Robert Moran: February 21, 2010 - February 27, 2010
Thank you Mark and Nate for the thoughtful critique of my post regarding incumbents polling under 50%. I think this is an excellent example of the kind of smart, civil dialogue we can have around polling data. And, I think it serves the public interest well.
The data is the data. It is certainly compelling. And it suggests that the old 50% incumbent rule may have fit the 1980s and 1990s more than the 2000s. As Mark noted in his post, those that gained their formative public opinion research experience in that era, absorbed the conventional wisdom of that era. I'm a product of that time. But, times change. Mea culpa.
I found Mark's "Four Pollsters on the Incumbent Rule" (December, 2006) to be the most interesting. Here I agree most with Hickman (incumbents are much more aggressive, not worrying as much about building their opponent's name ID) and Greenberg (nationalization and partisan consistency). I also think that microtargeting combined with a renewed focus on turnout machines can have the effect of both saving endangered incumbents and complicating the vote models.
It will be interesting to see how this old rule works in 2010. Like baseball statistics, we always have another wave of data to analyze.
In the case of the Strickland-Kasich race in Ohio, I remain VERY skeptical of the Governor's reelection prospects.
1. First, Strickland's ballot support average in the new year puts him within that perilous under 45% group that Nate's analysis of the 2006-2009 data highlights (67% average loss rate). Even throwing out the hard to believe OH Right to Life survey, Strickland's average ballot support is 42.5%. Moreover, of the last four surveys, Strickland has only led Kasich in one. This data reminds me somewhat of the DeWine reelect in 2006.
2. Second, Strickland's image looks to be in tough shape and approaching a 1-1 fav-unfav ratio.
3. Third, the President's job approval seems to have slipped into negative territory in the Fall of last year.
4. Finally, as I noted in my initial post on the Quinnipiac survey, Kasich is handily defeating Strickland 46%-34% in Central Ohio. This is the part of the state where voters are most likely to remember Kasich. It is also a key battlegound. This does not bode well for Strickland.
This morning, like every morning, I quickly scanned First Read from Chuck Todd, Mark Murray, Domenico Montanaro, and Ali Weinberg.
I enjoy First Read and very much like Chuck Todd, but I will admit that I threw down my Blackberry in complete frustration after reading the following:
*** Buckeye State Watch: Is the worm beginning to turn for Democrats in Ohio, too? Yes and no. A new Quinnipiac poll shows that Ohio Gov. Ted Strickland (D) now has a five-point lead over challenger John Kasich (R), 44%-39%. Back in November, the race was deadlocked, 40%-40%. Still, that Strickland is below 50% hints at his vulnerability in this swing state. Meanwhile, the poll also finds that President Obama has an upside-down approval rating in Ohio, with 44% approving of his job and 52% disapproving, which is essentially unchanged from November.
Most readers of Pollster already know what I'm going to write, but I feel duty-bound to write this.
While the Quinnipiac poll may show Strickland ahead of Kasich 44%-39%, that is in NO WAY the headline. The headline is actually that (a) Strickland is way below 50% and (b) incumbents under 50% in a two way race have a very poor track record in November. Why? Because voters already have had time to get to know the incumbent. The incumbent has had their term to close the sale with a majority of voters. If he/she can't close the deal now with voters, then it is very unlikely that they will when the contrast ads get started.
Further, the November, 2009 data showing the two "deadlocked" is in no way a deadlock. An incumbent at 40% is a clear signal to his political team that they need to start requesting cash up front and immediate payment terms. Years ago we used to refer to these clients as "wire jobs", because we knew they were going to lose and therefore were very keen for them to wire payment before their loss and the inevitable vendor scramble for payment.
Now, First Read does caveat things a little with this:
"Still, that Strickland is below 50% hints at his vulnerability in this swing state."
It doesn't "hint at" Strickland's vulnerabilty, it demonstrates it quite clearly.
As the trend data shows, Strickland hasn't been over 50% since last summer.
Barring some massive exogenous event, the next Governor of Ohio will be John Kasich. Strickland is a Governor in a swing, center-right state polling WAY under 50% with an economy that will not come back before November.
Interesting data nugget:
In the central part of the state (where the voters know Kasich) Strickland is getting crushed by Kasich 46%-34%. Ouch.
I've worked both sides of the challenger-incumbent chasm. When you're running the incumbent side campaign in this situation, this is your thought process:
1. Double down on the opposition research investment. Maybe something will turn up.
2. Find a way for (or hope for a way) a 3rd party to do most of the really tough contrast ads. I'm sure Strickland's team has this covered already and this will give them some plausible deniability when the negative campaigner stories proliferate ("Hey, they're not our negative ads and we've asked them to stop.") Standard practice.
3. How do I make this a respectable performance?
4. (If internal:) Begin circulating resume.
5. (If vendor:) Get upfront payment terms. Try not to waste too many flights. Minimize focus group swings.
Returning to the general problem for a moment, this misreporting is a constant source of frustration for those in political polling. I'm sure I wasn't the only one that didn't at least roll their eyes this morning upon reading it. Unfortunately, this is a symptom of a constant problem in reporting polling data relative to an incumbent. In a two way race, political professionals don't even bother to look at the spread between the incumbent and the challenger, they only focus on the incumbent's support relative to 50%. Incumbents tend to get trace elements of the undecideds at the end of a campaign. Sure, there is the occasional exception, but this rule is fairly ironclad in my experience.
