March 2009
Sun Mon Tue Wed Thu Fri Sat
1 2 3 4 5 6 7
8 9 10 11 12 13 14
15 16 17 18 19 20 21
22 23 24 25 26 27 28
29 30 31        

Authors' Committee

Chair:

Matt Blackwell (Gov)

Members:

Martin Andersen (HealthPol)
Kevin Bartz (Stats)
Deirdre Bloome (Social Policy)
Andy Eggers (Gov)
John Graves (HealthPol)
Rich Nielsen (Gov)
Maya Sen (Gov)
Gary King (Gov)

Weekly Research Workshop Sponsors

Alberto Abadie, Lee Fleming, Adam Glynn, Guido Imbens, Gary King, Arthur Spirling, Jamie Robins, Don Rubin, Chris Winship

Recent Comments

Recent Entries

Categories

Blogroll

Brad DeLong
Cognitive Daily
Complexity & Social Networks
Developing Intelligence
EconLog
The Education Wonks
Empirical Legal Studies
Free Exchange
Freakonomics
Health Care Economist
Junk Charts
Language Log
Law & Econ Prof Blog
Machine Learning (Theory)
Marginal Revolution
Mixing Memory
Mystery Pollster
New Economist
Political Arithmetik
Political Science Methods
Pure Pedantry
Science & Law Blog
Simon Jackman
Social Science++
Statistical modeling, causal inference, and social science

Archives

Notification

Powered by
Movable Type 4.24-en


« March 26, 2009 | Main | March 30, 2009 »

27 March 2009

Automated Text Analysis of Political Scientists

future2c.jpg

This is a volume of 100 essays by political scientists, each with less than about 1,000 words concerning one novel or insufficiently appreciated idea in some area of the discipline (edited by me, Kay Scholzman, and Norman Nie, and in honor of Sidney Verba). (I might have heard a rumor that if you buy two copies, your next article will be accepted on the first round and you'll get a great new job offer!) In any event, the last blurb above says nice things about the organization of the essays, which of course we appreciate, but I especially like that the essays, except for a little fine tuning, were ordered via automated text analysis (using an algorithm Justin Grimmer and I are working on). The number of possible orderings of 100 essays is enormous of course -- a tad less than the number of elementary particles in the universe squared -- and the idea that a "mere" human being can choose an optimal ordering is absurd. We've become accustomed to understanding that computers can do arithmetic far faster than we, but we also need to start to get use to the fact that (with some help from modern statistics) they can also "read" better than us too.

Posted by Gary King at 8:42 AM