Click here to go to the NEW College Discussion Forum

"Some find SATs don't 'define quality'
By Alvin P. Sanoff, Special for USA TODAY
Twenty years ago, Maine's Bates College, a highly selective liberal arts
institution, launched what was considered a bold initiative  it made
submission of SAT I scores optional for applicants.

Now it has examined two decades of data and found that Bates students
who did not submit their scores did as well academically as those who did.

Non-submitters had SAT scores that averaged about 160 points lower than
the 1,283 average of submitters. Yet graduation rates for the two groups
were virtually identical  86.7% for non-submitters and 86.6% for
submitters. The latter had a slightly higher grade-point average  3.11
vs. 3.06...."


Post edited to remove the rest of the text. Do not post entire articles because doing so violates copyrights.
Moderator DoveofPeace
-----

My only comment is that, to me, this is not news. Harvard has known for years that its lower-scoring SAT admits do as well as their higher ones. Again, as the Chancellor at the UCs has found, SATs do not accomplish the single and only thing that the CollegeBoard claims for it -- predict first-year college performance.

What they do rather well is serve as an indicator of highest educational attainment of one parent, and income level of families of students who attend the high school, or live in the area around the school - both important factors in top college admissions.

Have fun!

By Dadx (Dadx) on Saturday, October 02, 2004 - 10:07 am: Edit

I am not certain that what the article seems to say is, in fact, what the reality is.

It does not say, in fact that Bates found that students with lower scores did as well as students with higher scores. It says something very specific: That those who chose not to submit their scores 'seemed' to perform "as well" as those who did, despite their having lower scores.

I'd like to see this data. I was looking up the scores to ascertain the %tiles and found that their average group at 1283 is actually 88th %tile, while 120 points lower at 1163 is 73rd %tile.

One thing about using college GPA and graduation rates is that it lumps the physics and math and chemistry biology computer science economics and other more demanding majors in with sociology, politics, and in some cases, education. What courses did the non-submitting students take? I suppose that you could have differential requirements for different majors. Some schools already do that with SAT IIs, requiring engineers and science kids to send Math IIC and Physics or Chemistry.

I'd fight tooth and nail to keep the schools from throwing out uniform standard measurements as ONE tool to evaluate students. No one argues that you should ignore grades. There needs to be some standard measurement. Could be SAT IIs or whatever, but you need something.


"Harvard has known for years that its lower-scoring SAT admits do as well as their higher ones."

Are there any bases for these statements? I'd like to read their conclusions.

In any case, if, income is so important in admissions, colleges will simply ask about it.......won't they.

By Mini (Mini) on Saturday, October 02, 2004 - 12:54 pm: Edit

I thought they were very clear - they HAD the scores of the non-submitters, gathered after admissions. Those who submitted their scores as part of the admissions process averaged 1283; those who did not averaged about 1123.

There is no "despite" having lower scores. The point was that the score was "indifferent". It didn't matter when it came to college performance. Now this is not a difference between 1600 and 1440 (where the percentile difference is relatively small. The percentile difference between 1283 and 1123 is HUGE (roughly an entire quintile.)

Standard measurement is a reasonable idea. The initial impetus behind the SAT - to be taken once, and without any preparation - was to allow Harvard etc. to find students from schools usually thought to be non-feeders, and among students who were not wealthy, to benefit from what Harvard had to offer (and to benefit the wealthy ones with a wider range of social experience.) The impact now - statistically - is precisely the opposite of what they intended - it serves as a surrogate for peer group income and parental education. The problem is not that standardized measuerment doesn't work as intended, but rather that this one simply doesn't measure up. The CollegeBoard makes only one claim for SATs - they don't claim that it measures aptitude or achievement (they even changed the name to make that clear), but first year college success. And the colleges have known for years that it doesn't do that.

As for income being importantin admissions and colleges will ask about it - they do, don't they? (both directly, on the front of the application, in whether one is applying for financial aid, and what zip code you live in, and on the application itself, in ECs, and in contacts with guidance counselors. It's amazing how much they know, with Harvard and Williams and Brown and Pomona - all need-blind -- ending up with the same percentage of full-paying customers year after year after year. These are pretty smart folks we are dealing with.

By Reidmc (Reidmc) on Saturday, October 02, 2004 - 01:19 pm: Edit

Schools will use whatever criteria they want to admit students. Those that historically weighed SAT Is heavily have denied students that were as capable as those they admitted, and the resulting demographic biases are well-known. The Bates study does prove that, at least for Bates, and the Bates president does not appear to be asking all schools to throw out uniform standard measurements as admission criteria.

If a school wants a particular type of student, or as Dadx posts, wants a certain type of student in a specific area of study, SAT IIs and even Is are certainly viable criteria. But when they are used as a proxy for intelligence or general intellectual or academic capability they have limited utility.

Final note: Math, science, etc. coursework is more demanding only in a limited sense of the word.

