| Username |
Post |
| Kali Prasad |
Posted
on 27-Apr-01 10:03 AM
Source: Chronicle of Higher Education
A Study, Financed by the College Board, Bolsters the Reliability of the SAT
By JENNIFER JACOBSON
The SAT reliably predicts students' academic performance not only as freshmen but throughout college, according to a study by researchers at the University of Minnesota that is billed as the largest ever on the controversial exam. But critics said the fact that the study was financed by the College Board raises doubts about its validity.
The SAT has come under attack in recent years from advocates for minority students and others who say that the standardized exam is biased against women and members of some minority groups and that colleges weigh students' scores on it too heavily in the admissions process.
Court rulings and state ballot measures restricting the use of affirmative action in admissions, which colleges have used partly to overcome the lower scores by some minority students, have persuaded more institutions to limit the use of the tests. In February, the president of the University of California proposed dropping the requirement that applicants to the system's campuses take the SAT, increasing critics' calls for the test's demise -- and raising the stakes for defenders of the test to justify its existence. (See an article from The Chronicle, March 2.)
The new study, which is formally being released Saturday, is a "meta-analysis" of more than 1,700 studies on the subject and represents more than a million students. The researchers said it clearly shows that students who score well on the SAT have higher grade-point averages in their first year of college, especially in their first semester. The study also found that the SAT predicts G.P.A. during later college years, and also correlates to students' likelihood of staying in college and earning degrees, although those correlations weaken with each passing year.
"Our results clearly indicate that the SAT predicts important criteria both early and late in college," wrote the authors, who include six researchers at the University of Minnesota-Twin Cities and one at the College Board, which provided a grant to pay for the study.
That link elicited skepticism from some testing experts. "Knowing the College Board, I have a great deal of confidence it will turn out to be flawed or fraudulent," said John Katzman, founder and chief executive officer of the Princeton Review.
Regarding the study's finding that SAT scores were related to scores on the State Nursing Board Examinations and therefore predict success in entering the nursing profession, he noted that the nursing exams are given by the Educational Testing Service, which also administers the SAT. "What the study says to me is that people who test well on the SAT, test well on other E.T.S. tests," he said, adding that that was the one thing in the report he agreed with.
"The SAT is under attack because it predicts so little and costs so much," Mr. Katzman said. "The College Board is looking for any good news it can find right now."
Nathan Kuncel, a psychology research fellow at Minnesota who worked on the study, disagreed. "We go through cycles of people being interested and concerned about the test," he said. "It's healthy for people to periodically evaluate what we're doing. But the fact of the matter is the SAT does predict these things based on an overwhelming amount of evidence across different universities."
--------------------------------------------------------------------------------
|
| ashu |
Posted
on 27-Apr-01 10:31 AM
Unfortunately, I have seen too many smart and a lot smarter
Nepalis ending up at relatively-not-so-top colleges and graduate
schools mostly because their SAT or GRE or GMAT
VERBAL scores were/are relatively low.
[Conversely, I have also seen relatively weak Nepali
applicants end up at top places maybe because they had had fiercely devoted help on writing major parts of their applications :-)]
Personally, I have no doubts that these Nepali students,
ending up at relatively-not-so-top colleges and graduate
schools, would have done very well in any of the most
demanding program, but it's just that their low verbal scores keep them from being competitive applicants at top schools.
Unfair but true.
Richard (or Robert) Sternberg, a professor of psychology
at Yale, has developed theories (please read his excellent
book "Practical Intelligence") about SAT and other such
tests.
In the book, Sternberg argues, with solid evidence, that
your high or low scores on SATs/GREs/GMATs are NO predictors
of your professional or personal success in life.
In other words, you can score high on SAT, and graduate from Harvard. But twenty years down the road, you may find yourself working for a Quincy College graduate who has practical intelligence that s/he applies for success.
Law schools, I understand, swear by LSAT scores.
Have a nice week-end, everyone.
I'll be back late next week.
oohi
ashu
|
| Kali Prasad |
Posted
on 27-Apr-01 12:00 PM
One point about the graduate school: It is my experience that grad school does not only look at the GRE score for admission criteria. You have to have good GPA, Good GRE, and good evidence of scholastic research work, and hopefully presentation, and membership in different academic clubs for admission. Reputation of your undergrad school is a very important factor as well. I have a friend who graduate from Delhi School of Economics. She said her classmate who ranked number one in the class got a call from Princeton, harvard and MIT for her admission. They were practically bidding for her against each other. Also, this may be because at least in economics you have Indian professors in every top University. For example, the Department Head of the Department of Economics at Yale University, Dr. T.N. Srinivasan, is an Indian guy. We have nobel prize winner in economics and list goes on and on. My graduate coordinator used to swear if we admit students based on only GRE scores, we would have nobody but Chinese and Indians in US graduate program. May be it is happening that way but still let you all know that GRE is not the sole criteria in graduate admission and assistantship. There are other hidden factors which has more weight in regression than the standarized test. Again this is ceteris perebus argument.
