<p>Regarding Berkeley (and all the UC's): the MINIMUM GPA for eligiblity was changed to 3.0 this year. That's based on courses in A-G requirements taken in 10 & 11 grades. So, being admitted to Berkeley or any UC with less than a 3.0 is no longer possible. (In the past the minimum was 2.9 to be eligible). </p>
<p>The average GPA, by the way, for those admitted to Berkeley last year was 4.17, average SAT combined was 2009. Those figures are straight from Berkeley's office of admissions.</p>
<p>Graduation rates, and even freshman retention rates, don't really tell you how difficult or easy a school might prove to be. That's because people leave colleges for LOTS of reasons, not just poor grades. They run out of money, they have family problems, they have drug/alcohol problems, they get homesick, they have a bad break up, they decide they can't stand the social scene, etc. Poor grades are also a cause, of course, but I bet grades ALONE probably account for a relatively small percentage of the reasons why people don't return from year to year, or fail to graduate.</p>
<p>kyledavid80,</p>
<p>you've been misinformed in many ways:
[quote]
the university says it takes the highest single sections.
[/quote]
some do but many don't, at least not on their websites.
[quote]
many ask that you put only the highest single sittings, so they will get only the superscore.
[/quote]
i don't know where you got that from. for one, many schools use common application which asks you to list the scores and dates taken. for another, i have never seen an uncommon application asking you to cherry pick the highest from each section to fill it. besides, school receive scores through official reports and those are what go into their database.</p>
<p>
[quote]
As for the ACT data, we don't know whether they superscore that or not.
[/quote]
Yale says specificly say they don't superscore when considering applicants, let alone publishing their data. </p>
<p>Also, according to ACT official site,
45% of NU applicants from Illinois scored 28 or above (n=3,505)
26% of Michigan applicants from Michigan scored 28 or above (n=15,963)
19% of UCLA applicants from CA scored 28 or above (n=8642)
24% of California applicants from California scored 28 or above (n=6545)
<a href="http://www.act.org/news/data/07/states.html%5B/url%5D">http://www.act.org/news/data/07/states.html</a></p>
<p>Significant larger fraction of Northwestern applicants with 28 or above translates to higher ACT average of the enrolled students. I don't think there's any surprise. I am actually a bit surprised NU's average is only 2 points higher than Michigan, given this. Again, no trace of superscore whatsoever!</p>
<p><a href="http://en.wikipedia.org/wiki/ACT_test%5B/url%5D">http://en.wikipedia.org/wiki/ACT_test</a> shows a SAT-ACT conversion table:
NU's last year stats: 1401 on SAT and 31 on ACT, a perfect match according to the table. Same goes with Michigan (1320-1350 and 29).
[quote]
For one, some schools don't participate in the program. I think Berkeley used to enroll the third highest # of NMS students until it discontinued the program in 2002.
[/quote]
I was talking about number of non-college sponsored NMS. That Berkeley doesn't participate is a moot point.<br>
[quote]
Most importantly, why is this even still being argued? My god, it's just a few incremental differences in the SAT scores. It's like saying that Stanford students are on average more capable (or that the Stanford student body is of higher quality) than NU students because the SAT is on average 30 points higher. Or that Chicago's students aren't as smart as Duke's students, or that NU isn't as selective as Duke because of a difference in SAT scores. Honestly, the differences are so incremental it's really inconsequential.
[/quote]
</p>
<p>Not saying there's a huge difference. Just want to get the facts straight and hope you don't throw out that superscore theory again to explain the difference between public and privates without further new evidence. Your claim appears to be really really far-fetched given the data I've been showing. The real difference is due to what NearL mentioned in post #79: </p>
<p>
[quote]
Most public universities can't afford to cherry-pick their student bodies as much as their private counterparts. They have a duty as secondary public institutions to offer an education the citizens of their respective states
[/quote]
</p>
<p>
[quote]
The Common Data Set Initiative provides definitions for how the data should be reported.
[/quote]
</p>
<p>Exactly. Anyone not following to rule is gonna risk getting exposed sooner or later. (I actually have suspicion about one school but I am not going to say which :))</p>
<p>You say:</p>
<p>
[quote]
some do but many don't, at least not on their websites.
