Weird Internet Ideas: Global IQ again

More stuff about global IQ from various far right sources which I shan’t bother linking to. This a sort of whine about how the OECD’s PISA study of 15 year olds in several educational areas gets positive coverage but noted IQ nut Richard Lynn’s claims about global IQ don’t.

The reasons are straight forward but ignored by Vox Day et al who like to make use of global IQ claims to push their racist theories:

  1. These supposed global IQ studies are based on pulling together multiple data sets largely not designed for cross-national study.
  2. IQ tests themselves in general are not designed for this kind of study.
  3. IQ and IQ scores are based on core assumptions about how the trait is distributed across the population (e.g. an IQ score is directly related to the standard deviation of the target population).
  4. The samples for the data sets were chosen for the purpose of whatever the original study was and not for cross national comparisons.

Compare that with PISA.

  1. The tests are designed for the express purpose of cross national study. That involves extensive trialing of the test questions as well a heavily moderated translation process to try and ensure the people taking the tests are doing equivalent questions.
  2. The sample size per country are chosen to be at LEAST big enough to allow for cross national comparisons and are often larger per country to allow for internal comparisons.
  3. The scores are based on a completely different statistical model that makes fewer assumptions about the distribution.

Now having said all that there are still plenty of reasons to be skeptical about PISA rankings and comparisons. However, global IQ claims are necessarily WEAKER methodologically and statistically than PISA scores. Of the two, the PISA scores are far more sound and without some substantial evidence to the contrary, the PISA data trumps Lynn’s number fudging.

Fine to disbelieve both but if you have to pick PISA is far, far, far more sound.


9 thoughts on “Weird Internet Ideas: Global IQ again

  1. I’m not a fan of PISA at all and wrong conclusions drawn from the first PISA test have done a whole lot of damage to the German school system and a whole generation of students who had the misfortune of going through school during those “reforms”. But PISA is still a lot better than just comparing IQ test scores.

    Though PISA results can (and have) also be used to draw racist conclusions. In Germany, for example, the states that tend to have the best PISA results are inevitably the states with the most hierarchical schoolsystem and the most homogeneous population, whereas those states which have the worst results are the three city state Bremen, Hamburg and Berlin, which have the most diverse population. The conclusion drawn (beyond “City states are bad, why are we paying money for them?”) is often that diverse student populations mean lower test scores, hence diversity is bad. Now cities in general have lower PISA test scores than rural and suburban areas, only that in the surface states, there are a whole lot of rural and suburban areas clustering around their cities, where the scores are comparable to the three city states. Coincidentally, the surface state with the lowest test score is one with a lot of big and impoverished cities.

    Also, PISA test scores are a lot more dependent on family income than on race or ethnicity. Schools in low income neighbourhoods tend to score lower (though not necessarily – one of Bremen’s best schools is in a low income neighbourhood). Immigrants often tend to live in low income neighbourhoods, hence the lower scoring schools tend to have a lot of immigrant kids. But schools in predominantly white and German low income neighbourhoods get similar test scores. Finally, at least with the earlier PISA tests, the states that scored well only administered the test to academic track students, while the states with lower test scores administered the test to all students or had a lot of schools which have ditched the three-track system altogether. And if the only administer the test to more academically inclined students, you’ll get a higher score.

    Like

    1. I wasn’t aware of those patterns in Germany (I’ve only ever looked at PISA data for UK and Australia in any depth).

      Yes, it is a flawed tool and one into which much dangerous nonsense gets projected.

      Like

      1. Well, Australia and the UK do not have city states (and the three German ones exist for historical reasons and are also vehemently opposed to merging with the surrounding states), so the pattern would be less notable there. Though I suspect if you compared test scores for London, Birmingham or Manchester, particularly low income neighbourhoods in those cities, to test scores for wealthy semi-rural, semi-suburban areas like Kent or Oxfordshire, you’d get a similar pattern. And while the UK and Australia (as far as I know) do not separate students according to supposed academic aptitude early in their school careers, I also suspect that test scores would vary quite notably between a grammar school and a comprehensive.

        Another issue with the German city states is that they are generally poorer than the surface states because the German tax system disadvantages them. Because in Germany, you pay income tax where you live, not where you work. For surface states, this isn’t much of a problem, because someone who works in Munich, but lives in a suburb just outside the city limits still pays taxes in Bavaria. However, for the city states, this is a problem, because a lot of people who work there live just outside the city in a different state, so their taxes are basically lost. For example, I live five kilometres outside the Bremen city limits in the state of Lower Saxony and pay my taxes there. Now when I still was in traditional paid employment (I’m a freelancer now), I always worked in Lower Saxony, so Bremen did not lose my taxes. But most of my neighbours work in Bremen. There is some compensation, because the city states provide a lot of services such as universities, hospitals, airports, train stations, museums, sports stadiums, schools, which are used by the population of the surrounding states, but it does not make up for the lost income tax revenue. Coincidentally, the wealthiest of the three German city states is Hamburg, which had a lot of the surrounding rural area and suburbs incorporated in 1937. Bremen’s then-mayor Willhelm Kaisen got a similar offer in the late 1940s and refused.

        And since the city states are poorer, they have less money to spend on everything, including schools.

        Like

      2. Yes, our rural is probably highly populated by Australian standards.

        I remember seeing in an ancient kids’ TV show, probably Skippy the Kangaroo or something like that, how kids in the Australian outback were taught via radio. If that was ever real, they’ve probably upgraded to online lessons by now, but that sort of thing still can’t be good for test scores.

        Like

  2. I’m not sure if this qualifies as an Internet idea because the “research” of Lynn et al. dates back to the 90’s and and the book “IQ and the Wealth of Nations was published between 2001-4 or so.
    Also there simply don’t exist data for the majority of countries. Lynn estimates national IQs by combining those of neighbouring countries.
    Overall things are very fishy here.

    Like

Comments are closed.