Redebunking Ron

(1) On Ron’s Hispanic Flynn effect (a.k.a The Ron effect)

Ron’s at it again with his distortions and omissions. Here’s from his latest:

Naturally, such difficulties are non-existent for those individuals whose overweening self-confidence leads them to lob endless casual insults all across the Internet while regarding the evidence of their anecdotal personal experiences in New Mexico as vastly superior to my quantitative analysis

He notably failed to mention Unsilenced Science’s and my extensive rebuttals, in which his claim concerning a large Hispanic Flynn effect was utterly demolished. For my own part, I marshaled evidence from no less than 14 nationally representative samples ( PISA 2009, PISA 2006, TIMSS 2007, PIRLS 2006, PIRLS 2001, ADD Health, GSS, NLSY 79, NLSY 97, CNLSY, NAEP Main, NAEP Long term, High School Transcript Study, and NELS) and demonstrated beyond a doubt that the claimed Ron effect was a mirage. Anecdotal evidence, indeed.

Refer to the following:

Hispanic “Flynn Effect” in the NAEP

The Arnaz-Ball effect?

Hispanic performance by generation

Hispanics, the NLSY 97

I did, nonetheless, find a meager secular narrowing in intelligence differences (as opposed to mere achievement differences). As I noted in context to the NLSY 79, NLSY 97, and CNLSY study: “Taken together, we can conclude that there was at least a 2 point narrowing between 3rd + generation Whites and Hispanic between 1979 and 1997. And that the magnitude of this narrowing could have been as high as 4 points.”

But I also noted that there seemed to be an intermarriage effect, such that Hispanic who were deprived of having one non-Hispanic White parent showed little to no secular increase, relative to Whites. To further investigate this, I decomposed the NLSY 97 AFQT and PIAT achievement test results down by parental ethnic identification. A parent was said to be Hispanic if they identified with at least one Hispanic national origin. Multiple selections were allowed, so this was an inclusive form of categorization (i.e., a person who marked Irish, English, and Mexican was classified as Hispanic). Likewise, a parent was said to be Black if they indicated that they had any African origin. I included only European, African, and “Hispanic” (including Spanish) origin categories, so individual who did not select African or Hispanic origins were European.

Normed on the European IQ, using European standard deviations, the AFQT IQ for European-Europeans, European-Hispanics, and Hispanic-Hispanics was 100, 96, and 86. It can also be seen that the European-Black IQ was 95. The latter is 3 points above the figure that the adroit scholar Aaron Gullickson found using a likely more exacting method of identification (his PhD thesis versus my reply to ron). I’m also missing information for a number of subjects. My biethnic/racial, estimates, then might be upwardly biased. But the mono/bi difference is completely consistent with the high skin color-IQ correlation in the Hispanic (and Black) population (see: The Arnaz-Ball effect?).

As a side note: It will be noticed that the more heritable and more g-loaded AFQT scores show larger gaps for both Blacks and Hispanics than the more environmentally influenced and less g-loaded PIAT achievement score. The biracial/ethnic AFQT scores also show greater intermediacy.

Compare, now, these results with those discussed in in “Hispanics, the NLSY 97.”

Dis be bad.

All of the meager Ron Unz effect is attributable to the effect of cultural and/ or genetic assimilation. Hispanics aren’t narrowing the gap by themselves as implied by Ron. Given the accuracy of these findings, we can conclude that importing hordes of Hispanics without allowing for time for “assimilation” is a disastrous policy, regardless of the cause of the difference.

