Hacker News Reader

Your analysis is simplistic. When you break up TIMSS by ethnic groups, the US performs comparably to other counties across the board (Asians, Europeans, etc). [1] There's something to be said about the differences between attempting to educate a racially diverse population versus a largely homogenous one.

I've done more analysis on a similar test called PISA, but I think it's worth bringing up in the context of this debate. In standardized testing there are often significant underlying factors that have nothing to do with the schools themselves. From the executive summary of PISA: "[In the United States], after accounting for socio-economic background, the performance difference between students from single-parent families and those from other types of families stands at 23 score points.... Parents’ engagement with their children’s reading life has a positive impact on their children’s reading performance." Consider that the divorce rate in the US is one of the highest in the world, 5 times that of China. This problem along with any others is one that needs to be considered in the context of education.

Also, PISA was not done on China, but rather on two specific cities: Shanghai and Hong Kong (similarly, TIMSS was done in Hong Kong). Along with Beijing, these are the most advantaged areas in China in terms of both money and education. The US administers PISA to a wide range of schools across the country. I imagine we would see much different results if PISA only tested Boston and some other advantaged city. The PISA study itself even notes that scores were much higher in urban schools. Other countries are either testing exclusively urban schools or urban schools at a higher rate than the US.

There are serious, serious problems with education in the US. But it's important to look at these studies with a critical eye and avoid the temptation to go off on a rant on how the US is bad at math. Data doesn't lie, but analysis is often wrong and/or exaggerated. In sum, problems with education in the US are deeply rooted in racial, social, and geographic issues. Better management practices and policy reform, while good, doesn't change the fact that the US isn't Singapore.

[1] http://nces.ed.gov/pubs2009/2009001_suptables.pdf

I checked the link you kindly provided. Before I go further in replying to your reply, would it be all right to ask where the support for your assertion at footnote 1 actually is found in that publication? What I see, in the data tables there, is that some countries plainly outperform all subgroups in the United States. Several of those countries have lower spending per pupil than the United States (either by current exchange rates or by purchasing power parity), so I'd like to know what they are doing right. I claim, in my grandparent post to which you replied, that one thing other countries are doing better than the United States is simply providing better primary instruction, with better designed curricula. One other thing that I think they are doing right is giving students better-designed educational tests, aligned to those better curricula, that more realistically gauge whether or not the students are learning what they need to learn. (That's the point of the anonymous anecdote about the school board member mentioned in the article that was submitted to open this thread. Perhaps United States standardized tests given to tenth graders in some unnamed state have poor validity and poorly written item content. I actually think that is quite likely. But I don't think that the correct policy response to that is to stop giving students tests to find out what they know, but rather to write better tests based on better curricula. It's too bad that the article doesn't link to the actual test.)

I agree with some points in your reply. I don't think China as a whole is well represented by the schools in its most developed urban areas. The results from Shanghai in the most recently announced test to include Shanghai surely don't reflect what students from rural areas in China would do on the same test. But even agreeing with that point, I wonder if you've had a chance to take a look at what Ma's book

http://www.amazon.com/Knowing-Teaching-Elementary-Mathematic...

says about differing classroom practices and differing lesson content between the United States and China. China is very, very, very much poorer than the United States because of the lousy policies it had in the 1950s and 1960s. But its educational policies since the 1970s have been on an increasingly sound basis, and seem to be producing admirable results in economic growth with remarkably low school budgets. But please note that I never appeal to China as a country with country-wide results that are uniformly better than those of the United States. China is especially doing well on a resources-adjusted basis, while Singapore, Taiwan, and some other countries are just plain doing well nationwide, period. (I am most familiar with Taiwan, from much time living there.)

I also agree with the idea that it's important to look at education studies "with a critical eye" and it was with that in mind that I referred fellow participants on HN on several earlier occasions to the studies showing that United States schools are underserving the most able learners,

http://educationnext.org/teaching-math-to-the-talented/

missing opportunities to reach the top end of mathematics achievement reached by other countries. "Data doesn't lie, but analysis is often wrong and/or exaggerated," I agree, and what I find is that some forms of analysis are not even attempted by many commentators on education policy. I think writings that are good examples of good analysis

http://economics.stanford.edu/faculty/hoxby

http://edpro.stanford.edu/hanushek/content.asp?contentId=60

http://www.hks.harvard.edu/about/faculty-staff-directory/pau...

https://rowman.com/isbn/9781578866229

are food for thought for those of us participating on Hacker News who seek ways to improve education wherever we live.

