December 3, 2013

PISA Racial results for Americans on Math

With PISA results being released today, you are going to hear a lot about how stupid American 15-year-olds are, but smart analysts remember to always adjust for race.

From the federal National Center for Education Statistics, a breakdown of 2012 PISA math scores showing how Americans of different races compare to the world. PISA scores are much like SAT scores, with an intended mean of 500 for the OECD (i.e., rich) countries and a standard deviation of 100. On PISA, Math is America's worst subject (Reading is its best).

OECD average              494
Shanghai-China            613
Singapore                 573
Hong Kong-China           561
Chinese Taipei            560
Korea, Republic of        554
Asian-Americans 549
Macao-China               538
Japan                     536
Liechtenstein             535
Switzerland               531
Netherlands               523
Estonia                   521
Finland                   519
Canada                    518
Poland                    518
Belgium                   515
Massachusetts-All Races 514
Germany                   514
Vietnam                   511
White Americans 506
Connecticut-All Races 506
Austria                   506
Australia                 504
Ireland                   501
Slovenia                  501
Denmark                   500
New Zealand               500
Czech Republic            499
France                    495
OECD Average 494
United Kingdom            494
Iceland                   493
Multiracial Americans 492
Latvia                    491
Luxembourg                490
Norway                    489
Portugal                  487
Italy                     485
Spain                     484
Russian Federation        482
Slovak Republic           482
United States             481
Lithuania                 479
Sweden                    478
Hungary                   477
Croatia                   471
Florida-All Races 467
Israel                    466
Hispanic Americans 455
Greece                    453
Serbia, Republic of       449
Turkey                    448
Romania                   445
Cyprus                    440
Bulgaria                  439
United Arab Emirates      434
Kazakhstan                432
Thailand                  427
Chile                     423
African Americans 421
Malaysia                  421
Mexico                    413
Montenegro, Republic of   410
Uruguay                   409
Costa Rica                407
Albania                   394
Brazil                    391
Argentina                 388
Tunisia                   388
Jordan                    386
Colombia                  376
Qatar                     376
Indonesia                 375
Peru                      368

From the federal NCES, a breakdown of 2012 PISA math scores for American students by race:

Below
level 1
Level 1
Level 2
Level 3
Level 4
Level 5
Level 6
< 358
< 421
< 483
< 555
< 607
< 670
>= 670
Race/ethnicity
%
%
%
%
%
%
%
OECD average
8.0
15.0
22.5
23.7
18.2
9.3
3.3
U.S. average
8.0
17.9
26.3
23.3
15.8
6.6
2.2
White
3.6
11.5
25.0
27.8
20.3
8.8
3.0
Black
21.0
32.0
25.3
14.0
6.7
1.0
Hispanic
10.8
24.2
30.2
20.3
10.2
3.5
0.7
Asian
5.4
15.6
23.6
28.1
16.1
9.0
Multiracial
17.1
29.6
23.3
16.4
7.7
Cumulative: Bottom Up
OECD average
8.0
23.0
45.5
69.2
87.4
96.7
100.0
U.S. average
8.0
25.8
52.1
75.4
91.2
97.8
100.0
White
3.6
15.1
40.1
67.9
88.2
97.0
100.0
Black
21.0
53.0
78.3
92.3
99.0
100.0
100.0
Hispanic
10.8
35.0
65.2
85.5
95.7
99.2
99.9
Asian
2.2
7.6
23.2
46.8
74.9
91.0
100.0
Multiracial
3.0
23.0
52.6
75.9
92.3
97.1
100.0
Cumulative: Top Down
OECD average
100.0
92.0
77.0
54.5
30.8
12.6
3.3
U.S. average
100.0
92.0
74.2
47.9
24.6
8.8
2.2
White
100.0
96.4
84.9
59.9
32.1
11.8
3.0
Black
100.0
79.0
47.0
21.7
7.7
1.0
0.0
Hispanic
99.9
89.1
64.9
34.7
14.4
4.2
0.7
Asian
100.0
97.8
92.4
76.8
53.2
25.1
9.0
Multiracial
100.0
94.1
77.0
47.4
24.1
7.7
0.0

Tech note: I imputed Asian and Multiracial percentages below Level I and Multiracial percentage in Level 6, which were left out of report due to small samples sizes. 

For more postings devoted to analyzing PISA scores, click the Labels: PISA below. 