(2) On Ron’s Irish Flynn effect

Ron has rested much of his case on a supposed large secular increase in the Irish national IQ. In his most recent article, he pens:

my critics naturally focused their most intensive fire on the Irish side, denying any large change in the local IQ

To make his case, Ron focuses on an outlier of an outlier, an extreme score of a nation that has shown considerable IQ variability. The magnitude of variability itself is a non-issue for those familiar with the topic. Just as with individuals, the IQ scores of nations can bounce around considerably, despite being highly reliable on average. The table on the right gives an example of how IQ scores can vary for the same individual by test holding age constant. One can see that the ranges at the upper extreme approach 1.3 standard deviations. Since it’s already established that individual differences are, on average, highly genetically conditioned, it follows that such between test variability, per se, is not inconsistent with a “strong” genetic hypothesis. The same consideration holds at the national level. With national IQs, of course, one would expect even greater variability at the extreme as scores are derived from samples which comprise different age cohorts and which are often not nationally representative (for that particular cohort).


Kaufman, 2009. IQ Testing 101

Nonetheless, Ron claims to have found a definite pattern in the Irish national variability:

but an exceptionally strong correlation of 0.86 seems now to have decisively resolved this dispute in my favor

Here, Ron continues to claim without sound basis that the Irish have undergone a super-duper Flynn effect; as evidence he points to a purported statistically significant correlation between year and Irish test scores. Statsquatch and I have shown though, respectively, that there is no such correlation between year and IQ score and year and achievement (AQ) score, analyzed separately. Ron, no doubt, would maintain that there are no grounds for looking at IQ and AQ independently. After all, Richard Lynn combines them. However, Lynn properly groups the scores separately and only combines them to form a total score. The theoretical basis for this is that the general factor of intelligence is not identical to the general factor of achievement (Kaufman et al., 2012), though both are highly correlated. This is why it’s possible for some nations to have large consistent differences in IQ and AQ, as example Argentina, which averages a 10 point difference between the two indexes.

Regardless, even if we improperly amalgamate scores, Ron’s claim turns out to be, as typical, ungrounded. Immediately below is the full current list of Irish IQ and AQ scores. (Refer to Statsquatch’s and Chuck’s post for sources). Following are Pearson’s and Spearman’s correlation coefficients. Ron properly maintains that we should use weighted scores, but he neglects a more important issue. Since at issue is a Flynn effect, or a generational rise in intelligence, we also need to use birth years (calculated by subtracting the average age of the sample from the test year). This is the standard procedure when it comes to assessing the magnitude of a secular increases.

It can be seen below that there is, in fact, no significant correlation between birth year and amalgamated score, weighted or not. Certainly, a statistically significant correlation can be found between birth year and scores for some of the nations which Lynn has provided data for, but Ireland, at this time, is not one of them – not that we can expect Ron to admit this.

(3) Jensen’s Mexican research

In defense of his Mexican-White environmentalism, Ron now cites early research by Arthur Jensen:

Worse still for my critics, Eysenck’s discussion had been drawn directly from the work of Arthur Jensen, who had conducted detailed research studies on the question of Mexican-American intelligence and had similarly concluded that the IQ gap with whites was overwhelmingly “environmental” in origin. Now I claim no great expertise in IQ matters, but those who do seem to regard Jensen as one of the most towering figures in the history of their academic discipline.

Presumably, this might place my numerous and noisy IQ-activist critics in a bit of a quandary

Ron is referring to a study done by Jensen which was discussed in “Educability and Sub population Differences” (p.g. 305-312). To summarize, Jensen compared White, Mexican, and Black performance on PPVT and Raven’s matrices. He reasoned that since PPVT was more culturally loaded it should be less heritable than Raven’s matrices. Therefore, one could regress the more heritable Raven’s differences on the more environmental PPVT differences and see which were large for which group. Larger differences on more heritable tests would be more consistent with a genetic hypothesis while larger differences on less heritable tests would more consistent with an environment. Those familiar with the subject will realize that this was Jensen’s early investigation of the Jensen Effect (the positive association between a mean difference and heritability coefficients – discussed in section G, here).

