I remember hearing a few years ago that the SATs were racially/culturally biased, so that white suburbanites were more likely to get better scores–regardless of comparable academic abilities–than, say, inner-city black kids. What was the basis for this argument? I could never get my head around it.
-toadspittle, white suburbanite who took the SATs in the good old days (before essays, etc. … yikes! glad I’m not you, kids…)
The primary argument was that the questions, particularly the essay questions and analogies, were written in a way that used certain cultural touchpoints as references. The references reflected the background and biases of the test preparers, who were predominantly white professionals.
IMHO it’s virtually impossible to phrase questions in such terms that they would be understood equally by anyone from any cultural background, but now we’re getting into a Great debate.
This is going to turn into a Great Debate real fast but the SAT is not racially or culturally biased. It is designed very carefully using sophisticated statistical techniques to ensure that it is not.
Here is one technique that the test designers use: Look at all test takers that scored a given score (say 650) on the verbal or mathmatical portions of the test. Now compare the correct and incorrect answers for every single question among those test takers. Run statistics to ensure that it is equally likely that the test taker would have gotten that question correct regardless of race or sex.
Remember when you took the SAT, the proctors told you that some questions would not actually count and would only be used as test questions for future versions of the test. They don’t tell you which ones they are. That is one way that they ensure that there is no statistical claim for bias on any question or group of questions.
You will see articles all over the Web that say that the SAT is chock full of blantantly racially and sexually biased questions. Poppycock! Ask for a good example from a recent version of the test and all you will get is some grumbling. There used to be some moderately biased questions back when the SAT was first given (read 1940’s and 1950’s) but these have long since refined and continue to be.
One point that they make is true. Students from different minority groups as a group do not score as high as white middle and upper class students do. I will let somebody else get into some reasons why that may be although I will point out that those same students tend to not do as well in college either. Isn’t that part of what the test is meant to predict?
Whooo hooo - you asked a biggie. Not that you need any direction, but perhaps this is going to wind up in Great Debates soon because I’m pretty sure that’s what it is.
Yes - suburbanite white kids tend to score better on the SAT than inner-city black kids. The $64,000 question is why?
One theory that has been proposed is that the test is culturally/racially biased. If this were proven to be true, it would adequately explain the observed phenomenon, and would also force some very troubling questions and situations for us as a society to deal with. Thankfully, this is still just a theory.
However, there have been several other theories proposed which also explain the observed phenomenon. One of these is that suburbanite white kids tend to be smarter than inner-city black kids. If this were proven to be true, it would adequately explain the observed phenomenon, and would also force some very troubling questions and situations for us as a society to deal with. Thankfully, this is still just a theory.
So… is the test biased? Some people think it is, some think it isn’t, and it’s a pretty nasty argument all around.
Suggestion - read The Bell Curve by Herrnstein and Murray (available a megalithic online booksellers everywhere). Please note that I do not suggest you draw any particular conclusion from this rather controversial book, just that you read it. It addresses this topic.
I was going to say that it’s difficult to tell whether a test like that is biased because you need an objective standard to measure it. But the technique Shagnasty just described seems pretty clever. I’ll think it over. Do you have any references Shagnasty?
By the way, I used to teach SATs, and one piece of advice I always gave to my students was that they could safely eliminate any answer that said anything even remotely negative about a black person, a woman, etc.
You left out the more likely possibilities. You mentioned innate racial differences proposed in the Bell Curve as one possibility. Unless you are trying to touch off the uber-debate and cause the board to collapse in on itself, you might assume that that is also not the correct answer because several of our esteemed Great Debaters like Collounsbury have shown us that there are no clean racial distinctions to explain this at all. Anyway, that argument would’t work because we see it among many different minority groups, not just blacks.
There are a couple of much more likely possibilities. One is the quality of inner-city schools and poorer minority school districts in general. It is almost universally accepted that students from these schools are at a disadvatage to suburban, more wealthy school districts. I find it curious that minority activists would claim that the SAT is merely a biased instrument while at the same time, recognizing that students coming from these environments are at a severe disadvantage academically. If one recognizes the disparity in education given students, why would you be surprised is standardized test show a difference in educational standards achieved by students from the different environments.
