World Map of PISA 2015 Results

Comments

1. Anon says

   December 6, 2016 at 11:16 pm

   If you think PISA scores correlate so well with IQ that they might as well be equivalent to it, then why the hell do immigrant groups in Europe far far exceed those in their homeland despite not being selected for intelligence (in fact, you might argue many are selected for stupidity)?
   Don’t you think there are huge confounding variables that mediate this relationship, as opposed to the PISA test being some sort of IQ test?
2. Anonymous says

   December 6, 2016 at 11:18 pm

   Why omit science scores? As I understand it, the majority of PISA 2015 questions were on science.

   AK: They weren’t omitted.
3. Anatoly Karlin says

   December 6, 2016 at 11:30 pm

   … why the hell do immigrant groups in Europe far far exceed those in their homeland

   Because Europe has far better schooling and nutritional standards. You can call it an “accelerated Flynn effect.”

   Unfortunately, however, convergence is rarely total.

   Don’t you think there are huge confounding variables that mediate this relationship, as opposed to the PISA test being some sort of IQ test?

   The g-Factor of International Cognitive Ability Comparisons: The Homogeneity of Results in PISA, TIMSS, PIRLS and IQ-Tests Across Nations (Rindermann 2007)

   International cognitive ability and achievement comparisons stem from different research traditions. But analyses at the interindividual data level show that they share a common positive manifold. Correlations of national ability means are even higher to very high (within student assessment studies, r = .60–.98; between different student assessment studies [PISA-sum with TIMSS-sum] r = .82–.83; student assessment sum with intelligence tests, r = .85–.86). Results of factor analyses indicate a strong g-factor of differences between nations (variance explained by the first unrotated factor: 94–95%). Causes of the high correlations are seen in the similarities of tests within studies, in the similarities of the cognitive demands for tasks from different tests, and in the common developmental factors at the individual and national levels including known environmental and unknown genetic influences.
4. asdasd2 says

   December 6, 2016 at 11:52 pm

   Maybe time to concede that Slovenia doesn’t have an average IQ of ~95-96 as Richard Lynn & Co would suggest
5. Anonymous says

   December 7, 2016 at 10:28 am

   Чем аукнется Казахстану данная проблема?
6. melanf says

   December 7, 2016 at 11:01 am

   It takes into account the difference in relation to the tests in different countries? As far as I know (from stories of expats) In Japan and South Korea, the tests are very important – in many cases people have to pass a test for IQ when trying to get a job. As a result, people specifically trained to pass these tests.

   In Russia, the IQ tests are perceived as a funny curiosity – these tests are considered to be entertainment
7. Triumph104 says

   December 7, 2016 at 2:04 pm

   Every cultural group has a maximum level of achievement that can be obtained without giving up the values and behavior of their culture. Immigrant groups in Europe exceed their homeland because they moved to a more efficient education system. This is observed with different cultural groups in the US.

   States where Native Americans are primarily schooled on reservations show that those Native American students perform worst on the ACT than blacks. However, in states where Native Americans primarily attend non-reservation schools, they perform better than blacks on the ACT. Puerto Ricans on the island of Puerto Rico test worst than blacks Americans on PISA, but Puerto Ricans in the states test better than blacks on the SAT.

   US Hispanics outperformed all Spanish-speaking countries on PISA except for Spain.
8. AER says

   December 8, 2016 at 8:59 am

   Thanks for the map.

   Looking at the list, I laugh to myself seeing UAE and Romania being neck & neck. Iceland’s IQ is also laughable, i.e. unreliable. Sure, if you take a very broad view, you will find that there is some coherency to the general pattern but that is also true for TIMMS. It isn’t nearly as accurate to serve as a ranking seen in the table. If anyone seriously believes that UAE is on par demographically with Romania then they’ve disqualified themselves completely.

   Yet the PISA worshippers in the HBD blogosphere will continue insisting of its high value and sing its virtues. It sets out exactly what it aims to do: measure the education system. That performance is constrained by IQ but still allows for heavy fluctation. Iceland having 97 in IQ is not consistent with their development or past performance, or the plethora of IQ tests given to Icelandish students. Romania being at 90 is a joke.

   Kill the holy cow once and for all and be done with it.
9. Anonymous says

   December 8, 2016 at 11:49 am

   Re: the high Estonian ranking.

   Are ethnic Russians in Estonia increasing the score, or bring it down? From the Russian score, I would guess the latter.

   Which makes me wonder how high can Estonia go without being dragged down by Russia? Would they even beat Singapore?
10. Irving says

    December 9, 2016 at 11:48 am

    how are you converting pisa scores into IQ scores?
11. Anatoly Karlin says

    December 9, 2016 at 3:28 pm

    Transforming the PISA scale (m=500, S.D.=100) in the OECD countries, to the standard IQ scale (m=100, S.D.=15).
12. M says

    December 10, 2016 at 6:57 am

    Online at http://dx.doi.org/10.1787/888933433226 there are tables Table 1.74a, Table 1.74b and Table 1.74c which break down raw native-immigrant (second and first generation) score differences within participating nations.