35 comments:

  1. So white USA is basically an OECD country, the rest of the USA is third-world. And Generalissimo Francisco Franco is still dead.

    ReplyDelete
  2. http://conservativetimes.org/?p=14829

    'My Thoughts on Jack Hunter’s Mea Culpa at Politico'

    ReplyDelete
  3. So basically DuBois's talented tenth is approximately equivalent to the talented third of Whites and talented half of Asians.

    Yet the fall of Detroit is a mystery to the mainstream press.

    ReplyDelete
  4. OT: Australian PM is taking a hardline stance on illegal immigration. Awesome.

    ReplyDelete
  5. Korea > Japan

    Korea > Japan

    Korea > Japan

    That's about the only takeaway from PISA.
    And no, it's not "cheating". And it's not leading to demographic collapse on the "peninsula".

    ReplyDelete
  6. So white USA is basically an OECD country, the rest of the USA is third-world. And Generalissimo Francisco Franco is still dead.

    Actually, the United States appears to be doing an admirable job with its third-world raw material. And the beating of teachers will continue until morale improves.

    ReplyDelete
  7. I'm curious what the racial/ethnic breakdown is in Brazil. Does the report include that info anywhere?

    ReplyDelete
  8. What's up with Greece, Serbia, Romania, Bulgaria doing worse than Hispanics?

    And what's up with Montenegro, Albania, and Argentina doing worse than blacks?

    ReplyDelete
  9. ¿Are there any Argentinians who could comment on what is the problem over there?

    Argentina is not that poor so scores shouldn't be that bad on account of the quality of their schools. Regarding the students, they should rank similarly to southern European countries, maybe a bit lower if you consider a small amount of Amerindian admixture.

    Those scores are simply awful for a country of mostly European descent, to the point that they really make you wonder if HBD is truly the main factor explaining the performance of different countries in these sort of evaluations or technological and economic achievement in general.

    I haven't yet looked at the numbers in detail, but the best excuse that I can come up with of the top of my head is that the pool of students tested is not truly representative of the whole student population. I don't know how valid this is, probably not much.

    ReplyDelete
  10. Pat Hannagan12/3/13, 3:53 PM

    It would be interesting to see the racial breakdown of Australian results.

    According to this article:

    "According to the latest Program for International Student Assessment (PISA) results, the ''Australia'' of our most advantaged students is an education powerhouse.

    Teenagers in the top socio-economic quartile of this country are achieving in maths at a level that would put them in the top five countries in the world and on par with standout performer Korea, according to Sue Thomson from the Australian Council for Educational Research.

    By contrast, the Australia as represented by students in the lowest socio-economic quartile would be placed significantly below the OECD average, in about 40th position and on par with Greece.

    Students who are indigenous are performing so poorly on average that their maths, reading and science results are equivalent to being 2½ years behind at school compared with their non-indigenous counterparts, while teenagers who live in the country are about a year behind their city counterparts. Students in Canberra are two years ahead of those in the Northern Territory when it comes to mathematical literacy."


    Anyone who knows Canberra knows that it is White White White. It's so White it's Whiter than White. The children of the nation's elite live there in their own isolated terrasphere.

    The Northern Territory, on the other hand, is Aboriginal, *real* Aboriginal. Darwin is also the most multiracial of all Oz cities. The Whites of the NT are mostly alcoholics and misfits.

    The article asserts:

    Australians pride themselves on their egalitarianism but factors such as the colour of your skin, the state you were born in and the size of your parents' bank balance have a significant bearing on how well children will do at school, and beyond.

    According to the author, if you transferred all those Territorians to Canberra, and all those Canberrans to the Territory, the results would be reversed. And it's on that basis that:

    "It was this finding that was at the heart of the national model recommended by David Gonski, which directs loadings - or additional funding - to schools based on predictors of education disadvantage, such as the number of students who are indigenous, from poorer backgrounds or attend a school in a remote area.

    See, we're gonna fix all this inegalitarian evil through funding funding funding!

    And, it'll work you know, unless that bastard Abbott and his Catholic minions get in the way.

    Oh, wait a minute, an interesting interactive map.

    The snapshot of international education achievement by the Organisation for Economic Co-operation and Development (OECD) released on Tuesday night has revealed worrying trends for Australia, showing that while students in this country perform better than the OECD average, results are slipping and are bedevilled by equity issues.

    Ah, equity issues.

