Isn't College Admission supposed to be Getting Less Competitive?

Parents Forum
281

“In 2012 1,666,017 seniors nationwide took the ACT. In 2016, that number grew to 2,090,342, a 25 percent increase. If there was no score inflation, one would expect the number of students scoring a 30+ would also increase by 25 percent. Instead, in the same time period, the number of students scoring a 30 or more increased by 47 percent, so there definitely has been some change in the curve that inflates the number of high scorers.”

I think this came up earlier, not sure, but the demographics, at least in California has changed over the last 30 years, with more Asians applying and more of them taking the ACT. Typically Californians took only the SAT as it’s a SAT heavy state, but now many applicants (non-Asians as well) are taking each test once and submitting the better. Or retaking the one they did better on and submitting those. And California being the most populous state, will impact the national averages. And Asians typically won’t take the test more than any test twice.

282

@zinhead, I agree there’s some score inflation with the ACT. A 95th percentile score in 2016 doesn’t mean quite as much as a 95th percentile score in 2012 because the 2016 percentiles are calculated against a denominator that is inflated by state mandates requiring all HS juniors or seniors to take the ACT. This results in a lot of non-college-bound HS students taking the test, and presumably these are disproportionately low scorers. In short, 95th percentile as measured against all HS seniors is not as impressive as 95th percentile as measured against only college-bound HS seniors.

Nonetheless, your own figures suggest that more than half the increase in the number of 30+ scorers is not due to score inflation. So comparing ACT scores is not “meaningless,” though the change should be interpreted in the context of a changing pool of ACT-takers.

And it is not the case that every school has seen its 25th-75th percentile ACT scores rise. Some examples:
University of Minnesota-Duluth 2011 22-26 / 2016 22-26
Minnesota State Mankato 2011 20-24 / 2016 20-24
Bemidji State 2011 19-24 / 2016 20-24
Western Michigan University 2011 20-25 / 2016 19-25
Central Michigan University 2011 20-25 / 2016 20-25

Here in the Midwest, ACT scores are pretty much flat at second-tier state universities, but increasing at the flagships. .

283

I want to second bclintok on this. Across the midwest, the ACT range of the most notable privates and best known state flagships have gone up, while most of the rest have stayed flat. It is why general ACT creep doesn’t seem to make sense. A really clear case is the significant increase in the range at Michigan, and the flat range at MSU. I also think OSU has increased, while other state Us in Ohio have been flat. And we know Chicago, Northwestern, and Notre Dame have all increased. I have the impression most typical privates in the same geographical area have been generally flat.

284

BC – great data showing how things at the “top” (however defined) are working differently from the rest of the market…

Question – what does the data for the 2nd tiers look like if you go back a little further?

From 2011-2016, The U only went up slightly – 25/30 to 26/31. The big jumps happened earlier.

Could be that many/most schools got more competitive during the 90s and 00s. That’s when the overall pool was getting so much bigger (primarily because increasing percentages of HS kids starting going to college).

But that the competition in the last five years has flattened out. And (because of the flight to quality) you’d expect the competition to fade starting from the bottom and moving up.

285

@bclintock -

Well, at least that is something we can agree on, and the amount of score inflation is easy to measure. The following table has the same base data presented earlier. The difference here is that the % of Total column is the percent of test scores compared to the entire population of test takers (2012=1,666,017; 2016=2,090,342). The last column calculates the percent change between 2012 and 2016.

Score____2012% of Total2016% of TotalRelative Change in % of Total
36_______7810.05%2,2350.11%128.08%
354,4570.27%10,9930.53%96.58%
349,604 0.58%18,7450.90%55.56%
33__14,864 0.89%25,0311.20%34.72%
32__21,438 1.29%31,1441.49%15.78%
3128,154 1.69%37,2431.78%_5.43%
3036,675 2.20%45,9142.20%_-0.22%
30+115,973 _6.96%171,305__8.20%_______17.73%

As shown earlier, there has been a large increase in the number of 33+ scores. What is surprising is the relative decline in the percent of test taker who score a 30 or 31. So while the percent of test scores who scored more than 30 increased by nearly 18 percent, there was a relative decline in the number of 30 scores and a large relative increase in the 32+ scores as a percent of all test takers. The ACT has changed the curve between the two time periods, and have goosed high test scores upward. When you see schools with 25/75 percentiles in the low to mid 30’s, what is driving the increase in test scores is the change in the curve in addition to the larger pool of test takers.

The following table has the same data but calculates the change based of the cumulative number of test takers.
The “Cumulative” column lists the total number of test takers who scored at or above the test score for the row. The “% of Total” column represents the percent of test takes score at or above the test for the row out of the entire population of test takers.

Score2012Cumulative%ofTotal___2016Cumulative%ofTotal__Change
367817810.05%2,2352,2350.10%128.08%
354,4575,2380.31%10,99313,2280.63%101.28%
349,60414,842__0.89%18,74531,9731.53%__71.69%
3314,86429,7061.78%25,03157,0042.73%52.94%
3221,43851,144__3.07%31,14488,1484.22%__37.37%
3128,15479,2984.76%_37,243125,3916.00%26.03%
3036,675115,9736.96%_45,914171,305__8.20%_____17.73%

It shows the same trends from the earlier chart but tabulates the cumulative effect of the change in the curve.

