Sweden recovers in the PISA tables

The 2013 edition of the survey was a wake-up call for Sweden, which experienced a sharp drop in the results of 15-year-old students, pushing them below the OECD average and sparking debate about the country’s schools.

But Tuesday’s release of the 2018 Pisa ranking suggests that Swedish schools have further improved on a recovery that started with the 2016 edition, with above-average scores in reading, mathematics and science.

The report notes that Sweden is back at levels observed in early rankings, despite a rapid increase in the proportion of immigrant students, who tend to score below native-born students, in recent years. In 2018 around 20 percent of students in Sweden had an immigrant background, up from 12 percent in 2009.

Here is the full story.  You will note that way back when people blamed the 2013 PISA score plunge on Swedish school privatization and vouchers, and also it led to a lot of fury about immigrants.  Others blamed “progressive schooling” techniques.  In any case, the Swedish performance has recovered notably and is back to its previous levels.  Don’t overreact to particular data points, people!  Of course it remains possible that the next measure in three years will show renewed decline, but do not overtheorize based on very limited trends.


"Don’t overreact to particular data points, people! "

LOL, who is this aimed at? I don't recall the 2013 PISA results as being a watershed moment on this blog. Was it a really contentious issue 5 or 6 years ago?

Correction: The 2013 Swedish PISA results.

You need to get about 10 comments in from the 2014 link...

Wiki: "In 2011, with the total population being 9,562,556; roughly 15% of the population was born abroad, 5% of the population was born in Sweden to two parents born abroad, and another 7% was born in Sweden to one parent born abroad. Resulting in 27% of the Swedish population being of at least partly foreign descent.[11]"

Then Sailer chimes in and, well, you know...

I don't understand the logic.

Sailer references the differences between Swedish children and the children of immigrants. A difference that is true on every PISA report, so clearly not over-reacting to one data point. You can't really disagree with Sailer's data, because it's the raw data. You can disagree with the cause and plenty of people do, but he's clearly not guilty of what Tyler is charging.

Doesn't matter. He said something that was not allowed even if it was true so he must be chastised.

*even if it was false/irrelevant

if it were false/irrelevant then debate it. But they don't want to debate it or expose the truth. So it must be censored.

'I don't understand the logic.'

From the current post - 'You will note that way back when people blamed the 2013 PISA score plunge on Swedish school privatization and vouchers, and also it led to a lot of fury about immigrants. '

Seems like TC is referencing Sailer in terms of immigrants, even if this is only a single data point.

Admittedly, I have an extremely hard time figuring out who the 'you' or 'people' is supposed to be in posts like this.

"Admittedly, I have an extremely hard time figuring out who the 'you' or 'people' is supposed to be in posts like this."

+1, if feels as if Tyler has a weak point, knows it's weak and therefore refrains from specifics that would be readily challenged.

How do we know that they didn't just teach their students to do well on the exam?

So what? Unless they gave their students the tests'answers, who cares?

We actually know that this is exactly what they did.
Every country that receives a significant influx of intellectually impotent subhuman third-worlders can only do that to avoid a steep decline in any quantitative measure related to education.

I hope this is sarcasm mocking some of the regulars here. If not, I feel sad for you and am disgusted that you’d refer to other humans as subhuman. So are the Irish still on the list of subhumans because they were definitely on racists’ lists 150 years ago. And if they’re no longer subhuman, how did they overcome it all if genetics explains it all and there’s no hope for improvement among the subhuman classes? But, yeah, if your comment isn’t sarcasm, you need help.

The Irish pest are definitely the worst western europeans, just a tad better than their own travellers. Still white, though.
I'd take them over any non-whites and take any other non-muslim whites over them except southern italians and cypriots (to the extent they are white at all).

The one who needs help is you. Your grandchildren or great-grandchildren are going to be brown subhumans whereas the future of the white race belongs with my progeny. Which is why we will triumph.

Your progeny will become extinct due to opioid overdoses and accidental gunshots.

Both of these problems are specific to the USA (and, at that, to the USA of the last 2-3 decades).
In case you have not noticed, the majority of whites don't live in the states.