[/quote]
</p>
<p>yet then</p>
<p>
[quote]
Yale says specificly say they don't superscore when considering applicants, let alone publishing their data.
[/quote]
</p>
<p>You use Yale. What about the rest?</p>
<p>
[quote]
Just want to get the facts straight and hope you don't throw out that superscore theory again to explain the difference between public and privates without further new evidence.
[/quote]
</p>
<p>You haven't gotten any facts straight, I'm afraid, just inferences. My analysis from before still seems to work logically. While it isn't definite, neither is yours; you have been putting forth logic that may or may not be true. I admit that my analysis from before is not fool-proof. Yours does not "disprove" that, though -- what would disprove that is data from the school itself. Not averages from CB, not the compared data of schools across the country (which have very different percentages who take the ACT/SAT), the actual data. That's the only "positive" way.</p>
<p>By the way, the average GPA from Berkeley's site is a 4.25; the average SAT from the UC site is a 2040 or so. (However, an email from Berkeley was sent out that listed the average SAT of matriculated students as 2070. Not sure about the origin of this figure, though.)</p>
<p>
[quote]
Yale says specificly say they don't superscore when considering applicants, let alone publishing their data.
[/quote]
</p>
<p>It does? Where?</p>
<p>Swarthmore!</p>
<p>
[quote]
You use Yale. What about the rest?
[/quote]
</p>
<p>Most of the rest don't mention anything about superscoring or not "when considering applicants". I have already mentioned it. Furthermore, I have not seen any one school that said specificly they superscore in their class/stats profile page. Not a single one.
[quote]
My analysis from before still seems to work logically.
[/quote]
You never really have any analysis. One of the main things you said was privates had narrower score range and that indicates privates superscore. That's a twisted logic! The point IS Berkeley students have wider range. What I showed is that according to ACT official website, Northwestern (or many other comparable privates like Tufts, WashU) has a significantly larger fraction of students with high <em>unadjusted</em> ACT scores than Michigan (I know you said Berkeley but for now, let's use Michigan to make it easier for comparison since Michigan has almost the same SAT as Berkeley but geographically it's the same as NU which means ACT comparison is valid). You, on the other hand, <em>claims</em> that NU would have about the similar SAT score as Michigan (Berkeley) if NU doesn't superscore (which you were never able to prove they do). For your claim to be true, it means showhow the Michigan applicants that submit SAT scores have similar stats as NU while the applicants with ACT scores have noticeably lower stats! Very interesting and strange idea, IMO. But I guess I can't really say it isn't possible, lol!</p>
<p>
[quote]
By the way, the average GPA from Berkeley's site is a 4.25; the average SAT from the UC site is a 2040 or so. (However, an email from Berkeley was sent out that listed the average SAT of matriculated students as 2070. Not sure about the origin of this figure, though.)
[/quote]
</p>
<p>Where did you get 2040? <a href="http://students.berkeley.edu/admissions/freshmen.asp%5B/url%5D">http://students.berkeley.edu/admissions/freshmen.asp</a> says the range is 1800-2070. The mid-point of that is 1985, not 2040, and way lower than 2070. That 2070 is actually the average of admitted students. <a href="http://www.berkeley.edu/news/media/releases/2007/04/05_admissions.shtml%5B/url%5D">http://www.berkeley.edu/news/media/releases/2007/04/05_admissions.shtml</a></p>
<p>StatsFinder shows there's a significant gap between admits/enrollees; specificly, the % with 700-800 (average of M+V) is 47% for admits but 37% for enrollees. </p>
<p>By the way, how did you get that email? I thought you said you have no tie with Berkeley. :rolleyes:</p>
<p>Berkeley's colleges of engineering and chemistry are pretty tough to stick with. I know more than a few students who dropped to L&S after a semester. The rough part is that we (engineers) are required to graduate in 9 semesters (or petition), nearly eliminating the ability to experiment with classes. Also, engineering and chemistry students are required to choose their major WHEN they apply, so if one doesn't like his/her major, tough luck (most of the time).</p>
<p>Also, engineering classes (and from what my roommate tells me chemistry classes) are far, far harder and require more work than other courses. Last semester I was enrolled in an upper division political science class that was four units and was graded on one 40% midterm and one 60% take-home final. I was also in a materials science class that was three units and had weekly labs with 20+ page write-ups, homework sets, two midterms, and a final. Oy vey.</p>
<p>What Yale says: </p>
<p><a href="http://www.yale.edu/admit/faq/applying.html#3%5B/url%5D">http://www.yale.edu/admit/faq/applying.html#3</a> </p>
<p>
[quote]
In evaluating SAT or ACT results, does Yale consider scores from previous test dates? </p>
<p>Yes, in the sense that readers of the application will see all of the test results that are in your file, since you are asked to self-report your scores from all test dates. The formal admissions committee that meets to vote on applications, however, will see only the highest score you received on any individual test, if you have repeated any of the tests. For the SAT I, the admissions committee will consider the highest score from each of the test's three sections. For the ACT, the admissions committee will consider the highest composite score.
