Study Reaffirms Mt. Holyoke SAT Policy

Parents Forum
1

<p>SOUTH HADLEY, Mass. - Mount Holyoke College students who chose not to submit their SAT scores with their applications are succeeding academically, further bolstering the College’s contention that the SAT is not essential to making good admission decisions and has limited value as a predictor of an individual student’s success.</p>

<p>Mount Holyoke in 2001 made the standardized test optional for admission, convinced that the SAT had become a negative force in higher education, and committed to casting a wider net for applicants with strong academic potential and exceptional talents who may have been previously discouraged from applying because of their performance on the SAT. The College is now in the final year of a three-year study of the effects of that policy, made possible by a $290,000 grant from the Mellon Foundation.</p>

<p>Interim results from that study show no meaningful difference in academic performance between students who did not submit scores and those who did. The study shows that there is a .1 difference between the aggregate grade point averages of submitters and non-submitters. The difference is equivalent to approximately one letter grade in one course over a year of study.</p>

<p>“The fact is that the SAT does not add enough value for us to require students and their families to make such a large investment of time, energy, and money in this single, high-stakes test,” said Jane B. Brown, vice president for enrollment and college relations at Mount Holyoke. “We would encourage high school students to focus instead on activities that promote long-term intellectual and personal growth rather than on time-consuming and often expensive strategies to raise their SAT scores.”</p>

<p>One early result from the study confirms what has been widely assumed: As families’ income levels rise, so too does the likelihood that the student has had the advantage of SAT training classes or special tutoring. More than two-thirds of prospective Mount Holyoke students from higher-income families took an SAT preparation course, and one in three had private tutoring.</p>

<p>The new SAT that will be administered on March 12 has undergone a number of changes, most significantly the addition of an essay-writing section. Given that Mount Holyoke has historically placed great emphasis on the caliber of applicants’ writing and currently requires the submission of several essays and graded writing samples, this new section is not likely to add to the test’s value to our admission process, Brown said. In addition, “this test is a formulaic writing exercise and is likely to be quite coachable,” she said.</p>

<p>“We look at students as individuals and take into account their academic records, their leadership abilities, and their performance over the course of four years,” Brown said. The SAT-optional policy in fact is simply the evolution of the College’s traditional holistic approach to the applicant selection process that includes, among many components, a comprehensive review of a student’s high school record, rigorous writing requirements including several essays, and submission of a graded paper from a high school course. Admission officers also look for less tangible qualities such as intellectual curiosity, motivation, leadership, creativity, and a social conscience.</p>

<p>The ongoing study features six major elements, including an analysis of admission data; a survey of inquirers, applicants, and matriculants; the tracking of submitters’ and non-submitters’ academic performance; an in-depth persistence study involving student volunteers; assessments of admission committee ratings; and focus groups with guidance counselors.</p>

2

<p>This is an interesting study, and I am not surprised by the result. </p>

<p>I would assume that high school grades and significant extracurricular activities are reasonable (but far from perfect) predictors of success in college courses, just as the SAT is. If I were to choose a student to admit, I would rather have a student who earned straight A’s in a competitive high school than someone who had inferior grades and exemplar SAT scores. Someone with both excellent grades and test scores would be preferred. However, for my children, since we HOMESCHOOL, they have no choice but to do well on tests because no college will have any idea how one homeschooler grades compared to another homeschooler (most probably do not even assign grades until they need to create a college transcript). My kids have taken college courses and AP exams, but those are the only “grades” they have besides the SAT I and II scores, so we are stuck with the tests. I am curious how Mt. Holyoke would evaluate homeschooled students without test scores (if they do).</p>

3

<p>Can’t answer. My homeschooled d. took the SATs at age 12, and never again (scored extremely well), and at 16 was awarded Mount Holyoke’s largest merit scholarship (which she turned down.) But the numbers of homeschooled applicants are extremely small, so it is difficult to judge.</p>

<p>But the question they have about SATs would hold as true for homeschoolers as for everyone else. Their study found that as family incomes rise, so does the likelihood that a student will avail themselves
of SAT courses and tutors. Since so many other parts of the application (ECs, etc.) are often reflective of family income, the SATs represent a “double whammy” of same. The confounding aspects to do not lessen simply because it is more difficult to assess the rest of the “high school” record.</p>

