Saturday, September 15, 2012

East Germans are getting a lot smarter ... Really?

Unemployment rates in the West and East of Germany (source)
Did East German conscripts become smarter because East Germans were becoming smarter? Or were smarter individuals increasingly seeing the army as an alternative to unemployment?

In a recent study of conscripts entering the German army, Roivainen (2012) has found that the mean IQ of East German conscripts jumped eight percentage points between 1990 and 2006. That’s a huge increase! And a fast one! Far too fast to be due to genetic evolution. Sixteen years isn’t even one generation. The only possible cause seems to be a better environment, such as through improvements to the standard of living, education, and nutrition.

Specifically, the study found:

[…] strong IQ gains of 0.5 IQ point per annum for East German conscripts in the 1990s, after the reunification of the country. An analysis of IQ, GDP, and educational gains in 16 German federal states between 1990 and 1998 shows that IQ gains had a .89 correlation with GDP gains and a .78 correlation with educational gains. The short time frame excludes significant effects of biological or genetic factors on IQ gains. These observations suggest a causal direction from GDP and education to IQ. (Roivainen, 2012)

Until last year, Germany had conscription for all male citizens, so this finding couldn’t result from biased sampling of the East German population. The samples were the population.

Wait a minute …

Or were they? Germans could dodge the draft on several grounds: conscientious objection; service in international aid organizations; illness; drug abuse; delinquency; criminality, etc. In total, over half of all German men evaded military service:

The post-Cold War downsizing of the Bundeswehr led to a considerable decrease in demand for young conscripts. Of all men reaching draftable age, less than one half actually served. In 2005 about 15% served in the military, while 31% performed civilian service or some other form of alternative service. More than 36% were screened out for medical reasons. This percentage was lower in the past (15% in 2003), but to avoid drafting more men than needed, medical standards had been raised. The remainder includes those who were exempt for various reasons, but is mostly made up of men who were not drafted because the military had already reached its recruitment goals. This had led to discussions about "draft equality" ("Wehrgerechtigkeit"), which is the principle that the draft should have applied equally and non-discriminatorily to all men. (Wikipedia, Conscription in Germany)

So the possibility of sampling bias is back on the table. Is there reason to believe that this sampling of the East German population changed between 1990 and 2006?

There were in fact two changes:


Increased sampling of the highly educated

The end of the Cold War brought a steady process of deindustrialization to the former East Germany. Full employment gave way to chronic double digit unemployment. As a result, highly educated Ossies increasingly saw military service as a possible career.

This change has been commented upon in a recent New York Times article:

In a recent study, two experts said that while only 16 percent of the German population of 82 million lives in the former East Germany, easterners make up 30 percent of military personnel. At the same time, said Michael Wolffsohn, a professor at the German Institute for International Security Affairs, and Maximilian Beenisch, a social scientist, easterners with higher educational qualifications were drawn to the military because of a lack of alternative opportunities in eastern Germany, where unemployment is higher than in the west. (New York Times, 2011)

The same point is made by Roivainen himself:

In a Bundeswehr survey from 1993, attitudes toward the army were more positive among East German youth and those planning a career in practical vocations than among West German youth and those wishing to study further. (Roivainen, 2012)


Decreased sampling of troubled youth

With reunification, a more liberal social environment in East Germany brought about an apparent rise in IQ scores. Fewer low IQ individuals were present in the sort of institutional settings where IQ tests were routinely administered.

Many of these individuals were juvenile delinquents. In Germany, delinquency leads to exclusion from military service: “Delinquents sentenced to more than a year or charged with a felony against peace, democracy, the state or state security were not drafted for military service” (Wikipedia, Conscription in Germany).

As I noted in an earlier post, truancy was less common in the old East Germany than in the old West Germany because punishment was harsher. This factor had the effect of depressing IQ scores in East Germany, since potentially truant individuals were more likely to be tested for IQ in East German classrooms than in West German ones.

The same reasoning applies to juvenile delinquency. With reunification, East German youth entered a less controlled social environment. Relatively low-skilled factory work also disappeared. The result was a sharp rise in youth violence and delinquency:

Most of the increase in German youth violence has been encountered in what was communist East Germany before the reunification. Youth violence in the east is 70 percent higher than in the west, a factor linked to the exposure of eastern youth to greater poverty and unemployment than their West German peers. (Siegel & Welsh, 2009, p. 577)

From 1990 delinquency began increasing in the region of East Germany and has reached critical proportions. There are three general types of delinquency: traditional delinquency, new kinds of delinquency from outside the region, and delinquency associated with the unification of Germany. Traditional delinquency has increased in nearly all significant categories of crime: theft, fraud, robbery, murder (homicides generally) and rape. The new category of crimes consists of drug trafficking and abuse, car theft, and environmental crimes. Violence against foreigners in the former East Germany is also significant. Crimes committed in association with unification include profiteering, fraud, and financial manipulations (Buchholz, 1993)


Conclusion

There is little doubt that the mean IQ of East German conscripts rose dramatically between 1990 and 2006. This increase, however, most likely reflected the changing composition of the pool of candidates for military service. Because of rising unemployment, high IQ individuals were much more likely to consider military service in 2006 than they were in 1900. Meanwhile, because of a more liberal social environment, low IQ individuals were much more prone to delinquency and being excluded from conscription.

