This post gives a simple graphical view of the electoral vote maps put together at electoral-vote.com.
Any way you cut it, the graphs make clear that Obama is in a much stronger position than Clinton.
This is, in essence, a graphical view of a recent post from Markos about the electoral-vote.com maps. Quick synopsis of his post: the Clinton folks have been pointing to the apparent result that Clinton wins against McCain and Obama loses, but in fact what the data shows is that (1) Obama has a stronger base than Clinton, and McCain a weaker base against Obama than against Clinton, and (2) Obama puts many more electoral votes in play than Clinton.
The graphs and explanation of them are below the fold.
I wanted to see exactly how many electoral votes each candidate had with leads of at least a given size. For example, in Obama-McCain matchups, how many electoral votes does Obama have with a lead of at least 20 points? 5 points? etc. And the same for McCain. The key thing to know is that the leads have a margin of error of at least +/- 5 points (probably more), so anything within +/-5 points of zero is a statistical tie.
I took the data from electoral-vote.com from today, April 29, and here are the results (the horizontal line at 270 is the number of electoral votes needed to win):
A few things to notice.
--McCain has far more electoral votes with clear leads (at least 5 points) against Clinton than against Obama. Obama has a lead over McCain in electoral votes with clear leads, while Clinton is badly losing to McCain in such votes.
--Obama puts far more electoral votes in play (lead of at least -5 points, that is, electoral votes from states where either it is a statistical tie or he has a lead) than Clinton, while McCain puts far more electoral votes in play against Clinton than against Obama. Obama has a lead over McCain in electoral votes in play, while again, Clinton is badly losing to McCain by this measure.
--There's a small region, leads of at least 0-1 points, where Clinton is beating McCain, whereas Obama is losing to McCain in this region. This is why the Clinton folks say she is doing better than Obama. The electoral-vote.com site has numbers at the top of the Obama-McCain and Clinton-McCain matchup pages showing the number of electoral votes with at least a 1 pt lead (that is, where the candidate is ahead in the poll), and Clinton's numbers are better at that point. But given the 5-point margin of error on the leads, this is a meaningless number. What makes much more sense is to look at the states with solid leads and the states in play, and Obama is a big winner on both measures.
There are a lot of caveats on this data. First, there's two sources of error in this data, sampling error and systematic error. Sampling error means, if they did exactly the same survey over and over using the same methods and the same model of the electorate, how much would the results vary. This is the source of the "margin of error" reported on the polls, which for state-by-state polls is typically at least 3% for each candidate's number and at least 5% for the difference between them. (More precisely, this means that 95% of the surveys will give numbers within +/- 3% of the mean for each candidate and within +/- 5% of the mean for the difference between them.) Systematic error is the fact that, even if they sampled a zillion people so their sampling error was practically zero, they still might be way off in predicting the result of the election, because their model of the electorate is off or their method of sampling the electorate (e.g., land-line phones) doesn't correctly sample their model. Systematic error is the reason why polls from different companies, each with "margin of error" of only 3% say, can differ from one another by 10 or 20% -- the different companies use different models and methods.
The electoral-vote.com data considers only the most recent poll, and any other polls finished within a week of the most recent poll, for each state. Thus, for some states the data might happen to be from a company with systematic error in one direction, and for another state the systematic error might go in the other direction. Also, some of the data is old (six weeks, say), and things might have already changed. None of this is to fault electoral-vote.com, they're doing a great job with the data that exists, it's just to say that the state-by-state data that exists has its limitations and shouldn't be taken too seriously. But nonetheless, it points to some pretty clear trends.