The overall average composite ACT score in 2012 was a 21.8. This increased to a 21.9 in 2016, so there was no test score inflation for the universities you cited because they were not impacted by the shift in 32+ scores.

286

Do you have evidence for the statement that the ACT changed the curve? Without any evidence, it would seem at least as likely high stats kids were prepping more or better.

287

Many 2nd tiers post Common Data Sets going back only a few years. The data may be available somewhere, but I don’t really have time to hunt for it. But here are some schools’ data going back as far as they have CDS data publicly posted:

University of Minnesota-Duluth: 2009 20-26 / 2011 22-26 / 2016 22-26
Western Michigan University: 2006 20-25 / 2011 20-25 / 2016 19-25
Northern Michigan University: 2005 20-25 / 2011 19-25 / 2016 19-25
University of Wisconsin-Eau Claire: 2005 22-26 / 2011 22-26 / 2016 22-26
University of Wisconsin Stevens Point: 2006 20-25 / 2011 21-25 / 2016 20-25
University of Northern Iowa: 2010 21-25 / 2016 20-25
Ohio University: 2002 21-26 / 2011 21-26 / 2016 21-26
Miami University (OH): 1998 24-28 / 2006 24-29 / 2011 24-29 / 2016 26-31

Overall remarkably flat, except University of Minnesota-Duluth which had an uptick in its 25th percentile score between 2009 and 2011, and Miami University which had an upward bump at both the 25th and 75th percentile levels between 2011 and 2016. Clearly Miami U is more selective than the other schools in this sample, and perhaps doesn’t deserve the “second tier” tag; if I’m not mistaken, it was on the original list of “public Ivies.”

The contrast with the public flagship is these states is striking:

University of Minnesota Twin Cities: 1996 22-27 / 2001 22-28 / 2006 23-28 / 2011 25-30 / 2016 26-31
University of Wisconsin Madison: 2001 25-29 / 2006 26-30 / 2011 26-30 / 2016 27-31
University of Michigan: 1998 25-30 / 2001 25-30 / 2006 26-31 / 2011 28-32 / 2016 29-33
Ohio State: 2003 23-28 / 2006 24-29 / 2011 26-30 / 2016 27-31
University of Iowa: 1999 22-27 / 2001 22-27 / 2006 23-27 / 2011 23-28 / 2016 23-28

Minnesota, Michigan, and Ohio State show sharp increases at both the 25th and 75th percentiles, with the biggest increases coming between 2006 and 2011—but continued upward movement after 2011. The upward trajectory at Wisconsin is flatter but still discernible. Iowa shows only a slight and slow upward trajectory. In general, the flagships have strengthened the academic credentials of their incoming classes not only at the top but also at the bottom, while the second-tier publics have remained more or less constant.

288

@bclintonk

FWI - obviously this is a across all schools and not just competitive colleges:

https://www.wsj.com/articles/in-reversal-colleges-rein-in-tuition-1500822001?mod=e2tw

289

Just like the impression that “colleges are getting less competitive…” the notion that “tuition inflation has been tamed…” is also not quite applicable to elite/top colleges. Take Yale for example: 2007-8 tuition: $34530 and 2017-8 tuition: $51400. That’s 49% increase in the last ten years, three times the inflation.
Again, its the middle class that is getting squeezed. For the family making $250k in a Ca suburb where average house costs $1.5m, sending a kid to an elite private college like Yale (whose entrance is getting tougher each year) is increasingly out of reach. Pretty soon CC’s top schools are only reserved for either 1%ers or hooked kids.

290

For comparison, these supposedly “middle class” numbers are about triple the median income of San Jose, and about one and a half times the median house price in San Jose.

As usual, “middle class” here on these forums seems to mean something like top 5% excluding the top 1% in income.

291

“As usual, “middle class” here on these forums seems to mean something like top 5% excluding the top 1% in income.”

Even though the percentiles are way above 50 nationally, for purposes of this discussion, that really is an extremely reasonable definition of “middle class.” The median family income at Harvard, after all, is $170k. $190k at Duke; $200k at Dartmouth; $229k at Georgetown; $272k at WUSTL.

High end colleges that don’t do much merit aid are quite income stratified, which is not surprising. Those colleges often have full pay percentages of 40-50% of enrolled students. Writing checks totaling $275k per kid (the cost of a mid-range Ferrari) is obviously not something many families can do.

In fact, at the richest schools (HYPS) a “median” income kid ($50k) gets a full financial aid ride. From the fancy school’s perspective, they are correctly treated as extremely “poor.” For this purpose, it is completely irrelevant that statistically the Census Bureau would define $50k as “middle class.”

This layer of schools is particularly tough for the 2% ($300k) to 10% ($150k) donut holers, since they are mostly/entirely above where the financial aid peters out. Few of these folks are shopping for Ferraris. Or full paying at WUSTL. They are driving Volvos and taking merit money deals.

https://www.nytimes.com/interactive/2017/01/18/upshot/some-colleges-have-more-students-from-the-top-1-percent-than-the-bottom-60.html?mcubz=0

292

Indeed @northwesty if Harvard has median family income at $170k and 50% full pay then almost all of families above $170k are full pay. That’s is $280k after tax money and $500k before tax income.
If the question is “is college tuition supposed to get less inflation?” The answer is a resounding no.