| 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 |
« Workshop on Random Network Models | Main | Lok on "Bayesian Combination of State Polls and Election Forecasts" »
1 February 2009
Burt Monroe, Michael Colaresi, and our own Kevin Quinn have written an interesting paper (forthcoming in Political Analysis) assessing methods for selecting partisan features in language, e.g. which words are particularly likely to be used by Republicans or Democrats on a given topic. They have also provided a dynamic visualization of partisan language in the Senate on defense issues between 1997 and 2004 (screenshot below).
The most striking feature coming out of the visualization is that language on defense went through an unpolarized period leading up to 9/11 and even for several months afterward, but that polarized language blossomed in the leadup to the Iraq War and through the end of the period they examine, with Republicans talking about what they thought was at stake ("Saddam", "Hussein". "oil", "freedom", "regime") and the Democrats emphasizing the process ("unilateral", "war", "reconstruction", "billions"). (Link to visualization, a QuickTime movie.)

Posted by Andy Eggers at February 1, 2009 8:36 AM