[/quote]
</p>
<p>(some bold print omitted) </p>
<p>I read this statement as saying that Yale superscores everyone who submits multiple test scores, and that Yale reports in its Common Data Set filings superscored scores for all such applicants and enrollees. </p>
<p>This is standard operating procedure at most Yale-peer colleges.</p>
<p>token: depends on what 'is', is. sry, its an election season. :D</p>
<p>It just could be definitional, since "composite" on the ACT is typically means composite of the four tests (English, Math, Reading, Science) taken on one day. Or, it could be that Yale means composite of tests on any day.</p>
<p>In every info session I've attended (two dozen), the presenter said that while they mix and match SAT, they do not mix & match aCT. The only school that of which I'm aware that superscores the ACT is WashU (based on curmudgeon's research).</p>
<p>tokenadult,</p>
<p>note that in my post #83, I was talking about ACT when I referred to Yale. I read the same page as what you found. They take the "highest composite", not the highest score from each of the test's four sections, unlike SAT. There's an obvious reason; mixing and matching two sections is very different from doing that for 4 sections.</p>
<p>As for SAT, this is what I said about some schools saying they superscore to evaluate applicants. But it doesn't explicitly say they do so to fill their common data set. I agree that you may take this as yes. But note that even if they superscore SAT, if their ACT reflects equivalent high score (though the validity may be questionable if not many students submit ACT to Yale; that's why I used Northwestern vs Michigan), it means superscoring SAT does not have great impact on overall average. </p>
<p>In fact, you can run a spreadsheet using a hypothetical scenario of 1000 students and assume that 1/3 with 1 sitting (no effect from superscore), 1/3 improving in both sections upon second sitting (no effect from superscore), and 1/3 benefiting from superscoring. For every two of the former, it takes 90 point-difference from the last one to result a bump of 30 on overall average. 90 point-difference from superscoring is a very unlikely event and, in many cases, impossible event. You can try this exercise: if the first sitting is V 700 M 750, what kind of second sitting score would result a 90-pt difference from superscore? Note that even V 800 M 720 on second sitting, which is possible but probably not typical (bumping up 100 point on one but dropping 30 in another), would give you only 30-pt bump from superscore.</p>
<p>Also, Berkeley's average was 2070 for the admitted which translates to only about 1380 on M+V. NU's average is 1463 among the admits. That's over 80-point difference, not just 10, 20, or even 40. </p>
<p>Let me just make it clear I am not saying Berkeley is easier to get in. I think the primary reason their students have wider range in test score is that Berkeley admission emphasizes more on GPA than test scores. This fit their mission of a state university. By emphasizing GPA, students can be selected from a more diverse range of schools and no single high school could send ridiculous number of students to Berkeley. If Berkeley emphasizes on test scores like the privates, students from underperforming/non-competitive schools in poor neighborhoods could easily be shut out, no matter how well they do in their own schools if their test scores aren't competitive. This is why 15% of their admits were in the 200-599 range yet the overall GPA is so high and the fraction of students in the top-tenth of their HS is even higher than Harvard. It all makes perfect sense.</p>
<p>Sam Lee:</p>
<p>
[quote]
You never really have any analysis. One of the main things you said was privates had narrower score range and that indicates privates superscore. That's a twisted logic!
