The SAT, Test Prep, Income and Race

The announcement of the new, new SAT has created a lot of hand-wringing about SAT scores and their correlations with income and also race. Wonkblog, the New York Times and many others all feature a table or chart showing how SAT scores increase with income. Wonkblog says these charts “show how the SAT knowledgeorincomefavors rich, educated families,” and the NYTimes says about the SAT, “A Test of Knowledge or Income?” The consensus explanation for these “shocking” results is the evil of test prep as summarized by NBCNews:

…there is also mounting criticism as to whether students who can afford expensive SAT test preparation courses have an unfair advantage, especially given a strong correlation between family income level and test results.

Similarly, Chris Hayes blames test prep for inequality:

We’ve had…the growth of this tremendous testing and test prep industry in New York, along with the massive rise in inequality and it has produced a system in which the school is now admitting only three, four, five black and Latino students. The students they are admitting are almost entirely white, affluent kids with tutors or second generation, first generation immigrants from Queens and other places where the parents pay for test prep. You end up with a system where who you are really letting in are the kids with access to test prep, the kids with access to resources.

All of this is almost entirely at variance with three facts, all of which are well known among education researchers.

First, test prep has only a modest effect on test scores, on the order of 20-40 points combined for a commercial test preparation service. More expensive services such as a private tutor are towards the high of this range, cheaper sources such as a high-school course towards the lower. Buchmann et al., for example, estimate that private tutors increase scores by 37 points while a high school course increases scores by 26 points.

The average SAT score among those with a family income of $20,000-$40,000 is 1402 while the average score among those with an income $100,000 higher, $120,000-$140,000, is 1581 for a 179 point difference. Even if every rich family had a private tutor and none of the poor families had any test prep whatsoever, test prep would explain only 20% of the difference 37/179. If rich families rely on tutors and poor families rely on high school courses, the difference in test prep would explain only 6% (11/179) of the difference in score.

The second surprising fact about test prep is that it doesn’t vary nearly as much by income as people imagine. In fact, some studies find no effect of income on test prep use while others find a positive but modest effect. The latter study finding (what I call) a modest effect finds that in their sample a 2-standard deviation increase in income above the mean increases the probability of using a private test prep course less than whether “Parent encouraged student to prep for SAT (yes or no).”

Since test prep differs by income only modestly and since test prep increases scores only modestly, the effect of income on test scores through test prep is small, Modest*Modest=Small. Contrary to the consensus, test prep can in no way account for the large differences in SAT score by income.

The third fact is that test prep varies by race in the opposite way that people imagine. In the quote above, Chris Hayes suggests that whites use test prep much more than blacks. In fact, blacks use test prep more than whites, as is well documented among education researchers (e.g. here, here, here), e.g. from the first link:

…blacks and Hispanics are more likely than whites from comparable backgrounds to utilize test preparation. The black-white gap is especially pronounced in the use of high school courses, private courses and private tutors.

Indeed, since blacks use test prep more than whites and blacks have lower SAT scores than whites the effect of test prep is to reduce not increase the black-white gap in scores. Of course, the net reduction in the gap is small.


AlexT, race baiting, class warfare baiting LOL go Alex go! Speaking as a 1-percenter.

But seriously, if the stat about test prepping only adds 40 points max and if the 179 point difference from rich vs non-rich is true, then nutrition is perhaps the key to explaining why the rich are smarter. Unless you buy into the Clark thesis of the 1-percent somehow being the cream of the crop (I like to think so, but that's just conceit on my part!)

Or it could be that smart people earn more money because they are smart. I don't know about you, but I've never met an unintelligent doctor or engineer.

Since intelligence is almost entirely hereditary, smart people pass on their intelligence to their children. The fact that those smart parents are rich is immaterial, as the adoption studies have shown.

It's not that it's entirely hereditary, but that its large hereditary component often coincides with an amplifying environmental component. Thus clever people tend to have clever children, and to bring them up to be clever too.

Anyone who is familiar with large families knows of cases with two clever adults producing three clever offspring, and one rather ordinary - the genetic lottery explains that far better than complaints about environment or test prep.

Gene * Environment interaction likely has little effect, since narrow-sense heritability already gets you to ~50% (if not more) of explained variance, as demonstrated by heaps of twin adoption studies, and even genome wide association studies, e.g. Davies et al.

This leaves only another 50%, max, for environmental effects per se, measurement error, and random noise.

Nod to Jayman; I thought about this a bit more after he called me out making a similar point as you did.

50%, max, for gene * enviromental interaction alongside environmental effects and all those things, that is.

50% is a pretty big portion, isn't it?

Not really, when the 50% is a ceiling (could easily be as low as ~35%, as well) and when you also have to subtract out a lot of variance due to other terms to get to variance due to gene * environment.

35% is a pretty big portion, isn't it?

I have two idiotic, hard working doctors in my family.

I would bet all the money I have that they are above average in intelligence. You may consider them 'idiotic' compared to yourself or whoever else, but I can guarantee that they have IQs of 115+.

Tabarrok presented facts. You are the one who is baiting and who ought to get lost.

Your beef should be with Chris Hayes, whoever he is. Race doesn't exist, until it is useful to him.

Seriously, the Chris Hayes' of this world enable the HBDers of this world.

My intuition says that it's all about how much kids read for pleasure, and the difficulty of what they read. I suppose you could look at race/income effects on the math parts vs. verbal parts of the SAT to see if that holds up.

But what determines what they read?

I expect that is mostly cultural. My parents both read constantly, so it was a natural habit for me to pick up. If your parents watch TV all the time, you are probably going to grow up doing the same. It's not going to be easy to separate that habit from income/culture.

Achievement tests like the NAEP tend to suggest that better math education leads to higher scores, while reading scores are more recalcitrant. Few kids do math for fun (MR commenters no doubt excepted!), but lots of kids read for fun, so emphasizing math more at school has more impact on scores than emphasizing reading more at school.

Proving once more that telling blatant lies is fine so long as the lies support the left wing biases of the NYT et al. Were a right wing news outlet to publish such blatant, self-serving untruths it would be held up as an example of the ignorance and stupidity of conservatives. Instead it is just considered a statement of what educated opinion knows to be true.

You mean us !%ers, so known for our intelligence, misspent our money on test preparation?

Let's go to the FTC and argue that they engaged in false and misleading advertising, and hire Alex as our expert witness.

I'm not sure it's a waste of money. A few points could be the difference between, say, Princeton and the University of Illinois.

It might be cheaper to get your recruiter laid.

Princeton can use a guy like Joel.

Well done. (golf clap)

AND JUST WHAT IS WRONG WITH U OF I?!?!!?!!!!!!!!!!!!!!!!!!!!!!!!!!

SAT prep was debunked more than a decade ago. Prior to that, though, the test prep outfits were seriously claiming that they could increase your score by 100-200 points.

Debunked seems a harsh word. Granted, they were guilty of overpromising, but the 20-40 point gain is still an actual gain, though obviously fairly minor, and probably not worth the cost in most cases.

Income is a proxy for something else.

Why don't we have numbers based on the education level of the parents. Which of course is another proxy, but still...