The other possibility, which may be even harder to address is cultural weight given to educational achievement. In many Jewish communities, the social pressure to do well academically is usually great from day one. In inner-city minority neighborhoods, the social pressure can be just the opposite.
The Bell Curve is not entirely germane. The SAT is not an IQ test; it’s an achievement test, intended to convey to colleges an accurate impression of how much a student has actually learned in high school, and how well they can use that learning. The language questions, in particular, require actual knowledge as well as some reasoning, and it’s certainly true that language reflects culture. This unquestionably leads to some cultural bias, but to eliminate knowledge from the test would seriously undermine its value.
As Shagnasty suggests, cultural attitudes toward learning affect achievement in school. Moreover, a parent’s level of education affects a student’s achievement; the mere fact of having educated speech, books to read, and stimulating events in the home is a powerful predictor of academic performance, though far from the only one.
As a previous post mentioned, the test items that show racial / cultural bias are OLD. When you try to track down specifically biased items, you often here about one question from decades back that used the term “regatta.” Today’s questions are screened to eliminate that.
If the test WERE racially / culturally biased, then you would expect the gap between whites and blacks / hispanics / native americans would be much smaller on the math sections than on the verbal sections. That is, since math questions are much less culture-dependent, any difference due to cultural bias should disappear. That doesn’t happen. The gap is the same.
Lastly, the SAT is intended to measure readiness for college work. It does that well. It correlates with college success. When combined with high school grades, it is a better predictor of college grades than high school grades alone. That’s true regardless of the background of the testtaker.
I’m 41, and haven’t had to take the SAT in nearly 24 years, so I can’t comment on the nature of the questions in current versions of the test.
But even in 1978, it was laughably obvious that the test writers were trying like mad to be racially sensitive! In the verbal section, for instance, there’s a large section in whcih students are asked to look at sentences, and decide what (if anything) is wrong with its spelling or grammar. A typical sentence went something like:
“Mr. Martinez, the Mexican-American grocer, strongly supports his school’s free lunch program.”
Later, there’s a section filled with short essays by and about various famous authors. Students are supposed to read these essays, then answer questions testing their comprehension of what they’ve read.
Almost all the essays dealt with famous black Americans- frequently the cliched heroes of Black History Month (Harriet Tubman, Langston Hughes, George Washington Carver)!
MIGHT the S.A.T.s still be biased? Sure- but it’s clear that even a generation ago, test writers were making every effort to be politically correct and racially inclusive, even if the test questions sounded clunky as a result.
I’d really like to know on what basis such statements are made? They disagree completely with my direct knowledge.
Not too long ago, A Famous Tech School in the southeast did a tracking study of its “special admit” students. Minorities who hadn’t done well on the SATs. They did better than average than the rest of the frosh population. They promptly hid the study due to its political impliciations. That is, they should be accepting far more “special admits”. Their admission system was obviously discriminatory. (But showhow people think this is “reverse bias”. Nevermind the fact that how people actually do in college is the best measure.) I used to be on the faculty of said school btw, which is how I know about this.
I have taught many minorities (and women, which in my field are minorities) over the years. I got to know many over 4 years and 4+ classes. I know some really great hard working students yet somehow their test scores on the GRE were pathetic.
I have served on the graduate admissions commitees at 2 universities. All the members knew from first hand experience that you just have to ignore minorities’ GRE scores. Look at grades and letters of rec. We could have easily accepted far more minorities but they just weren’t applying in large numbers. Again, how they actually did in grad school proved us right.
While the answer to “What is the antonym of this word?” is the same for all cultures, there are tremendous cultural variations in preparing people for doing these kinds of tests.
A minority woman I worked with, 3rd generation college grad., had tremendous headaches with her daughters in public school systems. They kept enrolling her daughters in the “general ed.” level courses and she was always going to the school office to insist on them being placed into college prep. courses. BTW: One has a PhD and an MD from Harvard. Another graduated from Yale Law school.
(And this same woman was recently the object of obscene racial insults when she dared to drive her new Toyota thru a small town in rural Georgia. Racism is quite alive and well in America.)
Thanks. That’s the sort of thing I’m looking for. Stupid board ate my last post entirely.
Anyway–I was just looking for examples of why people thought the SATs were culturally biased. I’m NOT looking to take this into GD–so let’s please not get into a hijack re: the validity of that particular argument, or the reason why/why not minorities perform differently academically than white suburbanites.
