On a state-by-state level, Hillary more electable than Obama, polls suggest
There has been a lot of discussion about who is the more "electable" Democratic candidate. Hillary, Obama and Edwards all have claimed or suggested that they were the most electable candidate, using different arguments: experience, cross-over appeal and occasionally, how they are polling.
In blogs and the like, discussions about how the polls show that this or that candidate is the more electable are more explicit. We've had them here too. But usually, the only references are to national match-up polls. There is little reference to how the candidates' appeal against possible Republican contenders plays out region by region, state by state. How does Hillary do, in relative terms, out West? How does Obama do in the South? When you dig into the state-by-state polling that's been done, it looks like Hillary might, surprisingly, be the more electable candidate.
There are many ways to assess electability. Personality, ideology, political baggage. Questions about how the gender and race of the candidate plays out. And to be sure, a year out from the general elections, opinion polls are of modest significance. You're not going to go just on polls, at this point in time. Moreover, qualifications go both ways.
For example, Obama and Edwards supporters point to Hillary's high unfavourability ratings; but Hillary supporters counter that her negatives are only so high because she's already been in the crosshairs of the Republicans and the conservative media for a decade and a half. Once those folks are done with a nominee called Obama or Edwards, his negatives would be equally high.
On the other hand, Obama and Edwards supporters might claim that their candidate will still have a chance to present himself to the public that Hillary, already so familiar a figure, will not have. For now, they might argue, interest in the presidential race has been superficial at best, beyond political enthusiasts and those in the early primary states. Even taking the media blitz that surrounded Obama's arrival on the scene and Edwards' prior run in 2004 into account, public judgements of these candidates will for now still be superficial, and receptive to change.
Everyone has heard such qualifications of the significance of polls. However, those polls are often cited anyway. Notably national polls. We discussed how Obama claimed that "right now among all Democratic nominees or Democratic candidates I do better in a general election match ups than the other candidates." To evaluate claims like those, I tracked the national match-up polls since March. A set of resulting graphs from November
can be found here.
But state-by-state polling turns out to provide a different picture. I was already tracking how Hillary was matching up in state-level polls against the Republican frontrunners, I posted some tables here. Now I've gathered data for Obama and Edwards as well. Data on Edwards is too scarce to say anything much, but a comparison between Hillary and Obama surprisingly seems to show Hillary doing better.
Below is a first table, which lists all the state-level match-up polls I could find for the last three months pitting Hillary and Obama against Rudy Giuliani. Other tables for how they're doing against McCain, Romney and Huckabee are forthcoming. Giuliani, of course, appears to have almost become a non-factor in the Republican race; but since he was long the apparent frontrunner, there is simply much more data available for how the Democrats stack up against him than against any of the others.
Most of the polls are by Survey USA, Rasmussen or Quinnipiac, so that's how I've organised the table. Survey USA often has the Democrats doing better than Rasmussen. The Quinnipiac polls seem to mostly be in between. In addition to those three, I found nine other pollsters that each did one poll; I've inserted those data pragmatically in the tables where there was place, but each is indicated by a lettered reference in the cell to the left or right.
How to read this table: on the left are the states for which polling was available, ranked from the state where Kerry led Bush by the largest margin to the most Bush-friendly state. The margin with which Kerry or Bush led is given in the first column. The colours represent how "blue" or "red" the state was in those elections.
Then come the state-level match-up polls: the number given is the margin with which the Democratic candidate leads or trails. The cells are coloured in accordingly. If there were two polls by the same pollster in the same month, both numbers are given divided by a |, and the colour represents the average.
Finally, on the right of all the data for Hillary vs Giuliani and of all the date for Obama vs Giuliani, I've weighed out the colour that represents the balance of the polls for each state. Basically, the colour is based on the average of the polls, though in borderline cases I've taken into account that Survey USA may lean optimistically and Rasmussen pessimistically, and that polls from December are more significant than polls from October. Yellow represents a complete toss-up.
WARNING! You will notice that the variations from one poll to the next are large. When looking at match-up polls, keep in mind how the margin between two candidates is significantly more sensitive to volatility than the numbers polled for each candidate, as margin of errors double when looking at the margin between two numbers.
This is how I tried to explain it (but I'm no expert)...
This is a simplified side-by-side view of how the Hillary vs Giuliani and the Obama vs Giuliani numbers compare with each other and with Kerry's score against Bush:
)
