CAMPAIGNING IN POETRY: IS THERE INFORMATION CONVEYED IN THE CANDIDATES' CHOICE OF WORDS? *

We construct and analyze a database consisting of the words used in speeches made by the candidates in the 2008 Democratic and Republican Presidential Primary. We present findings in two key areas. First, we estimate which candidates are more negative by counting the number of times they mention their opponents by name. Both Obama and Clinton are among the least negative. John Edwards is 3.7 times more negative than Barack Obama and Mitt Romney is 4.7 times more negative than Obama. We also compare the speeches of the candidates to speeches delivered by famous political figures and orators such as Martin Luther King, Ronald Reagan, and John F. Kennedy. For example, we can ask whether a candidate's words are more like those of MLK versus Ronald Reagan. Using our metric, John McCain and Mitt Romney are the most like Ronald Reagan whereas (within the set of primary candidates) the words of Mike Huckabee and Barack Obama are the closest to those of Martin Luther King. Hillary Clinton is by far the candidate closest in oratory to Bill Clinton.

[1]  D. Holmes The Evolution of Stylometry in Humanities Scholarship , 1998 .

[2]  M. Laver,et al.  Extracting Policy Positions from Political Texts Using Words as Data , 2003, American Political Science Review.

[3]  Patrick Juola,et al.  Authorship Attribution , 2008, Found. Trends Inf. Retr..

[4]  James H. Stock,et al.  Who Invented Instrumental Variable Regression , 2003 .

[5]  F. Mosteller,et al.  Inference in an Authorship Problem , 1963 .

[6]  I.N. Bozkurt,et al.  Authorship attribution , 2007, 2007 22nd international symposium on computer and information sciences.

[7]  David Mannion,et al.  Authorship attribution: the case of Oliver Goldsmith , 1997 .