Seems like everything will be a proxy for everything else, unless a person believes there's one ur-factor from which all other achievements flow (the strong form of the nature over nurture hypothesis - which would probably eschew all but the most cursory attempts at studying).

SAT scores themselves are only intended to be a proxy for intelligence, diligence, and preparedness for education.

There is a reason that the company formerly known as the Washington Post Co. sold its least profitable major holding, while retaining the following - 'The Graham family, which had controlled the newspaper for eight decades, retained ownership of the Kaplan education company; the television stations WDIV-Detroit (NBC), KPRC-Houston (NBC), WPLG-Miami (ABC), WKMG-Orlando (CBS), KSAT-San Antonio (ABC) and WJXT-Jacksonville (independent); Cable ONE, a Phoenix-based company that provides digital video, Internet and phone service to homes and businesses in 19 midwestern, western and southern states; the Slate Group, which includes Slate, Slate V, The Root and Foreign Policy; Social Code, a social media marketing agency; Celtic Healthcare, a provider of home health-care and hospice services; and Forney Corporation, a global supplier of products and systems that control and monitor combustion processes in electric utility and industrial applications.'

Since nobody is ignorant at Marginal Revolution - I've posted Kaplan's earnings several times, so I'm sure everyone remembers them - there is no reason to highlight what makes Graham Holdings particularly profitable. Especially as test prep no longer needs to subsidize anything resembling real journalism.

But for the atypical reader who knows nothing about what kept the Washington Post operating for over years, here is a wikipedia link -,_Inc.

'First, test prep has only a modest effect on test scores, on the order of 20-40 points combined for a commercial test preparation service.'

Well, John Katzman would disagree with this entire post, but especially that point -

'ETS says it's not a coachable test...

ETS has refused easily 50 times to do an independent study that we both get involved in, because it would be so easy, right? We say, here's a list of our kids. Here's exactly when they took the course from and to. Go put it in a computer. ETS, you supply all the numbers. And let's look at the prior and the post and we'll know. Right? We know that as a country, kids who start where my kids start--which is around 1100 ah, better than the average--ah, go up about 20 points. And these kids go up this many points. And the difference is us. Right? It's an easy study. And they don't want to do it. The only time they do a study is by themselves, not asking for any help, not showing anybody their methodology or showing anybody the raw data.

Why won't they participate?

Let's say that they acknowledged, after 50 years of lying, that the SAT was very coachable. Number one, they've got to account for 50 years of lying. And number two, they've got to deal with, with the implications of successful coaching to race and gender bias issues. Okay. Rich kids all coach. Poor kids don't. There's a widening gap between rich kids and poor kids. We can no longer say it's because of the schools or because of the unfairness of society. Now we have to admit that we're part of the problem. There's a gap between men and women that's substantial--that they can just sort of laugh off. And you'll watch as you interview them and they do a dance. Why do men out perform women on the SAT? The SAT's supposed to predict college grades. Women do better in high school and they do better in college. What's the problem here? Ah, the more you use, the more you start accepting that the SAT's coachable, the more problems you have with it.' (The link has no date, unfortunately - though it reads much like things being written two decades ago.)

But what the founder of the Princeton Review know - well apart from what he details here, for example. And how to get rich from making a mockery of the SAT, of course.

Uh, so Alex says that the effect of coaching is 20-40 points, and you counter this by citing Katzman who says that the boost is 20 points. You aren't very good at this, are you?

Katzman is full of shit in that interview. I don't think he believes half of what he says.

+1 Hilarious retort.

And the simple version, just for you - if you take the SATs twice, a 20 point increase is expected between the first and second test taking. Compare the initial score with score of the second test - and see whether those using Katzman's methods of gaming the SAT have considerably higher scores on the second test than those who don't.

Katzman is more than willing to have empirical data show the effectiveness of his method, using a baseline score to compare with the score after using Princeton Review's methods - for some reason, ETS isn't.

The Katzman interview is decades old, I think. ETS has become much more open since then. There are many studies on the effectiveness of SAT prep, some of them mentioned by Tabarrok.

I enjoy the level of reading comprehension one finds here.

What Katzman said was that if you take the test twice, the second test results normally show an increase of 20 points. (At my high school, that was considered so elementary that you were always told to take the PSAT, and then the SATs twice). There is no way to see the empirical results of a test prep course without first taking having a score that does not involve test prep - Katzman has no problem comparing before and after scores of students using his methods, after subtracting the expected 20 point bump by taking the test twice. Strangely, the ETS does have a problem with such a simple empirical study.

Katzman claims that his student do considerably better than the simple 20 point increase expected by taking the SATs twice - and yet, ETS refuses to provide an opportunity to allow Katzman to support his claim that the SAT 'was very coachable'. And since I'm confident that Princeton Review has fairly reliable data on its own students, the reason for ETS's lack of interest in showing Katzman wrong is because Katzman isn't.

ETS still hasn't been able to respond to his other point either - 'Why do men out perform women on the SAT? The SAT's supposed to predict college grades. Women do better in high school and they do better in college. What's the problem here?'

ETS still hasn’t been able to respond to his other point either – ‘Why do men out perform women on the SAT? The SAT’s supposed to predict college grades. Women do better in high school and they do better in college. What’s the problem here?’ -

I think this can be explained by math and critical reading.

Men typically do better on the math section and women typically do better in the critical reading

4,760 Females got perfect scores in reading compared to 4,102 males

5,819 males got perfect scores in mathmatics
2,238 females got perfect scores in mathmatics

The gap between male and female in math produces such a large gap that the female advantage in critical reading is not able to cover.

As a result men have higher average SAT scores.

Math SAT scores probably correlate to success in STEM subjects.

Reading SAT scores probably correlate to success in humanities and social sciences.

More students go into social sciences and humanities than STEM.

Since women do better in critical reading they do better in humanities and social sciences thus have higher GPA's.

While for men scoring high in math is not very helpful for humanities and social sciences thus have lower GPA's.

Also in highschool majority of your classes will be humanities classes.

For most highschools you typically take 4 years of english, 4 years of social studies (aka history), 4 years of gym, 3 years of foreign language, 3 years of math, 2 years of science, 1 year of art.

There's also the fact that, controlling for student quality, grades are higher in humanities classes. I'd like to be able to just assume that people challenging the grade-predicting ability of the SAT are taking class difficulty into account, but I don't trust them that much.

Just because rich people spend tons of money on SAT prep doesn't mean SAT prep is effective.

Honestly if you have the money it is foolish to not try and get those 40 extra points.

Of course.

To be clear, I trust ETS and Kaplan each about as far as I can throw them. They are each selling very very hard to a certain audience, not to the truth.

I think the higher rate of test prep usage by blacks and Hispanics is related to the alleged 40 point increase. Students toward the low end of the curve would have more upside potential, hence would see a greater effect. I'm deeply skeptical that bright, well-educated students would see anything like a 40 point boost. I'd like to know what the effect of a cup of black coffee would be.

Studying for SAT can help your score by a lot, but the prep, in my opinion was useless if you were going for a perfect score like I was.

But Alex just showed that there is little link between income and test prep use...