Oddly, people rarely mention Canada when comparing the U.S. performance on the PISA to that of other countries, even though Canada is by far the country most similar to that of the U.S. among those who participate in PISA.

Canada was consistently ranked approximately fourth, a fact that seems to have been largely ignored.

Note: here is what tokenadult says about using cross-sectional TIMSS scores to make comparisons:

"The methodology of the blog post [comparing aggregate TIMSS scores] you point to repeatedly is laughable."

http://news.ycombinator.com/item?id=2773391

Incidentally, the data tokenadult links to suggests that differences in TIMSS scores are due to differences in ethnicity of the student body, not management practices. He deliberately ignores this data but the rest of us should not.

http://en.wikipedia.org/wiki/Trends_in_International_Mathema...

I have lived in one of the other countries, know hundreds upon hundreds of people from several of the other countries, and own and have read textbooks from several of the top TIMSS countries, including textbooks written in languages other than English. (I speak, understand, read, and write Chinese to the level of a professional translator and interpreter.) I reject the facile analysis you picked up from a blog post--so far not published in a peer-reviewed journal--by a graduate student who has yet to complete his degree because I have read better research on the subject by authors who have completed their Ph.D. degrees at better academic institutions, and who have published in peer-reviewed journals of high quality. I'm sorry for you if you are stuck on one lame explanation for the phenomenon of underperforming United States schools, but especially if you would take the time and effort to read good-quality dead-tree literature on the subject,

http://www.amazon.com/Vision-Elementary-Mathematics-W-Sawyer...

http://www.amazon.com/Learning-Gap-Schools-Japanese-Educatio...

http://www.amazon.com/Teaching-Gap-Improving-Education-Class...

http://www.amazon.com/Knowing-Teaching-Elementary-Mathematic...

you could learn something new that could help you better understand the other countries in the world and what the United States might learn from them.

Sir, your ability to cram as much fallacious argumentum ad verecundiam as possible into a single paragraph is unparalleled. I salute you.

When I looked up the Latin phrase I realized probably if I had to look it up, plenty of other people would too, so I copied the first couple of lines from Wikipedia here to save people the effort.

----

Argument from authority (also known as appeal to authority or argumentum ad verecundiam) is a special type of inductive argument which often takes the form of a statistical syllogism.

Although certain classes of argument from authority do on occasion constitute strong inductive arguments, arguments from authority are commonly used in a fallacious manner.

-----

Since I'm not logically proving anything, I'll refer to an authority, Strunk and White, who say:

"Avoid foreign languages: the writer will occasionally find it convenient or necessary to borrow from other languages. Some writers, however, from sheer exuberance or desire to show off, sprinkle their work liberally with foreign expressions, with no regard for the reader's comfort. It is a bad habit. Write in English"

+1

You are attempting to make the logical fallacy of arguing from authority. Hacker news is smart enough to see through this.

The only data I've cited is TIMSS data which is published and presumably peer reviewed. It's true that Tino Sanandaji PhD (note: he graduated, not that it matters) pointed this data in a blog post. So what? It's the same data you cite.

The data says the groups of people with top math performance worldwide (circa 2007) are:

    1) Taiwan 598
    2) South Korea 597
    3) Singapore 593
    4) Asian Americans 582
    5) Hong Kong 572

From this data, how do you conclude that American schools underperform?

I'm not asking if you speak Chinese or whether "hundreds upon hundreds" of your Chinese buddies agree with you. That is irrelevant to the topic at hand.

I'm asking you for a logical argument, based on data (in particular the data that you cited), that concludes American schools underperform. Any such argument needs to control for the quality of students, since student test scores are obviously a function of both the school and the students.

If you believe the dead tree books you hint have such an argument, please tell me which book and which pages.

Or feel free to continue making logical fallacies. I'm not really posting for the purpose of arguing with you, I don't expect to learn anything from that. I'm just trying to make sure others reading your posts are not misled by vague assertions that hundreds of unnamed authorities might agree with Chinese speakers like you.

P.S. To make myself appear smart, while providing absolutely no facts, I'll cite dead trees also:

http://www.amazon.com/Harry-Potter-Sorcerers-Stone-Book/dp/0...

http://www.amazon.com/Secret-Diary-Call-Girl-Belle/dp/044654...

Some more irrelevant facts: I have a PhD, I lived in Asia, I often have sex with Asian women, and I'm a good cook.

