Off-Topic Discussion from "Colorado College withdraws from US News and World Report rankings"

[circuitrider]

@circuitrider

From the Academic Influence website:

Our measure of concentrated influence looks specifically at the ratio of combined influence scores for faculty and alumni to size of the undergraduate student body. So defined, concentrated influence is relevant to undergraduates, with higher concentrated influence indicating greater accessibility of undergraduates to influential faculty. Concentrated Influence™ | Academic Influence

50 Best Liberal Arts Colleges and Universities Ranked for Students | Academic Influence

Interestingly, Wesleyan falls to #3 when small tech colleges (cough, Caltech) are added to the mix:

50 Best Small Colleges and Universities Ranked for Students in 2023 | Academic Influence

Hold that thought! Perhaps, it would be best to continue this string in this already off-topic thread. LOL.

@MWolf

I’m sorry, but their methodology sucks, and they don’t actually understand what they’re doing.

To quote:

Because our ranking methodology is not just objective but also driven by algorithms and databases, its results emerge from a computational process and therefore not by direct human intervention.

So my wife leads one of the largest data institutes in the country, and laughed long and hard at this.

First of all:
Every single ranking systems uses algorithms. An “algorithm” is just a set of rules that a person writes.

Algorithm for putting a giraffe into the refrigerator:

1. Open refrigerator
2. Grab giraffe
3. Place giraffe in refrigerator.

There, I wrote an algorithm. Now magically, if I program this to a computer is will put a giraffe into my computer, right? After all, it’s an algorithm, and that the same as “magic spell”, but for computers, right?

Second:
A biased person will write algorithms which produce biased results. Just because a computer follows the biased instructions doesn’t magically erase the bias inherent in the algorithm

Third:
Databases are mostly biased, and even an unbiased algorithm will have biased results,if it is used on a biased database. More generally, GIGO.

So no, just because they’re using AI or “machine learning” doesn’t mean that they are magically free of bias.

I could think of a dozen ways offhand by which their results would be far more biased than anything done by USNews.

I’m getting REALLY tired of what I will now call Computational Gaslighting.

“It’s not biased because it was done by a COMPUTER, and computers aren’t inherently biased, right?! So it’s scientific truth, and we have to accept it!”