Binning the data in that way produces a nice correlation of ~1 between SAT scores and income, but the individual-level correlation is only 0.4 (or less), i.e., something other than income explains at least 84 percent of the variance in the scores.

If you really want to talk about race, income, and SAT scores, here's something to think about: whites from families earning less than $10k per year outscore blacks from families earning over $100k per year on the SAT. One can only assume this is because of the overwhelming and oppressive privilege of those damn white kids.

Less than $10k? I think it's called trailer park privilege.

"The late John Ogbu, professor of anthropology at Berkeley, believed that broad cultural attributes among blacks — such as parental style, commitment to learning, and work ethic — bear a heavy responsibility for the black-white educational gap. Ogbu wrote in his recent book, Black American Students in an Affluent Suburb: A Study of Academic Disengagement, that black students in the affluent homes of doctors and lawyers are looking at rappers in ghettos as their role models. Students talk the talk about what it takes to be a good student, Ogbu wrote, but few put forth the effort required to get good grades. This type of behavior is typical, Ogbu said, of racial minorities adapting to oppression and the lack of opportunity. Ogbu, much as Bill Cosby has done recently, also placed the blame on black parents. He believed that many black parents are not offering sufficient guidance, do not spend enough time helping with homework, and do not pay adequate attention to their children's educational progress."

An alternative explanation is that because of regression toward the mean, the children of black professionals have IQs far below the children of white professionals -- even if black professional were as smart as white ones on the average (they aren't).

Does this explain the tendency of black professionals (especially the men) to marry white? Their children would have improved IQ and in theory more devoted parenting.

No. I would say that racial differences in masculinity/femininity are the main explanation of the patterns in interracial marriage, not IQ.

There is another possibility. David Henderson at EconLog talked about it

Short form: question about train schedules on standardized test. Poor inner-city kids ace it, rich suburb kids bomb it. So . . . the question was thrown out.

I'm skeptical of "SAT is racist" stories, but the above is something to keep in mind. Any errors in one direction are okay, errors in the other direction are corrected.

P, link actually an argument against your stance of disagreement.

The validity of a new test item is partially determined by seeing if those with high scores on the rest of the test were more likely to answer it correctly than those with low scores. Can you see the problem? Right, only questions that confirm the assumptions that were used to select the original question will be deemed valid. This is prima facie problematic when a new test item appears to have "low validity" but undeniably requires reasoning or semantic knowledge of some sort, such as the bus schedule question above or a hypothetical question asking for the precise definition of "sizzurp" or "crunk".

You can argue this is fine and dandy because the test has traditionally predicted GPA or competence quite well and the addition of the new questions, even if measuring some previously unaccounted for aspect of reasoning or vocabulary, would diminish its utility as a predictor, and I wouldn't disagree with you, but there is evidence to support a more nuanced interpretation of what these tests are measuring.

I'm very partial to the theory that people are more likely to discount achievement and success from out-groups and to place a false premium on in-group achievement. But it's also a pretty standard testing meta-evaluation that if the lowest-scorers do better on a certain question than the highest-scorers and middling-scorers, then ceteris paribus it's likely that question was poorly constructed.

Of course, one alternate explanation is that if the majority of a test has certain cultural biases and assumptions, including dialect bias, then a question with limited or reverse bias might look poorly formulate on paper. This seems the likely explanation for David's linked anecdote. It's why my "ceteris paribus" is a huge loophole.

It's impossible to construct a test of intelligence without making some assumptions - about what forms of intelligence matter (memory, spatial, linguistic, mathematical), whether it's better to be fast or right (test length and timing?), and about the language used to administer the test (even pure mathematical symbols are less than universal).

Here's a huge set of graphs on SAT scores over time:

The latest release of scores by race by income from ETS was apparently 2003, and that's no longer true. Blacks from six figure families now outscored whites from families up through around $25k.

Nice work in the comments section to rebut the racism:
Marmalade said...

"The graphs for income or education with race reach the provocative result that race affects scores more among the lower rungs of society."

That would be evidence against a genetic explanation.

Genetics only shows clear and strong influence at the higher rungs of society. This is because it is only after environmental problems are decreased that the genetic impact can be seen.

For the lower rungs of society, environmental problems are still the main influence on IQ. For poor blacks, you have the typical environmental problems of poverty along with the environmental problems of structural racism.

So, going by that conclusion you came to, it supports the data on continuing racial prejudice and bias.
October 28, 2013 at 1:32 AM

Also note that the master race for math is...the Asian. From Steve Sailer's link, top of the page.

Thanks Captain Obvious.

You really show your massive intellect every time you post, Ray.

No, as long as the environments are similar a genetic component can be determined. Similarly damaged, similarly enriched, it doesn't matter. They only need to be similar. The extent of the genetic component can change across environment qualities, and we may be more interested in some environments than others, and it may be more (vastly) difficult to ensure a similar set of disadvantages than advantages, but the actual words you just used imply something false.

There is a table of mean SAT scores by race (B&W) and income for 2009 here:

How do poor Asians perform?

Income correlates with IQ, parent's IQ correlates with children's IQ, the SAT is mostly a type of IQ test, so of course income correlates with SAT scores.

Different races have different average IQ.

If you just accept these facts instead of committing incredible feats of mental gymnastics to ignore they are true, all of our social "problems" fall right into place. People are animals and those best adapted to their environments perform the best, in our modern technological environment intelligence is an extremely highly valued trait, so much so that any minor differences can mean major differences in outcome. We could all stop pretending that education is going to be a panacea, because education doesn't change people's inborn abilities, just like going to the gym doesn't change a persons height or shoe size.

Finally some logic. Just apply occums razor to the situation. It's amazing the hoops people jump through to make themselves feel better. Why do many nba stars also have athletic children? Why aren't there more white basketball players in the nba proportional to population? It's easier than it seems guys.

The income gap in test scores has been narrowing, not widening:

It does amaze me that, with such clear public data available to everyone online, that the basic facts of racial differences in intelligence aren't widely known among the public. In the past, people had the excuse of never coming across the data, but now, it's just a click away. Though whether that's due to a lack of curiosity or a fear of knowing forbidden truths I can't say.

It's because they already know that their can't possibly be racial differences because that's what they've been told. Everyone is equally capable (except Cindy in my office, she's a moron) and if you don't believe that then you are a slip slide away from becoming Hitler.

Over the last 30 years the educational industrial complex has made it politically correct to believe that everyone can be educated to a high enough level to live very well. Problem! Average "is" over and spending big bucks on academic education for everyone didn't work! I like many college administrators lived a very easy fairly high income life.

I did read somewhere that the best correlation with high income is high SAT scores. Much better than education. Anyone ever see anything like this?

SAT test is mostly an IQ test so yes. The Bell Curve by Charles Murray is the best starting location.

asians are the smartest, but because they are smarter than whites and blacks they get oppressed by the leftist white/black liberals. Asians actually have to score 50-100 points higher on SAT just to breakeven with whites. Its appalling how race influences yur chances at a good education. Just because whites and blacks are dumber, they get supported by the government.

Asians are not smarter, they just are better prepared. Asians practice for SAT and ACT tests for years.