I took the PSAT and SAT just a couple years ago, and I can’t think of a single question that was culture-specific. I can only laugh when I hear accusations of social or racial bias.
The verbal questions involved either reading an essay and then answering questions about it on the next page, or correcting grammar in sample sentences - all the factual information you needed to know was right there on the page. The math questions were abstract or assisted with diagrams when appropriate, you didn’t have to know the shape of a baseball diamond, the number of quarts in a gallon, or anything like that.
Of course, I’m a white male, so maybe I’m just blindly assuming that students from other cultures learned the same rules of grammar and algebra that I did… :rolleyes:
I’m not quite sure how to parse the common sense logic of the claim you make above for this purported tracking study that was quashed. In virtually every real world circumstance I can think of in modern academia, a school would be trumpeting these results as evidence of the effectiveness and progressiveness of it’s programs. The school in this circumstance would usually be the beneficary of lots of positive media and academic attention and possibly significant grant monies if the effectiveness of the programs could be substantiated.
I can’t see the real downside of admitting that your current admission system needs to be changed because you’re doing so well with minority students.
The new example that I’ve heard thrown around is “Community Service” showing up in essay questions. The answers that are written are going to be different if you think of “community service” as volunteer work done to make the world a better place or think that “community service” is something you get sentenced to after commiting a crime.
Also, remember that cultural bias can involve patterns of thought as well as particular words or phrases. A student from a country where the schools place a lot of emphasis on rote memorization may have a tough time with a test that emphasizes problem-solving, and vice versa.
Or, to use a hypothetical example more specific to the SAT: When parents from culture X read stories to their children, they simply read the words on the page with little elaboration. Parents from culture Y, on the other hand, pause to ask the children questions about what’s going on in the story and what they think will happen next. Kids from culture Y have an edge when they first encounter reading comprehension questions in school – they’re already used to interpreting what they read and can anticipate these kinds of questions – and this advantage could very well carry over into the upper grades. Children internalize ideas about what they’re “good at” or “not good at” very early, and it’s often difficult to change their minds later.
Unfortunately, there probably isn’t any way to create a completely unbiased test; the best we can do is make sure all students have some previous experience with the types of questions they’ll encounter on the SAT. Right at the moment, kids enrolled in test prep courses – or even in schools that include some practice questions in the regular curriculum – have a definite advantage, and most of these courses and schools tend to be clustered in upper-middle-class areas.
I read somewhere (maybe here, although I can’t find it anywhere) that there was an SAT question that assumed familiarity with tea cups and saucers (the one-to-one relationship or something), and it was derided as being biased. Supposedly minority kids wouldn’t know this. I have no idea how old it was, but it interested me because it was the only concrete complaint I’d heard.
Bzzzzt!!! Ding-ding-ding!!! Honk-honnnnnnk!!! Whoop-whoop-whoop!!!
I’m sorry. Let me go turn down my bullshit alarm…
Ahhh, that’s better. Now, ftg, I’m afraid you’re going to have to give us a cite for that one. In fact some hard evidence to back up any of your claims would be nice.
Students from higher-income, more-educated families are more willing to spend money on SAT Study Guides, which raises their scores.
I too, am quite skeptical of ftg’s claims. Most of the evidence I’ve seen demonstrates the exact opposite. When the UC system was using affirmative action, the drop out rate of blacks at UC Berkeley (who had, as a group, much lower SAT scores than the white and Asian students) was astronomical.
Cites:
http://www.dartreview.com/issues/7.22.99/hart.html
http://www.time.com/time/magazine/1998/dom/980420/nation.lies_damn_lies_an2.html
http://www.frontpagemag.com/dh/older/civil.htm
Students from higher-income families are ABLE to spend money on test preparation.
Also, there is a theory called stereotype threat which derives from research conducted in the Stanford University psychology department. See http://www.theatlantic.com/issues/99aug/9908stereotype.htm
for more info. But basically, they found that when black students thought that black students did worse on a test, they did worse on the test.
And if you look at The Bell Curve , you must check out Inequality by Design: Cracking the Bell Curve Myth by Fischer et al. It breaks down the methods and data analysis of The Bell Curve authors.