Equations or ideas to rank potential professors

[coffeecoffeebuzzbuzz]

Ok, time for a bit of number crunching geekery here...

So grad school admissions committees get to compare students (at first glance at least) via numbers like GPA, GRE score, credits taken, and class standing. Any ideas on an equivalent number system to rate potential profs?

What equation, metric, or self designed indices do you use to rank grad school options and to find a mentor who is a good match? How does this rating system influence your decisions on where to apply and where to go?

[qbtacoma]

As all other methods of ranking have come under sharp criticism, I'd like to get the ball rolling. While number of publications may be tempting as a solid standard of comparison, I would caution that it could privilege male over female academics since women are the ones who not only physically create the children (and rightly need time off for that) but also have their time drained in order to take care of children. Generally even men who value parenthood devote less time to their children, because they aren't expected to give more and so they can.

So, with that in mind, here are a few ways to "measure" professors:

Number of publications (with caveat above)

-Especially number of publications in high-impact journals

Number of years in the field (which will hopefully translate to grad students as more networking opportunities)

Books produced vs. articles

Number of accolades (named chairs, teaching awards, etc)

For the sciences, number of NSF or other important grants earned

Number of existing grad students

Number of former grad students who went on to find jobs in academia (or wherever you want to work)

Honestly, except for the two metrics I'm not taking any of these into account. I'm choosing a lot of younger academics and academics whose work I like and publication productivity isn't terribly important to me as a grad student. I'm going to be looking at are what existing grad students say about a professor's personality (in order to weed out the dysfunctional ones).

[newms]

As all other methods of ranking have come under sharp criticism, I'd like to get the ball rolling. While number of publications may be tempting as a solid standard of comparison, I would caution that it could privilege male over female academics since women are the ones who not only physically create the children (and rightly need time off for that) but also have their time drained in order to take care of children. Generally even men who value parenthood devote less time to their children, because they aren't expected to give more and so they can.

So, with that in mind, here are a few ways to "measure" professors:

Number of publications (with caveat above)

-Especially number of publications in high-impact journals

Number of years in the field (which will hopefully translate to grad students as more networking opportunities)

Books produced vs. articles

Number of accolades (named chairs, teaching awards, etc)

For the sciences, number of NSF or other important grants earned

Number of existing grad students

Number of former grad students who went on to find jobs in academia (or wherever you want to work)

Honestly, except for the two metrics I'm not taking any of these into account. I'm choosing a lot of younger academics and academics whose work I like and publication productivity isn't terribly important to me as a grad student. I'm going to be looking at are what existing grad students say about a professor's personality (in order to weed out the dysfunctional ones).

Those are all great metrics. For fun they're always http://www.ratemyprofessors.com/ which has a 'hottness' (yep, that's how they spell it) factor.

[coffeecoffeebuzzbuzz]

Peer Impact Ranking...

I created an odd ranking system based on the number of times a "Potential Prof of Interest" was cited by other scientists in peer-reviewed manuscripts. I like this approach over the number of publications system because this ranking includes books and high impact papers that really influenced others within a specialized subject area. I also like to subtract the one paper with the highest citation number (some folks are one hit wonders...). To compare young prof who are up and coming to older established profs I then accounted for the number of years since the person was established in their field of study (with a grad degree).

Here is the equation:

(Total number of citations - maximum number for a single manuscript) / The number of years since the prof earned a PhD

As an easy way to tally the numbers go to Google Scholar, advanced search. Enter the exact name of the prof (specific first initials and last name) in quotation marks in the "Author" section. Next restrict the subject area (eg biology). Also limit the results page to only show "at least summaries". One interesting spin is to also compare citation of "all time" to specific time ranges (like just the number in the past three years). If you really want to geek out you can also look at the amount of variance in the number of citations also.

Anyhow this helped me to get a feel for which potential profs were rock stars, and which were so-so (at least in terms of peer impact...)