But why do these gains in IQ correlate so highly, on a state-by-state basis, with gains in GDP and also, to a lesser extent, with gains in education? These high correlations are actually not so surprising given the tight relationship between the positive changes to East German society (rise in GDP due to expansion of the market economy) and the more negative ones (rise in unemployment due to deindustrialization and rise in juvenile delinquency due to social liberalization). These changes were two sides of the same coin.

By way of example, there existed in the 19th century a significant correlation between the salaries of Protestant ministers in Boston and the price of rum in Havana. Is this proof that something underhanded was going on? Hardly. The two trends were part of a larger worldwide cycle of economic upturns and downturns.

As Roivainen himself points out, the correlations with GDP and education are simply an artefact of the east-west divide within Germany:

However, within western Germany, the correlation between IQ gain and GDP gain was .03 and between IQ gain and educational gain −0.12. Thus, the strong correlations based on calculations involving all states mainly reflect the east–west divide and its gradual narrowing (Roivainen, 2012)


References

Buchholz, E. (1993). Social changes, crime and police in the former GDR (German Democratic Republic), from Social Change, Crime and Police: International Conference, June 1-4, 1992, Budapest, Hungary, pp. 115-121, (eds., Jozsef Vigh and Geza Katon). https://www.ncjrs.gov/App/publications/Abstract.aspx?id=144804

Cowell, A. (2011). The draft ends in Germany, but questions of identity endure, New York Times, June 30, http://www.nytimes.com/2011/07/01/world/europe/01germany.html?pagewanted=all

Leipzig, a demographic analysis. Germany, Saxony and Leipzig http://demographicleipzig.wordpress.com/2012/06/13/germanysaxonyleipzig/

Roivainen, E. (2012). Economic, educational, and IQ gains in eastern Germany 1990–2006, Intelligence, 40, 571–575. https://lesacreduprintemps19.files.wordpress.com/2012/09/roivainen.pdf

Siegel, L.J. & B. Welsh. (2009). Juvenile Delinquency: Theory, Practice, and Law, Cengage Learning.

Wikipedia, Conscription in Germany, http://en.wikipedia.org/wiki/Conscription_in_Germany

6 comments:

  1. It occurs to me that former East Germany is also the western remnant of former Prussia, so could it be that there is also a stronger cultural affinity there to military service than there is elsewhere in Germany?

    ReplyDelete
  2. 'His team's study melds records from Scottish army units with results of national tests performed by all 11-year-olds in 1932. [...] "No other country has ever done such a whole-population test of the mental ability of its population," Deary says."'
    Here.

    I have read that in the UK during WW1 the regions of low unemployment had very low rates of volunteering for the army. In WW1 Scotland suffered a higher % of pop killed then any country apart from Serbia.

    (BTW, new book Sleepwalkers: How Europe Went to War in 1914 by Christopher Clarke, Professor of Modern History at the University of Cambridge agrees with your idea that important Serbians were trying to start a wider war)

    ReplyDelete
  3. Pete,

    Here were the 1990/1991 IEA reading results; East Germany was on parity with West: http://nces.ed.gov/pubs/96258-2.pdf

    As I noted in a previous post, concerning the "Irish Flynn effect," when it comes to secular differences AQ has to be compared to AQ and IQ to IQ. This is because the g-factor of IQ is not the same as the g-factor of IQ.

    As for the study mentioned, check out the original paper cited: "EUF-Testleistungen wehrpflichtiger junger Männer im wiedervereinigten Deutschland"

    Google it. And check out figure 6. All of the narrowing occurred between 1992 and 1995.

    ReplyDelete
  4. Ron Unz is now explicitly anti biologic in his thinking. It isn't possible to accept his arguments and still believe that human brainpower is a product of natural selection.

    ReplyDelete
  5. Alan,

    Perhaps. The young Ossies I knew in Russia didn't seem very militaristic.

    Sean,

    Smart people are probably more likely to respond to patriotic rhetoric. In my home town, we had "patriot clubs" that encouraged young men to enrol for the First World War. There was a blind willingness among many Canadians, especially English Canadians, to sign up for every imperial conflict, like the Boer War and the Crimean War. When I look at the memorials that have been erected in their memory, I can't help but shake my head. They fought nobly, yes, but what was gained?

    Sometimes, I wish we had been less idealistic.

    I'll have to check out that book. Small sub-state actors have much to gain and little to lose by triggering a major world conflict. This was true for Serb nationalists in 1914, and it's true for Coptic Christians today.

    Chuck,

    Then, do you agree that the rise in IQ was due to the changing composition of conscripts from the former East Germany? The steep rise from 1992 to 1995 (Figure 6) is not easily reconciled with a real increase.

    ReplyDelete
  6. this is ancient, so just for posterity:

    the paper looks at four chunks of data - conscription IQs for 1992 and 1998 and PISA derived IQs for 2000 and 2006 (plus GDP etc for those times). therefore any hypotheses about east-specific conscription sample biases between '1990 and 2006' are off. if anything they would have to be about biases introduced between 1992 and 1998 (or, as others have pointed out, 1995: http://www.mgfa-potsdam.de/html/einsatzunterstuetzung/downloads/ap122.pdf?PHPSESSID=92bb8).

    more importantly, i can't see how any of the hypothesised confounding factors would apply to the PISA data, which show the exact same trend as the conscription data (drastic gains in the east) *and* replicate the correlation between gains in IQ and gains in GDP.

    greetings from a german

    ReplyDelete