    ReplyDelete
  11. Pat Hannagan12/3/13, 3:53 PM

    Btw, Btw,

    Almost 10 million migrants over the next 50 years would swell Australia's population to more than 40 million people by 2060 and more than 50 million by 2100, under dramatically higher new projections by the Bureau of Statistics.

    The bureau's new projections, the first for five years, envisage tens of millions more people crowding into Australia's capital cities over the next 50 years, overwhelmingly due to migration.

    By 2060, Sydney would have 8.4 million, an increase of 80 per cent from now. Melbourne would overtake Sydney, doubling its population to 8.5 million. Perth would more than double to 5.5 million people, and Brisbane to 4.8 million. Both cities would be bigger than Sydney is now.

    Melbourne would overtake Sydney in size in 2053, the study estimates.

    ReplyDelete
  12. That's about the only takeaway from PISA

    Japan beat Korea in reading.

    ReplyDelete
  13. "So white USA is basically an OECD country, the rest of the USA is third-world. And Generalissimo Francisco Franco is still dead"

    Asia USA is not 3rd world either. They scored better than White USA.

    ReplyDelete
  14. What's up with Greece, Serbia, Romania, Bulgaria doing worse than Hispanics?

    About 500 years of Ottoman occupation, that's what's up. Look at what has happened to us after fifty years of integration and mass immigration. I can't imagine what we will look like in 500.

    ReplyDelete
  15. The previous PISA report implied that Argentinean officials were especially assiduous about rounding up 15-year-olds from "alternative" schools (e.g., reform schools) and making them take the test.

    In general, there are some sizable differences between countries in what % of 15-year-olds are represented in test scores.

    ReplyDelete
  16. Pat Hannagan12/3/13, 4:09 PM

    According to the author, if you transferred all those Territorians to Canberra, and all those Canberrans to the Territory, the results would be reversed.

    Sorry, buggered that up. The results *would* be reversed, but the theory is that the results would stay the same, because funding.

    ReplyDelete
  17. I read something on EdReal awhile back about how useless math skills are without verbal skills.

    She (?) said that a lot of her students got this simple word problem wrong:

    "The company started the year with 90 employees and finished the year with 10% more. How many employees did it have at the end of the year?"

    Any one of us can easily come up with the right answer, 99. And EdReal did say that her (?) students had the ability to figure 10% of 90 then in turn add 90 and 9. But she then said the big hangup is that they were so bereft in verbal and language skills that they were unable to figure out that that is the math that had to be done from reading the word problem.

    So what all these Orientals and Asians are cleaning our clocks on pure math ability? Even if you are great at math, you need to be at least above average on verbal for it to amount to anything more than a hill of beans.

    Besides, tinfoil hat time, who other than me thinks that Beijing is deliberately sending its best students to Shanghai just to be able to goose up the scores in Shanghai? Or maybe China isn't administering PISA to the lower half of its bell curve.

    ReplyDelete
  18. "So basically DuBois's talented tenth is approximately equivalent to the talented third of Whites and talented half of Asians.

    Yet the fall of Detroit is a mystery to the mainstream press."

    Haiti being the poorest country in the Western Hemisphere is also a mystery the mainstream press.

    It must be because of those French descendents, as to why Haiti is still stuck in the stone age.

    ReplyDelete
  19. Pat Hannagan12/3/13, 4:41 PM

    My own rough guide as to the racial breakdown of Oz results, on the state by state basis, from the interactive map:

    The ACT and WA score well above the national average. ACT (Canberra) as pointed out is Whiter than White. It's primary purpose is to congregate the nation's elite into one city, far away from all other cities, where they can fast rack their own pay rises.

    WA is again White, with a fairly large *real* Aboriginal population compared to other states. WA's migration is predominantly from the White British Empire nations: Saffas, Poms, Kiwis, Irish, Sweaties and Welsh, pretty much in that order.

    NSW is still majority White, scoring just above the national average. Victoria is below the national average departing from NSW in 2009 which is approximately when it's greater take of immigration kicked off. As pointed out prior, based on current immigration rates Melbourne will overtake Sydney in population purely due to immigration. There has also been a massive exodus of Whites from Vic to NSW, QLD, ACT and WA. Victoria's driving economic boom is in Indian and Pakistani taxi drivers as well as arts grant recipients.

    QLD is right on the national average, for reasons above. Though an historically illiterate state it holds its own out of pure spite towards NSW.