People generally don't do a lot of their own research. Generally they learn something new when it is taught to them. The teachers of the world don't make a concerted effort to teach people about racial differences. It is fairly easy to tell how intelligent people are, at least to a degree that outweighs racial predisposition, so why would people need to go around telling everyone that on average some races are smarter then others? The averages aren't really important. The people who really get excited about those differences usually want to push stupid, racist policy agendas instead of agendas that discriminate on the thing that some race is apparently lacking to begin with. So yeah, people generally don't know about those racial differences. Is that really a problem?

It is when you start making government policy based on the mistaken belief that the races are equal, to try to make the races equal based on the belief that they are equal, that it's just racial discrimination keeping them down. Racial discrimination is way overblown, because people deny that race exists, but it's a great cudgel to get votes and push policies liberals want. If they could just leave it alone then it wouldn't be a problem.

What is race a proxy for?

"Asians actually have to score 50-100 points higher on SAT just to breakeven with whites."

Not quite true. Go back and look at the work by Unz. Asians are the second worst group. The people with the fewest predicted admissions to elite schools based upon test scores are whites who are Protestant and Catholic. He didnt break it down, but I would bet that being male is also now a negative.


In the United States, race is sometimes a proxy for culture.

There's alot of things that we could know that's one-click away, but we don't. Curiosity, opportunity cost, stupidity, psychological barriers, and luck all come into play. How much of what we know do we really research?

For example, Terra can't spell the most common one letter word ("a"). How much time would it have taken her to figure out the correct spelling?

More time than it'd take to use "one" substitute?

Shorter Alex: correlation = causation

And you wonder why no one respects you!

No no, anyone with a tiny amount of sense must stay away from simply saying correlation does not cause causation. It is the most irritating and half baked argument possible, people who say it can't fathom that the entire base of personal and human knowledge is solely based on correlations.

On the contrary, it's an entirely sensible thing to say.

And the notion that the entire base of personal and human knowledge being based solely on correlations is incredibly irritating and half-baked. People are far more sensible than that.

I see what you did there. I'll drop the rhetoric.

It is not a sensible thing to say. It is a trite and stupid thing to say. Correlation often implies a relationship, especially when combined with some reasoning, either a mutual cause or a causative relationship. The way people use it, it's a carte blanche to throw out any relation they don't like, which is exactly what Benny Lava was suggesting Alex do. Anyone who says this would be better served explaining their reasoning for why the things are unrelated or why it's a mutual cause and what the mutual cause is.

The other half of the problem is that sooooo many social scientist (stop laughing) draw outlandishly wrong conclusions from paper thin correlations. The abuse of statistical correlations has caused the overuse of the phrase in question ;-)

You were earlier making the incredibly ambitious claim of "the entire base of personal and human knowledge is solely based on correlations". I replied, hoping you'd make some attempt to defend this claim. I'm badly disappointed. Firstly:
A) A relationship is not the same as causation.
B) You indeed have to combine correlation with some reasoning to get to causation. For example, is wind caused by trees waving around?

And it remains an entirely sensible thing to say that correlation does not cause causation. It may be that some people use it to throw out any relation they don't like, but people misuse any statement, no matter how sensible that statement is. Criticise the misapplication of the argument, not the argument itself. Personally I criticised Benny Lava below for totally misrepresenting Alex's argument.

Anyone who says this would be better served explaining their reasoning for why the things are unrelated or why it’s a mutual cause and what the mutual cause is.

On the contrary, that's the job of the person proposing the causal relationship in the first place.

To explore more generally the claim that "the entire base of personal and human knowledge is solely based on correlations".

For a start, where does mathematics come in to this? I know, for example, that the square root of two is irrational, and there is no largest prime. How could anyone possibly know this solely based on correlations?

Another problem with this claim, when the first European explorers reached Australia and NZ they left accounts of their first observations. Among these they included descriptions of things they recognised as people (Maori, Australian Aboriginals), and new species of trees, birds, etc. This makes sense if their knowledge of such things as people, trees, birds, is based in part on logical categories, not if it's based solely on correlations. (Note, the explorers may have been wrong on some classifications by modern biological measurements, but they still did classify in the first place).

Where do transgender people fit in, if all our knowledge is based solely on correlations? How could someone have first formulated the thought "I'll be much happier if I undergo dangerous and painful experimental surgery to change my appearance?" based solely on correlations?

Where do Newton's Laws come from, if all our knowledge is based solely on correlations? Why was the jet engine for planes ever built? You can't get to a jet engine by incremental improvements on propellers.

the phrase has become a favorite rhetoric crutch among the armchair experts

Because it's a good one, in and of itself. Cliches become cliches for a reason.

"correlation does not cause causation": eh?

Huh? Did you read the article? Alex is arguing against the idea that correlation is causation, noting that a) test prep doesn't increase scores that much, and b) the poor are also quite likely to use test prep courses.

Kids' SAT scores can't cause higher parental income, since the latter precedes the former. Furthermore, parents determine their children's upbringing, education, and genetics to a great degree; in fact, parents literally caused their children. It seems blindingly obvious that some factor related to parental income is causing higher SAT scores (note: no one is claiming that the income itself is directly causing the scores). If you want to argue that it is not a causal relationship, then go ahead and make an argument.

"no one is claiming that the income itself is directly causing the scores": no one? You are being too kind about the dim-wittedness of the arguments often used about this topic.

How would the actual income itself cause the scores to go up? Maybe the kid gets an extra confidence boost by thinking "My parents make loads of money! Yes!!!" Or maybe the College Board steals income tax returns and allocates scores based on that. Still, I don't think anyone is making such arguments. If anyone is, please pass along the link, because it must make for entertaining reading.

Income can easily affect test scores with the use of test prep being only a small part of that correlation. It's just that the other mechanisms by which income affects test scores probably run through making kids genuinely smarter/more conscientious, both of which are valuable skills in the workforce.

For that matter the ability to prepare for and take tests is becoming an increasingly valuable skill.

With all due respect, you're knocking down a strawman argument here. The people you cited never said test prep is the conduit through which income disparity ends up affecting SAT scores (and hence college admissions). Hayes particularly wasn't even talking about SAT prep, but prepping for admission into elite NYC high schools. And Hayes has noted the prevalence of that "cramming" to get into Stuyvesant or Bronx Science is prevalent by Asian students, with expected effects, regardless of income. Look this is a complicated issue that goes to the heart of how much America is a meritocracy or not, and where "resources to pay for SAT tutoring" is only one factor. Setting up people to make it sound like they're stupidly advocating a view that they're not really doesn't move the debate forward.

So your point is they are not talking about SAT prep but rather standardized test prep?

Among the New York Times' mostly white subscribers, there is a lot of resentment of Asian Tiger Mothers, but that has to get translated into an attack on the privileged white male one percent etc etc

What are the odds Chris Hayes knows what E[X|Y] means?

"What are the odds Chris Hayes knows what E[X|Y] means?"

Better to use a technical name (conditional probability).

Many readers will, perhaps, not know what conditional expectation value is. Using the name gives them a chance to look it up.

Some readers will, perhaps, have done Ph. D. work on Monte Carlo simulation in a field in which different notation is traditionally used. They can easily understand what conditional expectation value means but still have to make an (admittedly not terribly difficult) guess what your notation means.

