Rankings are awful. They give the impression of constant distance between the ordinal values, e.g., #10 is as far below #1 as #310 is below #300. In reality, things are much more tightly clustered. There are about 20 truly superior schools, and they are much closer to each other than a 1 through 20 ranking would imply. There's not much difference between a 20 and a 10 or even 1.
Then there's about 100 second-tier very, very good schools that are similarly clustered. After that, something like 1000 schools are good, solid schools with little tangible difference between number 900 and number 300.
Ordinal rankings throw all of this out the window, and students & families end up worrying about choosing between a school in the top 25% or top 27% etc.
Non-ordinal ratings instead of rankings would be better, and better show this sort of clustering, but it would still ignore extremely important differences. For example, there may be a school that's fairly middle-of-the-road, okay but not great on an all-around basis, but their proximity to New York City makes their performing arts programs a stand-out top 5 program in the region. But all you get from a US News ranking is that they're rated #768.
"They give the impression of constant distance between the ordinal values, e.g., #10 is as far below #1 as #310 is below #300."
I don't think most people interpret rankings like this. It's generally understood that the distribution of large ranked lists is not linear, and that there's more room between the entries at the top than in the middle.
No, with respect, you're incorrect. They do interpret rankings like this. My job is in higher Ed analytics. I have quite some years in it, and before that was my primary job I spent a decade of working in college admissions offices. My first role was in a 50/50 position handling technology and analytics needs, and the other half in customer-facing operations. My comment here is base on many hundreds of conversations with families going through the college selection and application process. This is indeed how most of them look at rankings. "Why should we choose your school when school X half an hour away is ranked 10 positions higher?" This is often the thinking.
Yep: In US News global rankings, Cambridge is #9 and UIUC is #59. Is Cambridge really 6x better? Is Cambridge as far above UIUC as UIUC is above Brown, at about #100?
This is often a problem with Ordinal variables, they don't convey the true distance between the "observations". Think of class rankings in High School: The Valedictorian and Salutatorian (2nd place) almost always have practically identical academic records: 4.0 GPA, challenging courses, etc. The difference can come down to hundredths of a point, or the fact that #1 took 8 AP courses while #7 took 7 AP courses and 1 Honors course. There isn't much meaningful distance between them.
Then there's about 100 second-tier very, very good schools that are similarly clustered. After that, something like 1000 schools are good, solid schools with little tangible difference between number 900 and number 300.
Ordinal rankings throw all of this out the window, and students & families end up worrying about choosing between a school in the top 25% or top 27% etc.
Non-ordinal ratings instead of rankings would be better, and better show this sort of clustering, but it would still ignore extremely important differences. For example, there may be a school that's fairly middle-of-the-road, okay but not great on an all-around basis, but their proximity to New York City makes their performing arts programs a stand-out top 5 program in the region. But all you get from a US News ranking is that they're rated #768.