Tuesday, October 28, 2014

Rankings Everywhere

Like many (theoretical) computer scientists, I spent far, far too much time this weekend thinking about rankings.  Well, really, it was more that I was being far, far too amused by the various posts and comments on the novel ranking site "Ranking of CS Departments based on the Number of Papers in Theoretical Computer Science",  which led to posts at the Computational Complexity blog and in theory, the first of which shut down the comments rather quickly, and the second of which is at this time nearing the amazing 100 comment line.  (Luca's post deserves it -- it was the funniest thing I've seen all week.  Kudos.) 

Now, to fan these flames to further untold heights, I was informed that new US News and World Report Computer Science Rankings are out.  Needless to say, I disagree with them.  How exactly can Harvard be second to MIT?  Ridiculous.

(For those who are concerned about such things, the methodology is also given, though I haven't checked to see what parts of it are reproducible.  And what is that "Web Of Science" thing?  Is it as good as the citation tools in Google Scholar?)

For the ultimate in ranking tongue-in-cheek-iness, and for those who haven't seen it, I also recommend the NRC Deranker, which will (eventually) provide you the ranking you personally agree most with. 


11 comments:

Anonymous said...

You are naughty Michael ;) You could not hold your temptation to write another post on ranking. Postpone the STOC deadline by one more week. This is too much distraction.

Michael Mitzenmacher said...

I admit, this ranking post is simply because the others have been so much fun to read, I don't want it to stop. Though it appears others (in particular those participating) have had more than their fill...

Anonymous said...

Well, Luca has shut down the comments in his blog, so probably those will be spread to your blog!

To start fueling the fire, let's hear your comment on the theory ranking :-)

http://grad-schools.usnews.rankingsandreviews.com/best-graduate-schools/top-science-schools/computer-theory-rankings

Anonymous said...

To help people to invent, prove/disprove new theory of ranking, here is a summary of the lessons we have learnt.

1. Never ever post a ranking in which your school stands high. UCSD beware!

2. Ranking is worthless. People are obsessed with worthless stuffs.

3. Write papers with 4 or 6 authors but never with 5. 5 is evil, it can suddenly reduce your SODA count from 27 to 8.32.

4. Posting anonymously is our birth right. Boaz cannot take that away from us.

5. If you are not from top-4 schools, be prepared to face selection bias in faculty hiring. Luca thinks this is bullshit citing UC Berkeley hired him from University of Rome.

This is a blog post by n anonymous, so the post should only count as 1/n.

Finally, a contribution from this blog: postpone STOC deadline due to war of rankings.

rweba said...

I found it more amusing than anything, although I am sure the people involved did not appreciate the more personal attacks.

I tend to agree with Boaz and Luca, that the number of papers metric is too easily gamable and probably already has too much influence. You can't completely remove human judgement or reputation out of the loop.

Perhaps a good idea is to take the average of the two rankings ... After normalizing for number of faculty, number of co-authors, etc ...

Or perhaps it is better to just not take rankings TOO seriously. Yes, MIT is probably a better place to do theoretical computer science than the University of Florida (to pick a random example), but is it really that much of a better place than Stanford, Berkeley or Princeton? My feeling is that a student or professor could more or less do equally well at any of these places, even if some are "lower ranked" than others.

Anonymous said...

To Anonymous@ 5:47 PM:

- I don't get your point 3. Where does the number 5 comes from?

- For point 5, I thought the argument was that unless you are a superstar and you are interviewing top school, you probably will face selection bias.

- I strongly agree that we should postpone the STOC deadline. There is a war going on!

Luca said...

Sariel's deranker is by far the best ranking system that I have ever seen, although it has the drawback that it takes exponential time to converge to the correct ranking.

Unfortunately, an average case analysis performed over a random user shows that this is best possible.

Anonymous said...

To continue on the hunorous side, more things we learned:

1. US News reputational rankings are perfect except when they are not, which only proves that reputational rankings are perfect.

2. Don't experiment with new ideas, instead we'll all just pretend that rankings don't exist.

3. If a star researcher at a top university has no need for rankings, then no one else does either.

4. If your first ranking attempt fails, quit and never try again.

5. When debating, if all else fails just point out that the comment was anonymous.

6. A single departmental ranking would be enough for people to profoundly change the way they do research.

6. If we were to have a paper count at STOC/FOCS ranking then people would try to get many papers there, whereas presently they don't.

7. If only I had a dime for every time I've heard someone say "not another STOC paper again, where did it all go wrong?"

Anonymous said...

CS rankings are terrible. I can not believe that Harvard is better than UC Berkley or Stanford in computer science.. no way..

Michael Mitzenmacher said...

Anonymous #9: Quit trolling. Of course Harvard is better. :)

Anonymous said...

they forgot to add a "0" behind "2" while ranking Harvard. Just because Harvard has other good departments, does not mean it has the best CS or Engineering department. Harvard engineering sucks even more. This USNews ranking is a complete garbage. Even TCS authorities admit that :)

Similarly when someone says, hey! Stanford theory group is great comparable to MIT, I just can't stop laughing. Do you even know how many active theory researchers does Stanford have compared to MIT? Go figure.