4

<p>I’m really glad to hear this!</p>

<p>There’s hope for us test anxiety racked students who suck at multiple choice tests and can’t afford the prep!</p>

<p>Kidding. XD I’ve know that all along.</p>

<p>But, I’ve never been a proponent of the SAT. I’ve seen this one test crush too many dreams… and, as someone who has really bad test anxiety and has never tested well (with math multiple choice mostly… I loved timed essays and long answer type tests) I can see that it isn’t a totally accurate predictor of anything. Of course, I might also just be a crabby student. ^_~</p>

5

<p>LOL! More yada-yada spin from Western Mass!</p>

<p>What a revelation from Mt. Holyoke -no pun intended. </p>

<p>“Interim results from that study show no meaningful difference in academic performance between students who did not submit scores and those who did.”</p>

<p>Why would there be a “meaningful difference” among the two groups of students at MHC? Since the group who decided to submit SAT scores scored in the 1250 range, isn’t it fair to assume that the group who did not submit scores SHOUD be CAPABLE of scoring in the same range. Do we know that their SAT scores were necessarily lower? Or is that simply assumed?</p>

<p>“One early result from the study confirms what has been widely assumed: As families’ income levels rise, so too does the likelihood that the student has had the advantage of SAT training classes or special tutoring.” </p>

<p>Really? Since when does the confirmation of an assumption of a likelihood carry any importance or relevance. Let me say that it is also widely ASSUMED that when families’ income levels rise, so is the LIKELIHOOD that the student had the advantage of selecting a better private school or benefit from a family’s upward move to a better school district. There is also a great likelihood that the families will order more pizza on a weekend. </p>

<p>Do we really need a study to tell us that rising income levels mean more disposable cash? Or that wealthier families have more resources? </p>

<p>A similar study might have been more relevant at an institution where the SAT scores play an important role, and where the admission rates are well below 50%. It might have been more interesting to see how students who do not present SAT scores fare at schools where the median scores are at 1400 and above than a comparative study of slightly above average candidates. </p>

<p>The Mellon Foundation is indeed very generous with his grant money.</p>

6

<p>A variation on Xiggi’s post: it would be interesting to find out more about the SES and educational history of the students who submitted SATs and of those who did not submit SATs.
I agree that non-submitters must have felt that their SATs were in the ballpark. Furthermore, a college that professes not to value SATs should have no qualm rejecting high SAT scorers, so we won’t know how the non-submitters would have fared against high scorers who were rejected (or does Mt Holyoke play it both ways?).</p>

7

<p>Xiggi - if you ANYTHING about the national distribution of SAT scores, you wouldn’t sneer.</p>

<p>“Do we really need a study to tell us that rising income levels mean more disposable cash? Or that wealthier families have more resources?”</p>

<p>From your note, it would appear that the answer is obviously “yes”.</p>

<p>If the test is not a predictor of college performance (extensively shown in California), but is an indicator of wealth (shown by the CollegeBoard itself), then there can be very good reasons to use, aren’t there?</p>

8

<p>The UCs should have dropped the SAT1 tests instead of changing them.</p>

<p>THe SAT2 tests, along with grades, were good predictors of freshman year college success and were less affected by income.</p>

9

<p>I think 0.1 GPA is a fairly significant difference.</p>

10

<p>So many people bring up the UC conclusion about the previous version of the SAT I that I have to comment that I haven’t seen anyone mention the issue of “restriction of range.” The UCs have used SAT I scores, along with a few other criteria, to screen for admission for a long time. The variance between a typically high-scoring UC admittee and a typically low-scoring UC admittee is not as great as the variance in SAT scores among all college applicants in the whole country. So the UC system was basically looking at data that were precooked when it concluded that SAT score variance accounts for a small portion of the undergraduate grade variance of UC students. I have no reason to doubt that UC students self-select harder or easier major subjects, and harder or easier courses in each major subject, the same way I observed students do that at my state flagship university. Was there a correction for remedial or “Mickey Mouse” courses in the UC review of admitted students that led to complaints about the former SAT I? Where is a link to a detailed description of the study? </p>