(This is related to a pet peeve of mine: instructors of ordinary nonspecialist courses who choose to test their victims on a mix of nonstandard notation and arcana that happen to be their personal priority, so that skilled practitioners in the field would have trouble with the test unless they had already become familiar with the instructor's personal worldview.)

Income could also affect SAT scores through better education throughout their lives. This is the real problem. Even if SAT tutoring works that doesn't mean that kids with high SAT scores shouldn't get in to the best colleges, as long as a tutored student is likely to do just as well in college as an untutored student with the same score. The problem isn't that rich kids aren't smarter but are getting into the best schools anyway, the problem is that rich kids are smarter do to a lifetime of opportunities to learn that poor kids don't have. The problem is neuroplasticity. Of course its more of an opportunity than a problem, but as long as we don't take the opportunity by increasing resources for the education of poor people, poor kids will remain dumb (on average by comparison).

but a poor kid today has much more opportunity than a poor kid many generation ago. Second to Switzerland, the USA spends more $ on education per pupil than any developed country in the world

It's quite clear, though, that schools in poorer areas are much worse--even when a great deal of money is spent.

Is it? I can think of a fair few anecdotal reports of lousy schools in rich areas.

Correlation doesn't imply consistency.

That's the catch of self-government. Also see: foreign aid.

On that note, should we penalize the wealthier family who chooses to invest their resources on educating their young? I mean, a tutor is just another way of improving a kid's overall level of education. Even teaching to the test has some benefit that accrues. Seems like something we'd want to encourage.

Equivalently, should we punish the poorer family who didn't have resources to invest in educating their young? "Your parents are poor, so you get fewer opportunities" isn't exactly the American Dream.

I'm not saying there's an obvious fix, but I think folks view the SAT as something of an equalizer:

"Sure, that guy's parents sent him to all those expensive, elite summer programs while I had to work at WinnDixie, but the SAT is a meritocracy."

When that idea is threatened, you get folks feeling like the system is rigged against them all the way up. And 20-40 points may sound trivial (it's not), but that's an average effect. For individuals, you're going to get wider swings. There's also the issue of improving scores over multiple takes. I mean, just by taking it more than once, you're score will improve even if you don't get any better at taking the test. (SAT reports your best scores from each section regardless of how many times you take it.) These are all things that are available to some folks but aren't available to others.

"the American Dream": -1.

The original purpose of the SAT was to find kids who were bright but poorly educated by testing their aptitude in a standardized way (hence the name).

As time has gone on, it was noticed that not everybody's aptitude was the same. And das racis.

All changes that have taken place since the SAT's inception have been towards more knowledge testing questions and away from intellect testing questions.

Knowledge testing questions are of course easier to teach, tudor and test prep for. The SAT has become less and less meritocratic with time.

"the problem is that rich kids are smarter do to a lifetime of opportunities to learn that poor kids don’t have"
Such as the opportunity to be the living breathing result of a law partner ejaculating into a management consultant.

This reminds me of the old bit about the Times claiming, "Despite increases in prison population, crime continues to decline." They struggle with cause and effect over there. As others have pointed out, the differences in income and SAT scores most likely have the same cause, given that test prep is not unique to one group and the impact is small. They ignore the fact test scores have been narrowing for a while now.

All of these points are filtered out by these people because they conflict with their list of axioms. For the cult, it is axiomatic that racism and income are the causes of education outcomes. Therefore all evidence that contradicts these truths is dismissed as invalid. They are no different than creationists when it comes to science.

As an aside, is anyone else looking forward to the war on testing that is about to boil over with the new SAT?

No surprise the left wants to make the SAT less like an IQ/reasoning test and more like a rote memorization/curriculum test. That way low scores among certain groups can be presented as a societal problem that can only be ameliorated with more useless funding instead of biological one, where no amount of $ will matter. Also, test prep may work for the math section but less effective for the verbal part. Prestigious schools will continue to use the SAT because it's better than all the alternatives.

Our daughter used a book she got from from Amazon and raised her total 100 points on the second test. They are CHEAP.

Key: You have to spend lots of hours mastering the different types of problems that are presented. She also used a Physics subject test prep book (she was just starting Physics in her senior year) and had a respectable score in the 600's.

She worked hard. Grit is the answer.

I agree with the idea about mastering the different types of problems that are presented. The only prep I ever did for the SAT and GMAT was using books like that (I only took each once, so I was an uncontrolled experiment) and I'm convinced that just not freaking out during the test is easily worth the 20-40 points they claim from test prep. The math isn't really high level stuff for most high schoolers and the verbal section is based more on a lifetime of reading/writing than on any vocabulary cramming one can do, but I have friends who experienced the test as enormously stressful and this caused issues like "freezing up" that surely can't have helped their scores.

While I never did an SAT, I raised my test day GMAT score by at least a 100 points (to 760/99%, which is perhaps slightly above my university and bschool GPA percentiles) vs. the early mock tests I did (only using books, not tutoring). In fairness, I always have tested well on test days.

Still, the test is VERY much teachable simply because there is a whole bunch of rather arcane stuff in there (data sufficiency anyone? oh and who ever used basic geometry on their job, I had not touched that stuff for almost ten years when I took the test) and after a while, you actually begin to see patterns. There is a certain way how the GMAC constructs the test and once you see that, about a third of the questions become easy and quickly solved leaving you more time for other questions...

I have a similar experience, I took my GMAT from 630 to 770 with a bunch of books. I was a bit lucky though because almost all my improvement was in the quantitative portion which I think is much easier to improve than the verbal one.

That was my initial impression, too.

But then I actually got more improvement out of the verbal piece (might help that I am not a native speaker).

I just realized that my GRE (the average percentile of both sections) is within 1% of my GPA. I'll spend the night meditating on that.

So a data point of one? Grit is definitely not the answer for a quasi-IQ test like the SAT. If it were, Asian scores would be a lot higher.

A gain as large as 100 points most likely strongly reflects regression toward the mean. She screwed up something when she did the test for the first time. Even without prepping, her score on the second test would probably have been higher.

A second data point -- my daughter raised her scores well over 100 points based on hard work and top-rated tutors. And as she steadily increased her scores, she started looking at higher level colleges, which kindled a desire for her to work (even) harder in high school.
Yes, it was expensive. Colleges are expensive, too. If you're going to spend 200K on college, it's worth spending 5K on SAT tutoring.

No, it isn't

It's not worth spending $200K on college or $5K on prep. Your kid is going to be who they're going to be - the college and high school can only add so much. Especially once your kid is out of college.

Another data point for what it's worth - I solely relied on the huge softcover study guide for the SAT and boosted my score 300+ points from the PSAT to the SAT. I went from low 1100s and 1200s to 1520. Those books are awesome - the $1000 prep course isn't any better than the book. This was back in the late 1990s.

Alex said:
"In fact, blacks use test prep more than whites, as is well documented among education researchers (e.g. here, here, here), e.g. from the first link:
…blacks and Hispanics are more likely than whites from comparable backgrounds to utilize test preparation."

Um, you know that black people and Hispanic people are, on average, from poorer backgrounds than white people, right? You might as well take data that shows "black people who signed up for Kaplan Test Prep use test prep more than whites" and use it to claim that "blacks use test prep more than whites."

