Mike Castle might lose the Delaware Republican Senate primary Tuesday night. That’s what a new poll suggests, anyway. The Public Policy Polling (PPP) survey shows "tea party" choice Christine O’Donnell holding a three-point lead over Castle, 47 to 44 percent.
That’s a shock. Representative Castle – he’s currently Delaware’s lone House member – is a Delaware political icon. He’s twice been elected governor and has long had high positive ratings in a state where voters in the past have valued longevity and personal contact.
So what’s going on?
One possibility is that Castle’s possible fall, and Ms. O’Donnell’s possible rise, reflects the depth of voter anger against incumbents and the power of the tea party brand.
Castle’s favorability ratings among state Republicans have taken a steep drop, according to PPP. A month ago the firm put his favorable rating at 60 percent among Delaware GOP voters. The new poll has a comparable figure of only 43 percent. Behind this slide seems to be a growing feeling among Republicans – 55 percent, according to the poll – that Castle is too liberal.
O’Donnell, a marketing consultant, has been running hard to Castle’s right. Her ads, for instance, talk about the “Obama-Castle agenda.” She got a surprise endorsement from tea party hero Sarah Palin last week, and perhaps that’s helped her. Only about 30,000 Republicans are projected to vote in tomorrow’s primary, and that means anything can happen. Palin-fueled enthusiasm, plus $250,000 in ads from the Tea Party Express, could propel conservatives to the polls and send Castle home.
After all, national polls show that right now political enthusiasm is almost directly related to partisanship – the most excited voters are conservatives, the next most excited are moderate conservatives, and so on, all the way down the line to unexcited liberals.
The other possibility here is that the PPP poll that has everyone in Delaware so excited is an outlier.
PPP is a respected firm, and side-by-side comparisons have found its polls to be among the more accurate in the nation. But it also uses automated phone calling to do many of its surveys, and it’s not clear whether that aspect of the Castle survey might have affected its results. It’s more difficult to make sure your poll sample reflects the characteristics of the population you are trying to model when you use automated polling equipment. You can’t adjust quickly to get the right number of young people, the right number of middle-class voters, and so on.
Plus, the new Castle poll attempted to gauge the opinion, not just of Delaware Republicans generally, but of Delaware Republicans who are likely to vote. This “likely voter” screen is notoriously difficult to get right, since the pollster in essence is making judgments about which combinations of personal characteristic make it more likely someone will actually get up and walk, drive, or bike to the polls on Election Day.
PPP is also often described as a Democratic-leaning firm. That should make little difference in the way it goes about framing surveys. But it is worth noting that Democrats would be overjoyed if Castle lost in tomorrow’s primary.
Delaware as a whole is not a conservative state, after all, and Castle presumably would do much better with the state’s overall electorate than with the subset of Republicans. At the Five Thirty Eight blog, polling expert Nate Silver predicts that Castle, if he wins the primary, would have a 95 percent chance of beating Democratic nominee Chris Coons.
If O’Donnell wins the primary, she would have only a 17 percent change of winning the general election, according to Silver.