[katerific]

Those are all great metrics. For fun they're always http://www.ratemyprofessors.com/ which has a 'hottness' (yep, that's how they spell it) factor.

Here's my current equation:

Professor Score = 100^h * (other things, like actual credentials and whatnot)

Where hotness, h, is defined

{ chili pepper on RMP: h = 1

{ no chili pepper on RMP: h = 0

Notice: non-hotness will not hurt their score, but hotness will significantly amplify the score.

I'm still working on the (other things) part, but I clearly got the most important part figured out already.

P.S. I jest.

[coffeecoffeebuzzbuzz]

Hmmm....well here is my score (with hotness factor)

Prof. 1 - 100^1 * 53

Prof 2 - 100^0 * 69 - not hot

Prof 3 - 100^1 * 158

Prof 4 - 100^1 * 118

Prof 5 - 100^1 * 62

Prof 6 - 100^1 *76

Prof 7 - 100^1 * 103

Prof 8 - 100^0 * 102 - not hot

Prof 9 - 100^1 * 23

Well... it still a tight race given how many hot profs are on my list... any other metrics to consider?

Edited January 8, 2011 by coffeecoffeebuzzbuzz

[qbtacoma]

Peer Impact Ranking...

I created an odd ranking system based on the number of times a "Potential Prof of Interest" was cited by other scientists in peer-reviewed manuscripts. I like this approach over the number of publications system because this ranking includes books and high impact papers that really influenced others within a specialized subject area. I also like to subtract the one paper with the highest citation number (some folks are one hit wonders...). To compare young prof who are up and coming to older established profs I then accounted for the number of years since the person was established in their field of study (with a grad degree).

Here is the equation:

(Total number of citations - maximum number for a single manuscript) / The number of years since the prof earned a PhD

As an easy way to tally the numbers go to Google Scholar, advanced search. Enter the exact name of the prof (specific first initials and last name) in quotation marks in the "Author" section. Next restrict the subject area (eg biology). Also limit the results page to only show "at least summaries". One interesting spin is to also compare citation of "all time" to specific time ranges (like just the number in the past three years). If you really want to geek out you can also look at the amount of variance in the number of citations also.

Anyhow this helped me to get a feel for which potential profs were rock stars, and which were so-so (at least in terms of peer impact...)

Wow. That is a great idea.

[waddle]

(Total number of citations - maximum number for a single manuscript) / The number of years since the prof earned a PhD

As an easy way to tally the numbers go to Google Scholar, advanced search. Enter the exact name of the prof (specific first initials and last name) in quotation marks in the "Author" section. Next restrict the subject area (eg biology). Also limit the results page to only show "at least summaries". One interesting spin is to also compare citation of "all time" to specific time ranges (like just the number in the past three years). If you really want to geek out you can also look at the amount of variance in the number of citations also.

Brilliant!

Hmm, can we assume that a citation is an event for which the occurrences are randomly distributed across time, and therefore follows a Poisson distribution? But then we'd have to account for time passed since publication, since papers get cited less as time goes on, and for the increase in researchers since the '60s ... oh man, this is becoming a complicated statistical problem ... any statisticians care to help us out?

P.S. One of my potential advisors has 404 publications according to Google Scholar. Not counting the most highly-cited paper (1500+), his 'C²B² index' is 160--and that's only counting the 25 most-highly cited papers! Yeah, I don't think I'm going to go through them all ...

P.P.S. coffeecoffeebuzzbuzz, I hope you don't mind me calling your index the 'C²B² index' ...

P.P.P.S. I found a pretty awesome metrics program, aptly called Publish-or-Perish, that can do some of this for ya ...: http://www.harzing.com/pop.htm

[repatriate]

Here is the equation:

(Total number of citations - maximum number for a single manuscript) / The number of years since the prof earned a PhD

This sounds a lot like the h-index: http://en.wikipedia.org/wiki/H-index

There are some tools (Scopus, Web of Science) that automatically calculate the h-indices for different fields. The linked Wikipedia article includes some of these.