[/quote]
</p>
<p>You're taking one of my claims out of context. The fact is, that was only one part of the analysis. I don't expect you to know, as you didn't even read my post. In addition, when you "rebutted" what I said, you didn't address my entire analysis.</p>
<p>
[quote]
says the range is 1800-2070.
[/quote]
</p>
<p>It's 1800 - 2170. It's almost as if you are intentionally changing the data to fit your preconceptions, lol!</p>
<p>
[quote]
The mid-point of that is 1985
[/quote]
</p>
<p>At least you calculated that right. But that doesn't work, and I explained this before (multiple times). You cannot necessarily derive the average SAT from the 50th percentile. Why? Because the average is based on three subsections that vary, and Berkeley will take someone who has a high score on two sections but not so high on another section. Therefore, this means not that 25% of the students got below an 1800 on the SAT, but that 25% got below a 600 on CR, etc. This also means that the average is NOT the same as the composite of the (presumed) 50th-percentiles. This is further supported by the fact that Berkeley's implied 50th percentile is 1990, yet the actual average SAT score (composite, from the real data, not from the sum of the supposed 50th percentiles) is 50 points or so higher. </p>
<p><a href="http://universityofcalifornia.edu/admissions/undergrad_adm/selecting/camp_profiles/camp_profiles_ucb.html%5B/url%5D">http://universityofcalifornia.edu/admissions/undergrad_adm/selecting/camp_profiles/camp_profiles_ucb.html</a> (for admitted, and the enrolled was about the same -- I'm finding the average for admitted, not the ranges)</p>
<p>Privates, on the other hand, will take the best single sections, and so the one section is not brought down. This means that the single section ranges won't be skewed because some students got low on one section but higher on the other two for one sitting; they'll take the two high sections on that one, and the highest score in the other section from another sitting, for example. Thus, 50th-75th percentile range of a given range is not pushed further than 100 points or so. Also what contributes to the skewing is that Berkeley does not emphasize the SAT as much (they are more willing to admit someone who scores high in two sections but not so much in another, etc.).</p>
<p>And notice USC's SAT ranges. They differ by about 100, they superscore, and the implicit average is virtually the same as Berkeley's.</p>
<p>
[quote]
That 2070 is actually the average of admitted students.
[/quote]
</p>
<p>Actually I wasn't talking about that. According to the link above from the UC site, Berkeley's average for admitted students was a 2040 or so.</p>
<p>And if that were true, it would definitely not make sense for the average to drop nearly 100 points. The differences between admitted and enrolled are simply not that high. The average for admitted is much more comparable.</p>
<p>
[quote]
By the way, how did you get that email? I thought you said you have no tie with Berkeley.
[/quote]
</p>
<p>I said I'm not a student at Berkeley. I have "ties," as in friends who go there.</p>
<p>
[quote]
In fact, you can run a spreadsheet using a hypothetical scenario of 1000 students and assume that 1/3 with 1 sitting (no effect from superscore), 1/3 improving in both sections upon second sitting (no effect from superscore), and 1/3 benefiting from superscoring. For every two of the former, it takes 90 point-difference from the last one to result a bump of 30 on overall average.
[/quote]
</p>
<p>A few problems with this: where in the world did you get 1/3? I am pretty sure that that isn't true. The College Board says that half of the students who take the SAT take it again. I would wager that at schools like Berkeley, more than half of the admits took the SAT more than once.</p>
<p>In addition, the average increase is 10-15 points on each section. I would also wager that Berkeley's admits on average increased more than that.</p>
<p>
[quote]
I think the primary reason their students have wider range in test score is that Berkeley admission emphasizes more on GPA than test scores.
