When pollsters interpret results for a race involving an incumbent, pollsters have typically applied the rule that incumbents rarely get a higher percentage in the election than they receive in polls, and that voters still undecided on the very last poll tend to "break" disproportionately for the challenger.
I wrote quite a bit on the incumbent rule before the 2004 election, but then the results at the presidential level defied the rule. I considered some of the reasons for what appears to be a recent weakening of the incumbent "rule" in this post in August, 2006.
These two posts from 2004 best explained the "incumbent rule:
Also, click on this link to display a listing of all posts concerning the incumbent rule on Pollster.com (in reverse chronological order - most recent first). This link display posts on the incumbent rule from the blog MysteryPollster.