I feel silly wading in after tokenadult has gone diploma waving and you've shouted expeliarmus, but it's worth noting that TIMSS data suffices to show a performance gap between two groups, but doesn't seem to suffice to show that, ceteris paribus, there's a performance gap.

It seems that tokenadult believes there's a mimicable or replicable non-genetic factor, and you believe there's a genetic or otherwise 'unique' factor that indicates a gap that can't be closed. It's probably some combination of the two, but it'd be nice if we could talk about that instead of this other stuff.

No, tokenadult believes the factor lives within the school system, and I'm arguing it does not (by pointing out that different school systems get the same result).

I made no argument as to whether that factor is genetics, tiger mothers, Kumon, etc.

But you don't need to make that argument. If the school system does not perform some trick that is able to obviate a purported genetic bonus, there is a school system factor that's holding back students that don't have this hypothetical genetic or tiger mother bonus.

If you don't give a wheelchair to someone who can't walk, of course they can't get around...

If schools in Singapore don't perform that trick (or if the trick doesn't exist), then Singapore schools are no better than US schools.

Since Singaporean schools have a student body that is roughly 100% Asian, we have no data on whether or not they perform that trick.

Are we failing able-bodied people by not giving them wheelchairs and handicapped parking spaces?

A different straw argument - we don't send dogs to preschool, we send them to obedience school, because we recognize certain things need to be different to achieve the same goal - not biting or urinating on everything.

edit: I'm realizing this might just be about effort/actions vs. results. I'm decidedly on the side that results are what we need to compare.

If the tests tend to emphasize memorization (either literally or memorization-of-algorithms), wouldn't that culturally put east asians at an advantage, and intelligent executives who manage a 3 billion dollar budget but are 20 years out of school like the author at a disadvantage?

Singapore is by no means a homogeneous country. There a re four separate historical groups (Malays, Tamils, Chinese and Europeans) with various subcultures within those groups (you can still find speakers of Chinese dialects in Singapore that you would struggle to find in China). There is also significant recent migration from pretty much everywhere else on the planet to Singapore. Asia isn't one block race where everyone is the same. Please, do some research before you go waving around random statements as fact.

I didn't claim Singapore was homogeneous, I claimed it was predominantly Asian. According to wikipedia, non-Asian groups make up no more than 3.3% of Singapore.

http://en.wikipedia.org/wiki/Demographics_of_Singapore

I don't think it's fair to blame a school system for failing to educate the children of a single mother on welfare as well as those with PHD's from Harvard. And I am not suggesting that it's genetic. If you took identical twins and adopted one into each of those family's I would expect to see a similar performance gap.

The US has some 'great' public schools, and some 'terrible' ones. But, the difference has a lot more to do with the students than people are comfortable admitting. The private / public gap in education is also a lot less than you might expect.

I agree, I don't think school systems deserve the blame. They're shouldering what might be considered an unrealistic burden, but this is a potentially major way to improve performance - by recognizing the differences that lie within students and getting 'single-mother-on-welfare' to perform as well as 'book mcbook III'.

I was put off by 'tokenadult's first post (no part of it seemed well argued; it also started out by citing Twain to argue that school board members were "dull"). And as always I appreciate your commitment to rational debate.

But even Asian Americans reliably outperform other ethnicities in American schools, placing in the top 5 in math performance worldwide despite American schools, it does not necessarily follow that American schools are performing adequately.

For instance, Asian families can obviously be supplementing education through private tutors, additional study and demands at home, &c.

It also doesn't follow that just because this system "works" for Asian Americans that it's a reasonable system to apply nationwide. Perhaps Asian families manage this because they're economically advantaged. Perhaps Asian families make unreasonable sacrifices in other areas of their life (by a measure of "reasonableness" along the lines of "benefit to child ends up being worth the cost to the family").

Also worth mentioning, but perhaps not worth dwelling on, that TIMSS includes private schools. How much more likely are Asian Americans to attend private schools than (say) Latino Americans?

Finally, I think this particular debate may be besides the point. We have the ethnic and socioeconomic mix that we have. We do not have the option of being South Korea. So the question is, does the education system we have serve the best interests of mix of students we have today? That's the question this WaPo story is addressing.

I didn't say American schools were performing "adequately" (whatever adequately means), I just said they weren't underperforming. I.e., US schools perform as well as Asian schools for all categories of student we have data on.