It doesn't matter though. Even if they persist and spread, having 100% white children leaves your offspring white.
Liberasts will eventually mix with the subhumans and exit the white gene pool.
So, yes, the white race in the future will be 100% far right and racist and, since only white lives matter, only my progeny will matter.
Yours won't. It's as simple as that.

But thanks for your concern.

Whatever gets you through the night, pumpkin

Greta Thunberg raises the Swedish PISA score well above the score of all other OECD nations. In a hundred years, nobody will remember the captains of industry or tech, they will remember Greta Thunberg.

True, nobody has ever forgotten Thomas Malthus and it's doubtful they'll forget Paul R. Ehrlich either.

Malthus, indeed. His theory was that population growth, left unchecked, would outgrow its resources, famine, disease, and warfare being the "natural" checks on population growth. What are the "natural" checks on climate change? Famine, disease, and warfare? Thank you for making Greta's point.

" and warfare?"

You and Greta are right. The coming Robot War in 2057 will be trivial in comparison to the Climate Change War in 2091.

rayward, your comment is nearly completely wrong. Yes Malthus thought that population growth would have to be checked otherwise it would outgrow resources. Of course, he (much like Paul Ehrlich) failed to take into account technological progress which would vastly increase the amount of resources available.

However, you are correct in that Greta Thuneburg is making the very same point / mistake. No one has pointed out to her the technology curve which indicates that solar/wind power backed up by battery power will become the cheapest source of electricity (and eventually transportation fuel) by mid century at the latest.

The world is already well underway to a translation to renewable power. The US has been converting over to renewables at 0.5% per year for well over a decade and is up to around 1% gain per year at this point. It's too early to declare this a long term trend, but there aren't many factors that would tend to slow it down by much, let alone stop it.

At current rate, you would expect the US to be getting over half it's electricity from renewables (including hydro) by 2055. However, it's quite likely that solar/wind/batteries will continue to drop in price which will accelerate their adoption.

Well, more than 11 percent of the Swedish kids did not participate (many of whom because they did not speak Swedish well enough to complete the test), i.e. lower participation than in any other country. It is quite possible that their improvement is mostly the result of data massaging.

Exactly! The OECD writes: "The massive inflow of immigrants in the most recent period, however, also led to an increase in student exclusion rates. In 2018, about 11% of 15-year-old students were excluded from the PISA test – the highest rate amongst all participating countries/economies. While limited information is available about excluded students, this increase is most likely the consequence of the large (and temporary) increase, between 2015 and 2018, of recently arrived immigrants in the school system."

"quite possible that their improvement is mostly the result of data massaging"

Why are you harshing Tyler's mellow? He doesn't want to know (or even think) that?!

It's kind of half hilarious and half tragic that the troll comes along to attempt to derail the thread. Purely because somebody might be talking about something he doesn't like and this is his best dullard's attempt at censorship. Then again the Heckler's veto is a real thing.

Kudos to whomever came in and cleaned out the trolls comments.

From the histogram of the PISA performance of Sweden and Denmark for Math, SWE has higher % in the low L1- and L2 proficiency categories and about equal in L3. However at the highest L5 and L6 levels SWE has higher % than that for DNK. SWE has a better smart fraction than that for DNK.


PISA also has a table of data where the various demographic factors for 2018 are projected back to 2015 and 2012, i.e. the adjusted score for 2012 and 2015 if the demographics were the same as that in 2018. Thus with the demographic effects removed, for SWE it can be seen that there are incremental improvements. For DNK relative to 2012, it performed better in 2015 but then decreases slightly for 2018.

Table I.B1.41: Mean mathematics performance, 2012 through 2018, adjusted for demographic changes


Err. higher % in the low L1- and L1 proficiency categories and about equal in L2.

Americans of Swedish descent seem to be on the dull side.

I'm curious what can be learned from any amount of data points about policy things that definitely improve k-12 educational quality above some very basic minimal standard. It feels like it's an awful lot "at home" variables and cultural attitudes / expectations and what kid peers are doing. I don't think there's a policy lever, for all I'm more than willing to adopt a "lets experiment" attitude.

