Harvard Professor Steve Pinker on the Ideal Elite University Admissions System

Parents Forum
141

<p>There was an interesting article in the New York Times in December, 2000, which I suspect has become increasingly applicable. The following quotation comes from the article, though it is not attributed to a specific person there:</p>

<p>"At Harvard, the admissions office has written a paper lamenting that students ‘seem like dazed survivors of some bewildering lifelong boot camp.’ "</p>

<p>I haven’t seen this myself among the Harvard admits, or even the Harvard applicants that I know. However, it’s coming from the admissions office. True, it was in 2000, but I doubt that Harvard applicants have become more “chill” since then.</p>

<p>Edited to add the link:
<a href=“Ease Up, Top Colleges Tell Stressed Applicants - The New York Times”>http://www.nytimes.com/2000/12/07/us/ease-up-top-colleges-tell-stressed-applicants.html&lt;/a&gt;&lt;/p&gt;

142

<p>What admissions policy - for 30,000 people trying to get into 2000 slots - wouldn’t “make the students appear like survivors of a boot camp”?</p>

<p>Moving to all-SATs won’t do it.<br>
Ignoring powerful EC accomplishments won’t do it. </p>

<p>Come on, what’s the Platonic form of admissions? (Which would be an awesome thread title)</p>

143

<p>Harvard apps have nearly doubled in the past ten years. The number of seats has stayed the same. Do the numbers. More and more kids, even in lower SES high schools, are being offered more chances to take more rigor and become more engaged in the world around them. So, none of the admissions work at any elite, is getting easier. To move back to a stats based scheme is to step backwards. As Rapelye said about Princeton, scores are not as predictive as the CC world thinks. And assumes. Combined with your grades, they only partially predict first-year performance in college. They do not predict, however, other values we hold in high esteem at the college level, such as motivation, creativity, independent thought, intellectual curiosity and perseverance.</p>

<p>Nothing tells us Pinker has any familiarity with the admissions process. Quoting his title doesn’t fill in any blanks, for me. He had an idea, he wrote it up. That’s all. And Juillet did a nice job in her post.</p>

<p>2000 is several ice ages away from today, in the admissions world. </p>

144

<p>I’m still trying to process and don’t feel I could be immediately onboard with the idea of purely stats based admission, but <em>technically</em>, to use the stats as a more effective tool to search for academically capable students (up to high school - obviously there are some late bloomers always), transcripts and grades cannot be looked at without the context of where they are obtained. Rigor in teaching and grading as well as the curriculum in the respective HS matter a great deal. If you are taking the straight A students from a poor school (over a B+/B student from a rigorous school), you are betting on the student’s possible potential alright, but you cannot use their grades to predict academic success in college can you? As for standardized tests, they should be re-designed to distinguish the top performers. When you rely on a 50 point difference in the current SAT tests to identify top performers, for example, it could just mean the difference of one math calculation mistake.</p>

145

<p>"That’s not to say students at lesser schools arent ambitious but when you read David Brooks’s “The Organizational Kid” we all know that type of kid is rarely created outside the walls of affluence. Or if a modest means student has those traits they typically picked them up in boarding or prep school. "</p>

<p>Isabella - since you said Wash U, I’m going to take a wild guess here – Burroughs or Mary I?</p>

146

<p>Re lookingforward, #149: I agree that 2000 is “several ice ages away from today, in the admissions world.” But do you think that the applicants to Harvard have become more relaxed in the interim?</p>

<p>Also, you mention that scores are not especially predictive. I could see this as a possibility, because max-ing out on the SAT in the current climate (books full of practice tests are available, at the least) is not like it was when I took the SAT, after reading the material they sent out when you registered. Period. There are many more students with 800’s, and not just scaled up by the increased student population since my era.</p>

<p>But this is not to say that a test could not have predictive power. The SAT seems to me to have a very low ceiling for max-ing out, and it got even easier to score high when it was re-centered. The question I have about that is: Cui bono? (“As a benefit to whom?” <a href=“Cui bono - Wikipedia”>http://en.wikipedia.org/wiki/Cui_bono&lt;/a&gt;)</p>

147

<p>East Asian university systems work based solely on standardized test scores. I’ve read a few articles about exam hell in countries like China, Japan and Korea (and India to a lesser extent) where test scores are everything to an admissions office and that would wreak havoc on the students at all levels.</p>

<p>You’d simply be shifting the stress of extracurriculars to more stress about test scores…</p>

148

<p>“But do you think that the applicants to Harvard have become more relaxed in the interim?”</p>

<p>“The applicants” to Harvard are not a monolithic group who all believe, act, think, do the same thing. </p>

