How accurate are college chance calculators like Cappex and Unigo? If you’ve used them, have their predictions turned out to be correct?
They 're super accurate.
Not at all.
For any college which practices holistic admissions, they aren’t very accurate. If the college depends mostly on GPA and testing, they’re better, but since they do not incorporate the rigor of classes, they can also be inaccurate.
They do, however, tell you how you compare, academically, to the students who attend the college.
There are also large differences between schools. So if you want to consider GPA/SAT scattergrams, use the ones on Naviance, which are created using the stats of students from your high school.
Their biggest weakness, though, is that they compare you to students who have been accepted. They do not compare you to students who have been rejected. The fact that your stats are better than 90% of the students who have been accepted doesn’t tell you how many students with your stats were actually accepted. So a student with a GPA of 3.95 and an SAT of 1580 may have better stats than 90% of the students who were accepted to Yale, however, they won’t tell you that 80% of all the applicants with those stats were rejected.
The calculator algorithms (at least some of them - Prepscholar for example) do take acceptance rates into account. So, two schools with identical stats for accepted students but different acceptance rates will yield different results. This gives a decent rough measure of “reachiness” in situations where the applicant in question is similar in other ways to the majority of applicants.
The problem comes when there are subgroups with different levels of competitiveness. For example, some public universities are much harder than others for out-of-state students to get into. Some have vastly different acceptance rates for Early Decision than for Regular Decision. The calculators won’t reflect those factors at all. If a particular major or division is tough to get into, that won’t be reflected. And so on.
And then, as stated above, there are all the factors not reflected in raw stats - whether the applicant attends a grade-deflated high school, whether the applicant has a “hook,” whether they are advantaged or disadvantaged by geography, how test-optional admissions come into play, and so on.
So, the calculators give a useful baseline but it takes more information to determine how the applicant may need to adjust expectations based on other factors.
I’d say they are relatively accurate in aggregate, but I don’t think individual percentages mean much.
MIT, for example, provides acceptance percentages by tier for several tests and GPAs. So given a GPA and test scores, it can probably calculate a somewhat accurate aggregate percentage.
But what does this mean? You have high stats, so rather than a 7% chance, you have a 15% chance. So?
There’s no way to factor in all of the other components that go into admissions. If you have those test scores but no ECs, poor essays, etc., probably not a 15% chance. If you have those numbers, have two IMO Gold Medals, great essays, and parents/grandparents as alumni, you’re probably a bit higher than 15%.
It just tells you the overall rate of the entire spectrum of students with just those 2-3 data points you put in.