a couple of months ago, greg cochran wrote about how a population (any given population) might raise its average iq quickly — like over two or three generations. one possible method, he suggested, was to quit inbreeding. here’s what he said:
“I’ve been thinking, off and on, about sudden changes in the cognitive abilities of populations: groups low suddenly scoring much higher or lower on a time scale too short to be explained by selection: say, three generations or less…. I can think of two perfectly feasible strategies that *would* cause significant one-generation increases in intelligence, in certain populations. Iodine supplementation, in places where it’s short, has a big payoff…. The other practical, low-tech strategy would be stopping cousin marriage. The next generation would be in much better shape, since the children of first cousins take a substantial IQ hit – maybe six points or so.“
i think i may have stumbled across an example of a population rapidly dropping cousin marriage and, also very rapidly, gaining iq points.
yesterday, i was working on a completely different post about japan and cousin marriage, when i rediscovered this [pg. 30] (click on chart for LARGER version):
that’s the decline in the average national consanguinity rates in japan from 1947 to the early-1980s. but the consang rates were even higher in the 1910s-1920s at 22.4% [pg. 29] (and who knows how high the rates might’ve been even further back?), so the chart above should look something more like this (pardon my crayola — and note that i just eyeballed it — click on chart for LARGER version):
when i rediscovered this yesterday, i remembered what greg had written, and got to wondering if there were any historic data for japanese iqs and if there’d been any changes in those iqs over time. so i googled (as one does) … and found this:
The Rise of National Intelligence: Evidence from Britain, Japan and the U.S.A. [pdf] – lynn and hampson.
from that article [pgs. 27-31]:
“It has been possible to find five studies providing evidence on the secular trend of intelligence in Japan for the post World War II period….
“(1) Ushijima’s study (1961)….
“Here the Ushijima intelligence test was administered to 1365 children in 1953 and to a comparable sample of 1370 children in 1960 with the objective of determining any change in the mean over this relatively short period. The children were aged 9-15 years.
“(i) All age groups show a rise in scores for all abilities. The overall mean increase was 0.66 standard deviations, the equivalent of 9.9 IQ points, and representing an IQ gain of 14.1 IQ points per decade. This is of course a very considerable increase and much greater than anything found in either Britain or the U.S.A.
“(ii) The IQ increases are in general greater among the younger age groups than among the older….”
see my crayola chart above. kids who were fifteen in 1953 would’ve been born way back in 1938 when the consang rates were above 15% — maybe 17 or 18% (remember, i just eyeballed it, so this is a complete guess really) — while kids who were nine in 1953 would’ve been born in 1944 when consang rates were hovering right around 15%, in other words lower. same for the later cohort from 1960: fifteen year olds would’ve been born in 1945, nine years old in 1951. the younger the kids in the cohorts, the less chances their parents were related.
more from the article…
“(2) Kaneko’s study (1970)….
“Kaneko’s invetigation of a possible rise in the scores on this test was carried out in 1963….
“Hence the mean IQ in these schools has risen 10.38 IQ points over the 9-yr period … represents an IQ gain of 11.4 IQ points per decade. This figure is evidently broadly similar to the rise of 14.1 IQ points per decade for 1953-1960 found by Ushijima and confirms a very considerable rate of IQ gain in Japan in these early post World War II years.
“(3) Sano’s study (1974)….
“It will be seen that in all samples there were considerable increases in mean IQ from 1954 to 1972. The increases appear to be a little greater among the city children than among those from the prefrecture. When the results for 10- and 11-yr olds are combined, city children gained 18.04 IQ points and prefrecture children 15.07 IQ points….”
that the city children had greater gains than the rural kids is not surprising if inbreeding is the factor making the difference here. even up to the 1980s, consanguinity rates have been quite a bit lower in urban areas in japan as compared to rural areas (see table in this post for example). (more on this soon in that upcoming post on japan.)
“The average of the two gains is 16.56 IQ points, representing a gain of approx. 9.15 IQ points per decade for the entire sample.
“Sano also considered the question of whether the IQ gains in Japan have been increasing at a constant rate. For this he used Kaneko’s 1964 data which were available for two of the schools. He calculated that the increase in mean IQ was 10.47 points for 1954-1963 and 3.42 for 1963-1972…. ([I]t will also be noted that the gain of 10.47 IQ points for 1954-1963 is closely similar to Kaneko’s figure of 10.38 for the same period.)
“It is apparent therefore that there was a considerable deceleration in the rate of increase in intelligence over the period 1954-1972….”
again, see my crayola chart. the decline in cousin marriage rates is much sharper in the decades preceding 1954 than during those preceding 1963. in other words, there was a deceleration in the reduction of cousin marriages over time, so perhaps we shouldn’t be surprised at a deceleration in the increases in iq over time.
“(4) Wechsler studies….
“Thus after making these adjustments we have Japanese mean IQs of 101.9 for 1951 and 107.4 for 1975. Hence over this 24-yr period (1951-1975) the Japanese mean IQ increased by 5.5 IQ points relative to the American IQ…. This represents a Japanese IQ gain of 5.75 IQ points per decade. This rate of increase for the period of 1951-1975 is somewhat less than Sano’s result of a 9.15 IQ point per decade increase over approximately the same period (1954-1972)….
