Friday, September 10, 2004

PCs, new folks, and new authors: some data

My post about SODA generated all kinds of comments: further discussion is in the comment threads. One point that merited another post:

A commenter complained:
I have noticed that most of the accepted papers in SODA/STOC/FOCS seem to be from one of the IVY league schools or from one of the established Labs...

Also, why in the world do we see the same set of names on program committees with little/no permutations.
I'll address both these points in some detail, accompanying my comments with statistics provided by Adam Buchsbaum (SODA 2005 PC Chair). First, let's look at the complaint about author diversity:

1. Diversity of authors:
What follows is a frequency chart of affiliations for a random sample of 27 (20%) of the accepted papers at SODA 2005:

6 max plank inst.
5 mit
4 u. illinois (u-c), tel aviv u.
3 u. bergen (norway), technion, hebrew u.
2 uc berkeley, u. waterloo, u. paris-sud, rutgers, ohio state u.
1 utrecht u., unsw (sydney), uc irvine, u. tokyo, u. leeds, u. glasgow, u. chicago, u. aarhus, stanford, simon fraser u., microsoft, kings college, it u. copenhagen, eth zurich, comenius u., cmu, christian-albrect u., charles u., acad. sci. czech republic

seems fairly diverse to me. It should not be hard to compile such stats for other years/conferences as well...

Next, let's look at the issue of recycling commitee members:

2. PC "freshness":

There is a perception that newer folks have a hard time getting onto S/S/F committees. To remedy this, David Johnson, at the SODA 2004 business meeting, introduced a list of 'neverbeens'. This list is defined somewhat roughly as
People who have never served on a S/S/F committee and plausibly could, where "plausibly could" includes things like
  • out of school for at least some small amount of time
  • some reasonable number (>= 1, more is better) of conference papers
  • regular attendance at such conferences
He also requested that people who want to be on this list should email him. He then provides this list to commitee chairs to do with as they please.

But in reality, how skewed towards "frequent members" are program committees in reality ? Let's look at some numbers in detail.

Freshness of PCs:
For SODA 2005, roughly 1/2 (my earlier comment had said 1/3) of the PC members are new i.e have not served on a SODA/STOC/FOCS committee before (disclaimer: this set includes me). For SODA 2004, the corresponding number appears to be 1/3. Both numbers are reasonable, and might demarcate extreme ends of the "right" ratio.

Historical PC composition (all years):

1. SODA
Served exactly once: 130 people
Served exactly twice: 34
Served three times: 13
Served four times: 1

2. FOCS
1 time: 140
2: 52
3: 24
4: 14
5: 9
6: 4

3. STOC
1: 139
2: 49
3: 19
4: 8
5: 5
6: 2
8: 2

Overall: (SODA U FOCS U STOC)
1: 200
2: 84
3: 55
4: 34
5: 29
6: 5
7: 7
8: 9
9: 11
11: 1
12: 2
16: 1

(My take: SODA is better than STOC/FOCS at integrating new people, but not hugely so)

Adam further points out:
Translating these statistics to reflect recidivism in filling slots reveals that for each of the three conferences, a majority of the PC slots were filled by then-first timers. For all conferences as a group, the majority was filled by 1st- or 2nd-timers.
It thus seems to me that the perception of incestuousness among
FOCS, STOC, and SODA PCs is in fact a mis-perception. Still,
we have to address the mis-perception by ongoing action, not
simply historical statistics. David Johnson's maintenance of
a list of people who have never served on STOC, FOCS, and SODA
committees and a continuing effort to include new people are
positive forces.
Data is a good thing... If you don't like the stats here, do your own digging and I will post the results here.
Post a Comment

Disqus for The Geomblog