Determining Rankings

[stackoverflow]

Hey Everyone,

I'm new here and been looking through the forums for a couple days, currently researching CS PhD programs for the Fall 2012 season. Everyone mentions tiers and rankings, but I was wondering what actually determines these rankings, or if it just an unwritten kind of system. I know about the US News rankings but I've heard from some professors to take these with a grain of salt. Are there any better resources out there?

Thanks!

[csperson]

I'd trust usnews only, but that's just my opinion.

[Eigen]

NRC Rankings are better to go by than US News and Report, imo.

That said, no ranking system shouldn't be taken without several heaping piles of salt. Rankings are exceptionally complicated in graduate school, especially being that you often find faculty that are top rank in their discipline at a program that might not be so highly regarded. And when all is said and done, who your advisor was and the quality of your work matter far more than the ranking of the school you went to for any future endeavors.

[mechengr2000]

NRC Rankings are better to go by than US News and Report, imo.

That said, no ranking system shouldn't be taken without several heaping piles of salt. Rankings are exceptionally complicated in graduate school, especially being that you often find faculty that are top rank in their discipline at a program that might not be so highly regarded. And when all is said and done, who your advisor was and the quality of your work matter far more than the ranking of the school you went to for any future endeavors.

I agree 100% with what Eigen said, but I would like to add something.

If someone is asking how the program is ranked, they usually only care about US News's ranking. That one carries the most weight since its the most popular one. All the other ones are really only known by "insiders" of academia. US News is considered "de facto." Also, if someone is asking about ranking, its because they dont know anything about the quality of the school. Keep that in mind.

Ranking of the school is the strongest indication of their reputation. It SHOULD NOT be taken as a measure of the quality of the program. Go on Yelp and read reviews of how awfully people have been treated at some of the top ranked schools. The most important thing about the PhD application process is to become "street smart" and lose the naiveity of an undergraduate. Do your own research into things and don't take lipservice as gospel.

[frenzydude]

Even after reading all the above comments, I would say trust USNews subjectwise/areawise rankings for the one rank number you are looking for. There is a ARWU ranking and TIMES (for overall univ ranking) but nothing else comes close. Of course, a match between the faculty and your interests and funding are probably more important issues than rank for a PhD program.

[edvolkov]

And when all is said and done, who your advisor was and the quality of your work matter far more than the ranking of the school you went to for any future endeavors.

If you want to go in academia after your graduate study, the ranking of your school is more important that your adviser or your work. But it pertains only to top-10 school. The placement of people outside of this top is pretty similar and depends on school ranking very slightly.

[Eigen]

If you want to go in academia after your graduate study, the ranking of your school is more important that your adviser or your work. But it pertains only to top-10 school. The placement of people outside of this top is pretty similar and depends on school ranking very slightly.

Having recently come out of watching a job search from the inside, I can say this is not true at all.... The big name of the school only helps if you actually worked with one of the well known faculty there.

And all either name will do for you is get you in the door... You still have to have the. Publications, funding and proposals to back it up.

All of the studies I've seen done end with the qualifying fact that should be obvios to anyone.... The high percentage of faculty with degrees from top schools implies a correlation, not a causation.

Also keep in mind that if you're going into academia... you'll be doing one or more post-docs, and the "ranking" of those is much more important than where you got your PhD.

Edited May 12, 2011 by Eigen

[Azazel]

Also keep in mind that if you're going into academia,, you'll be doing one or more post-docs, and the "ranking" of those is much more important than where you got your PhD.

Maybe. But if you are working in computer science, it's not uncommon to get a tenure-track position straight out of grad school.

Edited May 12, 2011 by Azazel

[Eigen]

Maybe. But if you are working in computer science, it's not uncommon to get a tenure-track position straight out of grad school.

TT at a highly regarded PhD granting institution?

If so, cool. That's a rarity/nearly impossible in the rest of the science/engineering fields.

[Azazel]

TT at a highly regarded PhD granting institution?

If so, cool. That's a rarity/nearly impossible in the rest of the science/engineering fields.

