One of the key competitive House races in the country is Florida's 22nd, held by incumbent Republican Rep. Clay Shaw for thirteen terms. Shaw is opposed by Democrat Ron Klein in one of the most expensive House contests of all time. Our observer in the Sunshine State, The Florida Masochist notes the latest poll in the South Florida Sun-Sentinel:

Shaw would get votes from 48 percent of those polled and Klein has 43 percent with 21/2 weeks until Election Day. With a margin of error of plus or minus 4 percentage points, the race statistically is a dead heat.

Even though the incumbent is widely known in the Broward-Palm Beach county district, he is being backed by less than 50 percent of likely voters. Eight percent reported they hadn't made up their minds. Undecided voters often move toward a challenger in the closing days of a campaign.

Read the rest, and TFM's comments, at the link above.

Notice also how the media reports this. Not to pick on the Sun-Sentinel, because it is all too common among the fMSM to simply report the lead when the Democrat is ahead, but to emphasize the margin of error if the Republican is leading.

Another bit of "conventional wisdom" is that the late deciders tend to go to the challenger. It isn't necessarily so. For example, in 1976 the late movement toward President Ford was so strong that both Gallup and Harris (the only national polls at the time) said if the election had been a few days later, Ford would have certainly won.

More recently, the Connecticut Democratic Senate primary comes to mind. Polls consistently showed Lamont with a strong lead - double digits in some polls. Yet it was Lieberman, the incumbent, who made the strong push at the end, closing to within 2.5% in the balloting. See the Mystery Pollster for details.

Perhaps this "wisdom" stems from 1994, when the late deciders overwhelmingly chose Republican challengers over Democratic incumbents. What most pundits have forgotten about that year is the overriding factor in turning the House. It wasn't reaction to Hilarity!'s health care plan or The Contract with America that made the difference with most voters. It was the House banking scandal, which revealed that dozens of House members, mostly Democrats, had been using the House Bank to kite checks without penalty, a sort of free loan from taxpayers. This led to many Democratic retirements, giving the GOP many open seats to challenge, and a weakened group of incumbents, many of whom lost.

Now, THAT was a "Culture of Corruption." And it illustrates just WHEN the "late deciders trend to the challenger," too: it is when the incumbent is damaged by scandal. Otherwise, they often go back to the familiar incumbent.

So, this means that Republican incumbents who have been touched by scandal - fairly or unfairly - are less likely to get the late deciders, although those who've "kept a clean nose" have at least a equal chance at the late deciding voters. This may not bode well for GOP incumbents like Sherwood, Weldon, or Burns, but it means that those like Shaw in Florida-22 have no reason to fear the "undecided" votes.

One last point: as Alexander McClure noted the other day:

Anytime the "undecided" portion exceeds twice the margin of error cited for the poll, that poll is garbage.

The Sun-Sentinel poll has a MOE of 4%, and shows 8% undecided, so by Alex' rule, it is "on the cusp" of being garbage.