<p>I think there are some competing and opposing trends. </p>

<p>Certainly it is true that there is an influx of both international and recent-immigrant families who often put a very, very high emotional price on getting into Harvard et al – and that certainly ups the tension. I mean, read the sad tales of kids on CC who are convinced they are losers in life because their parents are pushing them to get into one of a handful of schools. Those kids are stressed at big-time levels – because so much is emotionally riding on this, far beyond the actual life impact. I think it’s quite possible there are more of those super-stressed kids there vs 10 years ago.</p>

<p>On the other hand, there’s a competing trend where upper middle class families have begun to take a wider look at to what’s “acceptable” - which means a broader, deeper, more thoughtful list than “throw all the Ivies together.” Which means LAC’s in addition to universities, and it means Grinnell and Colorado College as opposed to just Williams and Amherst. That’s absolutely the trend among the more sophisticated upper middle class people that I see, which I think results in their kids being <em>more</em> relaxed – because they don’t have to see Harvard as the uber-prize and anything else as one-step-above-burger-flipping, they know they’ll be fine wherever they go. </p>

149

<p>Benley, when you look at a student’s whole app, you can see if the attitude and the “more” are there. When you look at the LoRs, you can see teachers who are focused and challenging, whose words astutely convey the right sorts of confidence in a kid. Or are lukewarm. This applies, believe it or not, to teachers at low ranked hs, too. CC has an idea lower hs are crap, but if you could see LoRs for their top kids, your perception could change.</p>

<p>An A student from an under-performing hs (and remember, that’s judged on the whole student body, not just the best college-bound kids,) may, in fact, be working much more successfully than a kid getting B’s elsewhere. Adcoms are only partly betting on raw or assumed “potential.” They’re also looking for the actual choices and performance that give them confidence there IS potential. B’s are still B’s.</p>

<p>Another thing people miss is that kids getting pre-college mentoring ARE getting the influences of successful people who have gone before them. Some of the programs are pretty phenomenal. </p>

<p>No, Q, I don’t think the H (or Ivy) applicants have gotten more relaxed. What I’ve seen, ime, is that the best of them are getting more savvy about what colleges like to see. I don’t know where it comes from and sometimes wonder if they are, haha, actually listening to some of us on CC. </p>

<p>Where the SAT can have predictive power-- let’s not call it predictive, let’s say, “where it can shine a light on a kid” - is that it is a test and all take it or the ACT. The more focused kids will work on it, they should get an attaboy or attagirl for taking it seriously and performing well. I don’t have much time to explain, but after you see a good score, after you say, good girl, you move to the rest of what you look for. Good scores are good. But not reflective of all these elites want a kid to be.</p>

<p>Agree with PG that the pressure on the Ivies- which aren’t looking to significantly expand their seats- is pushing a (new) mass of great kids toward more LACs. It would be an interesting thread, to look at this new power the top 15-40 LACs are accruing.</p>

150

<p>

</p>

<p>Also, with a few exceptions, unlike the US…students in East Asia along with many other parts of the world are expected to specialize much more earlier because what is considered “Gen Ed” courses taking in the first two years were dealt with at the high school or sometimes even middle school stage of one’s education. </p>

<p>In a sense, a bachelor’s program in most East Asian colleges…like those of the UK fall in between and lean closer to what a non-interdisciplinary academic Masters program would be here in the US. A reason why changing majors is exceedingly difficult or even impossible to do in such countries.<br>
Also, in the East Asian and some other systems, your major is also determined by how well you score on the national college entrance exam along with the college one ends up attending assuming one was admitted someplace.</p>

<p>This factored into parental misunderstandings when advising HS/college classmates as parents who attended college in their respective home countries were required in their colleges to take nearly all/all courses in their department/closely related departments. </p>

<p>Thankfully, my classmates ignored such parental advice as failing to do so in that case would have resulted in them taking too many courses in their major(s) and neglecting their distribution/gen ed and other non-major requirements. </p>

151

<p>

</p>

<p>I know exactly who he is, as I am a psychologist myself (although, of course, much his junior). But it’s not my position; it’s the position of established social science. There’s been years and scores of research showing that SES and test scores and race and test scores are correlated with each other, including data from the administrators of the SAT themselves. The entire second half of my comment presents that information. For example, he says that the correlation of SES and SAT is only about .25 when research from the College Board shows that it is only that low in very specific circumstances that confound the analysis, and when you look across all test-takers the correlation is actually more like .40.</p>

152

<p>

</p>

<p>I don’t know if this was done by twin studies or by comparing parents vs. any offspring, but even purely genetic traits won’t track 100% between child and offspring. The genetics of intelligence is not fully understood, but even far simpler traits such as the height of plants examined by Mendel did not track with the parents (e.g., as with recessive genes.) I don’t know what this 0.5 statistic means, but even if it did imply that 50% of IQ is inherited by parents, that doesn’t imply 50% is from nature. </p>