Jensen (2012) recently summarized the logic of this analysis:

“I had demonstrated in my research of the 1970s that mean Black–White differences in IQ were more pronounced on the more heritable, less cultural subtests. For example, Jensen (1973) cited a study by Nichols (1972) which found a correlation of r = .67 (p < .05) between the heritabilities of 13 tests estimated from twins and the magnitude of the Black–White differences on the same tests. I further demonstrated an inverse relation of r = .70 (p < .01) between the environmentality (the converse of heritability, that is, the percentage of variance that can be attributed to nongenetic factors) for 16 tests estimated from differences between siblings and the mean White–Black differences (Jensen, 1973)…
[...]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; (3) nature + nurture models predict both genetic and environmental contributions to group differences; while (4) culture-only theories predict a zero relationship between heritability and group differences. These results provide strong and reliable corroboration of the hypothesis that the cause of group differences is the same as the cause of individual differences, that is, about 50% genetic and 50% environmental (Rushton & Jensen, 2005, 2010) (Jensen, 2012.
Rushton’s contributions to the study of mental ability).

Those familiar with also recognize that this was a primitive forerunner to the structural equation modeling (SEM) studies using biometric decomposition (discussed in section H here). Unfortunately, these latter studies have not been conducted on Hispanics, as such there is only strong SEM evidence for a Black-White genetic hypothesis. (Which Ron curiously omits discussion of.) However, it so happens that heritable differences go hand in hand with g-loaded differences (Spearman’s effects). And Spearman’s hypothesis has been found to hold for 3.5 out of 4 studies involving Hispanics and Whites. (As an example of this Spearman’s effect refer back to the discussion above concerning the NLSY 97 AFQT and PIAT scores.) We can therefore infer that the White-Hispanic difference also positively correlates with heritability. If not, this would imply that Jensen and Spearman effects need not coincide, which, if so, would figure badly for environmental explanations for the B/W gap. (The g-heritability correlation is how environmentalists account for the B/W Jensen effect; it’s argued that the Jensen effect is collinear with Spearman’s effect, or a g-loaded difference, and that a g-loaded difference can readily be environmentally induced, and therefore that a correlation between a group difference and heritability can also readily be environmentally induced.)

This whole issue has been somewhat complicated of late by Dutch psychologist Kees-Jan Kan who has found, queerly, a strong correlation between cultural loadedness, g-loadedness and, presumably, test heritabilities. You can refer to his discussion of this in the comment section of my “Nothing is irrelevant” post. This finding seemingly undermines Jensen’s initial presumption that the cultural loadednesss of a test is inversely proportionate to its heritability.

In sum, Jensen’s early research project – and, in fact, his late research project – was and still is inconclusive. Where it otherwise, the contentious debate concerning the origin of the Black-White gap would have been resolved.

(4) On false dilemmas and reverse arguments from authority

Above are but a few instances of where Ron has either been grossly in error or intentionally misleading. Realizing the empirical shakiness of his position he has resorted to a handbag of fallacies. His remaining argument is some permutation of:

Ron’s False Dilemma
Either Lynn/Eysenck was right about the low Irish IQ in the ‘70s and there was a massive Irish Flynn effect or he was wrong and his judgment isn’t to be trusted. If the former, (b) it’s clear that national IQs are highly malleable, and therefore Lynn’s is incorrect about his genetic hypothesis; (b) if the latter, Lynn’s national IQ data can’t be trusted.

No matador is needed to handle this bull. It can be squarely grabbed by the horns.

(a) Firstly, heritability is a population level statistic. This applies to nations just as well as individuals. As such, even if it were shown that the difference between nations X and Y were completely environmental, this would imply nothing necessarily about the difference between nations A and B.
(High heritability only allows one to make probabilistic statements on the individual level. The quantitative expression for the relation between heritability and individual deviations from the population mean was given by Omri Tal:


Tal. 2010. From heritability to Probability)

Richard Lynn has claimed that approximately 50% of the between national variance in IQ is heritable. For this to be correct, it’s obviously not necessary, given what was said above, for 50% of differences between all (or any) nations to be genetically conditioned. The total amount of variance explained by genes just has to add up to 50%.