    Using these scores, I ran a PCA to pull out a general factor in the quality of natives performance:
    http://i.imgur.com/95IUqMl.png

    (Text version: https://www.scribd.com/document/333806359/PISA2015-Natives. Save as a .csv file to open as a spreadsheet)

    Graphically this is:
    http://i.imgur.com/stShIKy.png

    Using the first PC1 dimension should gives a slightly better measure than using the average of scores, because it isn’t thrown off by outliers in any particular one of the 3 PISA subtests (Math, Reading or Science), like Viet Nam’s outlying science score.

    Convert these into “IQ” on the assumption that the UK is the “Greenwich Mean” of 100 and Japan is 104.3 and scores are as follows, for natives:
    http://i.imgur.com/O8CKpZ4.png

    (Text version: https://www.scribd.com/document/333806403/PISA2015-nativesIQ)
13. notanon says

    December 10, 2016 at 11:24 am

    Pure guess but Iceland’s score looks to me like a perfect way to test if low-level inbreeding depression can happen in an enclosed population even when they’re actively trying to be exogamous (Iceland’s ancestry book).

    idea being

    height and IQ take energy so where it’s not needed a seemingly “deleterious” allele may not be deleterious (at the time) and so can become homozygous among a small population i.e. selecting for minimum necessary height/IQ (at least in the past).

    If so making that homozygous allele go away would(?) be a good job for CRISPR.
14. M says

    December 10, 2016 at 12:21 pm

    My thinking about Iceland was a bit more prosaic, that in such a small culture, where you’ve got what, 300,000 people total and a fairly small percentage are schoolkids and teachers, then you’ve got more opportunity for short term random cultural fluctuations to swing the performance. While in a big population that can’t happen as much.

    Like the opposite of how they excelled in the Euro 2016 football tournament, where they’d applied a lot of specialised knowledge to hone their performance. Small fluctuations and intensive interventions can shifting the mean in a small population, while they just average out in large nations.

    I don’t really know specifically about the idea of alleles with big tradeoffs being selected in smaller populations with fewer competing mutations. I do see what you’re talking about with inbreeding depression, but it seems like they should be all f*d up in lots of different ways if that’s true, and it seems like not really, like they have the highest male life expectancies.

    The Iceland natives converted “IQ” score by my method is kind of low, but then if you look at Slovenia, Czech Republic, Slovakia which are all these really close by European nations, they go 101.8 to 97.8 to 93.3, so quite a big swing is possibly for closely related countries, and there’s not necessarily a genetic explanation.
15. notanon says

    December 10, 2016 at 12:45 pm

    more opportunity for short term random cultural fluctuations to swing the performance

    like they have the highest male life expectancies.

    both of those plausibly point in the other direction

    and there’s not necessarily a genetic explanation

    Agreed but it would be interesting to test if there was.
16. Andrei kuznetsov says

    December 13, 2016 at 2:19 pm

    How come countries like Russia,South Korea gets a lower score in your chart? The average Russian native is smarter than the average immigrant while in south Korea case they should have the same score like Japan and China because they barely have any immigrants to affect the score. In booth your chart and karelin one Japan and China has exactly the same score so its strange for me that Korea is lower.
17. M says

    December 13, 2016 at 3:59 pm

    @ Andrei kuznetsov, the scores should generally go down for countries with fewer migrants compared to Anatoly’s because these are only considering natives, all the scores have been normed relative to the British->Japan natives gap, and Britain gains on most other countries with few migrants when only its natives are considered (because across most nations immigrants perform less well and are of a significant number, with the biggest gaps in Germanic North-Central Europe, who also have fairly large migrant populations).

    So I think it’s mainly due to the difference in the scores that’s anchor the mean of 100.

    I was also considering there might be a slight difference where I am calculating the first PC of the scores as a sort of g factor of the PISA results, compared to Anatoly using the average. As the different PISA subtests have a floor and ceiling and average at different levels, the Principal Components Analysis is just on the correlation of the PISA variables.

    But that said, the two measures of Average for Natives and PC1 look essentially perfectly correlated – http://i.imgur.com/B7CsN8X.png, so I think any differences from method there is very minimal as an effect and using the average essentially generates an identical outcome.

    Just for comparison, using Anatoly’s method of the average, for native scores, and then converting to IQ based on the same assumption of Japan 104.3 and UK 100:

    http://i.imgur.com/CHuluNu.png

    (Here’s the raw numbers in CSV format – https://www.scribd.com/document/334120436/Untitled-1.

    If you reckon there might still some other differences, I’d recommend to check out the spreadsheet I linked to as it’s easy to pull out the natives PISA scores from that and you can compare to see if there’s anything gone wrong with the values for natives I’ve copied and used, and if you find anything then please comment back in the thread.
18. Winston says

    December 19, 2016 at 12:33 am

    Anatoly-please note West is aging rapidly. Germany in fact has outpaced Japan! Germany is the engine of Europe. France has big problem with rural areas left behind, which is fueling swing to hopeless Le Pen. France needs to create something like Germany’s Fraunhofer.

    Canada and Australia stuck in trap of commodity/resource dependent economy. PISA results meaningless for them until govt wakes up from coma-before too late as already aging. In both countries immigrants leave because limited job market for skilled immigrants. Probably just use passports to settle /work elsewhere.

    China- very impressive! You should check out Joseph Needham’s books on China’s innovations.

    China also only country where farmers make working submarines, robots, etc from scrap metal!

    Poor people improve over time, unless education inequitable.