METHODOLOGICAL ISSUES IN THE STUDY OF NEW PARTIES' ENTRY AND ELECTORAL SUCCESS
Peter Selb and Sandrine Pituctin
Party Politics, 2010
https://kops.ub.uni-konstanz.de/xmlui/bitstream/handle/urn:nbn:de:bsz:352-144777/selbparty.pdf?sequence=3
Forecasting Congressional Elections
Carl E. Klarner
Department of Political Science, Indiana State University, 2009 paper
The first time a collection of scholarly forecasts were published in academic journals before the election was held was in the October 1996 issue of American Politics Quarterly (Campbell and Garand 2000), although other academics had published their point-predictions before elections (Fair 1978).
Abramowitz (2006) published his congressional forecasts for the 2006 election, only two published statistical models forecasting the 2008 congressional elections were publicized. One of these was published by Lockerbie (2008) and the other by Klarner (2008).
Lockerbie predicted that the Democrats would pick up 25 seats in the House, while they actually picked up 20 or 21. The dependent variable in his model was the number of seats picked up by the Democrats. His predictor variables were survey respondents’ prospective evaluations of the economy as well as a variable measuring how long the party of the President has been in the White House. A final predictor variable measured the number of open seats, adjusted for which party is at advantage during the election year. This adjustment is done by multiplying the figure by “-1” if survey respondents judge it to be a good year for the
Republicans, and “1” for a Democratic year. The variable takes a value of “0” if the public forecasts no clear favorite. Lockerbie theorized that open seats present the biggest opportunities for an advantaged party to pick up seats since incumbents so often win.
The theoretical motivation of the Klarner (2008) House model is based on variables that have been identified by scholars in the field, such as incumbency, the previous office holding experience of non-incumbents (Jacobson 2004), the partisan disposition of the district (Highton 2000), and incumbent ideological moderation (Erikson and Wright 2005). The national level factors used (midterm penalty, presidential approval, change in real per capita income, vote intention) are close to those used by Abramowitz (2006), although he uses no measure of the health of the economy.
Pollyvote.com publicized fifteen forecasts for the 2006 House elections, seven of them purely quantitative in nature. . . Authors of six of these models fully explained the methods by which they generated their forecasts.
Most tellingly, when the proportion of seats that are competitive in an election (i.e., won by five
percent or less) is regressed on the proportion of seats that are open, only nine percent of variance is explained (analysis not shown).
In an election year that favors one party, that party campaigns heavily in opponents’ districts that are closest to swinging toward the favored party.
http://www.apsanet.org/~lss/Newsletter/jan2009/Klarner.pdf
Successful re-election strategies in Brazil: the 48 electoral impact of distinct institutional 49 incentives
Electoral Studies (2001) –
PARTY ORGANIZATION, IDEOLOGICAL CHANGE, AND ELECTORAL SUCCESS
A Comparative Study of Postauthoritarian Parties
Grigorii V. Golosov, Working Paper #258 - September 1998
http://kellogg.nd.edu/publications/workingpapers/WPS/258.pdf
