Prolific IQ researcher Richard Lynn together with two Russian collaborators have recently published arguing that multiple aspects of socio-economic development – infant mortality, fertility, stature, and literacy-as-a-proxy for intelligence were significantly intercorrelated in late Tsarist Russia.
Here is the link to the paper – Regional differences in intelligence, infant mortality, stature and fertility in European Russia in the late nineteenth century
And here is a summary by James Thompson – 50 Russian oblasts.
To the right: Here’s your map, JayMan. You’re welcome.
The main potential sticking point:
There are no data for regional intelligence in the nineteenth century and we have therefore adopted rates of literacy as a proxy for intelligence. This is justified on the grounds that a high correlation between literacy rates and intelligence have been reported in a number of studies. For example, a correlation of .861 between literacy rates for Italian regions in 1880 and early twenty-first century IQs has been reported by Lynn (2010); a correlation of .83 between literacy rates for Spanish regions in the early twenty-first century has been reported by Lynn (2010); (Lynn, 2012); and a correlation of 0.56 between literacy rates and IQs for the states and union territories of India in 2011 has been reported by Lynn and Yadav (2015). There is additional support for using literacy in the nineteenth century as a proxy for intelligence in the results of a study by Grigoriev, Lapteva and Ushakov (Григорьев, Лаптева, Ушаков, 2015) showing a correlation of .58 between literacy rates of the peasant populations of the districts (uezds) of the Moscow province in 1883 and the results of the Unified State Exam and State Certification on Russian Language in the districts of the contemporary Moscow oblast.
The methodology at first struck me as being rather problematic.
I’ve read a bit about Russian state literacy programs in the 19th century (National Literacy Campaigns and Movements) and one of their main features is that they tended to spread out from the central European Russian provinces due to cost effectiveness reasons, hence the low literacy rates of e.g. Siberia in Lynn’s data set. However, there is no particular evidence that Siberian Russians are any duller than average Russians. To the contrary, some 3% of Siberian schoolchildren become “Olympians” – high performers who qualify for highly subsidized higher education. This proportion is lower than the 15% of the central region (which hosts Moscow, Russia’s main cognitive cluster with a 107 average IQ), and the 14% of the north-west region (which hosts Russia’s second cognitive cluster with a 103 average IQ Saint-Petersburg, plus the Russians there are probably slightly brighter in general on account of Finno-Ugric admixture), but is considerably higher than in any other Russian Federal District: The Urals and Volga (both about 2%), and the Far Eastern, Southern, and Caucasus (all considerably below 1%).
In other words, would such a historical literacy – modern intelligence correlation apply to Russia as it does to Italy, Spain, and to a lesser extent, India?
Fortunately, we don’t have to postulate, since we do actually have PISA data for many Russian provinces that I revealed back in 2012.
This allows us to test if Lynn’s assumptions apply.
There are difficulties, to be sure. Not all Russian provinces were tested in PISA, and there is, needless to say, no data for any of the Ukrainian and Polish oblasts, or for Belarus. As such, only 20 Russian provinces could be tested in this manner (26 if you also include now independent countries excluding Russia itself).
In some cases, names have changed, typically to honor some faceless Soviet bureaucrat; in more problematic cases, borders have changed significantly (e.g. the five provinces of Estonia, Livonia, Courland, Kovno, and Vilna have become the three countries of Estonia, Latvia, and Lithuania – I have tried to average the literacy figures between them in a common sense but back of the envelope way). The Moscow Governorate has been split into the City of Moscow (with its 107 average IQ) and Moscow oblast (with a modest 96 average IQ). Which of those should be attached to Moscow’s 1897 literacy rate of 40%? (As it happens, I went with just the City of Moscow instead of figuring out how to weigh the populations and adjust and so forth. I’m not trying to writea formal paper, after all).
There is an exponential correlation of R=0.75 between average PISA derived IQs of Russian regions and of now independent countries, and their literacy rate according to the 1897 Census. Therefore, this bears out Lynn’s assumptions.
The two downwards outliers – more relatively intelligent than literate – are Moscow, Tatarstan, Tula, Samara, and Tambov. Moscow is easily explainable – the city itself in Tsarist times would have been more literate than the Moscow Governorate, while its average IQ was artificially boosted in Soviet times since it became not just the empire’s political but also its cognitive (artistic, scientific) capital. Getting a Moscow propiska required considerable intelligence.