At least as far as blacks go, your explanation does not hold. This is what it says in the article Alex quoted:

BCR demonstrated an advantage for blacks in test preparation, conditional on an array of background characteristics. The results in Table 1 show that this advantage exists even when we do not control for family and academic background. Blacks used test preparation more than all other groups: 84 percent of blacks used at least one form of preparation compared to only 68 percent of whites, and were also more likely than whites to be engaged in multiple test preparation activities. The black advantage is notable in regard to all forms of test preparation except in the use of self-preparation materials such as books, videos and software. The black- white gap, therefore, is mostly incurred by the utilization of private preparation (private courses and tutors). In that light the gap is even more fascinating because blacks are generally a more financially disadvantaged population than whites.

I didn't say anything about the reasoning in the articles, as I haven't had time to read them. What I criticized was the reasoning provided by Alex. If the articles provided information that supported his conclusion, he probably should have quoted that instead of the irrelevant information he did quote.

The first two links both show that the absolute rate of test prep use was higher for blacks than whites (the third isn't clear without access to the full text). But you would have had to look all the way to the first picture in the first article to see that.

In this century, average SAT and ACT scores by racial/ethnic group have been pretty stable ... except for Asians, who have gone up about 60 points (on the 1600 M+V scale):

The Emperor of China introduced civil service testing in 595 AD, and Tiger Mothers presumably started paying for test prep around 596 AD. I'm not sure that the old kind of test prep that has been studied the most is all that comparable to the new multi-year kind that Asian Americans parents pay for.

Has anyone studied how retaking the test alters scores? Students seem to take the SAT multiple times, in an effort to improve their score. I'm assuming this has the desired effect. Otherwise why bother? I'd be curious to know if simply retaking the test has a greater or lesser impact than test prep.

From the College Board:

Just glancing at the tables, it appears the biggest jump is between the first and second attempt. It appears to be greater than the uptick from test prep/ Maybe the right answer to prohibit whites and asians from taking the test a second time.

Isn't there some thing to be said about the quality of test prep. More money buys better test prep and better tutors.People who can afford go to to better test prep schools (where of course you get a better peer group to compete and practice with) and those who cannot go to the lousy ones. So the scores simply reflect that. No?


I hear of $6000 per year contract tutors for jr and sr year. Would market believers really believe that their marginal utility is 20-40 points?

Right, just like lots of golfers buy these energy-leveling wristbands with holograms in them, which must mean they work.

Heh, have you regretted going to "golf" yet? It is rife with productive investment, right? One does improve one's score with better clubs (to a point) and coaching (to a point). There are just declining returns (at some point). I don't believe anyone wins big tournaments with $99 Walmart clubs ("Complete Set with Cart Bag!")

Ah, there's a thought! SAT score-increasing wristbands with holograms on them.

It is possible, but without some evidence you are just making up a story to fit your prior beliefs.

A randomized trial is impossible. No one would risk their kid.

But I am arguing that "price" while not equal to "value" must be a clue. Surely parents would have figured it out if $X thousand were no better than a prep book.

(In my day (70's) and area (suburban LA) a prep book was all we had. Pre arms race.)

But prices are just a chaotic random walk, and 'value' is like the Easter Bunny. Or so I've heard.

Actually you heard that it was a three part mix of value, emotion, and attention. It *is* possible that these prices are part emotion, but I'm asking if they can be all emotion.

How exactly would they figure it out? They are lots of overpriced expensive things. Have people "figured out" that a Rolex is no better than a normal watch?

An interesting argument, but not really pro-market.

Expensive tutors are the new luxury good?

I mean... maybe?

Seems obvious to me! I knew a girl who was in a very prestigious PhD program at a top ranked school (since graduated). She was hired at like $60/hr to tutor some rich couples' middle schoolers. She freely admitted that she didn't create $60/hr of value, and I'm sure her middle schoolers would've been able to pass pre-algebra just fine with a tutor from Kansas State. But the parents' attitude was "nothing but the best" and a PhD student from that university was "the best" so there you have it. Anyway $60/hr may not have meant that much to them.

Well, another story I hear is that these kids get high prep, extended test times, do well on the SAT, to fall flat in college.

This might both reinforce and explain the "grades are a better predictor than SATs." Over-scoring, paid for by over-protective parents.

The SAT is primarily a test for college entrance, which for most people involves lots of opportunities to study. It's only natural that the SAT should test people post-studying, if it's to be a proxy for how well people will perform in college.

Frankly, I find the kind of comments (NYT and co.) linked from this post highly offensive to all of us "non-diverse" (understand white or Asian) immigrants. A lot of us come from non-privileged working class backgrounds (unless you count a stable functioning family as a privilege these days) and we really do not care about your white guilt because our forefathers were not enslaving African-Americans (and I do not believe in hereditary sin anyway). We just, you know, work to overcome the obstacles we are facing here and there are many: language, different cultural conventions, visa/greencard applications and the distance from family being some of them. We hear all this blabber about diversity but for us it means only working even harder to outmatch the positively discriminated groups (or should we say the privileged groups).

"unless you count a stable functioning family as a privilege these days": maybe so. Who knows what "privilege" means in the press? Wealthy? Likely to be subject to advantageous string-pulling, perhaps? I don't believe I've ever seen an attempt to explain what's meant. I'll except obsolete examples - the privileges of the landed nobility, that sort of thing.

Wow, this discussion turned racist really fast. And from there to "poor white kid with no NGO to look out for me". Way to go, guys.

I know, right! I mean, wow, just wow!

I am shocked, shocked, to find that argument-free outrage is going on here!

Will we next talk about he huge basketball and football coaching advantage black Americans much have that allow them to get so many of the most covet of all jobs college and professional basketball and football players. Yuppy engineers and programmers may protest that college basketball players are poorly compensated for their work but those jobs are still more highly pursued positions than any others, and when one gets one of those position one talks about it for the rest of his life it is so highly regarded.

Will we next talk about the huge basketball and football coaching advantage black Americans must have that allow them to get so many of the most covet of all jobs,l college and professional basketball and football player. Yuppy engineers and programmers may protest that college basketball players are poorly compensated for their work but those jobs are still more highly pursued positions than any others, and when one gets one of those position one talks about it for the rest of his life it is that highly regarded.

There is definitely a 'nurture' component which has been neglected here regarding parent involvement and encouraging academic success from an early age. Trying to play catch-up in the few scant months between the PSAT (when most parents start panic-buying as much extra help as they can) and the SAT is too little, too late. I would argue the lack of parental engagement at the lower end of the socioeconomic spectrum plays a much larger role in the disparate outcomes than any inherent difference in potential intelligence or cultural biases latent in the testing material. Parents must create a culture of success, an environment in which young minds can flourish.

So are you saying poor people are not good at parenting? Who is to say that children should not enjoy their early years more and study less. Who has the right mix of educational activities and just plain fun, I do not know.

Isn't the salient takeaway then that equality of opportunity is clearly not enough in a just and fair society?

What would be enough Steve? Complete government forced equality?