    Of those well below the national average, NT is already explained, Tasmania became a Green/Left wasteland many years ago and has been gutted. It's sole purpose these days is to oppose any kind of national economic progression and ensure the nation is swamped with refugees and terrorists.

    South Australia is a Protestant state never having taken convict transport. Though overwhelmingly White it is barren of intelligence producing nothing of note except for excellent shiraz and our best serial murderers. It stands against all HBD theory that Protestants are inherently awesome due to them being of Northern European stock.

    ReplyDelete
  20. "Asian Americans: 549"

    Are this "Asian Americans" ethnically Chinese, Koreans and Japanese only, or include as well Indians, Bangladeshis, Pakistanis, Filipinos, Thais, Indonesians, Srilankans, Pacific islanders, etc. etc.?

    ReplyDelete
  21. And Vietnam scored almost identical as Germany did, in all 3categories! Whatever will be China's average, surely it would be better than Vietnam.

    ReplyDelete
  22. Or maybe China isn't administering PISA to the lower half of its bell curve
    .


    Like this country, maybe?

    ReplyDelete
  23. Someone asked about why Argentina's scores are so low, despite being a white country.

    Well.

    I'm an American that has lived in Buenos Aires for more than 8 years now, so I have a lot to say. (Plus I'm a big fan of Steve's analysis for a long time.)

    The best way to summarize it is....

    Argentina has the fame of being the only first world country to have converted itself into a third world country. It's a country of white people who think and act like--to use my parents' word!--schwartzas. That is the key insight to understand Argentina.

    Interesting related point: the saving grace of Argentina is that cultural capital declines very slowly. There are still small numbers of highly highly educated, sophisticated people. I walk down my block and hear kids practicing the oboe, and I have young friends who still study Latin in high school. Some world-class universities, and the faculties are shockingly, incredibly smart (including one of the only bastions of smart capitalist thought in LatAm!). But this is 0.01% of the population... the rest are people who (despite their white skin) think and act like the worst scum of the US.

    The root cause, insofar as I can determine it, is the politicization of EVERYTHING since the time of Peron (1940s). Argentina formerly had a culture that valued more hard work and the traditional virtues. But for 70 years, the slow triumph of leftist thought over every single aspect of culture has left a culture where every single individual, rich or poor, educated or uneducated, feels a deep sense of entitlement and no desire to try to actually do anything other than get money for nothing from other people. Welcome to the future.

    M

    ReplyDelete
  24. MattW: I couldn't find the raw scores by race for Brazil, but based on the scores by states on previous years and the census data about states racial composition, I estimate the white and black score to be 433 and 333, a difference of one SD.

    This a surprinsingly wide gap considering the black and whites are much closer genetically in Brazil than e.g. in the US.

    It would be nice to do the calculation again with the recent results to check how reliable the estimate is.

    ReplyDelete
  25. Argentina is not that poor so scores shouldn't be that bad on account of the quality of their schools.

    It may not be that poor, but it's poorly run and disorganized. I can easily imagine its school system not being run well enough to adequately prepare its students.

    ReplyDelete
  26. Level percentage can be easily made into bell curves.

    ReplyDelete
  27. Here in New Zealand the media has been full of the news of us having plummeted in the rankings. This, of course, they blame on our current right wing government led by John ‘John the Jew’ Key. Unlike for the 2009 results they have neglected to publish the New Zealand numbers broken down by race (in 2009 Whites scored about the same as Finns, Asians scored lower, about the same as Japan, Maori scored about the same as Israel). For the 2012 results in mathematics they do say Asians stayed the same (they got 529 in 2009), whereas Whites declined (they got 537 in 2009). Immigrants apparently did as well as non-immigrants.

    They also say “New Zealand has a relatively large proportion of both top performers (Level 5 and 6) and low performers (below level 2) in mathematics.” and
    “New Zealand students demonstrated relative strength in the mathematical area of uncertainty and data (statistics) and weaker achievement in space and shape (geometry and measurement).” which doesn‘t sound very Asian so I guess Whites are still doing relatively well in Mathematics.

    ReplyDelete
  28. So amazed by so many sour-grape comments.

    ReplyDelete
  29. "What's up with Greece, Serbia, Romania, Bulgaria doing worse than Hispanics?"

    "And what's up with Montenegro, Albania, and Argentina doing worse than blacks?"