The three very major upwards outliers – more relatively literate than intelligent – are the Finno-Ugric Baltic states: Finland, Estonia, and Latvia. This can’t have been a non-Orthodox/Muslim thing: Both Poland (On-The-Vistula Governorate) and Lithuania (Kovno and Vilna) lie neatly on the correlation curve. Nor was it something Finno-Ugric; Karelia (then Olonets) is not an exception either. It must have been something specific just to them and the most obvious explanation is Protestantism. There is a lot of literature on the independent literacy-raising effects of Protestantism and I see no reasons why Estonia, Latvia, and Finland should have been exceptions to that.
Another outlier, though this one is at the bottom of the IQ scale, is Moldova. To be fair I think Moldova’s PISA-derived IQ is artificially lowered by a third to half of an S.D. due to the massive brain drain it has experienced after the collapse of the Soviet Union (something like half the working age population are Gastarbeiters in the EU and Russia). We see similar drops in other countries so afflicted, such as (possibly) Puerto Rico, and (almost certainly) Ireland during most of the 20th century, when it repeatedly reported IQs in the ~90 range (and ironically one of the reasons Richard Lynn himself abandoned it to move to Northern Ireland, thus getting stuck in the most depressed region of the UK and missing out on the rise of the Celtic Tiger a few years later).
The correlation improves further to R=0.80 when we consider only those Tsarist-era provinces which are still part of the Russian Federation. This is accomplished (more than) entirely just by removing the Protestant Baltic nations (Finland, Estonia, and Latvia) and Moldova (whose current day average IQ is depressed due to massive brain drain as per above).
As usual Lynn does his north/south IQ gradient analysis, finding it to be a real thing but diminishing to nothing once the Baltic states of Estonia, Livonia, and Courland are accounted for.
Quoting from Thompson’s summary:
The Russian provinces differed significantly by geographical location. The positive correlations with latitude (r= .33, p<.05) and the negative correlation with longitude (r=−.43, p<.01) show that the rates of literacy were higher in the northand west than in the south and east. These trends were partly determined by the rates of literacy being highest in the north-western provinces of St. Petersburg and the three Baltic states of Estland, Livland and Kourland (correspondingapproximately but not precisely to contemporary Estonia and Latvia; Livland consisted of southern part of contemporary Estonia and eastern part of contemporary Latvia). Removing these four regions makes both correlations non-significant (.21 and −.23).
One additional issue worth bearing in mind: The influence of the Jews. Namely, their concentration in the Pale of Settlement, which correlates to modern day Poland, Belarus, and right-bank Ukraine (west of the Dnieper). There were more than 5.2 million Jews, and their literacy rates were very high (according to the 1926 Soviet Census, Jews over the age of 50 – i.e., who had been educated under the Empire – had a literacy rate of 63% versus 28% for ethnic Russians).
This must have “artificially” raised the literacy rates in this area – as pertains to those regions’ 21st century average IQs, anyway, since the vast majority of those Jews are no longer there due to the trifecta of the Holocaust, Jackson-Vanik, and Aliyah. The effect would probably be to reduce the “indigenous” literacy rates in Lithuania and Poland closer to those of European Russia, while pushing the already low literacy rates of strongly ethnic Malorossiyan and Belorussian provinces considerably lower still. Not a single province of modern Ukraine outside historical Novorossiya (with its strong Great Russian admixture) had a literacy rate above 20% in 1897, despite highly literate Jews helping them out with the statistics.
Unfortunately, there is a severe paucity of usable psychometric data from Ukraine – for instance, it is one of the very few European countries that doesn’t participate in PISA. So its average IQ has to be estimated through generally more indirect means. It does the converted equivalent of 9 IQ points worse than Russia on the TIMSS standardized test. Ukrainians spend less than half as much time as Russians reading, and those from the western parts at least spend a lot more time participating in torchlit processions and chanting “Putin Khuylo.” Some of those activities are considerably more g loaded than others. The low literacy rates in late Tsarist Malorossiya, coupled with the finding of a close correlation between those literacy rates and modern day average IQ across both Russian provinces and today’s independent post-Soviet states, constitutes further evidence of a modest average IQ in Ukraine. Higher than in Moldova to be sure, but probably closer to the level of the Balkans than to Poland.
|IQ||Literacy in 1897|
Literacy and Social Development in 1890s Russia (from Grigoriev et al. 2015)
Incidentally, I am not surprised to see Yaroslavl being the top non-Baltic/non-capital Russian region by literacy rate in 1897. It struck me as by far the cleanest and most civilized provincial Russian town on the Golden Ring when I visited it in 2002 (a time when Russia was still shaking off the hangover of the Soviet collapse). Curiously enough, it also hosted one of the most vigorous insurrections against the Bolshevik regime in central Russia. Although it was not one of the regions covered by PISA, I would not be surprised if Yaroslavl oblast was to get a 100-102 score on it should it be carried out there (and as would be implied by the correlation curve).