Comment by _delirium | Hacker News Reader

Comment by _delirium | Hacker News Reader

Parent Full thread

I haven't found academia exceptionally hierarchical in computer science (though I'm not full-time in it at the moment). Even quite famous people tend to be pretty accessible, people take your talks at conferences seriously regardless of what university you come from, etc. There is definitely some hierarchy, but if anything it's precisely because of the "capitalist structure" you mention, which big universities are definitely not immune to. There are some power differentials especially around money, with grad students needing to work for a well-funded lab to get paid, which gives power to professors who have lots of money (depending on the area, that usually comes from either grants or corporate funding).

In fact at AI conferences, the non-academics are way more hierarchical about academia than the people actually in academia are. Journalists, investors, recruiters, and industry consultants who show up to a conference like AAAI or ICML seem to really care about people's pedigrees, especially impressed by names like MIT and Stanford. Professors are more likely to choose what talks to go to based on sub-area, talk title, or knowing a specific person, rather than going by institutional affiliation.

I have hovered in EE, CS, and math circles. I find CS and math to be made up of more genuinely curious folks than EE. But this comment and my parent comment are just a few points of observations, too few to generalize from.

But, once you get into the grants and peer review game, connections and credentials, unfortunately, matter a lot. It is not explicit in that no one says "hey you are from XYZ university, hence I am rejecting your paper." What happens commonly is that for peer-reviewers opinions are colored by your credentials. My hypothesis is this: if one were to control for quality of paper, a credentialed or pedigreed author's paper is likelier to get accepted than the non-credentialed ones.

There are many exceptions and there are deliberate programs like EPSCOR (for grants) to prevent such things from happening, but the inequality exists nonetheless. The inequality is sometimes justified, sometimes not.

I obviously can't speak for other fields, but at least in my field, "peer review" means that the people reviewing your papers are from all kinds of academic (and sometimes industry) settings; it's not like it's just a bunch of MIT professors criticizing all submissions and discriminating based on the authors' credentials. For instance, here is a program committee that I'm on right now for a top conference in my sub-field ... you can see for yourself which institutions people come from:

http://www.educationaldatamining.org/EDM2016/committees.html

The chances of your paper being reviewed exclusively by "top credential" people (whatever that means) is astonishingly low; reviewers come from all sorts of institutions around the world. Even if someone exhibited an "elitist" opinion in discussions, they would be summarily shot down by the majority of people on the committee, who are not from top-ranked institutions. Again, N=small personal anecdotes, but in paper/grant review meetings I've sat on, I've never seen this to be an issue simply because most members are not from top-ranked institutions.

The same is true for NSF grant reviewing in the U.S.. Again, different funding agencies have different models, but the NSF is a peer-based model just like papers.

I didn't claim anything about the peer reviewers' credentials! I said that if a person is from a big shot pedigree or from a top school, it certainly helps in a situation where a paper or proposal is as good. It is a common flaw in academia to pay too much attention to one's pedigree or affiliations. In my experience it is quite common to hear people say admiringly, "XYZ is Big_Shot's student" and so on. It is sometimes justified to attach importance to pedigree, sometimes it's not.

... though I will concede that one legitimate source of inequality is due to larger institutions simply having more financial and material resources, so they can perform more "high-impact" sorts of research. e.g., if I'm at a small school without a $100m piece of state-of-the-art lab equipment or a staff of 20, i can't do certain kinds of research and thus can't publish certain kinds of Big Science sorts of papers as easily.

Reply on news.ycombinator.com