<p>

</p>

<p>The example he uses is the people who score high in the SAT in the 7th grade–basically people who were prodigies. These people usually take the SAT as part of a talent search to qualify for classes. </p>

<p>

</p>

<p>The correlation is much greater when you look at people who are truly poor. Once you get to middle class, there is rapidly diminishing returns with SES. Standardized tests have historically been the great equalizer of SES. </p>

153

<p>No time to read thru all the threads but I do have two questions about this idea.</p>

<ol>
<li><p>If admission is based on test score only, why have an admissions process at all? Simply send in your score and attend the next Fall. A 1600/2400 student is guaranteed admission so no need to fill out any paperwork, write any essays or get any recommendations even though some recommendations may be warnings.</p></li>
<li><p>If schools give up their discretion and rely solely on tests written by third parties, how will that affect the testing game? Will greater effort go into gaming and cheating the system?</p></li>
<li><p>What happens to the life of the school? Will there be a football team or a basketball team (what if nobody taller than 5’10" with scores over 2300 apply to the school?). Will there be any writers and poets, artists, musicians, activists and thought leaders? Obviously, with test scores being the sole criteria for acceptance and the competition to test well growing ever fiercer, could it be that parents and their kids devote less energy and effort on their ECs? Football, art, music, etc., may compete with precious test prep time and suffer as a result.</p></li>
<li><p>Are their statistical and qualitative differences between a kid that scores a 2400, 2350 or 2300? The actual difference is only a few test questions. The 2300 scorer probably knew the answers and filled out 2 or 3 bubbles incorrectly. Maybe one kid took his Ritalin before the test and one didn’t. Are the differences between the scores so great as to rule that one is superior to the other based on that sole determinant?</p></li>
<li><p>Is their statistical evidence that the highest scorers are the best kids on campus? Are there any 2200-2300 scorers that perform academically better than any 2400 students? If so, why didn’t the tests predict that?</p></li>
</ol>

<p>Not that holistic is a perfect system but I am doubtful that this one is any better. I’m open to data demonstrating that it is. It’s definitely a worthy research project.</p>

154

<p>juillet #157: Here is the difficulty that I have with arguments about correlations between average SAT scores and family income: The data I have seen cut off at an SAT score that is too low for admission to top schools (by quite a way); this corresponds to a cut-off in the family income data somewhere around “$200,000 and up” as the top bracket. For example, this article from the Washington Post shows that students within the family income bracket “More than $200,000” average 1714 out of 2400 on the SAT I.<br>
<a href=“http://www.washingtonpost.com/blogs/wonkblog/wp/2014/03/05/these-four-charts-show-how-the-sat-favors-the-rich-educated-families/”>http://www.washingtonpost.com/blogs/wonkblog/wp/2014/03/05/these-four-charts-show-how-the-sat-favors-the-rich-educated-families/&lt;/a&gt;
Isn’t that too low to say, “University of Illinois, here I come!” to say nothing of Princeton? (Inside joke for CC regulars and others who have seen the film) Students whose parents have a graduate degree average 1689 out of 2400. That’s really not useful for admission to the CC top schools, nor the Ivy league. Three scores of 560 will get a student a 1680, and three scores of 580 will get a student a 1740, which is higher than the average score for students coming from families with incomes higher than $200,000.</p>

<p>It is a bit hard to find data on the median family income of Harvard students, but this 2012 analysis argues that about 45% of Harvard students came from families making more than $200,000 at that time:
<a href=“Diversitas? Take a Closer Look | Opinion | The Harvard Crimson”>http://www.thecrimson.com/article/2012/1/26/diversity-lack-figures-evidence-harvard/&lt;/a&gt;&lt;/p&gt;

<p>The critical question for me would be: How far does the income to SAT score correlation extend, as you go further up the income range? Within the “normal” income range, there is a monotonic increase in SAT score with family income. Does this persist all the way up? (Slightly different question from the first one in this paragraph) </p>

<p>I don’t know whether CB has collected the data that would make it possible to answer this. But if they have, perhaps you have access to it?</p>

<p>Personally, I think it would be quite awkward for CB if SAT scores continue to rise with family income for a considerable distance beyond $200,000. My guess is that at some point it must level off. I don’t believe that there is a family income beyond which all the children score 2400. </p>

155

<p>QM - You are considering the curve linear with a 45 degree slope or something when scores vs income are plotted which is not what you should be looking for. When we discuss averages, it is an average, just like CB says their 50%ile is normed to a 500 score. If 200k income only takes you up another 70-80 points, then the income has not done much for you.</p>