[/quote]
</p>
<p>Why don't you explain the logic behind that? My analysis is logical. Just assuming that the wider range is due ONLY to an emphasis on other factors doesn't make sense. Plus, you honestly think that a 120 or 130 range is due solely to a de-emphasis of the SAT? That's a total of about 70-80 points skewing. Yet you have difficulty believing that superscoring increases the average score by, say, 40 or so points?</p>
<p>I agree that GPA is more emphasized, but have you looked into Berkeley's admissions? Berkeley emphasizes essays as much as it does GPA and rigor of course load (they're all "very important"). In addition, standardized test scores (along with a myriad of other factors such as work experience, volunteering, extracurriculars, etc.) are "important."</p>
<p>So there's an emphasis of other things, but far from what would be classified as "state university" standards. Berkeley's standards for admissions are very much like a privates. How do you explain MIT's 97% in the top 10%? Are they adhering to state university standards? Or is that evidence of a high degree of selectivity?</p>
<p>
[quote]
Let me just make it clear I am not saying Berkeley is easier to get in.
[/quote]
</p>
<p>Then what is your point? Why do you go to such ends to argue endlessly on another thread, then continue that on another thread, then continue that on this thread? There must be a reason, yes?</p>
<p>^^ thank you for that -- I kept trying to find that (wasn't in my bookmarks).</p>
<p>If you look at it, the average increase for those who take it twice is, surprise, 41. And three times? 78 points.</p>
<p>Now, applying this to a school like Berkeley or Michigan is tricky, but I think the general idea is there.</p>
<p>^ On glance from this data, it looks like each time a test was taken, scores improved by approximately 10 points in each category....~20 points on second attempt...there seems to be some diminishing return, as would be expected.</p>
<p>
[quote]
How do you explain MIT's 97% in the top 10%? Are they adhering to state university standards? Or is that evidence of a high degree of selectivity?
[/quote]
very simple. MIT has high stats in both because it's that selective and it doesn't need to adhere to anything to achieve that.
[quote]
here in the world did you get 1/3? I am pretty sure that that isn't true.
[/quote]
relax, it's clear that was just a hypothetical guess. actually my wild guess wasn't that bad 1/3 vs 1/2. i don't see why you would expect me to know the actual number. it's very likley most people improve in both sections but i didn't do that in my hypothetical situation. i was actually making it more conservative on that end. funny how you didn't see it and just kept thinking i was fudging numbers to suit my position. </p>
<p>
[quote]
Then what is your point?
[/quote]
i've already stated my point very clearly: your theory that says berkeley SAT is about the same as schools like Tufts, Georgetown, or Cornell... is very very far-fetched. just keep going around and telling others that doesn't make it true. i tried to reconcile by saying it has different emphasis and it has higher GPA, suggesting they are all about the same in selectivity; yet you seem to be very interested in selling the idea that berkeley is more selectivity (i.e. "same" SAT but higher GPA). </p>
<p>by the way, you seem to be rather confused with when superscore actually applies. when a candidate improves in both sections (a lot of privates don't look at writing and it's very clear i have been focusing on M+V), there's nothing to superscore. so don't get all that excited about the increase shown on collegeboard site. it doesn't support your "theory" in any way. if anything, it actually weakens your argument because superscore means something only when one improves in one section but drops in another. superscore brings no benefit when one improves in both sections and i can imagine this applies to most who took twice. the fact that the increase in M and CR are almost identical (14 and 13) suggest the increase goes hand in hand for both sections for most students. the college board link shows that 86% of the students took it either once or twice.</p>
<p>
[quote]
MIT has high stats in both because it's that selective and it doesn't need to adhere to anything to achieve that.
[/quote]
</p>
<p>Double standard, no? (You're letting your own assumptions change your explanations; you should let the data mold your ideas, instead.)</p>
<p>
[quote]
relax, it's clear that was just a hypothetical guess. actually my wild guess wasn't that bad 1/3 vs 1/2. i don't see why you would expect me to know the actual number.
[/quote]
</p>
<p>I didn't. But to try to make an even casual argument based on such an unfounded number would be pointless. Thus, I thought you had a point. If you didn't, you could have just said so. =p</p>
<p>
[quote]
your theory that says berkeley SAT is about the same as schools like Tufts, Georgetown, or Cornell... is very very far-fetched.
