To some degree we conflate intelligence again with propensity to prepare, especially at the highest levels of scoring. Think of it this way. A 145 IQ kid spending 10 hours preparing will score higher than a 100 IQ kid who spends 10000 hours. Of course, everyone taking the test needs some familiarity with format etc. You can get this from a library book. Anyone who thinks low ses kids don’t know how to do this have never been low ses themselves or known low ses kids. The smart ones will then teach themselves what they need to know.
Moreover, if schools really wanted to address this preparation issue, the answer is to make the tests more like g- loaded aptitude tests, which are much harder to game.
In fact, all the standardized tests have been going the other way for decades. So, perhaps, preparation will matter a little more today than twenty years ago, but I am skeptical.
^ Same basic comment about parent educational attainment, especially when you consider that heritability of intelligence is roughly 0.8, which is about as high as you are going to get. Intelligent parents generally have intelligent kids who then score well.
All this is really a fascinating area to learn about. The recent textbook by Hunt, “Human Intelligence” is the gold standard imo, and the much maligned “The Bell Curve” is also very good especially for SES issues. Just ignore the literally 5% of the book that dealt with racial differences if that offends you. The science in that book is sound and recent evidence and data from the last twenty years have only strengthened its conclusions.
"More intelligent people will test higher, and on average they will be of higher SES, but we should be aware that SES does not “cause” the higher scores, nor does test preparation. Intelligence does.
Agree that it’s correlation and not causation but still I’m not sure whether SATs measure wealth or intelligence. They may measure both but you can’t say it’s a straight up IQ test.
As for CUNY, not lowly at all, my manager went there in the 60s I think and he said that there were 15 students who entered the electrical engineering program and one graduated - him!
^ There is an enormous amount of literature out there, check it out! SAT score is just about uncorrelated with SES, and it is certainly not measuring wealth. Sorry if this is insensitive, but this IS the affirmative action thread, after all. Black students from families earning over $160K per year score on average lower than white students from families earning less than $30K per year, with no discernable test preparation or coaching factors at play (in fact, the black group tends to be higher in test preparation than the white group at those income levels). See the research published by the Journal of Blacks in Higher Education here: http://www.jbhe.com/features/49_college_admissions-test.html. (The income levels have been updated since those mentioned in the article - I’ll try to find the cite, but I’ve used most recent numbers available.) Hard to square that with a wealth correlation.
It’s true that the SAT is increasingly becoming less an IQ test than it used to be, especially after the 2005 format change, but it is still remains highly correlated with IQ (and of course IQ is not some magic thing - it is just an attempt to quantify “intelligence” and has some issues, but it has been validated as a useful predictive tool for over 100 years now, and no one - I mean no one - who specializes in intelligence doubts this; besides, it’s the best thing we have).
My argument has always been that the SAT should return to a more g-loaded IQ measure, as this would remove much of the small preparation effects that are no doubt there, at least for people in the middle of the intelligence distribution. At these middle levels, there is enough motivation to engage in concerted preparation, but not enough intelligence to undertake it on one’s own. At higher intelligence levels this is unnecessary. At lower intelligence levels, the old adage comes to mind, “You can lead a horse to water, but you can’t make him drink.”
If you have specific research on SES and SAT correlations, I’d love to see it, especially peer-reviewed work.
CUNY was a fantastic place in the 1960s. Open admissions arrived by the early 1970s I believe and nearly destroyed the entire system. They have since moved to a more selective model I think within the last 15 years or so, but I’m not positive.
While older SATs attempted to be more like IQ tests, “scholastic aptitude” or IQ cannot be as easily separated from environmental effects as you seem to believe.
Older SATs had a verbal section that was nearly all an English vocabulary test. Because of that, some high school English departments really emphasized learning vocabulary words, though others may not have, resulting in automatic advantages or disadvantages based on one’s high school. The math section used high school algebra and geometry, so those attending high schools with better math instruction had an automatic advantage.
^ Good discussion, @ucbalumnus. Fortunately, we don’t need to settle the argument and answer all the questions!
Last word on this specific SES issue from my end, but I will just reiterate that we need to be very careful not to conflate intelligence and SES. There is tons of research now that intelligence is basically unaffected by environmental factors by the mid-teenage years, even though there are some small non-durable effects in the younger years that later fade by mid-adolescence. Intelligence is of course correlated with SES and test scores, as you would expect, but the causality is intelligence -> SES and intelligence -> test scores. SES will of course capture school environment, and I bet that the correlation between mean IQ and any measure of school quality for any given school will be extremely high. (I’m not aware of any specific research, and current data are unlikely to be found in the US because for various reasons IQ testing has been ruled discriminatory and illegal in some states.)
So, to say that some schools will teach something, and others do not, etc., is simply stating the truism that every institution needs to work with what it has. With a test of basic general reasoning like the SAT, this is unlikely to be a big deal when the school has median IQ or higher anyway. The real tragedy, and the one we should be most concerned with identifying, is the high-IQ but low SES student within a school with very low mean IQ (say, below 80 or 85, and there are many examples of these). Early testing offers the best hope of identifying these kids.
Just a few cites for people who are interested in SES versus testing and the general discussion of testing fairness – someone has pulled together a great list of sources, including the 2009 College Board study (Item 8 linked in the source) that basically found very little influence of SES on SAT scores insofar as SAT scores are valid as predictive tools (all these guys dance around the basic truth that intelligence predicts almost everything, so they say the same thing in different ways - it’s frustrating that’s for sure).