<p>In real life terms here is what it means.</p>

<ol>
<li> Someone with an income of 100k or more is willing to spend $1000 for princeton review class for their kid.</li>
<li> Someone with an income of 150-200k is willing to spend $5000 on tutors and additional resources.</li>
<li> Someone making $500,000 or more and sending their kids to private schools is willing to plunk down the 10-30k it takes to get a battalion of people around their kid to ensure not just their scores are fine but the app meets that halo required for holistic admissions.<br></li>
</ol>

<p>But then if you take an Asian family, these numbers are out the window. Someone making $20,000 or even 10,000 will spend 1000 and someone making $500,000 is willing to sign the contract with that guy in California for 100,000 or whatever to guarantee that vaunted admission.</p>

<p>None of the above will take away the expectation that the reading and writing porition of SAT is inherently biased towards caucasians and their heritage and no amount of retreading of the messaging by CB will change the fact.</p>

156

<p>

</p>

<p>Can you explain why you think this is the case? A lot of people think the math SAT is also biased toward caucasians.
Are you saying that the gap between the math and CR scores is wider for other races than for caucasians? That would seem to be a natural corollary.</p>

157

<p>Also, rather than using a grand total score, you should really focus on the CR/M total, when talking about admissions. </p>

158

<p>

</p>

<p>Asians are making the Math curve really bad for everyone else. 25k+ scored 750 or above last year and 7800 scored 800. This is one reason someone can no longer make a single mistake in Math and still score 800. Used to be that on some tests one can make a mistake or two and score 800 until 5-6 years ago.</p>

<p><a href=“http://media.collegeboard.com/digitalServices/pdf/research/SAT-Percentile-Ranks-By-Gender-Ethnicity-2013.pdf”>http://media.collegeboard.com/digitalServices/pdf/research/SAT-Percentile-Ranks-By-Gender-Ethnicity-2013.pdf&lt;/a&gt;&lt;/p&gt;

<p>I see a clear advantage in VERBAL if 61% of whites are scoring 500 on a 50%ile normalized score especially when they represent 50% of the test taking population.</p>

<p>Juillet might be a lot better than me at explaining why African Americans find the test biased against them. I think it is quite bad that only 24% are meeting the 500 score.</p>

159

<p>lookingforward #156, The Washington Post article gives the breakdown of scores on each section (CR, M, W). It looks to me as though there would be relatively little difference if one took just CR + M.</p>

<p><a href=“http://www.washingtonpost.com/blogs/wonkblog/wp/2014/03/05/these-four-charts-show-how-the-sat-favors-the-rich-educated-families/”>http://www.washingtonpost.com/blogs/wonkblog/wp/2014/03/05/these-four-charts-show-how-the-sat-favors-the-rich-educated-families/&lt;/a&gt;&lt;/p&gt;

<p>The math scores are slightly higher than the other two scores, across all socio-economic groups. The writing score tends to be tied with CR, or very slightly lower.</p>

<p>I don’t think this alters the analysis in any significant way, really.</p>

<p>texaspg, the plot of the total score vs. family income is clearly nonlinear, though it is monotonic. I am not looking for a slope of 45 degrees!</p>

160

<p>There are clearly cultural differences that influence the CR SAT score, but it’s not exclusively a matter of race. Some time ago, Harper’s published an informal study in which they presented just the answers connected with the reading passage to a group of American-educated people in the editorial offices and to a group of British-educated people in the editorial offices. The Americans were able to pick out the correct answer a very high fraction of the time without seeing the reading passages on which the questions were supposed to be based! However, the British-educated people had a much harder time doing this.</p>

<p>There was an instance a while ago on CC when someone competing in something like the Siemens/Intel or some similar competition was being asked about parents’ occupations on the application form. That poster assumed that it was beneficial to have parents who were scientists, because he/she expected the judges to react more favorably to someone who was the child of a scientist or scientists.</p>

<p>In my estimation, the judges were more likely to be critical in scrutinizing projects from children of scientists. You may disagree about this latter point, but I don’t think that a multi-generation American would think that it would be an <em>advantage</em> to have parents who were scientists, in terms of the judges’ reactions. They would most likely think it was an advantage overall, in terms of parental help, connections, educational environment in the home, etc.; but not in terms of that fact impressing the judges.</p>

<p>There are many underlying cultural values that are not universal in the US, but that are implicit in the CR section (and perhaps in the writing section, too).</p>

<p>Incidentally, I think the analysis of the effects of taking the PSAT and its timing (in the Washington Post article) missed the major effect.</p>