New York 23rd Congressional District — Special Election to replace Rep. John McHugh
The Candidates: Bill Owens, Democrat
State Assemblywoman Dede Scozzafava, Republican
Doug Hoffman, Conservative Party of New York
The Polling: There has been a relative paucity of independent polling in this race, particularly given that the dynamics have shifted quickly. What the polls agree upon is that Scozzafava, once the favorite, is slipping badly; there is some disagreement about the relative standing of Hoffman and Owens. A Research 2000 poll for Daily Kos puts Owens 1 point ahead of Hoffman; two polls for conservative interest groups which had endorsed Hoffman gave him leads of 4 and 5 points, respectively. The only truly nonpartisan group to have surveyed the race is Siena Univerisity; they gave Owens a 10-point lead over Hoffman with Scozzafava still in second place and 4 points behind, but that poll is now somewhat dated.
Earlier this week, I criticized one of the interest group polls, that issued by the Club for Growth on behalf of Basswood Research. I think people are too caught up in the specifics of the criticism and not in the general principle: the incentives of an interest group to issue a poll on behalf of a candidate or position they are actively invested in are radically different than those of an independent organization. Moreover, the numbers are often profoundly different, favoring their preferred candidate by an average of 3 points versus a non-partisan poll.
This is just a huge difference: in a two-candidate race, with relatively few undecideds, a 3-point difference in support for one candidate will usually translate into a 5-6 point difference in the margin between the candidates (obviously this gets more complicated in a multi-candidate race). Considering that the long-run difference between the better pollsters and some of the worst ones is on the order of only about 2 points worth of accuracy, a 5-6 point swing is huge. Although sometimes the swing may wind up being in the right direction, the 2003 study I linked to above found that nonindependent, partisan polls made the wrong call in the race 35-40 percent of the time, versus 16 percent for the independent polls.
Basically, between their perverse incentives and their poor track record, I think these polls provide virtually no informational value versus what an informed read of the independent polls can tell you. That is, I think you are literally best off ignoring them in most circumstances; we virtually never cite them around here and they weren’t included in our Presidential or Senate race modeling in 2008 (and they likely won’t be in 2010, unless I can find some reasonable way to strip out their bias.) When I see one of these polls substantially altering the media narrative, and I don’t see the appropriate amount of hedging and caveating, I’m going to call it out.
By the way — are the Research 2000 / Daily Kos polls something that we should consider independent? The bright line that we drew during the 2008 campaign was that a poll was considered to be a nonindependent, internal poll if it was conducted on behalf of a registered candidate, campaign committee, or PAC. A poll conducted by the Club For Growth, by a union, by the DNC, or by the Scozzafava campaign would qualify as nonindependent under that criterion; a poll conducted by Daily Kos or by Fox News would not. Obviously, all polls, independent or not, need to be monitored for bias (we have generally not found much of a “house effect” associated with either Daily Kos or Fox News polls, incidentally.)
With that long diatribe out of the way, let’s talk about this very interesting Special Election.
Analysis: Suppose this were literally a two-candidate race between the Democrat Owens and the Conservative Party Candidate Hoffman. Who would win?
I think Owens would be favored. Although this district most recently had elected Republican John McHugh to the Congress, it also voted 52-47 for Barack Obama last November, and Obama has a 56-40 favorability rating there according to Siena. It is basically a middle-of-the-road district. And in a middle-of-a-road district, a moderate Democrat like Owens should beat a very conservative, uh, Conservative like Hoffman more often than not, according to a standard one-dimensional model.
But this isn’t a two-candidate race; there’s also Dede Scozzafava, although she’s fading fast. Scozzafava is a Republican. But she’s a moderate Republican in a district with a long history of ticket-splitting, and is probably much closer to Owens than Hoffman on the political spectrum. Arguably, she is even to his left; Kos (Markos Moulitsas) flirted with endorsing her. It’s not immediately clear whether Scozzafava ought to have been taking more votes from Owens or Hoffman, and therefore, it’s not immediately clear who might benefit from her collapse.
I actually think this Hoffman surge is something a little different. Undoubtedly, he has benefited from Scozzafava’s problems — although Owens should conceivably be benefiting too. But a lot of Hoffman’s votes may not be coming from Scozzafava, so much as from people who were previously outside of the likely voter universe. The Daily Kos poll asked Hoffman voters who their second choice was; only 11 percent mentioned either Scozzafava or Owens, whereas 89 percent said they weren’t sure or they wouldn’t vote. Those are very unusual numbers and suggest that Hoffman is drawing a lot of disillusioned conservatives into the race who would otherwise sit things out. Keep in mind that special elections are low-turnout affairs. Were this 2010 or certainly 2012, and there were a number of races on the ballot and a more robust turnout, I don’t think there would be enough conservatives in the district to give Hoffman much of a chance. But in the 2009 electorate, with his voters being significantly more motivated, and unspectacular alternatives being offered by the two major parties, he has quite a good chance.
Another advantage for Hoffman is that the pull on Scozzafava voters has been somewhat asymmetric. The Hoffman campaign understood early on that it needed to injure Scozzafava, whereas the Owens campaign seems to have been caught in more of a two-front war between Scozzafava and Hoffman, in part because conservative organizations in the district have been engaged in the “gadget play” (some might say “dirty trick”) of misdirecting potential Owens voters to Scozzafava.
With all that said, Scozzafava’s support is unlikely to collapse entirely. Being the only one of the three candidates to hold elected office, she’ll have engendered some loyalty among some voters in her district, and others who aren’t particularly tuned in to the national news will vote for her because she’s a familiar face or will have the Republican label attached to her name.
The Odds: This is a race that Hoffman “shouldn’t” win; he’s too conservative for his district. But sometimes, the “wrong” candidate does win, if he runs a superior campaign, motivates stronger turnout, or is the beneficiary of unusual circumstances, all of which apply here. I wouldn’t call Hoffman the favorite but it’s close — I’d give him a 45 percent chance of winning, Owens a 50 percent chance, and Scozzafava no more than a fleeting, 5 percent chance. Subject to further revision, of course, as I hope we’ll get some further polling from Siena or another trustworthy source before the weekend is out.
There’s also the question of just who Democrats should be rooting for, but we’ll table that until later.