I have no doubt that Asian families (in the US, Singapore, Japan, etc) send their kids to Kumon and are terrorized by tiger mothers. Net result: in all these nations, the children of tiger mothers score about 570-600 on TIMSS. So US schools systems educate Asian students pretty much the same as Asian schools.

There is not enough data to compare the results for non-Asian students, since we lack data on (Asian School, Black/Hispanic/White student).

Similarly, people of European descent tend to get scores of about 470-530, with Finland being an outlier at 546. Among this group, Americans of European descent are #6 (at 524). Again, data on (European school, Asian/Black/Hispanic student) student is lacking.

http://super-economy.blogspot.com/2010/12/amazing-truth-abou...

So the data suggests US schools do not significantly underperform either Asian or European schools, at least for the categories of student we have data on.

I.e., if US schools are inadequate, then so are the schools of most of the world.

You are attempting to make the logical fallacy of arguing from authority. Hacker news is smart enough to see through this.

You field logical fallacies of your own on a regular basis. Indeed, your gratuitous rudeness and sexism borders on ad hominem and is inappropriate here.

From this data, how do you conclude that American schools underperform?

If American schools with heterogenous student bodies* are preparing only one ethnic group to be competitive in international math comparisons, then they are certainly coming up short in other areas. The 'melting pot' is not a new concept in American education and while disparities in language ability, cultural mores, and economic condition present an additional challenge for American educators, addressing these disparities has long been a part of their mission and resources are allocated appropriate.

* It's reasonable to say that American schools serve a more ethnically and culturally diverse student body than the other examples cited here, notwithstanding the relative diversity of Singapore and Hong Kong by regional standards.

> I often have sex with Asian women

Now I'm jealous!

You are comparing Asian 8th grade scores to the Asian-American 4th grade score. If you compare 8th graders to 8th graders (Asian-Americans scored 549[1]), you'll find that Asian-Americans came in a distant last place among all ethnic East Asians.

Still, somewhat ironically, you are both right. Underperforming ethnic groups in the US bring the average down, and the US educational Prime Directive of making sure the high-performing groups don't get ahead brings it down even farther. While Asian schools work hard to push Asian kids ahead, US schools refuse to do so, because that would just widen "the achievement gap" they're trying so hard to close. So US schools, first in spending, are last in effectiveness for ethnic Asian kids. That "doing the least with most" is school underperformance. The only reason Asian-American kids do as well as they do is that, like their cousins in Asia, they get a lot of their education outside of school.

[1] 2007 TIMSS results: http://nces.ed.gov/pubs2009/2009001_suptables.pdf Asian scores are on p. 3, Asian-American scores on p. 15

Thanks, my mistake. That's embarrasing, I totally misread wikipedia. Regardless, I do stand by the claim that the bulk of the gap is still explained by ethnicity.

I do agree with you that US schools are not cost effective - I've long been a proponent of cost cutting. In my view, the biggest problem we have with US schools is cost, not quality, and we should focus our efforts on making school cheaper.

The tone of this post is horribly condescending, I feel embarrassed for you. This is the kind of stuff I read HN to avoid.

Other than that, the author of the submitted article simply describes the school board member as a "success" who makes money.

The person was described in somewhat more detail than that:

He continued, “It seems to me something is seriously wrong. I have a bachelor of science degree, two masters degrees, and 15 credit hours toward a doctorate. “I help oversee an organization with 22,000 employees and a $3 billion operations and capital budget, and am able to make sense of complex data related to those responsibilities.

I'm in favor of standardized testing and am as annoyed as you by the shallowness of this article, but you're undermining your own argument here.

Um, "limited academic ability?" The guy has a bachelors in science, two masters degrees and is working on a Ph.D. He's not stupid, regardless of Mark Twain's wit.

Now, I completely agree that a single anecdote isn't cause for policy overhaul. But if a guy who has been successful in college can't pass a test designed to weed out students who shouldn't go to college, that's a clue that maybe something, somewhere isn't working right.

Minor correction: he has a bachelor of science degree, which does not necessarily require learning any science. Or, more to the point, any math.

In the followup article it's revealed that all his degrees are in education: http://www.washingtonpost.com/blogs/answer-sheet/post/reveal...

That does not bode well for him having a lick of intelligence

So, if a smart guy gets elected to the school board, he becomes stupid? Does he have to quit his job since he's too dumb to do it now?

Did you read the article? Sounded like a pretty smart guy. I've served in local office, there are plenty of idiots on school boards but there are smart people too.