@PurpleTitan I haven’t found the numbers in one place in an easy format, but just grabbing some self-reporting by schools:
FWIW -
Harvard: http://features.thecrimson.com/2015/freshman-survey/makeup/
(for some reason they break out averge SAT by public/Private, but they do 3 years, so I’ll use this:
Class of 2017 (so, 2013 admits)
Math Public SAT Av.: 745.4 (96-97 percentile)
Math Private SAT av.: 745.2 (96-97 percentile)
Class of 2018 (so, 2014 admits)
Math Public SAT Av.: 741.4 (96-97 percentile)
Math Private SAT av.: 741.4 (96-97 percentile)
Class of 2019 (2015 admits)
Math Private SAT av.: 753.5 (97 percentile)
Math Public SAT av.: 744.5 (96-97 percentile)
Class of 2017 (so, 2013 admits)
Reasoning Public SAT Av.: 746.4 (98%)
Reasoning Private SAT Av.: 736.7 (97-98%)
Class of 2018 (so, 2014 admits)
Reasoning Public SAT Av.: 740 (98%)
Reasoning Private SAT Av.: 744.1 (98%)
Class of 2019 (so, 2015 admits)
Reasoning Publics SAT Av.: 744.2 (98%)
Reasoning Private SAT Av.: 746. (98%)
Class of 2017 (so, 2013 admits)
writing Public SAT Av.: 738.7 (98%)
writing Private SAT Av.: 751.1 (98-99%)
Class of 2018 (so, 2014 admits)
writing Public SAT Av.: 742 (98%)
writing Private SAT Av.: 748.5 (98%)
Class of 2019 (so, 2015 admits)
writiing Publics SAT Av.: 743.9 (98%)
writing Private SAT Av.: 753.4 (99%)
I’ll see if I can get better numbers, but looking at this small 3 year window, some fluctuation but not much of a trend line. Percentage range essentially flat.
@CaliDad2020: Going steadily from 96-97 percentile to 98-99 percentile sure seems like a trend to me.
I think you’d find even more pronounced trends below HYPSM and definitely at the Near-Ivy level. Harvard was skimming off the cream then and is skimming off the cream now, but look at the improvement of the UMich or ND student bodies.
@PurpleTitan I don’t think you are reading those figure right, but it’s a very small sample anyway and pretty useless. But in those 6 catagories over 3 years you had only 1 net increase in % by .5% (in a window. Sat doesn’t break out these average scores because they don’t exist in individual test scoring.) You had 2 net point losses across the 3 years and 4 net point increases. Had we simply done 2017 to 2018 the data would suggest it was getting less competitive. We need full numbers by year from 2009/10 to present to make this useful.
But I was wondering where you got your figures as I’ve not found a good “central clearing house” yet and don’t really love picking here and there as you never know where folks pull their numbers from.
As far as trend line, I’ll line them up by catagory per year to make it easier to read:
Math Public: (class of)
2017: 745.4 (96-97%)
2018: 741.1 (-4.3) (96-97%)
2019: 744.5 (+3.4) (96-97%)
(net point decrease -.9: % flat)
Math Private
2017: 745.2 (96-97 percentile)
2018: 741.4 (-3.8) (96-97 percentile)
2019: 753.5 (+12.1) (97 percentile)
(net increase raw +3.2: +.5 %)
Reasoning public:
2017: 746.4 (98%)
2018: 740 (-6.4) (98%)
2019: 744.2 (+4.2) (8%)
(Net point decrease -2.2: % flat)
Reasoning Private:
2017: 736.7 (97-98%)
2018: 744.1 (+5.4) (98%)
2019: 746. (+1.9) (98%)
(net raw increase +10: % flat. Note that 2017 10 pts below Public. 2018/19 number “reverted to mean” perhaps.)
