June 17, 2008

New SAT Is as Predictive as Old SAT, College Board Finds

Recent changes in the SAT did not substantially alter how accurately the test predicts first-year grades, according to new research released today by the College Board. Whether that’s cause for celebration or concern depends on whom you ask.

Officials at the College Board, the nonprofit organization that owns the test, say the research affirms that revisions to the exam — including the addition of a writing section — did not decrease its value to admissions offices. The research also found that the writing section was the most predictive of the three SAT sections. “This is very important and positive news for colleges,” said Gaston Caperton, the College Board’s president.

Not so, said Robert A. Schaeffer, public-education director for the National Center for Fair & Open Testing, a watchdog group. Mr. Schaeffer noted that the College Board’s research revealed that the new SAT, like the old version, is a better predictor for some subgroups of students than for others. “Maybe the College Board’s slogan should be ‘Meet the new test, same as the old test — only longer and more expensive,’” Mr. Schaeffer said.

The College Board’s research comprised two “validity studies,” the first to include the full cohort of students who took the new SAT, introduced in March 2005, and who finished their first year of college this spring. The studies evaluated data on approximately 150,000 students at 110 four-year institutions. —Eric Hoover

Comments

1. Somehow, I am not surprised that the college Board found that the new SAT is just as good a predictor as the old and that the new writing section is the best predictor.

   — M    Jun 17, 02:39 PM    #

2. Sort of like a rancher evaluating his own cows…Interesting particularly because many admissions officers of selective schools state straight out that they do not even look at the writing section.

   — Z    Jun 17, 03:44 PM    #

3. If we expect the College Board to speak negatively about their own test, then we are the idiots. A real evaluation of the SAT’s value MUST come from an independent source, and ALL subgroup results should be evaluated together & separately.

   — R.D.    Jun 17, 03:47 PM    #

4. Until the validity study was completed — it was not valid nor appropriate for admissions officers to use the writing test.

   — P.H.    Jun 17, 04:05 PM    #

5. Big surprise! The College Board found its biggest cash cow to be a good predictor of FYGPA.
   When researchers reviewed clinical drug studies from four major psychiatric journals, they found that eight out of 10 studies paid for by the pharmaceutical company making the drug being studied showed favorable results. When pharmaceutical companies competing with the firm making the drug being studied paid for the studies, only three in 10 studies showed positive results.
   Without outside objective research these SAT studies are meaningless.

   — Brad MacGowan    Jun 17, 06:07 PM    #

6. I hold a Ph. D. from Brown, but due to a low combined score of 852 on the SAT had to start my academic career in a “C” ranked school, then transfer to a public Ivy.

   It took years to rebuild my self-esteem over the low score on the SAT. Beyond that, I’m caucasian.

   The primary function of the SAT and its parent comapny is to provide employment for the people who develop and grade the exam. The SAT is bogus, and should not be used.

   JR

   — JR    Jun 17, 07:31 PM    #

7. There’s two sides to every story. I would never have gotten in to graduate school without standardized testing. I was working in undergraduate school, and my GPA was only a 3.3. With standardized testing (I scored in the 99th percentile on the SAT and in the 93rd percentile on my GMAT), I received an MBA from a top-30 program.

   I support standardized testing. It offers a second chance to smart students who may have not had the time to get the best grades.

   I find JR’s story very hard to believe. Noone who scored only an 852 would have gotten through an Ivy-League program-especially not in math, finance, or any of the sciences. The tests are VERY predictive in these areas.

   Either JR’s program was very slanted towards liberal arts, or he was drunk when he took the test, or he’s simply lying for political reasons. In fact, his school should be very insulted when he implies than someone with that low a score got through. All colleges I’ve heard of are proud of how high their students score on the SAT. They usually are not very eager to admit they gave PhD to someone who scored in the lower 30th percentile on what’s basically an IQ test.

   There’s a lot of politics behind the drive to eliminate the SAT. Moreover, the SAT’s predictive power has been diminished in part due to Affirmative Action. After all, the test does not YET ask a person’s race before tallying the score-although I’m sure in the current political environment, that option has been considered.

   Affirmative Action proponents, in their efforts to bury the evidence of their social engineering, should not be allowed to eliminate standardized testing. The tests exist for a reason-that schools and curriculums vary widely in difficulty. In addition, just as some people are better in classroom work than at preparing for standardize tests-others are much better at preparing for the test than at the daily grind of schoolwork. Noone would suggest eliminating GPAs simply because they don’t predict performance on the SAT very well. The standardized test should be used for the same reason that GPAs are used-they are a valid indicator of a person’s academic performance. If GPAs help measure what was learned in high school, standardized tests help measure what was retained.

   — Test Test    Jun 17, 08:09 PM    #

8. Dear Test Test,

   I have nothing to gain by lying about my low combined SAT score. Indeed, I did earn that score. I also graduated Phi Beta Kappa.

   Sorry, test test, but some people just aren’t meant to take standardized test, which by the way have absolutely nothing to do with IQ tests.

   As for Affrimative Action influencing any part of my education, at age 59 within months of turning 60, in additoin to being the wrong color, I’m too old.