“A second Japanese study using the WISC and WISC-R is also available (Anon, 1981)…. This gives a rise of 20.03 IQ points over the 24-yr period, making 8.34 IQ points per decade….”
lynn and hampson conclude:
“The conclusion of the Japanese studies is as follows. Two studies of the early post World War II period show substantial IQ gains of 9.9 and 11.4 IQ points per decade. Three studies of a longer period from approx. 1950-1975 show lower IQ gains of 9.1, 8.3 and 5.7 IQ points per decade, giving an average gain of 7.7 IQ points per decade. Since the early part of this period was characterized by a greater rate of gain, it appears that since around 1960 the IQ gains in Japan have decelerated to approx. 5 IQ points per decade….
“It is not particularly surprising that the Japanese gains should have been the greatest of the three countries. Japan was a relatively underdeveloped country in the 1930s with a per capita income about one eighth of that of the U.S.A.”
but the country was probably relatively underdeveloped because the japanese had a lower average iq at the time, and it’s since increased phenomenally … due to the sharp and rapid decline in inbreeding in japan? and so has their economic success obviously.
i’d be curious to know two things:
– why did the japanese quit marrying their cousins? was there some official policy out there discouraging it — as part of their general modernization, “be more like the westerners” move? did the people just naturally adopt it as part of that westernization package? was it connected to christianity in japan? (japanese catholics historically had lower first cousin marriage rates than other religious groups — i’ll get to this in my upcoming post on japanese cousin marriage — the one that i was working on yesterday!). was it related to industrialization and urbanization? all of the above? none of the above?
– has something similar happened in china and/or korea (or elsewhere for that matter — i mean in modern times)? the cousin marriage rates for china are reportedly very low — today — but i’m pretty sure they were much higher in the past — even the recent past (see mating patterns in asia series below ↓ in the left-hand column), but i don’t have as good data for china as what i’ve presented here for japan. but inquiring minds want to know!
see also: Reversal of Fortune
previously: historic mating patterns in japan
(note: comments do not require an email. cool japanese people!)
“Lynn refutes my evidence for a low Ireland IQ during the 1970s by referring me to the more extensive data in his latest book, saying it debunks my claim. However, when I examined the Ireland IQs in that book (p. 402), I discovered that he had inexplicably failed to include the massive 1972 study of 3,466 students which established an Irish IQ of 87 and which had appeared in all of his previous books. When I asked him why he had excluded the largest early Irish IQ study, he said he had no answer, and that perhaps ‘this omission was a mistake.'”
well, that is weird, i agree.
i still wanna know, though — has anybody ever looked at the actual data from this 1972 study? i mean, evaluated the research — the testing and how it was conducted and so on. i’m not saying that the finding (average iq of the irish in 1972 was 87) is wrong. i just want to know if anybody’s — you know — double-checked it. did richard lynn actually check it personally?
as far as i can figure out, the only people who have seen the original data are the authors of the master’s thesis for which the data was collected — one enda byrt and one peter edward gill — plus jonn raven (note: not john c. raven). the data has not been published anywhere afaics — only in the master’s thesis itself, a copy of which can be found in the reference section in the library of a university in ireland. i discovered the reference for it in this paper by john raven: The Raven’s Progressive Matrices: Change and Stability over Culture and Time [opens pdf]. (<< you might want to read that paper, ron, if you haven’t already.) i can’t find any academic traces of enda byrt, but gill seems to be teaching at a university in sweden.
here’s what john raven had to say about the irish data in his paper:
pgs. 5-6: “In 1972, Byrt and Gill (1973), working with the author [i.e. john raven], collected data from a nationally representative sample of 3,464 primary school children ages 5 to 11 in the Republic of Ireland. The urban norms seemed to corresponded [sic] to the 1938 Ipswich norms, although the figures for the rural areas lagged behind.”
pg. 9: “[N]orms for rural and isolated communities are typically lower than others. The previously mentioned norms for the Republic of Ireland and Newfoundland can, in this context, be seen to confirm this.”
pg. 32: “Thorndike suggests that television may have had an effect [on rising scores, a la the flynn effect]. However television was widely available in Ireland when what can now be seen to be low Irish norms were collected.”
so, according to raven, the data was nationally representative and so should be ok. maybe it is. i would feel a little better about it, though, if it had been publically published somewhere so that others — people who had not been involved in the data collection — could’ve had a look at it.
who cares? well, an argument is only as good as the data on which it’s based, right? (that statement is gonna bite me in the *ss one day, i just know it! (~_^) ) was the average iq of the irish in 1972 really 87? i’m leaning towards maybe/probably, but i’m not certain about it because i don’t feel secure about the data.
another argument entirely is whether or not a score of 87 in 1972 tells us anything about the average iq of the irish in 1840. or 1890. don’t think ron can extrapolate backwards from that 1972 score. i mean, if the current scores for the irish are correct, and say we didn’t have any iq scores for the irish from the 1970s, we never would’ve guessed the score back then (in the ’70s) had been so low. (if, indeed, it was so low.)
i still think that my — and anatoly’s — suggestion that there was a ca. 130 year brain drain in ireland that resulted in that low score, which just happens to coincide with the lowest point in their population stats btw, is pretty good. i suppose ron will continue to ignore that possibility. that’s certainly his prerogative.
(note: comments do not require an email. hi there!)