A very strong safety net (healthcare, minimum income, etc.) for those at the bottom. If enhanced opportunity won't be enough to help people who are doomed to the bottom due to their mistake of being born with the wrong genes and to the wrong parents, I would think those born more fortunately would see the need to help guarantee a good life nonetheless.

At what point do these people have responsibility for their own life?

And what if they continue to make more who have the "wrong genes" as you say? How will society function then? How much are they allowed to take before those of us with the "right genes" can not give any more? Right now we are subsidizing people with the wrong genes to continue to make more people with the wrong genes. We've so far avoided a Malthusian trap, but subsidizing the population growth of people who can't take care of themselves is a sure way to get there.

I think one reason "rich liberals" are suspicious of test prep is that they live in areas with massive test prep going on. They feel an arms race. They hire a tutor, but hear about one neighbor who lines up an aderal prescription, and another who has a psychologist write a letter asking for more SAT time.

In that environment it is very hard to believe that only 20-40 points are in the balance.

Serious? With a $140/hr tutor?

$150/hr in my neck of the woods. 90 minute sessions once/week for 10-14 weeks.

I would like the see the distribution of scores at every income level. I'm curious about the right tail, how many low-income kids hit near 1600 versus how many high-income kids do....I think that reporting mean test scores tells you almost nothing here

Maybe SAT test prep only has a modest effect, but what about the effect of the 12 years of schooling that preceded that? It's no secret that in America the rich get a much better education than the rest of us, whether through private schools or by living in the good public school districts.

I was shocked to discover the ability public schools in rich areas (CA) have to fund raise. We had school carnivals. They pull a couple hundred grand at a charity auction.

Ability yes, but necessity for sure. Arms race vs the private schools.

The mean and median cost per child at a private school is far below the median and median cost per child at a public school, so try again.

Depends on which kind of 'private schools' you mean. Parochial schools are pretty low cost, for example, and I can see where per-pupil costs there might be below public costs. But the hoity-toity private schools in big cities that the rich scramble to get their kids into? Yeah, they cost way more per pupil than public schools. That's just common sense.

If test prep is worth 40 points, and repeating the test is worth 40 points (dont know if they are cumulative, but would expect that to be at least partially the case), you really are talking about the difference between an elite or second tier school. Of course that just gets you into the pool. There are way more kids with high SAT scores that there are positions at elite schools. It is the extracurriculars that then matter. In that case, going to Peru to raise llamas for poor people counts way more than working part-time to help pay the bills while making your way through high school.


"There are way more kids with high SAT scores that there are positions at elite schools." Then the exams are too easy: make 'em harder.

"It is the extracurriculars that then matter. In that case, going to Peru to raise llamas for poor people counts way more than working part-time to help pay the bills while making your way through high school." An indefensible state of affairs: shame on the USA.

I had the same thought as commenter P. Knowing the mean score by income does not tell you what the "correlation" is. By definition correlation includes not just the regression line, but the variation around the line. All the graphs and tables I've been seeing in the popular press show that the correlation is positive, but they do not tell us anything about how *big* the correlation is.

So then you get a bunch of people who think they are sophisticated saying things like "college applicants should just send in their tax returns instead of their test scores to save time." But they ignore that the vast majority of variation in test scores is *not* explained by family income.

Rich people spend lots of money to send their kid to private school, or to live in a neighborhood with good schools. Their kids get relatively good education for 12 years, and then take a test and do relatively well.

Poor people can't afford private school for their kids, or a house in a neighborhood with good school. So, their kids get relatively poor education for 12 years, and then take a test and to relatively poorly.

@ AOD: While I am sympathetic to your environmental argument, you must in your model account for cultural factors, since rich white kids do worse than rich Asian kids, and rich black kids do worse than rich white kids.

So the cultural argument would be: Asian kids study harder than white kids study harder than black kids.

Maybe that cultural argument is correct. My Asian-American wife (who went to a prep school and took SAT prep courses) definitely worked harder than I (public school, no SAT prep course) did based on her recollection of her schooling.

Since Asians do far better than Whites and Blacks (btw all these groups are self-identifying as a 'race', since technically there's no such thing as race), we can conclude, if indeed prepping is only 40 points of more better SAT score, one of the following:

(1) either Asians are a master race when it comes to math (shout out to 陶哲軒, i.e. a "genetic" argument,


(2) "prepping" is not sufficient to raise your SAT score, and you need more intensive efforts, like studying hard for years on end, the way the Asians do.

It is clearly (2); I rest my case.

"since technically there’s no such thing as race": aw, bless.

Self-identification is a big part of "race".
Ex: in south american we find from hilarious to offensive the fact that you put creoles and mestizos under the common label of "latinos", which you use as a synonim for "hispanic", while for a Spanish-speaker the former is a cultural and linguistic label that includes portuguese, italians, etc... and the latter a different label which includes spaniards and spanish-speaking latin-americans, but not portoguese or brazilians (they are "lusos" or "lusophones")
So, what is race?

I agree, Asians have a greater genetic propensity to study hard.

How about ethny, demos, extended family, tribe, phyle, or common descent group?

Do you at least believe there's such a thing as sex? Then consider that gene tests can identify "race" as accurately as they can identify "sex".

Money has nothing to do with it. In high school, I spent many hours going through a couple of books with names like Ten Real SATs and Ten More Real SATs. Did no other test prep. Those books are available FREE at local libraries. I took the actual test only once and ended up with a perfect 800 on the math section and a 760 on verbal. And I'm not even that good at math! (I'm a girl, after all....)

"Case is not base", which is a cute way of saying one personal anecdote (yours) does not prove the base (average).

It could well be that you, Anon Girl, are in fact a girl Einstein. Speaking of which, for the Brownian motion paper by Albert it has been said his wife contributed a lot of material, and I think even his 1905 Special Theory, which is an extension of the earlier 1887 Lorentz transforms, was done first by others (namely Lorentz), if not his wife (speculative).

Still, an anecdote is worth so much more than "it has been said".

Much of the math was done first, yes. However, the physics, which is what he won fame for, was not (this is all well documented, there's a massive cottage industry writing scientific biographies about him, Feynman, a few others).

And none of it is related to the topic of the post. Anongirl is right, it's the hard work. You are right, there's a certain level of intrinsic intelligence. Now getting back on topic...

I am afraid Alex is missing the point here with all these "facts." This is all setting the stage for a retreat from objective criteria for admission towards a more subjective system. Since explicit racial quotas have been rejected legally, at least in part, what is needed is a subjective system so our betters can be sure the right folks get in.

More than a tad annoyed that, with all those links, you don't mention this essay, here:

Pretty sure that no one had dug up those studies until I did. And you're linking in two of the three studies I mentioned--even using the same construct (here, here, and here).

Total coincidence, I guess.

@Education Realist - Would love to hear what you think about the new SAT and the affirmative action stuff in California with SCA 5 and Asians opposing it.

For whatever it's worth, there was a study recently that found standardized test scores (ACT, SAT) aren't all that good of predictors of future collegiate success. High school GPA was the best indicator.

Anecdotally, that was definitely true for me. Did well but not spectacular in high school because of laziness, aced the ACT, got into a great undergrad engineering program and got my ass kicked for two years because I had essentially zero study and time-management skills.