   If you are an educator, I feel sorry your students who just aren’t programed to sit for standardized tests. I’m guessing you don’t have the wherewithal to design your own instruments.

   JR

   — JR    Jun 17, 10:34 PM    #

9. Test Score Variation —- Predict What?

   My daughter took the ACT her math was dismal —- somewhere in the 40 percentile.

   I made her retake the test, and Lo! Behold, the Math improved, but English dropped —- if I recollect from the 99 or so percentile to somewhere in 60s

   I made her retake now English was backup but another section took a nosedive in percentile terms —- so I just simply gave up.

   By the way my daughter completed HS in three years, had excellent grades, class rank, etc., she was the recipient of Rotary scholarship for study abroad, —- got accepted by top schools, —- and decided to attend a sister college (graduating with honors).

   On test scores —- she admitted after graduating from college, that she did listen to me and pay attention to the “Section/Area” about which I complained —- but simply found the rest of the test too tiring, so she simply skipped over the questions in another Section (or randomly checked the answers).

   My thanks to those Admissions Officers/Committee’s who in all probability ignored those scores (for she was not a legacy, —- nor did she in any way even qualify as a special admit)

   — zahid    Jun 18, 12:33 PM    #

10. The problem with the study lies, probably not in its methodology or objectivity — although such issues may exist — but in the interpretation by admissions personnel of the findings. While it is well enough understood that there is a positive correlation between SAT scores and college GPA, the use of this relationship to state that high SAT scores are predictive of high GPA is inappropriate. As any introductory statistics textbook will emphasize, correlation does not imply causality. In other words, the only statement that can be made about the relationship between high SAT scores and high GPA is that the relationship exists to such a degree. The ‘proof’ that high SAT scores are not predictive of high GPA is that there are inevitably those who score well on the SAT but who subsequently earn relatively poor grades. This may seem like splitting hairs; however, relying on one instrument to predict future performance in another area, which appears to be how the SAT is often used with respect to academic performance, fails to take into consideration other tremendously important factors that mediate the relationship between test performance and grades.

    — WEO    Jun 18, 01:45 PM    #

11. I find the article interesting since the lit review I conducted on the subject of entrance exam predictive ability showed that high school grades were a better single predictor of college success (grades, retention) than SAT scores (The addition of SAT scores to HS GPA added very little additional predictive value). I am fascinated by the insistence of college admission departments to continue subjecting students to the pressures and costs of the SAT or ACT when the current research says they add very little value to the process.

    — GN    Jun 18, 03:28 PM    #

12. The tests aren’t great, but for the vast majority of kids a pretty good predictor. The ACT is probably a little better and a little more relevant. I’d rather be without them, but what would we use without them that is really feasible. As long as they are used correctly by college admissions officers as a piece of the equation then I have no problem with them. I know that not all colleges use them that way.

    — jake    Jun 18, 04:19 PM    #

13. Test test,

    I’m glad the test worked for you, but it obviously didn’t check your ability to spell “no one” with a space, or to use the dash (always two dashes together, friend). Humans prevail in social settings for numerous reasons, and none of the standardized tests assesses all of these items; in fact, they test a tiny fraction of what one can use to be successful. I too received a very low SAT score, mainly because I had no idea how to take the damn thing. For graduate school, I spent hours teaching myself how to take the test, and did much, much better on the GRE. But still, I labored too long on the logic problems, because I didn’t want to leave them unfinished. As it is, in my field of study, that turns out to be a good thing.

    One more thing: your comments to JR assume you understand graduate programs outside of MBA programs. Do you? They couldn’t care less what one’s SAT score was, especially if you have a strong grading record, great recommendations from known academics, and very strong writing/research samples.

    You are just flat out wrong.

    — Chris K    Jun 20, 02:56 PM    #

14. SAT and ACT are somewhat predictive for white males, but less so as you look at other populations. It is amazing we have clung so tightly to using these tests in higher ed admissions. The jokes on us, as ACT and SAT are laughing all the way to the bank.

    — Kyle David    Jun 23, 04:33 PM    #

15. As a parent of high school senior, I don’t have any problems with SAT and ACT. Just like GPA and class rank, it is simply another measurement that help colleges to get a sense of students’ strength. I do think these test scores are overly emphasized by some college admissions offices – perphas driven by the desire to improve US News rankings.

    I also believe that these tests should be free to students (perhaps colleges should pay for the scores). Therefore, all students will be able to afford the tests.

    — WC    Jun 23, 11:35 PM    #

16. The SAT and ACT exist for only one reason: to create an equal playing field for everyone to strut their stuff. GPA’s and extracurriculurs can be unreliable, for example a smart child in a underfunded inner-city school may have found it easier to get A’s and become the Valedictorian, whereas a child of the same intellectual caliber could have found it difficult to get B’s in an upper-class private school.

    The goal of the SAT is to even out those geographical and social differences with a test where everyone receives the same difficulty nationwide. While they in no way perfect (or even good) they are the best we can come up with. Unless you want to give it a go?

    — Gaz    Jun 24, 11:35 AM    #

Previous: Berkeley Looks to a Wall Streeter for Help in Finding New Partnerships
Next: Animal-Rights Militants Say They Torched a UCLA Van