After seeing how well these chaps at UMich Admissions did in the EA round at my D’s high school, it has me wondering “How did they do it?” That is, how did they make so many smart decisions—almost too smart somehow—and do it so quickly?
Then, I reviewed my notes about what UMich Admissions said about their admissions process, took another look at the unusual Naviance chart for my D’s high school, stumbled across the surprisingly high freshman retention rate and overall graduation rate percentages for UMich, and the fog started to clear.
Now, I cannot prove a thing here; this is pure speculation on my part. But, I think UMich has a fairly sophisticated predictive data model under the hood that assists in admissions decision making. Consider:
– They collect and process a lot of granular data on each applicant (recalculating GPAs, etc.);
– They obviously know how the students who matriculate from a particular high school perform once they get there;
– As admissions people go, the UMich staff are unusually data and metrics focused (just listen to them talk about metrics the next time you get the chance);
– And because of their applicant volume, they will almost certainly have significant ongoing investments in upgrading their admissions databases and workflow technology.
Now, given the number of deferrals that they have out of EA, I don’t think that they have it completely wired down. Yet. But there sure are some data science fingerprints visible from where I sit. The schools that are still doing all the admissions tasks manually are at a disadvantage, I think.
There is speculation that umich defers applications they don’t get to but I think that only applies to out of state because they get 4 times more out of state applicants than in state applicants. Umich says they use a holistic review and they do have an evaluation sheet posted on their admission website which shows you how each of the 3 readers evaluates each students application giving a final grade at the end.
@Eeeee127 I think they absolutely do all the things you mention. However, I think they are doing a whole lot more.
Consider, they actually do have the data to determine which high schools are the top 1,000 high schools in the country as measured by the common denominator of the Michigan curriculum. If true, this would mean that they would know quantitatively know how much better school #25 is from school number #275. They would be able to make a prediction that they ought to admit prospective student A from school #25 with GPA 3.4 SAT S1 and Rigor R1 to LSA over student B from school #275 with GPA 3.8 SAT S2 and Rigor R2. And, if their models are really good, they would be able to set the line between admit and defer for EA not based on how many apps they processed, but rather on the calculated strength of the admit pool and the probability that each admitted student will matriculate.
Why else would they be asking about grandparents who attended Michigan? I’ll wager that “1 or 2 grandparents who attended Michigan” is actually a statistically significant variable for predicting yield–and that they know the coefficient in the yield equation for that variable.
Yeah legacy is def something they consider especially for out of state applicants because the yield is only 25% for out of state whereas it’s more 75% for in state. The only reason umich acceptance rate dropped from 50% to 28.6% since 2010 was because of the huge increase of out of state applicants. Even though they consider each high schools rigor, I don’t think they will admit someone with a 3.4 gpa unless that was the highest gpa in their school. Umich average recalculated gpa is 3.87 but they may admit ppl with 3.7 Gpa if their school is really rigorous but I don’t think they will admit anyone with less than a 3.6 gpa unless they are recruited athlete.
Yes they do but 3.4 gpa is too low regardless of situation. Ppl who come from crime infested areas and r first generation may have more room for a lower gpa like 3.5 at the lowest but not any lower unless they r recruited athlete. U were comparing two different students with the same high gpa but if the poor one had a 3.4 and the rich one had a 3.9 who would admissions admit? even questbridge finalists have an average 3.8 something gpa even though they faced hardships first generation and came from family incomes of only 35k @KneeDeepClunge
I think it is useful to compare Naviance charts between schools to help get a sense of what the admissions departments are up to.
These charts have a familiar look to them–green boxes in the upper right hand corner (admits), a thin band of purple (deferred), and a sea of red (denied) across the rest of the page. What differs between schools is the relative size of these regions. So, no surprise, the green boxes on Dartmouth’s Naviance diagram are concentrated in a tiny region in the upper right hand corner where the highest GPAs and SAT score applicants live. On Kenyon’s chart, the green region is a tiny bit larger. For large universities like Penn State, Ohio State, and Wisconsin, the green region is larger still. There are also the odd data points here and there for the recruited athletes and legacies. But, the key take away is that there is a general pattern across all of these schools. It suggests the admissions departments at schools with “conventional looking” Naviance charts all have a broadly similar, let’s call it a traditional admissions decision-making process, albeit with different applicant pools.
So, the dog that didn’t bark in the night is UMich’s Naviance chart for my D’s high school. It looks nothing like a conventional Naviance chart. It looks more like a block of swiss cheese that has been expertly carved up. There are green boxes where you would least expect them–places on the chart that the competition had ignored. And, there are red xs where you would least expect them–places where their competition had a field of green. I ask myself, “What kind of decision process leads to a ‘swiss cheese’ Naviance diagram?” Now, one possibility is that they just don’t know what they are doing. But, I have met these people. They do not give me the impression that they do not know what they are doing. Indeed, they strike me as the most logical, disciplined, and structured admissions office out there. So I am going with the hypothesis that these chaps at UMich have some extra information that their competitors do not. Remember, they have 55,000 apps they have to get through, and they really, really cannot afford to mess around. Hmm. What might that extra information be? Then I take another look at that chart and realize – wow, these guys at UMich have a ton of data on the kids who apply and ultimately matriculate at UMich. Bing! Does anyone remember the Brad Pitt movie “Moneyball”? In that movie, Brad Pitt’s character Billy Beane used a quantitative approach to help make scouting decisions for his ball club (great movie). I suspect that UMich has made some quant tools of their own, and the evidence is the swiss cheese Naviance diagram for my D’s school.
I would have to say being a legacy absolutely carries a lot of weight in the EA decision process just based on who was accepted and who was deferred from the 2 high schools in my town. The only people who got accepted were either legacies or were enrolled in our district’s special science research program. If you did not have either one of those credentials it didn’t matter if your GPA and ACT scores were higher and you had more extracurriculars, you were still deferred.
@jbtcat r u in state or out of state? legacy doesn’t care much weight for in state applicants. Many in state without legacy like me were accepted early action. My friend had a 3.2 gpa and even though she had legacy she didn’t bother applying. Legacy only gives u a small boost but your stats have to be in the range. Of course legacy is considered more for out of state because many 75% out of state ppl who r accepted dont even choose umich so what’s the point of accepting them early action. If out of state ppl have legacy there’s a much higher chance they will enroll.
Perhaps it is simply that UMich goes fishing OOS in fishing holes where they have had a track record of generating both good candidates and good yields. For example, suppose that a high school generates 100 apps/year, UMich accepts 20, and 10 matriculate for a 50% yield. That might be a good OOS fishing hole. If so, that might also explain @jbtcat’s observation as well. That is, suppose it were true that getting accepted to UMich OOS is a function of both the individual candidate and the historical productivity (from UMich’s perspective) of the fishing hole that the candidate applied to Michigan from. In hat case, it would be much harder for a candidate from an historically lower-yielding fishing hole to stand out from the rest of the applicants.
@RustyTrowel , your Naviance chart for Michigan sound’s like my son’s (in-state feeder) chart for HYP. On the other hand, his HS Michigan Naviance chart is a solid block of green in about half of the upper right quadrant. All it shows is that Michigan is using an holistic admissions process for OOS, and a more relaxed GPA/ACT based process for in-state. It’s a pretty easy decision considering the low tuition cost for in-state families which results in the extremely high in-state yield.
I thought Urbana Champaign is easier to get into than umich engineering out of state? Also isn’t it easier to get into the uc schools like uc Davis and San Diego than umich out of state? Also agree that many in state students at umich come from Michigans top high schools. Only 4.5% of Michigan high school students go to umich. @KneeDeepClunge