    Well thats basically because South eastern europe was under Turk muslim rule for hundreds of years, theres was some mongrelization of the blood that happened of course, and some people who call themselves greeks are really turks. Albania is the Same thing, its a muslim country full of a bunch of arab mongrels, its one of the poorest countries in europe as well. Argentina is strange though, they do have a large italian and spaniard populace, as well as mestizo and sephardic jewish, so its not unlikely.
    I thing people should note is that mongrelization does eventually lead to the downfalls of civilizations, look at Greece, Rome, Persia, Assyria, and Egypt for instance, once great, and now gone.

    ReplyDelete
  30. Someone who, for some bizarre reason, didn‘t even provide a fake name so we could refer to him or her wrote:

    “So amazed by so many sour-grape comments.”

    There was one comment that could possibly be described as sour grapes. WTF are you talking about?

    ReplyDelete
  31. The root cause, insofar as I can determine it, is the politicization of EVERYTHING since the time of Peron (1940s). Argentina formerly had a culture that valued more hard work and the traditional virtues. But for 70 years, the slow triumph of leftist thought over every single aspect of culture has left a culture where every single individual, rich or poor, educated or uneducated, feels a deep sense of entitlement and no desire to try to actually do anything other than get money for nothing from other people. Welcome to the future.

    Is the situation better once you get out of Buenos Aires, e.g. up in Salta?

    ReplyDelete
  32. Turk muslim rule for hundreds of years, theres was some mongrelization of the blood that happened of course, and some people who call themselves greeks are really turks. Albania is the Same thing, its a muslim country full of a bunch of arab mongrels

    Turks aren't Arabs. Albania was poor and backwards even before the Ottomans, and there are other factors that can explain that, like clannishness and inbreeding as well as limited natural resources.

    mongrelization does eventually lead to the downfalls of civilizations, look at Greece, Rome, Persia, Assyria, and Egypt for instance, once great, and now gone.

    Egypt was always "mongrel."

    ReplyDelete
  33. Thanks Steve for carrying on, I think I was the first to think of comparing races versus countries.

    The percentile chart shows percentiles at each level
    Top 3% can get into schools like Cornell or Brown, top 1% might get into MIT or Harvard. Notice that 0% of blacks fall into top 3%, but 3 times as many Asians qualify, which is why Asians who are 5% of the population can be expected to be over 15% of applicants who qualify to get in, and why African Americans can be expected to be underrepresented if they are held to the same test score standards (which they aren't due to affirmative action)
    Black top 0% =
    White top 3% =
    Asian top 10%

    The top 1% of African americans are equal to the top 12% of whites or 25% of Asians

    Level 5
    Black top 1% =
    Hispanic top 6%
    White top 12% =
    Asian top 25% =

    Interesting the comments show that Asians in New Zealand, and I would assume in Australia probably also peform comparably to Asian nations.

    Shanghai is an anomaly in China, it's like an elite suburb of DC or Palo Alto, not at all representative of the average peasant who only recently were lucky to finish a k6 education at public expense.

    ReplyDelete
  34. Last year I might have been the first to notice whare all the Arab nations are clumped together (hint, it's not at the top or middle)

    ReplyDelete
  35. How do we explain the remarkably low score of Argentina, when their racial majority are from Spain and Italy who score 96 points higher (388 vs. 484)? Might it be because 97% of them are Catholics? How do we explain why African Americans score 14 points higher than Costa Rica, who are also mostly Spanish and Italy? Because they are 76% Catholic?

    How do we explain why African Americans score 30 points higher than Brazil who has the most Catholics on the planet? Only half of them are Whites [or, specifically 47.7% are Portuguese], and more than half (50.7%) are mulatto or Black. If the Whites and Asians in Brazil score as high as Portugal at 487, then for Brazil to score 391 would require the mulatto and Black to score 298:

    .507 x X + .493 x 487 = 391
    .507X = 391 - 240 = 151
    X = 298

    Another possibility is that less than half of African Americans are in school at the age that this test is given, and only 4% graduate from hilgh school prepared for college. If they were included, could their average score be as low as 298?

    But back to the point above. It would appear that religion has an even larger effect on test scores than race.

    My bet is that if "White Americans" who scored 506 were to be broken down by "White Catholic Americans" and "White Protestant Americans", the gap would be just as big as these cross country gaps.

    ReplyDelete

Comments are moderated, at whim.