Live data from Hacker News

Most US professors are trained at same few elite universities

nature.com

171–180 of 207 posts

Re: Most US professors are trained at same few elite universities

#171
post #156

Earlier quoted context omitted.

Have you been on a university hiring committee? My anecdotal evidence is that yes prestige very strongly matters and coming from a very prestigious university often outweights other factors such as publication record. I haven't read the original article yet but one can easily test if e.g. the publication record of a graduate student from a non-and prestigious uni differ to the extend that is indicated by the hiring.

All else being equal eg publication record, a candidate A from a less prestigious place is ranked higher than candidate B from a more prestigious place. This is because the candidate A did the same stuff with much less support than B. The issue is that prestige is highly correlated to research outputs, so it is difficult to disentangle.

While it maybe should be that way it definitely is not in my experience.

In my experience if you have done a postdoc/PhD at e.g. Stanford people will look much less thoroughly at your publication record. I've actually argued in committees that a candidate from a less prestigious school should be reranked higher than a candidate from a prestigious school, because they had a much better publication record. The interesting thing was once I brought this up in the discussion, most people agreed. However in their own ranking they had put the person from the prestigious school higher, which indicates to me that this is an unconscious bias.

Re: Most US professors are trained at same few elite universities

#173
post #149

Earlier quoted context omitted.

> Harvard is really good at finding the "Harvard quality" people among applicants. How hard is that sifting process? How good is Harvard at doing that? How was it tested? I certainly think Harvard could have made an incoming class out of rejected students and they would all be the same caliber.

This study is talking about doctoral training not undergraduate. Selection is very different.

I guess I assumed that doctoral training would be more likely to draw from a fully qualified pool than undergraduate studies.

Re: Most US professors are trained at same few elite universities

#174

Earlier quoted context omitted.

There is some truth to the heuristic that "if you got into Harvard" you have to be pretty good. Because Harvard is really good at finding the "Harvard quality" people among applicants.

Maybe I'm jaded but I've met a lot of brilliant people in my life. They come from all sorts of backgrounds. I've often in turn seen how those backgrounds can become safety nets, not equally distributed. The benefit of the doubt is extended very far, and when you have knowledge on the other side of the curtain it starts to look self-fulfilling sometimes. It's hard not to look at scatterplots of things like SATs, GREs,…

> It's hard not to look at scatterplots of things like SATs, GREs, and GPAs, and see a lot of randomness. Not pure randomness but it starts to be hard to justify to yourself when correlations of .1-.3 begin to be interpreted in the general public as if they were .9.

Do you have a link? Be keen to read about this.

Re: Most US professors are trained at same few elite universities

#175
post #94

Earlier quoted context omitted.

Totally anecdotal, but I've been in a number of tenure-track hiring meetings at multiple universities and someone always brings up the idea of "pedigree". As in, "X has a good pedigree" etc., as if that's an argument for hiring the person above and beyond their track record. So at least in my personal experience, these things do enter into the process. Of course, the process involves a lot of other variables too; the…

There is some truth to the heuristic that "if you got into Harvard" you have to be pretty good. Because Harvard is really good at finding the "Harvard quality" people among applicants.

That's true. "Harvard quality" means either:

- Very smart

- Connections to wealth and power

A bit over 2/5s of the white undergrads are admitted for the later. That came up in the Asian law suit. Similar mechanisms apply elsewhere, but we haven't had legal discovery (note: not speculation; insider knowledge).

Re: Most US professors are trained at same few elite universities

#176
post #156

Earlier quoted context omitted.

All else being equal eg publication record, a candidate A from a less prestigious place is ranked higher than candidate B from a more prestigious place. This is because the candidate A did the same stuff with much less support than B. The issue is that prestige is highly correlated to research outputs, so it is difficult to disentangle.

While it maybe should be that way it definitely is not in my experience. In my experience if you have done a postdoc/PhD at e.g. Stanford people will look much less thoroughly at your publication record. I've actually argued in committees that a candidate from a less prestigious school should be reranked higher than a candidate from a prestigious school, because they had a much better publication record. The interest…

I guess it depends on the publication record. But a bad publication record from someone who came from a prestigious place gets an extra negative from someone who came from an average place.

On the other spectrum, someone from nowhere solving something big makes huge headlines (mathematician from unknown place solves millennia-old problem). At that level it doesn’t matter where you came from, you’re in.

It’s the muddy middle that becomes difficult to assess. Two people who have published decently but not great. Nothing remarkable about them. Then prestige may factor in to their benefit.

Re: Most US professors are trained at same few elite universities

#177
post #144

Earlier quoted context omitted.

The problem is that hiring committees are busy, have imperfect information and therefore rely on easy metrics, i.e. prestige of candidate's university and journals where he has published. Many universities are now signatories of the SF Declaration, where they promise to value research based on ideas, not on prestige of journals: https://sfdora.org I have noted e.g. UMass goes one step further and asks tenure-track ap…

How does that work if anyone actually reads the publications? They'll have the journal name and probably current institution of the authors, right? Am I being naive with the question? Is reading the publications unusual?

It would be very interesting to have that studied... but my general impressions (as somebody who does read the publications) is that the answer is usually no.

Re: Most US professors are trained at same few elite universities

#178
post #153

Earlier quoted context omitted.

>If pedigree was a useless signal, it wouldn't be used. The notion of good pedigree benefits anyone who might fit in that category, so they're incentivized to convince people that this matters. For example, if someone on a hiring committee sees an applicant coming from their alma mater, they can call that a "good pedigree" and reinforce the value of their own background.

I mean, only if it actually is a good pedigree. I used to work at a company where one of the HR directors went to Harvard for undergrad. He could say that and people probably would have nodded their heads in agreement. You can't say that if you went to Ohio State. Nothing against the Buckeyes but it's just not in the same tier, to the point where that "pedigree" -- I can't even imagine anyone referring to their Ohio…

The problem though is that exceptional people are scattered all over the system, since 1) not everybody can get into Harvard (class size) and a significant percentage who do are selected from a small monied elite. The number of non-Harvard graduates is much, much larger.

So statistically it is actually far more likely that the Harvard person is relatively mediocre to a neutral sampling of the entire system.

Re: Most US professors are trained at same few elite universities

#179
post #58
post #46

Earlier quoted context omitted.

> less prestigious schools that graduate a ton of professors, like UC Berkeley How in the world can UC Berkeley be "less prestigious". It constantly ranks world top-10 in all rankings.

I was thinking ivy league. smaller, more selective schools.

Sure, but as a non-American I see it as more prestigious than e.g. Dartmouth, at least in STEM. Maybe some employers would prefer the Ivy based on selectivity though?

Re: Most US professors are trained at same few elite universities

#180
post #75

Earlier quoted context omitted.

> Would be like saying that 50% of all Americans who become teachers come from just 20% of the states - but not adjusting for the fact that 50% of the population lives in the top 10 largest states. There's data here for number of doctorates awarded by school in 2020: https://ncses.nsf.gov/pubs/nsf22300/data-tables #1 was... Walden University. Which I'd never actually heard of before. It's a private for-profit online…

The list you cite undermines the article's claim about elitism because it looks like 4 out of the 5 schools that it lists are also on your list. So the conclusion is that the 5 schools that produce the most professors...include 4 of the top 10 schools in terms of PhD graduates. Not much of a surprise there!

You can't just look at supporting evidence and ignore contrary evidence. That's profoundly unscientific thinking. What about the 6 schools — the majority! — on my list that are not elite?
Post reply on HN