[rintense]

He is the ONLY pollster that I've ever heard that has said the undecideds break for the challenger. He may have 20 years experience, but that doesn't mean he's right. John Kerry has 20 years as senator, and we know he is a frickin liar.

[kingattax]

undecideds generally break for the challenger in congressional races,

not in presidential elections despite what Mr "20 years" says.

[HawaiianGecko]

	Frank Newport of the Gallup organization says undecides break for the challenger.   Boston, Mass.: From a historical standpoint, where do Bush's numbers need to be going into election day? I've heard that he needs to be above 51% to offset undecides who break for the challenger. If this is the case, it seems his people might be a bit concerned at this point. Barring any major news or gaffes, I suspect we'll see these same polling numbers come Nov. 2nd. Frank Newport: Interesting point. There is a consistent conventional wisdom that "undecides break for the challenger"; i.e., that "what you see is what you get" for an incumbent. The idea is that incumbents are well known, and those who remain undecided have in essence already decided that they are wary of voting for the incumbent.

[OneTimeLurker]

I think most pollsters say they do. Zogby has and Charlie Cook has said so too. I have heard people here say they don't but I have never seen any backup.