By Dadx (Dadx) on Saturday, October 02, 2004 - 01:40 pm: Edit

Yes, in the application of logic. I do understand that that has its limits at universities as it gets in the way of certain things they want to do.

By Interesteddad (Interesteddad) on Saturday, October 02, 2004 - 02:11 pm: Edit

Be careful when calculating percentile differences.

This research was done over a period of 20 years, meaning that the bulk of the scores were SAT scores before recentering. In these ranges, Verbal scores would have been increased by 60 to 100 points with the recentering used today.

Two considerations emerge:

1) If these scores were "re-centered", they would be in the very high range for an average SAT figure -- probably somewhere in the neighborhood of 1350 for the score reporters. I don't believe that many educators believe that differences in SAT scores at such high percentiles mean a whole lot.

2) The gap between the non-submitters and the submitters in this research would tighten with re-centering. In recentering the verbal portion of the test, original lower scores increased more dramatically than original higher scores.

Add these two factors together and the difference between the submitters and non-submitters is not as signficant as it may initially seem. That being the case, it would not be surprising to see roughly comparable results in college.

By Xdad (Xdad) on Saturday, October 02, 2004 - 03:30 pm: Edit

The SAT test is far from being perfect, but it still provides the best equalizer among all other criteria. Without the validation of a national standardized test, elements such as GPA and relative difficulty of the curriculum become worthless, or subject to the sole recognition and reputation of the high school in the college circles.

The rejection of the SAT by certain schools reveals more about their own ineptitude to interpret the scores than true issues with the test. Nothing precludes schools to apply their own weighing system to correct geographical, ethnic, or income differences. Schools or systems that have opined that the SAT-II are better predictors can simply continue to request them from their students, and rely on the scores without feeling compelled to attack a test that has served colleges well for decades.

I also continue to reject the notion proposed by Mini that a single and unprepared test would serve the disadvantaged population better. First, it would be impossible to control or monitor the level of preparation. Insisting on a single test score does not help at all, as it simply would reinforce the need for additional preparation and reward the ones who pursue special courses. Further, rather than equalizing the test scores, it would exacerbate the differences between schools that effectively prepare their students for the SAT and others where such preparation is inexistent. Do you truly believe that a student who attends a school like Andover or Marite son's does not benefit from the design of numerous classes towards test prep? That is why colleges, rather than rejecting or faulting the process, should continue to evaluate the scores in their own context.

It is that individual interpretation of the standardized scores that allow schools to build the type of entering class they really want. It is obvious that a school like Harvey Mudd considers the SAT scores more important than, for instance, Smith or Wellesley. If this translates into different levels of school selectivity in admissions is also a matter of personal interpretation, but this time by parents and students.

By Jamimom (Jamimom) on Saturday, October 02, 2004 - 06:16 pm: Edit

How on earth are you going to get the SAT scores of non submitters or those who did not take the test? Having dealt with this age group, they are not that good in submitting anything unless it is absolutely required with dire consequences if it is not done and possible great rewards if it is I have trouble and highschool counselors have trouble getting kids to return postcards telling a college that they WILL NOT be coming, and getting a look at the letters of accept, reject and waitlist even though I pay a bounty for them. Most colleges directly get the final transcripts from the highschool and I doubt they would get many of those unless they did it that way. I doubt very much if students who never submitted the SAT,would take the trouble to order one up for Bates after they have been accepted. Plus I know several kids who did not take SATs who went to Bates.

Though some schools are not considering the SAT1s at times, such as the UC system, many want extra SAT2s or count them more heavily. So instead of a one sitting SAT1, you end up having to take 3 subject tests. Also if you even sit for the SAT1, those scores are not blocked on the reporting form and once you see something, it is difficult to ignore. Also many highschool transcripts include unofficial SAT scores.

By Reidmc (Reidmc) on Saturday, October 02, 2004 - 11:09 pm: Edit

Dadx - Sorry. . .I forgot that lab = world.

By Ariesathena (Ariesathena) on Sunday, October 03, 2004 - 12:23 am: Edit

I will not debate whether math and science are more demanding overall or just in logic areas, but at many schools, the courses are graded on a strict curve. 1/2 of the class will get C+s or below, and only half get As and Bs. That is a radically different grading system from that of liberal arts subjects. At my alma mater, about 1/3 of the engineers graduated cum laude, which requires a 3.2. Roughly half of the liberal arts students graduated cum laude - which is a 3.5 for that school. Huge difference. Empirically, the engineers come in with higher overall SATs and much higher math SATs, and a slightly higher average class rank.

Just throwing that in there.

Also - what does Bates have for support systems? Are those four-year or six-year graduation rates? If minorities, the poor, and non-native English speakers have tremendous support systems, lots of tutoring and help from professors, then they will probably be able to succeed in college. It does not undermine their achievements, but they do not have an identical college experience to their more privileged peers. Also - I would be VERY interested to know if those are four or six year graduation rates, and if there is a change from one to the other.

By Ariesathena (Ariesathena) on Sunday, October 03, 2004 - 12:26 am: Edit

double post, sorry