Here's 5 more data points:

Male testing in mid 70's: moderate amount of solo prep, no classes/tutors: tested once: mid-1500's
Female testing in early 70's: moderate hours of solo prep, no classes/tutors: tested once: mid-1500's
Their offspring:
Male testing in mid 00's: no classes/tutors, minimal solo prep before 1st test (score 2180), about 30 hours before 2nd test (score 2340).
Female testing in late 00's: no classes/tutors, minimal solo prep before 1st test (score 21400), >100 hours before 2nd test (score 2380).
Female testing in early 10's: no classes/tutors, minimal solo prep before 1st test (score 2150), about 50 hours before 2nd test (score 2320).

The BEST way to prepare for the test is to study on your own from practice tests, analyzing your mistakes carefully. A private tutor is 2nd best. A class is hardly worth the time you spend in it.

Test scores are a proxy for a mix of IQ and determination. Elite colleges are for the most part perfectly happy with this. They don't really want the one without the other, so they don't admit based on test scores alone. Test scores + GPA tell you a lot for white kids. For Asians, high test score + high GPA might still be the result of more determination than IQ, so they get discounted. For African-American kids, black culture (somehow) apparently interferes with scores + GPA reflecting IQ + determination, so they get a boost.

Why do the wealthy (and the hard-up but scared) pay for tutors and classes? a) information asymetry & lack of discoverable real information about individual tutors'/classes' effectiveness; b) it's scary to trust your sheltered 17-year old baby this much, parents would rather believe that there's something they can do (==$$) to make it happen.

By the way, the current political struggle in Turkey in which prime minister Erdogan is being undermined by a corruption investigation by the shadowy Gulen cult that dominates the Turkish National Police is due to ... test prep!

The Gulenists dominate test prep in Turkey, and Erdogan ordered private test prep centers shut down last year, so the Gulenists raided his cronies' offices and bugged the PM's phones:

From a leaked Stratfor email based on a coversation with a Gulenist defector:

"The schools are central to the movement's efforts. They keep this list close hold.

"The source described for me how the Gulenist recruitment process works. In Turkey, preparing for your university exam is a huge deal (XXXX has explained this to me in a lot of detail). You essentially have to sign away a couple years of your life to study for this. The Gulen schools are known to provide the best preparation, offer the best resources, etc. A lot of students will attend them, regardless of their political/religious affiliation. Yout start out going a couple times a week, then 3 times a week, then every day of the week by the time you're done. They make it a gradual process and earn your trust.

"The movement will then take the brightest students from the class and will focus on them."

Maybe richer kids have access to a form of test prep over the course of their whole lives. Perhaps predominantly rich schools, (however unconsciously they maybe accomplishing this) groom kids for performing well on the SATs when compared with private schools. When I read this I assuming its a test prep you sit for sometime in highschool, and is for a certain defined relatively short period of time. Maybe this could be tested by comparing a rich kid who happens to go to a poor school vs a poor kid who by some luck of the draw was able to go to a rich school.

What is it about this post that brings out the humblebrag in people?

"the humblebrag": is that a Harry Potter character?

I never cracked a book but scored a 1600 hungover while also banging the proctor during lunch break. I didn't think it was a big deal.

I bet the proctor wasn't even that hot.

The raison d'etre for the SAT is to help sort students among schools. Does it do that well? I've read it correlates well with 1st year success, but not with graduation or subsequent income. I've also read (WAPO) that high school grades better predict success. If that's so, what good is SAT?

The correlation between high school GPA and first year college grades is indeed slighly higher than the correlation between the SAT and first year grades. However, if you combine HSGPA and the SAT, you'll get an even better predictor -- one is not redundant when the other is used.

The claim that the SAT does not correlate with success beyond the freshman year is false. The SAT can be regarded as an ordinary IQ test, and IQ correlates with college graduation, earnings, job performance, etc.

I love the complaint that the SAT favors educated families. Pure gold.

True, there are comedians that cuss a lot, but you got to earn that right before you do
that. Celebrity impersonations, musical gags, and the use
of props are just some of the things that you could do in your routine.

Well, it can easily be because the audience has been watching
her in the TV for the past 6 years, but nevertheless, her FACTS
are easy to remember and apply to everyone.

What does the SAT measure? Education quality? Inherent intelligence? Motivation?

I understand that we need some mechanism to discriminate between people for practical purposes of allocating training opportunities to people of varying skill levels and skill sets. But anyone who thinks that a standardized test can offer a broad representation of one's intellectual abilities has got to be kidding themselves if they want to make generalizable conclusions with recourse to perhaps just one or two or three other variables.

Come on. We're probably all pretty smart here. Is there one person on this board who went to a sub-par school, got no extra help, no access to supplementary learning materials at home, and still got an extremely high score on an SAT or IQ test? These tests are as much a measure of whether one's learning fits a white view of what intelligence is ... but what will they ever tell us about wit, or mirth, which so readily bubble to the surface throughout diverse ethnic neighbourhoods of the world.

"First, test prep has only a modest effect on test scores, on the order of 20-40 points combined for a commercial test preparation service"

From Color and money dot blogspot:

Briggs concludes that SAT test preparation increases scores on the math portion of the test by just 10 to 20 points, and on the verbal portion of the test by just 5 to 10 points. He does not base that conclusion on any new research, but on a review of a tall stack of past studies of the impact of test coaching. Actually, to be precise, he bases his conclusion on just three studies in the stack. He discounted the rest--some of which found score increases from coaching of 100 points or more--as based on small samples that were not representative of the nation's population or as otherwise methodologically flawed.

Two of the three studies that Derek C. Briggs characterizes as valid and pointing to "a consensus position" on the effects of SAT test coaching were performed by none other than Derek C. Briggs. (If his name is otherwise recognizable to people in the field, it is because, far from being a neutral arbiter of such research, he already has established himself as a prominent critic of the idea that SAT coaching works.) Briggs not only put himself in a position to pass judgment on his own research and (surprise surprise) declared his own work rock solid, but he also has declared a consensus based on me-myself-and-I vote counting. The third study that he counts toward that consensus, by Donald Rock and Donald Powers, unsurprisingly reaches the same conclusion he had.

About that only other study in the stack that Briggs found methodologically acceptable: It was sponsored by two organizations which are highly invested, financially and otherwise, in the idea that SAT scores cannot be raised significantly by coaching--the College Board, which owns the SAT, and the Educational Testing Service, which administers it. Powers was a principle research scientist at ETS, and, as the book Color and Money shows, both ETS and the College Board have a history of promoting research that makes their case and squelching research that doesn't.

the article said. As the only superpower in the west, the United States has long been the "begging target" for the Dalai clique. For its strategic needs, the U.S. has used the Dalai clique as a "chess piece" to contain China and is ready to embrace the clique, the article said.

I was wandering over the internet to get these informations. Thanks so much for such an informational post. It really helped! and along with this I would also like to recommend one other post on online sat tutoring to those who are seeking information. This site also helped me, I hope it might help others too.
If any one interested visit http://

Excellent willing synthetic eye to get fine detail and may foresee difficulties prior to they
will take place.

Comments for this post are closed