Writing public:
2017: 738.7 (98%)
2018: 742 (+3.3) (98%)
2019: 743.9 (+1.9) (98%)
(net raw increase +5.2. % flat)
Writing Private:
2017: 751.1 (98-99%)
2018: 748.5 (-2.6) ( (98%)
2019: 753.4 (+ 4.9) (98-99%) (Input % incorrectly in original post, btw)
(net raw increase +2.3: % flat)
Again, I think this is way too small a window and only one school, but most metrics were basically flat - although class of 2018 at Harvard are dumbells! lol…
I agree with this up to a point, but I wouldn’t overstate it. It’s still not a uniform national market. All the Ivies still get 40-50% of their entering class (or more, in some cases) from the Northeast, a region that represents just 18% of the nation’s population. And kids in the Northeast are generally much more inclined to attend private colleges than those in other parts of the country, and because so many of the elite private schools are right in their backyard, they’re essentially “staying home” when they do so. There’s also a huge state-by-state variation. In the Midwest, for example, Illinois produces about 25% more HS grads annually than Michigan, but Illinois sends roughly twice as many students to top private colleges. Many of the top stats HS grads in Michigan just don’t even bother to apply to elite private schools. They figure the University of Michigan is plenty good enough, they apply early and get their acceptance by December, and they’re done.
This is also the flaw in @CaliDad2020 's argument. He assumes that all (or at least a fixed percentage of) the top-stats kids apply to the most selective colleges. Not so. There’s a huge untapped domestic market out there of top-stats kids from states and/or other demographics in which the top HS grads don’t apply elite private colleges. Look at it this way. Suppose a highly selective college considers only the top 2-3% of high school grads serious contenders for admission.
(In fact, their 25th percentile test score medians are generally below that level, so that assumption is conservative). But 2-3% is 70,000 to 105,000 students. No elite private college get anywhere near that many applications in total—and we all seem to be assuming here that many of the application they do get fall outside that range. So that means there’s a huge number of potential contenders out there who they’re not hearing from. It doesn’t take much of a shift in the number of applicants they get from those previously (and still) underrepresented states/demographics to move the needle on competitiveness of admissions.
I’m also not buying the argument that the increasing number of ACT takers just reflects test-shopping by people who previously would have taken only the SAT. The biggest part of that increase for the ACT has come from state requirements–now in 20 state—that all HS juniors take the ACT, not even so much to increase college participation (though that may be a secondary goal), but because the states use the ACT to measure school performance. And if I’m not mistaken, the number of SAT-takers has also declined somewhat, losing market share to the ACT as more states flip from SAT-dominant to ACT-dominant. And there’s always been test-shopping. One clue is to add up the percentages who submitted scores from each test to a given school; if the total is great than 100%, you know at least that many took both tests. ( It’s not a perfect measure because some students may submit only the scores from the test on which they did better, but at least it gives us some idea). At Yale, for example, in 2006 91% submitted the SAT and 32% submitted the ACT, for a total of 123%—so we know at least 23% took both tests. In 2015, 69% submitted the SAT and 43% submitted the ACT, for a total of 112%—a much lower number. This shftt is partly due to broader national trends–the SAT has been steadily losing market share to the ACT, to the point that more students now take the ACT than the SAT. But it also might reflect a geographic broadening of Yale’s entering class. The Northeast is still heavily SAT-dominant. the South and Midwest are mostly ACT-dominant. If Yale is now enrolling a smaller percentage of Northeasterners and more Midwesterners and Southerners, you’d expect the percentage of ACT-takers to rise and the percentage of SAT-takers to fall. and this is consistent with a strengthening of Yale’s applicant pool as it taps more deeply into traditionally underrepresented regions.
BC – the market for top colleges still has a strong regional component. But the market is much less regional than before, which added a lot of kids to the pool for top college seats. The market will never be fully national. But I’m not seeing any reason why it will become significantly more national than it is now. I think that wave has pretty much happened.
Purple, CaliDad – I do think there’s been a lot more competitive growth in the top 15/20 than at the vey top. But even there, things are starting to taper off. At some point, it is mathematically hard to have stats continue to increase absent large new increases in applicant in the top 20 pool. But the stats are still inching up and not yet coming down.
@bclintonk still very little from anyone in terms of real, comprable numbers.
You can look at Ivy apps for instance and see for 6 of 8 or something (going from memory, could be 5/8) there was a drop in apps from 2014 - 2106)
I think absent real numbers this discussion (on my side as well) is just how we feel and anecdotes.