Now this theoretical point has practical importance. To make the case that the difference between B and C is not genetic, it’s not sufficient to merely show that the difference between A and B is (probably) not. One has to show: That some environmental factor(s) X plausibly causally explain the difference between A and B, that these factors could causally explain the differences between B and C, and that B and C differ in these factors by a magnitude enough to explain the B,C gap.

So, for example, Ron has argued that the Irish English gap has narrowed (but see above). He has postulated that “ruralness” was the cause of the supposed original difference. The next step, in his chain of arguments, would be show that a bucolic lifestyle could plausibly be the cause of the English-(Hispanic) gap (e.g., rural deprivation can cause a g gap) and that English and Hispanics differ in ruralness, on average. With the plausibility of that theory established we can then check if ruralness does, indeed, statistically explain the gap. After this, we can try to control for genetic covariance.

Of course, Ron can always postulate environmental causes – just as Lynn can postulate genetic causes. But he has not demonstrated anything.

(To help him out –
Here is a link to cross national data on urbanity.

http://esa.un.org/unpd/wup/CD-ROM/Urban-Rural-Population.htm

Step 1 is to show the strong positive international correlation between National IQ and urban living).

(b) Secondly, one need not solely rely on L & V’s National IQs. One could alternatively use those of Rindermann (2007) or Altinok and Murseli (2006) or Altinok et al (unpublished). Here’s the correlation matrix:

For a list of sources refer to the first part of my “National IQ references.”

The scores aren’t perfectly matched. For example, Altinok et al. give an (equivalent) National IQ of 54 for India compared to L & V’s 82. But, as discussed above, large individual range differences are consistent with high group reliability. And not inconsistent with genetic differences, so long as one appreciates than an individual point scores in an imperfect representation of “true latent ability,” which itself is an imperfect representation of genetic potential.

The point here is that Ron’s reverse argument by authority fails right off the bat because Lynn is just one of the many researchers that are investigating the established national differences in cognitive ability, human capital, or IQ — whatever one wished to call them.

So far everything Ron has presented is utterly dismissible.

That said, a strong case against global hereditarianism of even a modest amount can be made (forthcoming, perhaps). Ron just isn’t making it.

About these ads
This entry was posted in Uncategorized. Bookmark the permalink.

24 Responses to Redebunking Ron

  1. statsquatch says:

    Of course, correlate Irish IQ with birth year! I wish I had thought of that.

    When you mean global heriditarianism do you mean IQ difference between various ethnic groups or the general within group hereditary nature of IQ? If the later, good luck.

  2. x says:

    he most undoubtably means the former

  3. Amibiguous says:

    Good point about the weighted mean, why doesn’t Lynn weight data like that? Btw I calculated a weighted mean for Thailand’s IQ = 97.

    • Chuck says:

      Re: Thailand. Thanks. So, using Lynn’s method: IQ = 97, Quality = ~25; AQ = 91, Quality = 12. National IQ = 95. Looking at it this way, Lynn was only 6 points off. It’s interesting that other South Asian nations are turning out to have relatively high national IQs (e.g., Malaysia’s AQ). I don’t see that this, though, necessarily completely supports an environmental position. Latin American nations which have comparable HDIs do much less well. To me, it, perhaps, suggests a more generalized Mongoloid advantage. It will be interesting to see Vietnam’s 2012 PISA scores when they come out.

      • Amibiguous says:

        I’m not at all familiar with Lynn’s new method but it seem very questionable to use academic tests like PISA to derive an IQ. What was his IQ estimate for Thailand just based on IQ tests? I think he was more off if you just look at those. Given the massive national IQ survey over 70,000 pupils gave an IQ of 99, and Bangkok and Central Thailand scored around 104-108, and given that a massive proportion of the population lives in that region, I think 99 is likely an underestimate.

        • Chuck says:

          (1) All tests are g tests. They merely differ in degree of reliability. So using PISA tests, which correlate with g at about 0.8 — my estimate based on the subtest inter-correlations — is as valid as using a number of so called “true IQ tests,” for example PPVT. What’s the correlation between Raven’s and g in Thailand?