[/quote]
</p>
<p>The data presented, as well as the logic behind it, makes sense. (You act as though your proposals in the past posts, even recent ones, are not far-fetched in the slightest.)</p>
<p>
[quote]
by the way, you seem to be rather confused with when superscore actually applies. when a candidate improves in both sections (a lot of privates don't look at writing and it's very clear i have been focusing on M+V), there's nothing to superscore.
[/quote]
</p>
<p>No, I'm not confused. I'm talking about all three sections in my explanation.</p>
<p>
[quote]
so don't get all that excited about the increase shown on collegeboard site.
[/quote]
</p>
<p>Again, the above. And the data seems to support my theory. (Not prove it, of course, but support it.)</p>
<p>
[quote]
if anything, it actually weakens your argument because superscore means something only when one improves in one section but drops in another.
[/quote]
</p>
<p>If there are two sections. But there are three, you see.</p>
<p>
[quote]
superscore brings no benefit when one improves in both sections and i can imagine this applies to most who took twice. the college board link shows that 86% of the students took it either once or twice.
[/quote]
</p>
<p>Again, there are three sections. And on top of that, superscore is applied to scores that all improved. The university is still looking for the highest single sections. That they are all in the same sitting doesn't matter -- it's the highest sections that matter. (That's the whole point of superscoring -- that sitting doesn't matter.)</p>
<p>The impact of superscoring is focused on those whose superscored SAT would be higher. This would bring up the 25th percentiles (and to an extent the 75th ones). Again, refer to my explanation above of the effect that not superscoring has on the ranges. In order to understand it, look at all three sections.</p>
<p>
[quote]
superscore is applied to scores that all improved. The university is still looking for the highest single sections. That they are all in the same sitting doesn't matter -- it's the highest sections that matter. (That's the whole point of superscoring -- that sitting doesn't matter.)
[/quote]
</p>
<p>You are still confused. Let's try this exercise:
First sitting: 670 V, 680M; second sitting 700 V, 700M. So what does Berkeley record? 1400. What does superscoring school take? 1400. It's the <em>same</em> for both. The only time superscore makes a difference is if the second sitting gets something like 700 V but 660 M, which is less likely to happen since the whole point of retake is to improve in both sections or at least to maintain one and improve on another. Please don't bring in the writing section to try to make your theory look more plausible. You brought up that theory before there were even 3 sections. When you were trying to say Berkeley SAT was about the same as top privates, you meant M+V, not three sections because most top privates don't even care about writings or report them. Last I checked, US News still use M+V. </p>
<p>anyway, the following should finally clear things up:</p>
<p><a href="http://statfinder.ucop.edu/statfinder/default.aspx%5B/url%5D">http://statfinder.ucop.edu/statfinder/default.aspx</a>
1. Freshmen 2. Fall 2006 3. All students 4. Berkeley 5. Click all 6. Complex table 7. Type of high school, Highest of SAT/ACT</p>
<p>shows how stats profile for in-state and OOS candidates differs (score below is average of M+V from the highest sitting):
the profile of applicants/admits/enrollees from CA goes like this:
200-499 12%/4%/5%
500-599 27%/13%/16%
600-699 42%/40%/43%
700-800 20%/44%/36%</p>
<p>note how 12% of applicants have 200-499 and 39% have below 600. it means the group with under 600 is almost twice the size the group with 700-800. also note that 21% of the enrollees scored less than 1200 in their highest sitting. i don't think schools like georgetown or tufts would have 1/5 of its students body with less than 1200 on SAT. </p>
<p>now look at the out of state applicants:
200-499 3%/1%/3%
500-599 14%/3%/8%
600-699 45%/22%/31%
700-800 38%/74%/58%</p>
<p>The 58%/31%/8% split is similar to that for schools like Georgetown or Williams, except the 3% in the 200-499 range which probably comes from the recruited athletes. It's no surprise the out-of-state applicants for Berkeley is very much like the top privates. The caliber of OOS applicants/admits/enrollees is what one would expect for Berkeley. on the other hand, berkeley, being a public that needs to take 90%+ from instate, is sorta "stuck" with the in-state pool with lower average and wider range of test scores. this explains the difference i've been talking about. </p>
<p>this should bring the debate to closure, IMO.</p>