@northwesty - again. 2010 is the date. There is no question, I don’t think, that pre-2009/10 college was getting more competitive, both because of rising domestic #s and international. Vandy from 2010 - 2017: ACT rises 2 pts at bottom. 1 pt at top (love to know the average but obviously it’s fairly close to 32/33.5 - but be interesting to know where it falls.) while ACT test taking increase by like 20-25%.
(additionally those numbers are clearly rounded in some way. A true average would be more nuanced.
And we need to decide what schools we are talking about. I think it’s pretty likely that the Vandy, UWashSTL and USC of the world continue to get more competitive, or at least find ways to seem more competitive, but I’m still unconvinced at the top end. But I really don’t know:
Hard to tease out what affects what without more numbers.
@CaliDad2020:
“I think it’s pretty likely that the Vandy, UWashSTL and USC of the world continue to get more competitive, or at least find ways to seem more competitive, but I’m still unconvinced at the top end.”
Well, your initial statement was “i think it’s most likely that college admissions competitiveness has bascially flatlined since 09-10 or thereabouts.”
So are you now conceding that competitiveness has increased outside of “the top end” (which is HYPSM? Certainly above the Vandy/WashU/ND/UMich level)?
It is not clear what these numbers include. Do they include kids taking it multiple times in one year? Does it include non-senior taking the test? In our state, gifted kids in middle school take the SAT and ACT. Are these scores included in the curve? One would expect a general increase in the number of high test scores if you have more people taking the test which has been the case with the ACT.
After reading through the thread, it seems that college admissions have gotten more competitive largely due to the increasing pool of kids applying that were not part of the system before. These include:
1 - Colleges enrolling an increasing number of internationals students who were previously outside the system.
2 - Changes in admissions that expand the number of students with hooks. Princeton’s renewed emphasis on first-generation students (going from 6% to 15% in 10 years) is a good example of this. That is 9% of the class that is crowded out by first generation kids that previously would not have previously considered Princeton.
I said that is probably true a bunch of posts ago (although I still haven’t seen any place with good numbers) But my guess is that the “2nd tier” of top schools has continued to get more compeitive.
But I think it’s pretty clear that the fastest rise was pre-2009 and it has leveled off since then.
This is not to say it’s now “easy” or anything. It’s still incredibly difficult. But for the top schools, my guess is it has leveled off.
Obviously a decent % has to be “dual test takers” - I would guess a certain amount of any small rise in top school scores has to do with kids that previously would “only” have taken SAT or ACT now “test shopping” to see which they score higher.
Also, the schools themselves calculate their own SAT/ACT ranges differently: some take only best single sitting, some take the superscore, some seem to average all reported, so it’s hard to tease that out as well, especially if a school has changed it’s calculating method at some point. (most aren’t very transparent, far as I’ve been able to find out.)
Absent someone collecting good, consistently calculated numbers it’s going to be hard to give the “increased” competition a value. Even increasing 1st gens at Princeton can overlap with URM TAGs but might be expressed as unique admits. Can’t say without seeing how they collect and calculate.
But raw numbers suggest that if nothing else the “rate of increased competitiveness” has certainly leveled off, and I’d bet, in reality, if we could find some way to test it, kids applying in 2010 found it about as competitive as those applying last fall.
It occurs to me that in this whole discussion on drivers of selectivity nobody’s mentioned US News, which initially rewarded blunt measures of selectivity in its rankings. They don’t anymore, but the measures they use are still pretty blunt, and have IMO fueled rather an arms race amongst elite and aspiring colleges toward **appearing/b highly selective.
@CaliDad2020: I think the fastest rise in competitiveness is still going on, but it is moving down the food chain.
While no one besides a special few had been able to count on HYPSM as a safety for generations, 20 years ago, an “average excellent” kid with top 1 percentile stats could count on UChicago/Northwestern/Cal as safeties.
Around 10 or so years ago, that stopped being true, but that type of kid could still count on UMich/USC as safeties.
Around 5 years ago, that stopped being true about UMich/USC but they could still count on Tulane/American as safeties.
