What statistics tells us about the Presidential race
My last post about the electibility argument for/against Obama & Clinton included the total number of votes cast for each of the remaining candidates both republican and democrat. What this tally shows is that both democratic candidates individually enjoy more current support from people who have actually voted than both republican candidates combined.
This indicates that both Obama and Clinton hold a massive advantage going into the general election over McCain. This is not the same thing as saying they will win. And, frankly that wasn't the point of my post. The point of my post was that both enjoy a massive structural advantage going into a general election.
Anon pointed out that Kerry also enjoyed a massive advantage over Bush and lost. To which I responded specifically to in comments but also wanted to elaborate on here. (And I am not sure Anon disagrees with the following, these are just clarifications that Anon's comment brought up in my mind).
1st: as I stated above, having a massive advantage now does not mean that either Dem candidate will win. There is plenty of time for either Obama or Clinton to manage to screw it all up by the general election. My point is that there is nothing inherent in either candidate that would indicate they can't win.
2nd: I don't remember any of the resources I use (mostly economists like Krugman and DeLong) actually asserting Kerry had an advantage. IIRC during the primary the consensus among economists who analyise this stuff is the Dean had the massive advantage becuase he was stronger on issues people would vote on. On the otherhand they felt Kerry was weak because the data indicated Democrats were primarily voting for him because of percieved electibilty rather than enthusiasm for his postisions, many of which likely dem voters did not feel comfortable with (like the vote to authorize force).
3rd: I don't really pay much attention to the polls. It seems like every election I can remember involves pundits pontificating on polls that turn out to be off by around 10% and then pontificating on what went wrong with the polls? How are they off by so much? Well, there have been several studies and analysis (as opposed to polls which aren't) as to why this is. The bottom line is that the GOP nationally has a 6% structural advantage in polls on average meaning that polls have consistently under-estimated the republican's share of the vote/support by 6%. by themselves, individually polls are historically all over the place. The studies also found that this is primarily due to voter suppression efforts (which, again, the GOP happily admits to so don't yell at me) and a powerful GOTV program. It doesn't hurt either that much of America is informed by talk radio which is decidedly republican in flavor. Will this hold true in 2008? well, see, that's the problem with polls. It's hard to figure out why they were right or wrong until after the fact and the foundation keeps shifting every election. It might be telling that Obama has consistently out-performed polls (except NH) by around 10% and has a massive GOTV operation.
4th: Polls are a snapshot of who people who are polled will vote for when the polls is taken. A large number of people make up their mind at the last minute and they generally vote party line despite identifying as independents or undecided. These people also tend not to vote based on TV ads so its hard to say how they'll vote until those ads hit the airwaves.
Having said all that, I think Clinton's recent statements trying to rebrand super-delegates as automatic-delegates and apparent strategy of giving up on the popular vote and aiming to win based on these super-delegates *is* exactly the kind of thing that could make her unelectable. Democrats have a strong advantage because their declared partisan support is so freakin high. But I think a lot of those partisans will not vote for Clinton if she pulls that stunt because it is precisely what they don't like about the republicans. And if Clinton looses that group I don't think that independent/undecideds will break democrat at some crazy rate. But that last part is my personal suspicion and should be considered in light of my incredible prediction rate (Patriots win Superbowl, Hawks beat GreenBay, GreenBay beats Giants, Edwards wins nomination, ect). The reality is there is no data that can tell us how undecided/independents will break beyond historic patterns because, by definition, the decide at the last minute.