October 22, 2004
Jim Miller looks at the question of whether Undecideds do break for the incumbent President or Do Undecided Voters Break For The Challenger?
Posted by Orrin Judd at October 22, 2004 8:08 PM
Do you know of any studies that examine this phenomenon by party? I did one myself, using SurveyUSA's data for 18 elections held in 2002, where an incumbent was running. Do you know of anyone who has done a larger study with more data?
My own results from those 18 recent elections were pretty clear: undecideds break to a Republican challenger, but not a Democratic one. In fact, they just break to the R's, regardless of incumbency, suggesting that a polling bias is all that's at work here.
Given that the D's & R's have recently switched positions as to who's the dominant political party, it is possible to interpret recent trends as reflecting voters' favoritism of the challenger, I guess, but it's pretty sloppy to do so.
That sounds like what may be at work.