Yes. Although CS is gradually moving toward a more post-doc centric system, TT after grad school is still a fairly common career path. None of my advisors/mentors have done post-docs, and I don't get the impression that they expect me to do one either (speaking way too soon, but if I were to go into academia...).

Edited May 12, 2011 by Azazel

[Eigen]

That's cool. It's interesting to see how post-docs factor into different fields...

Chemistry usually does 1-2 post-docs 1-2 years long each.

Biology/medical fields, on the other hand, are now sitting at 2-3 post-docs, each 3-4 years long, on average. That's a LOT!

[newms]

The NRC ranking isn't thought highly of in some (most?) quarters in Computer Science. Here's a statement from Washington CS department about errors in the NRC rankings for Washington, which is almost universally acknowledged as a top-10 CS school, but NRC ranked them ridiculously low. Here's another statement from the Computing Research Association about the errors in the NRC data for Computer Science. I'd stick to the USNews rankings for what it's worth. It's far from perfect and should only be used as a guide not a definitive ranking.

Edited May 12, 2011 by newms

[csperson]

The NRC ranking isn't thought highly of in some (most?) quarters in Computer Science. Here's a statement from Washington CS department about errors in the NRC rankings for Washington, which is almost universally acknowledged as a top-10 CS school, but NRC ranked them ridiculously low. Here's another statement from the Computing Research Association about the errors in the NRC data for Computer Science. I'd stick to the USNews rankings for what it's worth. It's far from perfect and should only be used as a guide not a definitive ranking.

I agree. If you look at NRC's rankings for CS based on research productivity, it totally does not make any sense for some of the schools.

[edvolkov]

Having recently come out of watching a job search from the inside, I can say this is not true at all.... The big name of the school only helps if you actually worked with one of the well known faculty there.

It is true in your area, but not in CS. If you check up the professors in top-10 you will see that vast majority of them are having degrees from top-10 institutions.

[csperson]

Ranking certainly matters if there is a big gap between two schools.

I have studied at both a school usnews ranked somewhere in #75-#100, and a top 15 school. There is a huge difference in the quality of research, professors, peers, job opportunities, and everything between the two schools.

There was a small number of the PhD graduates from the lowly ranked school who ended up in places at Google/Microsoft. But almost nobody was able to secure a faculty position at a school ranked higher than #100 (most faculty positions were at non-PhD granting schools). A few of the PhD graduates had trouble finding *any* job for a while (yes, there were pretty much willing to take almost any job).

At the top 15 school, PhD graduates were able to get faculty positions at other top 15 schools (including Berkeley). When they got jobs at places like Google right after graduation, they got very high starting salaries. Nobody had trouble finding a good job.

Edited May 13, 2011 by csperson

[edvolkov]

At the top 15 school, PhD graduates were able to get faculty positions at other top 15 schools (including Berkeley). When they got jobs at places like Google right after graduation, they got very high starting salaries. Nobody had trouble finding a good job.

It's a bit exaggerated to say "were able". They "are able" to get position at school with similar or lower rank. But getting position at top-10 will be very difficult even for people from schools that are outside of top-10, but inside top-15.

In UCB there are only 12% percents of professors from non top-10 schools. And half of these profs are from schools outside of top-20.

If we will take something more lower-ranked from top-10 like UT Austin, there are 19% percents of profs from non-top-10 schools, quarter of them are from schools outside of top-20.

In Princeton there are 27 percents from non-top-10 schools and half of them received PhD outside of top-20.

[csperson]

There is certainly a number of professors at top schools who did not get their PhDs from top schools. But, if you look closely, most of these people are not like in their 20s or early 30s and some have non-CS degrees. A lot of them got the positions when there weren't that many CS PhD graduates, like in the 80s or earlier, or in the early 90s in some cases.

The faculty job market today is different for our generation. Just look for young people who are professors are top schools but did not get their PhDs from top schools, you will not find many.

The bottom line is, ranking matters (like *size matters*). Just go to the best school you get into.

However, there is not that much difference between the top 5-20 schools

Edited May 13, 2011 by csperson

[Eigen]

It is true in your area, but not in CS. If you check up the professors in top-10 you will see that vast majority of them are having degrees from top-10 institutions.