Edit - I’m having trouble linking to another blog, so if anyone is interested just google “No the SAT doesn’t just measure income” and you’ll get to it. Well-sourced and worth reading imo.
Whether or not that is true, the measurement of intelligence is heavily affected by environmental factors. As mentioned previously, the older SAT was basically an English vocabulary test and a math test using algebra and geometry. These proxies for intelligence are highly affected by school quality.
And it is unlikely that environmental factors have no effect on mental abilities or other mental aspects such as behavioral tendencies. For example, someone growing up in a very poor environment is more likely to be conditioned to favor the short term over the long term, because a poor environment means that if the short term needs are not satisfied, there may be no long term. This can negatively affect intellectual progress.
Well, I hear you @ucbalumnus. Do take a look through the source links at the blog I mentioned. I have no argument that we would do better to substitute IQ tests for SAT and other test proxies, to remove any residual SES effects from the middle of the intelligence distribution, even though I think they are small (which is consistent with what the College Board research says as well, but again we need always to be critical of self-interest when testing companies publish research - and when test prep companies promise results).
On testing generally, the earlier we do it the better, before any time preference crystallizes. Even if one believes that time preference is largely genetic, there is no reason to delay testing in order to identify those kids who have the best chance of succeeding in a different environment, in case it turns out that it isn’t genetic. The kids deserve at least that from this monstrously large and expensive education system we have. Good discussion.
Starting standardized testing earlier will not eliminate environmental influences (e.g. do the parents have time to read to the kids before they are school age?), and may introduce irrelevant randomness (some kids may start talking late, but are not later disadvantaged in mental abilities).
“There’s a lot of stuff out by fairtest.org that shows the opposite - a linear increase in sat scores as income rises.”
That’s not the opposite, and of course that makes perfect sense. Higher SES (which can include income or wealth variables) implies higher intelligence, as a general matter, and higher intelligence leads to higher SAT scores.
However, when controlled for IQ (the preferred proxy for intelligence), SES becomes almost irrelevant as an explanatory variable, which is the point of the College Board research. This can be most easily seen in the comparisons among groups that have different mean IQ. Some of those group data are presented right there in the fairtest.org link.
As mentioned upthread, the Journal of Blacks in Higher Education published data showing that blacks from families earning from $160K to $1MM (and beyond) score lower on the SAT than whites from families earning from $0 to $30K. It’s hard to square that data with any SES causality variable.
The ship has really sailed on all of this. SES is not what is driving most of what people are concerned about with testing. My personal belief is that we need to face the question head-on if we wish to make real progress. Articles like the fairtest.org cite simply perpetuate wrong assumptions and make any solutions that much more difficult.
Once again, in the real world, both IQ and the measurement of IQ are influenced by one’s environment, so how do you find an IQ that is completely free of environmental influences?
@ucbalumnus - “both IQ and the measurement of IQ are influenced by one’s environment”
What makes you think that? What are your sources and assumptions? Be careful that you are not using nonscientific “common sense” or engaging in the moralistic fallacy that because we wish something to be, then it must be. Nature does not care whether it offends anyone’s sensibilities.
If you are seriously interested, and you seem to be, please spend some time in the academic literature. The Hunt “Human Intelligence” text I mentioned upthread is a great place to start, and for the latest thinking from the neuroscience world, the Haier “Neuroscience of Intelligence” book is also excellent. It’s a fascinating area, and I promise, absolutely nothing like what the public has been led to believe from traditional media. I’m happy to provide additional avenues for exploration by pm.
In terms of IQ itself, trivially obvious cases such as traumatic brain injury, lead exposure, undiagnosed PKU at birth, and general malnutrition are obvious environmental factors that can reduce someone’s intelligence to lower than what his/her intelligence could be in an optimal environment. Other factors like childhood exposure to a multilingual environment may affect brain development so that future language learning is enhanced (but at what cost to other types of brain development?).
In terms of measurement of IQ, all IQ tests and test that purportedly proxy IQ (like older SATs) have to measure something as a proxy for IQ. Such things, such as (in the older SAT example) English vocabulary, algebra, and geometry, do have dependence on one’s environment.
^ Ok, @ucbalumnus , I hear you. Do you think those questions and influences haven’t occurred to the literally hundreds of academics who have studied this area for the last 100 years, literally on every continent? (Also, those are individual factors, and are unlikely to influence aggregate statistics - especially with regard to defined groups and when the studies have been carefully constructed.)
By the way, the most intensive research is currently being undertaken in China, and we can only hope that in the spirit of academic openness, the Chinese researchers do not treat their research as providing a competitive advantage.
At this point, we are simply talking past each other. If you don’t have the time to read the academic texts (who does these days - we are all so busy!), perhaps a short, reasoned article by one of the prominent researchers might tempt you?
If your main concern is that education be tailored to the individual so that a person’s full potential can be maximized, then you and I have no argument at all!
Someone noticed the “skilled worker and PhD student immigrant” phenomenon. Up to now, most people seem to assume that the associated academic achievement is an Asian racial/ethnic thing, due to race/ethnicity being more visible than other aspects of someone, and that white and black immigrants of this type are relatively few in comparison to already existing white and black populations in the US (in contrast, the existing Asian population in the US was small before the increased influx of skilled worker and PhD student immigrants from Asia after 1965).
Very good article though I think that human capital is physical labor and supports the article’s thesis on structured racism in this country meaning lower-wage jobs. Asians haven’t been in this country as long and have done much better in that regard also reinforcing the visibility that ucb has noted a few times here.