          (2) I don’t see the point about “Bangkok and Central Thailand”; there are regional differences in every country. And the Thai scores seemed to be n-weighted.

          I can’t say more about the Thai score until I read the paper and see how scores were calculated.

          But, ya, the results are interesting because Thailand isn’t super developed.

          If you want you can check the Vietnam Raven’s and PPVT scores from the large Young lives survey. I have a copy of the results floating around but I can’t locate it — Ethiopian, Peruvian, and Indian scores can also be found for that survey.

          http://www.younglives.org.uk/our-publications

  4. JL says:

    That said, a strong case against global hereditarianism of even a modest amount can be made (forthcoming, perhaps).

    Do you have new data, or do you mean in a conceptual, psychometric sense? Lynn’s global hereditarian thesis is of course not really supported by his data because the psychometric meaning of the international gaps is currently indeterminate.

    • Chuck says:

      JL,

      (a) I have some new international immigrant data — but I’m still collecting it and the results are not without ambiguity, so I won’t comment on it at this time.

      (b) In the latest book L & V more or less acknowledge the psychometric issue. Refer to page 9:
      “We regard national IQs as measures of general intelligence defined as the totality of cognitive abilities. These include Spearman’s g … and other cognitive abilities that are independent of g. Thus we do not regard national IQs as necessarily measures of Spearman’s g…

      (c) Did you get a chance to look at the CNLYS?

      I tried to the other night but didn’t get anywhere. The problem that I’m running into is that year cohorts (e.g., “96 and ’02) are given separate weights. So it seems that one has to look at each of the dozen or so cohorts separately. Is it possible to combine weights across years? I really don’t have the patience (or energy) to go through the results year by year.

  5. gcochran9 says:

    I think Ron was probably talking about me, since I have lived in New Mexico for about 20 years. I noted that the academic convergence he is talking about has not occurred in New Mexico, even most of the Hispanic population here has been living in the US for 160 years. I didn’t go around the Internet insulting him, although I did say that I thought he was wrong.

    • Chuck says:

      He weaselly phrased his statement in a way to suggest that others were also relying merely on anecdotes. Notice the plurals:

      “Naturally, such difficulties are non-existent for those individuals …while regarding the evidence of their anecdotal personal experiences in New Mexico.”

      So, in turn, I phrased my statement in a way to suggest that he was erroneously accusing me and others of not relying on hard evidence.

      No misunderstanding.

  6. Kiwiguy says:

    OT. Intellectually disabled athletes are allowed to compete again at this Paralympics.

    “All competitors in these sport classes have to fulfil the World Health Organization (WHO) definition of intellectual disability:

    An IQ below 75

    Impairment in adaptive functioning – for example, social, domestic and communication skills
    The disability must have occurred before the age of 18″

    http://www.bbc.co.uk/news/magazine-19371031

  7. Toddy cat says:

    “Given the accuracy of these findings, we can conclude that importing hordes of Hispanics without allowing for time for “assimilation” is a disastrous policy, regardless of the cause of the difference. ”

    Very true, and this is exactly what Ron Unz does not want to hear, and will, seemingly never admit. Unz is a smart guy, and he’s done a lot of good for conservatives in various ways, but he is simply not rational on this issue. He simply will not admit that unlimited immigration to the U.S. of people from a very different culture is not, cannot be, a good thing for the American order as we know it. He just has a blind spot on this issue, and nothing will convince him. Kind of sad, actually.

  8. Steve says:

    “That said, a strong case against global hereditarianism of even a modest amount can be made (forthcoming, perhaps). Ron just isn’t making it.”

    Why don’t you make it then?

  9. Pingback: linkfest – 09/09/12 « hbd* chick

  10. Pingback: Unz on Race/IQ: The Entire Series and Debate | Ron Unz – Writings and Perspectives

  11. Pingback: Color Differences: Corrections and Further Analysis. Part 2 | Human Varieties

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