This may well be true, but you need to draw your opinions from people on the SCs at those top schools, not just the relative percentages of top-10 institutions among current professors. That's a correlative relationship, not a causal one.

It's harder in this case, because often the graduates of top-10 schools are excellent students in and of themselves- but unless you're hearing from SCs at these schools with how heavily they weight the PhD institution (which was what my comment was from), you can't imply a causal relationship. Otherwise, it's impossible to separate the fact that the person was a top notch candidate from the fact that they went to a top-10 school.

[newms]

Just go to the best school you get into.

As you say just below this statement, there isn't much difference between the top 5-20 schools and that's an important caveat, because while it is important to go to a school that is well regarded in the field, it probably is more important to go to a school that's a good fit for your interests. So there has to be some balance between prestige and fit. I can absolutely imagine a scenario where someone gets into a top 5 school and a top 15 school, but the top 15 school is a much better fit for the person's research interests and because of that the top 15 school is a better choice than the top 5 school.

[timuralp]

As you say just below this statement, there isn't much difference between the top 5-20 schools and that's an important caveat, because while it is important to go to a school that is well regarded in the field, it probably is more important to go to a school that's a good fit for your interests. So there has to be some balance between prestige and fit. I can absolutely imagine a scenario where someone gets into a top 5 school and a top 15 school, but the top 15 school is a much better fit for the person's research interests and because of that the top 15 school is a better choice than the top 5 school.

Good point. Also, bear in mind that overall rankings are not as helpful either. What matters is how good of a fit and how good of research is being done in the particular sub-area the applicant is interested in. Personally, I didn't end up paying much attention to rankings, since I've found they were quite deceptive in my case: the two "better" schools had either 1 faculty member interested in my sub-field or none. Going into the "best" school, I believe, would have been a worse choice in my case.

[csperson]

Good point. Also, bear in mind that overall rankings are not as helpful either. What matters is how good of a fit and how good of research is being done in the particular sub-area the applicant is interested in. Personally, I didn't end up paying much attention to rankings, since I've found they were quite deceptive in my case: the two "better" schools had either 1 faculty member interested in my sub-field or none. Going into the "best" school, I believe, would have been a worse choice in my case.

Well, go to the "best" school only if there is a big gap between the rankings (i.e. a top 15 school and the #90 school). I had a friend who got into a top 15 school and a school ranked ~90th. He chose the lowly ranked school just because they offered him more money. After graduation, he ended up getting a not-so-good job...

Edited May 13, 2011 by csperson

[achapiro]

A bit off-topic here, but which resource would you guys use to check the ranking of a non-US program?

[newms]

There are a few worldwide rankings of universities out there, such as

Shanghai Jiao Tong University's Academic Ranking of World Universities

Times Higher Education rankings (they don't have a CS specific ranking)

Quacqarelli Symonds' rankings (they don't have a CS specific ranking)

Those are three that I've come across, I'm not sure if there are others. I can't speak to the value of these rankings and some of their rankings look a bit suspect, so use it FWIW.

[OH YEAH]

Word of mouth or the "Google Scholar test" (find someone who is famous/teaches at your dream school, and then compare their publication/citation statistics to faculty members from the school you are unclear of the ranking of). After a few iterations of the Google Scholar test you can pinpoint about how good the department is, I like to think of it as Newton's method for CS rankings

[symbolic]

Yeah, in general the NRC rankings are supposed to be more accurate than US News', but in the CS community people seem to ignore the NRC rankings. The NRC rankings in general tend not to take into account conference papers; CS profs convinced them to make an exception for CS, but they executed the exception poorly. In fact, a lot of the data across the new NRC rankings has been called out, so the NRC has finally decided to start fixing its errors... after basically telling the CS community to shut up. So they said they wouldn't fix the problems with the CS rankings, but now have decided to fix their errors since profs across many disciplines have complained.

Basically, NRC shot themselves in the foot by trying to make a very complex methodology and have lost a lot of credibility since their last rankings report in 1995.

Regarding faculty from top-10 schools, this is relevant:

http://pages.cs.wisc...mni_matrix.html

Edited May 15, 2011 by symbolic