Cheers,
|
| Hari |
Posted
on 27-Apr-01 07:09 PM
Granted, this issue is very touchy and quite complex. My personal opinion is that yes, these tests are not a FULL measure of "intelligence", but, nonetheless, they are an essential part of the screening process. Yes, these standardized tests are not a predictor of personal or professional "success", but they never fail to directly correlate the "success" the prospective student would have in the first year at the program. Sheer volume of applications makes standardized tests a good option to OBJECTIVELY screen them. This is not to say the admissions process should be completely objective or that these tests should be the only yardstick. That is certainly not the case, since none of the top colleges and universities rely SOLELY on them: they always take into account prior preparation in terms of classes, GPA, etc.
I applaud the University of California announcement, but wonder how they would translate that into practical terms. How is the biggest public university system in America going to comb through the thousands of applications it receives every year? It will be tough, and already, faculty members are lamenting the fact that they might have to do a lot of extra "overtime" in screening these applications, at a time when university teaching positions are decreasingly lucrative.
>ashu wrote:
>Unfortunately, I have seen too many smart and a lot smarter >Nepalis ending up at relatively-not-so-top colleges and graduate
>schools mostly because their SAT or GRE or GMAT VERBAL scores >were/are relatively low.
Do you agree then that the testing itself is not a problem? That the 'foreigners' are suffering because of the VERBAL component of the tests? If it is of any consolation, most schools do take that into account and have specific TOEFL requirements. But, it does not mean that you can score pathetically low on the verbal section and get away with it.
>[Conversely, I have also seen relatively weak Nepali
>applicants end up at top places maybe because they had had >fiercely devoted help on writing major parts of their >applications :-)]
This only strengthens the argument that top places routinely look at the whole picture, not only the standardized tests. Those weak Nepali applicants must have landed at those top places, not because their standardized scores were low, but as you mention, they had help in other parts of their application. So, again, the process now, even though it definitely USES the standardized scores, does not mean that it discards everything else.
>Personally, I have no doubts that these Nepali students,
>ending up at relatively-not-so-top colleges and graduate
>schools, would have done very well in any of the most
>demanding program, but it's just that their low verbal scores >keep them from being competitive applicants at top schools.
I fully agree. I would even go one step further that its not ONLY the verbal scores, but other scores too. There will always be a few people that will score low on EVERY section of the test, yet perhaps would have done very well in the most demanding of programs. But from the standpoint of the program, are you willing to bet on a statistics of a 25% success probability in the program as opposed to a 75% success probability? I'm sure the universities would REALLY like to use a better indicator without compromising speed and objectivity, and simply put, such an indicator (better than the current standardized tests) does not yet exist.
>Unfair but true.
>
>Richard (or Robert) Sternberg, a professor of psychology
>at Yale, has developed theories (please read his excellent
>book "Practical Intelligence") about SAT and other such
>tests.
>
>In the book, Sternberg argues, with solid evidence, that
>your high or low scores on SATs/GREs/GMATs are NO predictors
>of your professional or personal success in life.
Absolutely. These tests are NOT predictors of real-life personal or professional successes. The statistics that the Educational Testing Service presents are merely the prediction for first year "success" in the program.
And yes, Sternberg is perhaps correct. There is this whole argument about "Multiple Intelligences" pioneered by Howard Gardner. And througout the nineties, people all over the US have been pressing for a "holistic" approach to testing: portfolios, verbose and descriptive (rather than numerical) report cards, etc. But just as these methods might be better, and would recognize the talents of, say, a musically inclined person that a standardized test will not, these methods are also VERY time consuming, expensive and would be very hard to implement when screening thousands.
>In other words, you can score high on SAT,
>and graduate from Harvard. But twenty years
>down the road, you may find yourself working
>for a Quincy College graduate who has
>practical intelligence that s/he applies for
>success.
No disagreement here.
>Law schools, I understand, swear by LSAT
>scores.
And they have a reason to. Top law schools like NYU, HLS, and others admit a class with the strongest LSATs. Sure, they might miss the 2% of the otherwise great law student that turns out on the bench of the Supreme Court, but, quite frankly, the other 8 judges on that bench perhaps came out with very strong LSATs.
Dui Paise Musings
Hari
|