I think we should remember the People's Republic leads the world in educational outcomes.

Yeah..... I really doubt anything is a "wake-up call" for Sweden.

"but do not overtheorize based on very limited trends."

Wise words, which will never be followed.

+1, if it hasn't been happening for at least 5 periods and more properly 10, it's as likely to be noise as it is to be true signal

Why is there such an onus on Sweden, or any other nation, to accept immigrants?

Why is immigration sacralized?

I sense there is much more acceptance of immigration by the upper classes. Is it simply a desire for cheap labor?

I don't know if sacralized is the right word, but there is an ethical principle in play. For people who don't draw large distinctions between the moral worth of someone standing in Sweden and someone standing just outside of Sweden, and to the extent acceptance of such a person as an immigrant could dramatically increase their well being, like way more than giving them aid in some other form, and to the extent that such people can also positively contribute to the societies into which they are accepted, a reasonable moral calculus could suggest letting them in. I understand some hold higher value for near relations than far relations. I understand some don't feel there is a necessity of charity, but I think there is a reasonable case to be made that is not based on cheap labor supply. That case would be quite weak given the nature of global trade.

"between the moral worth of someone standing in Sweden and someone standing just outside of Sweden"

There are clearly genetic differences between Swedes and non-Swedes.
The moronic Swedes who treat those two groups the same are being evolutionary maladaptive and will become extinct. (Yes, group selection is real).

The future of the white race rests with far-right racists.
Everyone else will either become extinct or a mongrel.

"For people who don't draw large distinctions between the moral worth of someone standing in Sweden and someone standing just outside of Sweden"

But that logic leads moving the whole world into Sweden (or at least far more people than Sweden can maintain). It's a clearly unsustainable model.
The other fact is that the upper classes, being shielded from the consequences of bad choices they might make tend to be as dumb as a box of rocks about anything that happens outside their protective bubble.

Extracted from PISA's table (https://www.oecd.org/pisa/Combined_Executive_Summaries_PISA_2018.pdf):

Country * Migrant Share % * Non-Migrant * 2G Migrant * 1G Migrant * Gap (Afters SES Controls)
Hong Kong * 37.9 * 529 * 533 * 502 * 9 * 24
Canada * 35 * 525 * 535 * 508 * -1 * 26.2
Switzerland * 33.9 * 503 * 453 * 448 * -25
Australia * 27.7 * 504 * 523 * 501 * 7
New Zealand * 26.5 * 510 * 518 * 500 * -8
Singapore * 24.8 * 546 * 587 * 554 * -9
United States * 23 * 510 * 512 * 479 * 16
Austria * 22.7 * 500 * 446 * 421 * -33
Germany * 22.2 * 519 * 477 * 405 * -17
Sweden * 20.5 * 525 * 471 * 410 * -54
United Kingdom * 19.8 * 511 * 493 * 488 * -4

Above subset ranked by non-migrant performance:

Country * Migrant Share % * Non-Migrant * 2G Migrant * 1G Migrant * Gap (Afters SES Controls)
1) Singapore * 24.8 * 546 * 587 * 554 * -9
2) Canada * 35 * 525 * 535 * 508 * -1 * 26.2
3) Hong Kong * 37.9 * 529 * 533 * 502 * 9 * 24
4) Sweden * 20.5 * 525 * 471 * 410 * -54
5) Germany * 22.2 * 519 * 477 * 405 * -17
6) United Kingdom * 19.8 * 511 * 493 * 488 * -4
7=) New Zealand * 26.5 * 510 * 518 * 500 * -8
7=) United States * 23 * 510 * 512 * 479 * 16
8) Australia * 27.7 * 504 * 523 * 501 * 7
9) Switzerland * 33.9 * 503 * 453 * 448 * -25
10) Austria * 22.7 * 500 * 446 * 421 * -33

Unfortunately, PISA only give Reading as migrant/non-migrant difference in 2019 results, as far as I know.