These days, even Tulane/American can’t be considered safeties.
Note that the increase in competitiveness is concentrated at the national research U level.
Small schools with a gender imbalance or in rural or depressed areas have not seen a rise in competitiveness, but most people on CC don’t concentrate on the Ohio Wesleyans or WPI’s of the country.
How much of the increase in test scores is due to students taking the test multiple times or overlapping the SAT/ACT? 20 years ago, people frequently took one test once or twice. Now some kids, particularly high scoring kids, take each four or five times. If the kids churn the test, you will see higher test scores with no change in actual competitiveness, placement or ranking; the kid that got a 30 in 2010 will be the same kid who scored a 32 seven years later.
The only real change in competitiveness will be if the pool of kids applying is expanded, like with non-resident aliens or low income, first generation kids.
For Tulane, its 2012-2013 common data set does not say, but its 2011-2012 and 2013-2014 common data sets say that “level of applicant’s interest” was “considered”. See http://www2.tulane.edu/oair/common_data_set.cfm .
Purely on stats, American and Tulane may still appear to be “safeties” for top-end students, but their use of “level of applicant’s interest” indicates that such “overqualified” applicants cannot assume that they are actually safeties. This is a somewhat different phenomenon from the rising admission competitiveness of Chicago, USC, etc…
The focus on SATs/ACTs may be a little misguided. I think the ECs, essays, LORs that are required for admission to the selective schools have gone way up. Twenty years ago, if you won or placed at Intel or Siemens, you were admitted to the top stem programs pretty easily (MIT, Stanford, CMU, some ivies), but now many of them are not. Essays now more often than not need to be attention-grabbing (BLM, costco, papa johns) rather than display critical thinking.
Sure everything matters and matters more now. But it is pretty hard to make assessments on how much better today’s winning essays are as compared to 15 years ago.
But it is pretty easy to compare test scores over time.
Those are very few numbers and very small increase (in range btw) on a test that has increased # of takers (or unique tests taken, not sure which number was quoted) by 25% in that same period.
(and it doesn’t cull out internationals, which could easily be getting more competitive withtout affecting US based apps if the # of Internationals accepted are not increasing.)
Again, I think without a really good study with solid numbers we’re all just guessing a bit.
Certainly the idea of “safties” has changed, as the # of top-shelf applicants applying to a larger # of schools has increased. But USC and UMich, while perhaps no one is calling them “safties” are still (outside of USC CA) safe for “1%” I would guess, absent yield protect, the vast majority of “98% standardized test” students who apply seriously get accepted to USC and/or UMich.
And Tulane, according to common data, still takes 65-70% of it students from SAT 700 and below. You can’t really argue this is because they are turning down large numbers of 750 students. (2100 is 96th percetile for SAT. That’s not top 1 or 2%. It is probably, with multiple tests taken, top 5%. Still awesome, but I doubt they are rejecting 2200’s except to yield protect. And Tulane’s a great school by the way, with wonderful honors program etc. It should be a goal of elite students who want to live in NO.
Anyway, enough of this fun, I’ve avoided my crushing deadlines long enough… After I clear my desk I’ll try to find a good place to look at a wider and more precise range of #s.
“How much of the increase in test scores is due to students taking the test multiple times or overlapping the SAT/ACT? 20 years ago, people frequently took one test once or twice. Now some kids, particularly high scoring kids, take each four or five times. If the kids churn the test, you will see higher test scores with no change in actual competitiveness, placement or ranking; the kid that got a 30 in 2010 will be the same kid who scored a 32 seven years later.”
Sure some of that is going on. But that still says to me that competition is increasing. Even if the ratio of applicants to seats is constant, isn’t the market more competitive if you have to do more to stay in the same place?
If 32 is the new 30 and 4.75 W is the new 4.0, the game is still escalating.
@northwesty the game is only escalating if the percentiles are changing. If the kid now showing a 32 ACT didn’t take it 7 years ago, but showed a similar % SAT there is not change in the “pool” just a different expression of quality.
Ok, now I’m really, really going back to work… after I pay some bills, paint my nails, take out the trash… does the garage need cleaning…