<p>I don’t expect MIT to become SAT-optional any time soon.</p>

11

<p>“Xiggi - if you [knew] ANYTHING about the national distribution of SAT scores, you wouldn’t sneer.”</p>

<p>Mini, I assume that “knew” should complete the sentence, but I’ll accept understood as well. Further assuming that I really do not much about the national distribution of the SAT, I invite you to show me what the finer points of the distribution, and point to my naive eyes all the elements that I should understand. I have an inkling that one of those elements will involve the relative strength of the average scores of MHC. But I rather let you explain your point in greater detail. </p>

<p>“Do we really need a study to tell us that rising income levels mean more disposable cash? Or that wealthier families have more resources?”
From your note, it would appear that the answer is obviously “yes”.</p>

<p>I have to admit that I do not get that remark, but so be it. Let’s talk about our mutual agendas. You have been on a crusade to establish that the SAT is a bad test, and that it exacerbates the unfairness of admissions by having a negative impact on the poorest test-takers, all the while failing to predict the success of students in college. My rebuttal has been pretty simple: the test is not responsible for admissions; the policies of the colleges and the interpretations of the adcoms are responsible for building the classes of freshmen. It is the job of school officials to interpret the relative strength of a 1300 score from a Bhutanese son of goatherder and the relative weakness of a 1450 from an Exonian who lives on Fifth Avenue.</p>

<p>And that is where the theory behind your recurring posts falls flat on its face. The SAT measures the differences but it does not create the problems. The SAT mandate is not to eliminate or analyze the differences, but simply to report them. To repeat what I have said many times, you can see the correlation between income levels, but it does not establish any causality. If the correlation would be so great, how do explain the stratospheric scores of the numerous foreign applicants who come from impoverished country? Obviously, since quite a few foreign applicants believe that Kaplan is a relative of Woody Allen and that Princeton Review is the school newspaper of the New Jersey’s powerhouse, we cannot support the notion that test prepping could have played a role. </p>

<p>Also, could you not attribute the remarkable scores earned by the “Bubble Queen” to the level of parental education, and recognize the lack of correlation with income? How do you reconcile the differences in scores among recent immigrants who show similar low income, especially between asian and hispanic families? That was one is, in fact, very easy: you only need to look at the education levels of the parents. </p>

<p>It is undeniable that wealthier families send the kids to better schools. It is also undeniable that, with or without the SAT’s impact, the better schools are providing a better preparation to their students. Should we not blame the better schools for providing such an “unfair” advantage? And should we not blame our politicians for allowing the great differences in education between the have-a-lot and the have-less? Given the differences in school systems, the SAT might very well provide one of the best opportunities for poor students to show they deserve a shot at a selective school. For admissions, scoring a 1200-1300 when your peers swim in a sea of 1000 is much better -and much easier to attain- than scoring a 1450 that is just average among your privileged peers. </p>

<p>Oh well!</p>

12

<p>

</p>

<p>I question the latter part of that statement. Is there evidence for that? It seems to me that the SAT 2s, which are content based, are much more likely to reflect the level of the school the student attended. For similar reasons, I predict that adding the essay to the SAT 1, and eliminating analogies (whcih had a smaller gap between URMs and others) will only exacerbate the economically-based differences.</p>

13

<p>For what it’s worth, I was an SAT tutor a number of years ago. My observations were that, although the math part had a better correlative value for college success in math courses, both parts really weren’t that good at predicting college success. If flipping a coin generates a 50-50 shot at guessing correctly, I would estimate, based on my observations of my students that the SAT had about a 55-45 predictive rate, which is slightly better than flipping a coin. If you check the college board web site statistics, they note that the correlation is about 60%, which is close to my observations. I guess colleges feel that this extra 5-10% over simply guessing justifies the hassle incurred by the students.</p>

14

<p>Middlebury would be an interesting one to study in this regard.</p>

<p>Candidates must submit standardized tests in at least three areas of study. The requirement may be met by submitting either the ACT; or the SAT I administered on or after March 12, 2005; or three exams in different areas of study which may be selected, mixed and matched from either the SAT IIs, APs or International Baccalaureate exams.</p>

15

<p>Very interesting thread!</p>

<p>Both Mini and Xiggi seem to be making very valid points. I’d love to see or hear more information on the subject. </p>