Taking overall reading score of a subset of randomly chosen countries where immigrant share is low enough that they didn't bother (or couldn't) estimate gap:

Finland * 520
Poland * 512
Japan * 504

Sweden overall performance - 506.

Though, despite local:migrant gaps in Maths and Science not being present, if we for the same of a proxy assume those are about the same size as in Reading, then Sweden, Germany and Switzerland (and probably some other nations) locals largely converge on country scores with Japan, South Korea, and Taiwan (i.e. East Asian mostly democratic and "normal" regimes)... though still quite a bit short of China's chosen testing areas (HK SAR, Macao SAR and the four city region) and of Singapore.

I ... was not aware of the depths being plumbed in the MR commentariat these days.

"I understand some hold higher value for near relations than far relations. I understand some don't feel there is a necessity of charity ..."

Well, just to to put a bow on your shock, I'd submit that some people hold a mix of values, and more babies, more "far relations", are only one of them.

But I was only presenting a version of the case that's not contingent on "we want low cost labor". I agree there can be a basket of values that is neither of those things, but the claim to which I was responding was that it was all about cheap labor and other explanations were improbable.

Oh, I see - though it must be said, a "reasonable moral calculus" is hardly what would be operating in that case; rather, a lazy, unreflective recital of one of the more ethically questionable vestiges of Judeo-Christianity to have passed into post-Christian "intellectual" life.

Since several commenters have brought up Steve Sailer, I thought it useful to link to the new version (revised for the new Pisa) of his incredibly depressing graph.

Why is that depressing? It looks as if US blacks score better than most African countries, US hispanics score better than most hispanic countries, US Asians score better than most Asian countries and US whites score better than most European countries.

Overall, it looks like the US education system is doing a pretty damn good job.

The Sailer take is probably a little *bit* too positive on the USA.

Co-plot of Local and Migrant Reading scores: https://imgur.com/a/N9WP4hd

US is at the top again, but breaks less considering local vs migrant differences.

Plus this is probably less considering the US's relative Reading advantage (about +10 points, where most European nations are neutral or - on reading compared to overall).

And US Asians are likely quite heavily selected for education compared to Asian countries' averages (so the advantage is not "same people, better system").

But yes, that said, of course it is quite positive - consider that US Blacks match Malta and the Balkans, and that US Hispanics match the weak performers in Western Europe (Iceland and the like) without even adjusting for 1st vs 2nd generation status (which would likely improve them)!

You can interpret the differences in the negative light, that there is a persistent gap... but the positive spin on it is that between race gaps in the US are quite compressed compared in the grand scheme within race, between different country gaps, and that in particular US Blacks (very much so) and Hispanics have a large premium of US advantage on co-ethnics that is not explained by selection (as does probably explains things for Asians, where the average Chinese that ends up in the US is probably quite a different person to the average Taiwanese or Japanese).

Yes, that is a reassuring way to look at it.

But behind that reassuring ... A totally naive "man from Mars" implication of the graph is that, on average, "Asians" are smarter than "whites", who are smarter than "hispanics", who are smarter than "blacks". Why is that depressing?

1) I don't want to believe it.

2) Most Americans don't believe it, or are very ready to publicly deny it. This means that either most Americans are wrong or they are liars.

3) American public policy and law are based on it not being true. For example, any "disparate impact" is treated as caused by illegal, immoral discrimination in public discourse. In law, an "over-representation" or "under-representation" of any of the above groups sets up a rebuttable presumption that whoever causes the "disparate impact" is breaking the law.

But if it is true, "disparate impact" is natural and to be expected (at least to the extent that smarts have an impact on outcomes).

"1) I don't want to believe it. ... But if it is true, "disparate impact" is natural and to be expected"

Cognitive dissonance always causes stress. I try to avoid disagreeing with the data. That's not always possible, because the data doesn't always agree or craft a discernible narrative. But I try to avoid worrying about things beyond my control.

If there is a racial difference, American society will do a decent job of getting everyone a pretty good chance at success and happiness. It won't ever be perfect, but it will tend to get better over time.

