Sunday, February 22, 2009

It's Good to Know the Odds are Against Me... Not

Like most standardized tests, the GRE claims to measure general knowledge acquired by the individuals who take it.  However, most people think this is an inaccurate description of what it really measures: an individual's "thinking skills."  In this blog entry, I will talk about the research I found on the bias of the test... and complain about it.

The GRE is used to measure applicants against each other and each individual's potential for success in graduate programs.  Some schools rely on the GRE to make decisions about applicants because it is a numerical value.  But looking only at these standardized test scores, schools might pay no attention to other determinants of success among graduate students like drive, passion, determination, and charisma, and whether or not each applicant has them.  Ultimately, these characteristics seem vital to success.  How does one expect to be successful if they don't care about or want to use their knowledge?

Discrimination started against the individuals taking the GRE when it became a computer-based test.  The computer-based version zeros in on a person's score within the first few questions in each section.  This is unlike early paper versions which offered every student questions weighted with the same value.  You're now probably thinking what I'm thinking: how in the world is this fair?  Well the truth is, it's not.  

Despite what the ETS says, the GRE doesn't measure an individual's critical skills associated with competence, it measures the students ability to pace and guess strategically.  This means that exceptional GRE scores can be prepared for by learning techniques for answering questions that have nothing to do with the amount of knowledge and individual has obtained.  I bet you know what that means: students with the money and access to test-prep materials have an advantage and these advantages are, of course, associated with a high socioeconomic status.  Just because someone has the money to go to test-prep classes and buy every book on the GRE in the book store, it doesn't mean they will do well on the test the first time they take it; they just have the money to take it again.

The GRE further discriminates against females and minorities.  Research shows that most females receive better grades in undergraduate studies but lower scores on the GRE and other standardized tests.  On average, white men tend to score around 100 points better on the GRE than females of various cultures and all minority males.  Within ethnic groups, men received higher scores than females.  Overall, people who proclaimed themselves as white scored higher on the verbal and analytical sections that all of the other ethnic groups studied.  At this point I'm thinking, "WTF?!?!"  This kind of information is seriously disturbing... at least to me.

Figure 1: What graduate schools picture when they see my GRE scores
One good thing about the ETS, however: they discourage universities from using GRE scores to discriminate against applicants.  This might be because GRE scores were only responsible for explaining a 9% variation between students' first year graduate grades.  Hey graduate schools, you know what would be a better way of screening applicants?  Undergraduate grades, which were responsible for explaining 14% of the variation in first year graduate grades between students.  Furthermore, graduate schools that placed a high emphasis on GRE scores accepted mostly white male students.  So for all you minority females with a low socioeconomic status, good luck!  The odds are against you.....


  1. I have no interest in defending the GRE, per se, but I think you're being rather unfair here. You write, "the GRE claims to measure general knowledge," but on the very site you link to it says "Graduate Record Examinations (GRE) General Test measures verbal reasoning, quantitative reasoning, and critical thinking and analytical writing skills." In other words, you begin by badly mischaracterizing the GRE, and ignoring the claims that it makes about itself.

    You also write, "schools might pay no attention to other determinants of success among graduate students like drive, passion, determination, and charisma" but you don't seem to realize that all applicants claim to possess those qualities. Put yourself in the shoes of an AdCom. You have 2000 applications for 200 positions, and every single one of those 2000 personal statements begins, "I am a unique person with drive, passion, determination, and charisma..." How do you distinguish? All 2000 applicants seem to be driven, passionate, etc.

    "The GRE doesn't measure an individual's critical skills associated with competence, it measures the students ability to pace and guess strategically." Even if this is true, the ability to pace yourself and make educated guesses is an absolutely essential life skill. Everyone's most valuable asset is their time, so being able to pace yourself is an essential time management skill, as is being able to cut corners ("guess") when necessary. The entirety of most medical training can be boiled down to an education in how to guess accurately. No medical professional can act with complete knowledge; they all must have the judgment skills and reasoning skills to do their best under less than ideal circumstances. The GRE seems to simulate this well. If you can accurately guess on a question with seconds to spare, you will probably be good at making quick decisions as a professional.

    Of course the GRE, like all tests, rewards those who prepare for it. Being able to prepare for a major event (whether a test or a surgery) is an important life skill, as is being able to perform under pressure. The argument that the test prep industry favors the wealthy is just absurd. You can buy four reputable GRE prep books on Amazon for less than 100$ (not each, total). Most are between 15 and 30 dollars. The test costs between 115 and 140$ to take, so it's ridiculous to complain that those "rich enough" to take it more than once are somehow privileged.

