Bush's win was a historical anomaly-- is this evidence of fraud?

Initially I wanted to see if the 11.5 million votes Bush got compared to 2000 was unusual for an incumbent president.

What I found was that his increase in votes, compared to the overall percentage increase in voters, was not unusual.

Thus, Bush got 11.5 million new votes from 2000-- but this was really an increase of 9.5% over his totals from 2000.

What is strange is if you compare Bush winning percentage, 2.5% over Kerry, with his increase in new votes, Bush's win really stands out.

Thus looking at incumbents winning the presidency in the past 75 years, and I count Truman winning in 1948 and Johnson winning in 1964, there have been 10 elections where the incumbent has won (FDR, FDR, FDR, Truman, Eisenhower, Johnson, Nixon, Reagan, Clinton and Bush). Of these 10 elections, 7 times the incumbents received more votes than he did 4 years earlier.

Now look at this list of incumbents who increased their votes from the previous election:

1st number is % margin of victory
2nd number is % total increase votes (compared to previous election)

FDR 1936 / 24% // 5%
Eisenhower 1956 / 15% // 2.3%
Johnson 1964 / 23% // 12.8%
Nixon 1972 / 23% // 20%
Reagan 1984 / 18% // 11.4%
Clinton 1996 / 8.5% // 2.6%
Bush 2004 / 2.5% // 9.4%

So Bush's increase in votes is right in the middle-- but when you look at his margin of victory, it is way way far off from the other president's margins.

While this clearly doesn't prove anything, it makes Bush's 2004 vote increase very strange. Of course, one explanation is fraud-- switching votes from Kerry to Bush.

In fairness, if you go back to Wilson's win in 1916, he had only a 3% margin of victory and he picked up 15.6% more votes than in 1912. But the 1916 election had a 29% increase in voters compared to 1912, which was one of the highest increases ever (I'm not sure why-- maybe because of the start of WW I). In 2004, there was a 16% increase in voters, high but not unusually high.

So I think Bush's win can either be seen as fraud, or due to some world-transforming event like 9/11, which could have caused to the electorate to behave like 1916.

But for many other reasons, I think fraud is the most likely explanation.

Update: A simple way of explaining this finding is this: normally when an incumbent picks up a lot of new votes, he wins the election by a very large percentage margin (e.g. Reagan's 18% margin and 10.6 million new votes in 1984). In the case of Bush in 2004, he picked up a very large number of new votes (11.5 million) without much of a winning margin (2.5%).

Because we know Democrats were very united against Bush, and the electorate was very polarized, there can't have been much switching from 2000. This means Bush got something like 60% of the new voters who didn't vote in 2000. But given Bush's overall vote percent was 51%, this means new voters behaved very different from the electorate as a whole--- and it is very hard to believe this was the case, especially given that new excited voters tend to go 1) for Democrats and 2) for the challenger.

The question has remained as ever since Nov.2, 2004-- where the hell did Bush get all his new votes from?!?!