If there's no racial difference and it's largely a cultural issue then once again American society has a pretty good track record of ameliorating those differences over time.

And frankly, starvation in Africa, or Chinese concentration camps for Uighurs or civil strife in Venezuela are all bigger current problems, than a 20% difference on a standardized test.

Granted, I think the American education system would do even better if it just admitted there were differences and didn't try to fit square pegs into round holes. But still the overall test results indicate it's not a huge issue. I am glad we actually have concrete data now.

If there is a racial difference, American society will do a decent job of getting everyone a pretty good chance at success and happiness. It won't ever be perfect, but it will tend to get better over time.

Perhaps. But this particular bit of data (and the several previous PISA results) suggest the "racial difference" is big. That there will be not just a small "disparate impact" that people can deal with. It may well be a "huge issue".

A complication is that on present trends, we will never know just how much, if any, inherent difference exists. There is lots of data and lots of second-tier scholarship saying it exists but also counter-arguments that "no, this doesn't prove ..." And there the matter hangs. No one with any desire to be successful in academia wants to seriously investigate it.

Which leaves us where we are today. The continuing performance gaps lead to increasingly bitter diagnoses of American society. If there really are no differences, then America must indeed be a "terrible, horrible, no good, very bad" place--and perhaps the only answer is a thoroughly quota-ized society with lots of official power to the good and true and woke.

Almost everybody believes that Asians are smarter. And they have believed it for at least 30 years now.

Pauline Kael found it hard to believe that Richard Nixon had won a landslide victory in 1972 because she didn't know anyone who had voted for him. Perhaps everyone you know "believes that Asians are smarter" but 99% of the American people will not say such a thing publicly.

"...but 99% of the American people will not say such a thing publicly."

There are a litany of articles about how common it is to stereotype Asians as smarter or good at math. So yes, people do say that all the time.




I defy anyone running for office to say, "Asians are smarter." Hell, I defy anyone who is a commentator in the respectable media.

What would happen if someone took me up on it? My guess is that within an hour they would be issuing an apology.

Well of course the PC media would crucify anyone publicly saying that, but Andrew Yang has come pretty close:

"“The opposite of Donald Trump is an Asian man who likes math.”"

"“Now, I am Asian, so I know a lot of doctors,”

And yes, even that drew criticism from the PC press. I also suspect that if an Asian Republican has made the same comments, he or she would have been excoriated.

You posit an interesting situation, where "Almost everybody believes that Asians are smarter" but no one will say it publicly.

Paging Timur Kuran.

Why would anyone running for office, or any other public figure, need to say that?

No politician needs to say anything, and part of being a good politician is deflecting questions with non-responsive "answers" but having said that ... if the politician was asked, "how come Asians are over-represented at New York's exam schools, and at Cal Tech, and at Harvard?

The idea of "over-representation" and "under-representation" of certain groups assumes that the groups are inherently equal.

Racial IQ hokum suffers from the same problem as the more woo sociology experiments, in that there is no way to replicate or see evidence for them in the real world.

If racial differences were true, wouldn't we expect to see a recurring pattern over time?
Just as we do with the natural world, wouldn't a superior race emerge over time, and subjugate all others?

Yet nowhere is any of this evidence presented.
A century ago, Englishmen were the obviously superior race, while Chinese were subhuman apes;
Today China looks to eat Britain's lunch.

How to explain all this?

Bad social policies? Listening to Swedes?

Ian Morris tries to do just that in Why the West Rules--for Now: The Patterns of History, and What They Reveal About the Future. It's an interesting, well-written book. I would summarize it as "Some times the East is ahead; sometimes the West is ahead. It's a matter of technology, institutions, history.

Cultural effects are more significant.

Eh, that came out wrong. I was almagating "technology, institutions, history." into cultural effects.

So, I was agreeing with you (and/or the book), not disagreeing.

The socialist paradise needed reforming. Who knew?

My son received a Masters in Engineering from Swedish school...now I'm wondering... good they got back on track.

Comments for this post are closed