    Finally, your argument that the GRE discriminates against females and minorities is illogical. The GRE may be biased, for all I know, but your argument is a logical fallacy. The mere fact that minorities and women perform worse on the test does not prove that the test is biased. Women and minorities (especially minorities) are usually given inferior educations as children, or not pushed as hard in high school. I believe that women and minorities are just as smart, naturally, as white men, but that does not mean that their lower GRE scores are not valid.

    In other words, women and minorities likely perform worse because they actually are less qualified. This is not a racist thing to say, because I blame the failing city school systems for poor minority performance, and I blame a patriarchal culture for females being pushed to under-perform. White men have access to better schools, and a culture that reinforces their ambitions. This gives them an unfair advantage, but it also means that they tend to be better at academics, and the GRE reflects this. You can blame "the system," but I don't think you should blame the GRE itself.

  2. Most of the information (if not all) that I used for my blog came directly from my resources. So I did not make this stuff up. There have been actual studies that support the conclusions made in my blog entry.

    First of all, the whole test is biased (though not intentionally) toward females because it is in a multiple choice format. Studies on any multiple choice exam show that men will do better than females (especially during timed tests) because of their impulsive personalities compared to women who tend to over-analyze: causing them to not be able to pace themselves correctly. This fact has nothing to do with education, just the physiology of the brain, which for the most part, can not be changed.

    Second of all, I'm not saying the GRE isn't valuable in other ways, but it's main purpose is said to test general knowledge. It is true that being able to accurately guess the answer to a problem is a vital life skill, but this is not what the GRE is said to measure. As for the comment about "drive, charisma, etc..." this is definitely true that all individuals applying to a certain school might have these qualities on their application, but I am talking specifically about the GRE. Some schools rule out applicants on only their GRE scores. Yes this is an easy way to narrow down 2000 applicants, but why not use undergraduate GPA or another factor? Personally, I think the whole process of narrowing down applicants would be a pain, but by using strictly GRE scores to disqualify applicants, universities may not be accepting the next Albert Einstein.

    Third of all, while I agree that minorities may receive an education worse that that of whites while they are children, you can't make the argument (especially in today's society) that women do not receive the same education as men. How is this possible? I've had coed classes all of my life and I haven't noticed any difference at all between the way males and females are taught.

    Fourthly, the test is expensive. I know that the cost of the test is $140 but then you have to consider the cost of test prep (some exceptional courses cost upwards of $1000), and the cost of ordering additional score reports (which can add up if you are applying to a numerous amount of schools due to the fact that the field you want to go into is very competitive), and last but not least there's the cost for the application fees of the numerous amounts of schools you are going to be applying to. I don't know about you, but I am not made of money. There is absolutely no way I would spend $140 to retake an exam that made me want to hurt myself unless I felt adequately prepared (even more so than I did before). Becoming adequately prepared is definitely a time-consuming event, in which case it may be too late to do because your graduate school application deadlines are already passed.

    My blog merely reflected the flaws in the GRE, and the GRE only. Yes there are other successful components of an application, but the GRE does not reflect those directly. And if you're planning on applying to a school that weeds out individuals solely based on their GRE scores, then good luck.

  3. I think being able to "guess" at the last minute does not mean that people always make good decisions. Yes they were able to pick the correct answer, but that choosing may not have been based of any knowledge at all. They could have just been picking the letter "C" for all of the ones they didn't know, and most of the answers happened to be "C". In this case they got lucky, they did not have the ability to "guess" on impulse.

  4. The multiple-choice format simulates many real-life situations. You have limited time, multiple options, and one right answer. I don't think you can admit to women being worse at multiple choice tests without admitting that women are worse at real-life emergency situations. I don't believe that, and I don't think you really do either.

    I don't know anyone who thinks that the GRE measures knowledge. That seems to simply be a misconception on your part. Every source I've ever seen says that the GRE measures thinking skills. For instance, the Wikipedia entry says, "the exam is primarily focused on testing abstract thinking skills."

    The reason schools favor the GRE over GPA is that GPA can change depending on the school and the courses took. The GRE is an objective measure. GPA can be affected by life circumstances (dead grandma during finals) much more so than the GRE. The GRE is a common factor, and is in the fact the only common, objective, factor that schools have to look at. It's easy to say that we should use other factors, but no one has yet designed a better factor. They can't hook you up to a machine and measure your passion or drive.

