Correlation between ranking and selectivity.

[Mechanician2015]

I was recently reviewing some of the stats cited to build different "top graduate schools" rankings, and I realized that the link between a program's ranking and it's selectivity is not as direct as one might think.

 

For instance, I was comparing Cornell(purported #8) with Duke( about #30) and their selectivity is very similar( 12% and 19% ).

 

Aside from "research fit", this seems to coincide with anectodal evidence of people claiming to gain admission in "much better" or "better ranked" programs while being rejected by others.

 

Any thoughts on this topic?

[random_grad]

Some lower-ranking schools might have low % because very many people used them as safety school but there s still only that many people whom they can admit; while not all apply to the top schools (if you know your app is not that great why pay the fee) so it evens out.

I got into some top schools but not into the lowest ranked to which I applied.

Edited April 24, 2015 by random_grad

[victorydance]

Cohort size has a pretty big effect. Obviously if a lower ranked school has a smaller cohort size, but the number of applicants remains similar or even less than higher ranked schools the selectivity may actually be greater at these schools. 

 

As a general rule, the higher ranked schools often have larger cohorts than lower ranked schools. What makes schools higher ranked in any given field is more to do with their resources, professors, and pedigree not necessarily their selectivity in the admissions process. 

[TakeruK]

Another way to phrase what random_grad said is that the same population of applicants do not apply to all the schools, so there's another reason you cannot directly compare two school's acceptance fraction to measure their "selectivity".

 

In addition, different schools have different criteria. As an international student, I got into some top private schools but rejected from many top and second-tier public schools since international students cost a lot more at a public school!

 

Overall, I do not think "selectivity" (as measured by fraction of applicants accepted) is a terribly useful metric and I would only expect weak correlation between prestige and "selectivity".

[shinigamiasuka]

There isn't quite. I only applied (for MS only) to the universities I found my research interests (and location) match and got accepted to two of them: UMich and UIC. I wasn't keen on applying to too many places. Both of these are very different in terms of their ranking and selectivity (UIC was actually a safety school, along with KAUST in Saudi Arabia), and I don't have a stellar profile to impress any of UMich, Purdue, or OSU for a PhD; yet UMich took me in for Master's (and some others with GPAs lower than mine ).