With Internals like this, Bush should be up by 10 points. It makes no sense to me.
Trust me, as a history professor, I can tell you that most elections in U.S. history are quite close. Either the pop vote or the EC is usually close. FDR had an electoral blowout in, I think, 1940, but the popular vote was much closer than you would think. Three elections (1884, 1888, and 1892) were decided by a few electoral votes and a small pop vote margin.
I makes perfect sense. Polls regularly understate Republican strength. FOr example the 1996 polls had CLinton up by 13 against Dole - he won by 8 ...
A Bush lead of 3 in the polls is no "neck and neck", his real performance could be 6-8 point victory.
JA of 53% translates into 53% voting for him -> 6-8 point victory.
JMHO.