    It's true that schools maybe rejecting good applicants because of their reliance of the GRE. What you don't seem to comprehend is that schools are always going to reject good applicants, because there are more good applicants than there are openings. The odds of the next Einstein getting rejected due to his low GRE are lower than the odds of the next Einstein getting rejected because someone with a lower GRE was admitted instead of him. Letting people with low GREs in means keeping people with high GREs out, and that is just a losing game. With so many people chasing so few spots it's all zero-sum.

    There are many studies showing that women are discouraged from achieving in modern academia, or are otherwise discriminated against. I can find them later if you really want them (I believe that Dr. Isis posts on this...) but I don't think anyone can argue that our culture encourages women to be airheaded.

    Applying to graduate schools may be expensive, but that has no direct bearing on the GRE. The expense of score reports does not actually affect a person's score, so it's hard to see that as a bias in favor of wealth. Prep classes are expensive,yes, but also unnecessary. The textbooks used in prep classes are available for under 100$, and a driven student should be able to study without the prodding of an instructor. It may be wise to take a year off to study for a standardized test; there's no law requiring you to apply this year.

    Brandi: Of course some people get lucky. But the test should be designed with enough questions and false answers to minimize this impact. Luck plays a big role in any human event, but the GRE, like all standardized tests, is designed to minimize that effect. The odds of someone significantly improving their score based on luck alone is very small.

  5. If the GRE is such an objective measure, then how do you explain that it only explains variations in grades of first year grad students only 9% of the time, while undergraduate GPA explains variations in grades of first year grad students 14% of the time? Obviously under these circumstances undergraduate GPA better predicted the grades that a student could achieve in graduate school over the GRE. If graduate schools really wanted to be accurate in their decision making process... maybe they should be informed about the profoundness of the undergraduate school an applicant is coming from. Don't you think a student is more likely to be successful if they have received say a 3.9 undergrad GPA from a profound college and scored low on their GRE than an student with a 3.9 GPA from a moderate college and a high GRE score?

    Even the ETS warns about the limitations of the GRE. The two main limitations, according to the ETS website, are: it does not take into consideration all of the components that predict success in graduate schools and can not correctly access undergraduate achievements; also, it is an inexact measure and only significantly different scores can predict differences in students' academic knowledge and abilities.

    And I do believe men and women respond differently in emergency situations. Men are more impulsive which I think would lead them to respond faster than women who would want to think about a situation longer before acting.

    Even if, through all this, the GRE isn't biased toward any individuals at all, the bottom line is white males score higher than any other group of individuals. This is a proven statistic. Claiming the GRE doesn't have any flaws could possible never be proven.

  6. Objectivity and predictive value are two entirely different things. Height and weight are objective measures, but neither predicts grad school success. Obviously GPA is an important part of any application, but it is impossible for every school to be familiar with every UG program. It is even more impossible to know the exact difficulty of every course and every instructor. The fact is that a GPA is less than reliable as a comparative tool. If an applicant has a low GRE but high GPA it casts doubt on the credibility of the high GPA. Someone with a lower GPA but a higher GRE should be preferred, because a high GRE can compensate for a low GPA (that is, a high GRE suggests that the GPA deficiencies were not due to a lack of academic ability).

    Here's the main idea: GRE scores are comparable across applicants. GPA is not. It is impossible to say what constitutes an "impressive" GPA. Many elite colleges are notorious for grade inflation, and many state schools grade on a harsh curve because of their large class sizes. Even within a single school two instructors can teach a class differently. For instance, many people think that Fenster is much more difficult than Brauner in the BIO 303 (Genetics) course. A "B" from Fenster may be an "A" from Brauner. Or it might not. Who can say? That's why we need an objective and comparable measure.

    I believe you about the 14% GPA statistic, but I question its meaning. I have no doubt that low (sub-3.2) GPAs correlate with equally mediocre graduate performance, but the question that most AdComs face is choosing between many high GPA (above 3.6) applicants. Does a 4.0 really predict better grad school performance than a 3.6? Your statistics need to be more detailed in order to be meaningful. You need to find out where the GPA "ceiling" is. Just because lower GPAs correlate to lower performance doesn't mean that we can assume that higher GPAs are equally predictive.

  7. I don't see a compromise happening in this argument so I'll just say, agree to disagree