<p>Taxguy, do you know if there is a variance in predictability between the high/mid/low scores?</p>

<p>I know that the SAT seemed to be a big add-on in junior year for our daughter–providing for a lot of extra studying and anxiety for the test itself, at one of the more anxious times in a child’s life (science projects etc.).</p>

<p>“the test is not responsible for admissions; the policies of the colleges and the interpretations of the adcoms are responsible for building the classes of freshmen. It is the job of school officials to interpret the relative strength of a 1300 score from a Bhutanese son of goatherder and the relative weakness of a 1450 from an Exonian who lives on Fifth Avenue.” (Xiggi)</p>

<p>If SAT’s scores correlate to income, do SAT scores also represent “ability-to-pay” to adcoms? I ask because I wonder if that might mitigate some of the Solomon like wisdom we might expect from an adcom who offers to “cut-the-baby-in-two.”</p>

16

<p>If SAT’s scores correlate to income, do SAT scores also represent “ability-to-pay” to adcoms? I ask because I wonder if that might mitigate some of the Solomon like wisdom we might expect from an adcom who offers to “cut-the-baby-in-two.”</p>

<p>I would say no. There are other ways of guessing at income: by zip code (eg. Scarsdale), by prep/vs. inner city school; by ECs (Mini’s example of fencing), history of travel and summer internships vs. jobs, parental professions, and so forth.</p>

17

<p>Bowdoin made SAT scores optional years ago but gave extra scrutiny to course selection and grades, e.g. number of APs taken and student’s performance in those classes, especially for junior & senior years.</p>

18

<p>Garland, I am getting that stuff from the Berkeley website. If you put SAT test in the search you get lots of info. <a href=“http://www.berkeley.edu%5B/url%5D”>www.berkeley.edu</a></p>

<p>Here is 1 result. http.<a href=“http://www.berkeley.edu/news/berkeleyan/2001/11/07_sat2.html[/url]”>www.berkeley.edu/news/berkeleyan/2001/11/07_sat2.html</a></p>

19

<p>Woodwork asks,“Taxguy, do you know if there is a variance in predictability between the high/mid/low scores?”</p>

<p>Answer: I don’t remember what was posted by the College Board,but I don’t trust their statistics. They are quite biased. </p>

<p>As for my experience, I can’t say there is a variance between high, mid and low scores necessarily. For example, kids with high scores usually did well and some didn’t. However, there were a number of kids who got under 1300 on the SATs and even under 1200 who had very decent grades. I know of two cases in particular whose kids scored under 1100 and yet they achieved 3.5 or better in their first year! Likewise, I know of several cases of kids that scored 1500 or better that didn’t breat 3.0.
I just didn’t see a lot of correlation.</p>

20

<p>taxguy,</p>

<p>Obviously there is not NECESSARILY any difference in GPA performance “between high, mid and low” SAT scores. One can always give anecdotal evidence and compare one student who had a bad day on the SATs (and therefore “overachieved” in GPA) versus someone else who had high SATs but got involved in drugs or just stopped going to class (and therefore had a low GPA). </p>

<p>I am not sure what you are saying about in your experience the SATs having a “55-45” predictive rate. For 1000 students with SATs in the range of 1500 to 1600 compared with 1000 students with SATs in the range of 1200 to 1300, are you saying that there is only a 55 percent chance that the former group will have higher average GPAs (at the same schools)?</p>

<p>Certainly within a small range of scores, SAT will not be very predictive. A student with a 1350 SAT compared to a student with a 1380 is not likely to be statistically different in terms of GPA. But I cannot believe that when comparing a student with an SAT score of 1000 with one who scored 1500 that there is not some VERY significant difference of predicted GPA (at the same school, in the same major, etc.)</p>

<p>“If you check the college board web site statistics, they note that the correlation is about 60%.” “I guess colleges feel that this extra 5-10% over simply guessing justifies the hassle incurred by the students.” What is the reference for this number? (I could not easily locate it on <a href=“http://www.collegeboard.com%5B/url%5D”>www.collegeboard.com</a>) Are you saying that SAT score can account for 60 percent of the scatter in GPA when a linear model is assumed (why would the relationship necessarily be linear, anyway)? Or what does your statement mean?</p>

<p>Thanks.</p>