetic re-engineering is inevitable. The conflict of the 21st century will not be between racial preservationists and racial eliminationists — but between biological liberals, who will embrace genetic artistry, and biological conservatives, who will cling to their images of old types and to relative human similarity.

    • Xsilent says:

      “Hey Chuck he is right about the diversity part though.”

      Again, go back to the ice cream analogy: vanilla, chocolate, strawberry into the blender. Turn it on and tell me what you get after five minutes. Now I agree that a bowl of ice cream that has vanilla, chocolate, and strawberry is more diverse, in a sense, than one that has just vanilla. And that if you begin mashing up the ice cream it looks like you have a lot of diversity — a unique bowl of vanilla, chocolate, strawberry swirl. But this is swirl on the way to uniformity. And if you are swirling all of the ice cream then you are headed for one big boring mixer of unvariegated, blah, ice cream. Is that a bad thing? Surely if you value conserving global genetic diversity.

      • Xsilent says:

        You cannot compare human genetics to ice cream. That is not how it works. That is a very bad analogy.

        Each human is a unique combination of genes from one gene pool and that gene pool gets bigger when people mix. So the more people mix the more combinations become possible. The old ones like white people are still going to show up. Mixing greatly increases diversity, each person technically becomes more unique to the other not the same.

        I thought you knew this already? Are you joking? I can’t tell.

        • Xsilent says:

          Races cannot go extinct via mixing. There are no unique genes for races, not even any combinations of genes, even the frequency of genes per race is not fixed for that race. Two blacks can give birth to a white and two whites can give birth to a black, which technically means not even the race of someone is unique or exclusive to that race.

          • Chuck says:

            In zoology, racial extinction is recognized; as such, endangered species acts typically include conservation measures for subspecies. Here is an excerpt from a typical discussion in the literature: In our view an allopatric subspecies has four possible fates; it may: (i) go extinct; (ii) exchange genes with another subspecies and become a new “mixed” subspecies; (iii) by genetic drift, selection, subdivision, or other demographic processes change its genetic character over time to become one or more new subspecies; and (iv) if effectively isolated, become a new species by acquiring genetic isolating mechanisms. ((O’Brien and Mayr, 1991. Bureaucratic Mischief: Recognizing Endangered Species and Subspecies.)

            Subspecies either (1) go extinct (in the way that species do), (2) merge (with other subspecies), (3) change (in the sense of temporal differentiation), or (4) speciate. A racial conservationist wants to maintain racial diversity (i.e., diversity of unique patterns of genes and phenotypes within a given species) and so prefers (3) but is not opposed to a moderate amount of (4); that is, she doesn’t mind evolution, just the absence of interesting patterns. I presume that a transhumanist would prefer (4), a misanthrope would prefer (1), and a make-the-world-flat idiot would prefer a radical version of (3). Personally, I don’t understand the latter. Perhaps, egalitarians want uniformity for the sake of equality; and some mixed race individuals are ressentful> on account of some perceived inferiority.

        • Chuck says:

          “The old ones like white people are still going to show up.”

          Intentionally or not, you’re engaging in fallacious reasoning. Race is defined in terms of the correlation — i.e., the unique patterns — of genes and the resulting patterns of phenotypes (i.e., a concordant distribution of multiple, independent, genetically based traits). Race mixing breaks down this unique pattern. Overtime, if reproductively isolated, the hybrid population will develop it’s own unique patterns (correlated genes and phenotypes). But it will be a different one. If you don’t like the ice cream analogy, we could compare races to paintings painted on the same type of canvass using the same pigments. The difference between paintings emerges from the different patterns of colors. A mixed race population is an incohate race.

  10. Chuck says:

    Steve said: “Chuck, last question: I just looked it up and apparently twin studies have shown approx a .75 correlation between iqs of identical twins reared apart. Is that twins both reared in middle class homes or does the same high correlation hold when there is a big difference in class, eg one twin is working class and the other is upper class.”

    Steve, concerning MZA see these opposing critical reviews:

    http://humanvarietiesdotorg.files.wordpress.com/2013/05/the-effects-of-shared-environment-on-adult-intelligence-a-critical-review-of-adoption-twin-and-mza-studies.pdf

    http://lesacreduprintemps19.files.wordpress.com/2011/07/lee-2009.pdf

    MZAs and DZAs and URTs generally don’t capture the full range of environmental variance — but this isn’t the problem that environmentalists make it out to be since this is just “range restriction” and you can correct for this by adjusting for the restricted standard deviations. I mean this is an obvious point. So, much of the criticism in this regards is just dishonest. Whatever the case, GCTA is rendering all of this debate moot. http://www.complextraitgenomics.com/software/gcta/ I mean, when GTCA is applied to nationally representative samples in the UK — look up TEDS — you get about the same results as from classic twin studies. More generally, “equal environments” no longer needs to be assumed. Also, there is a multivariate method under development which will soon resolve the issue that KJK brings up concerning active COV(GE).

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s