fbpx
Politics Foreign Affairs Culture Fellows Program

Reading the Polls

I have to disagree about the electoral map. The state and national results are highly correlated: when someone goes up in the national polls, he goes up in the swing states as well. If Obama is +6 in the Rasmussen and +8 in the Gallup (as he is today), he’s going to win almost all […]

I have to disagree about the electoral map. The state and national results are highly correlated: when someone goes up in the national polls, he goes up in the swing states as well. If Obama is +6 in the Rasmussen and +8 in the Gallup (as he is today), he’s going to win almost all of those close states. The state polls are not taken as often – you’re really comparing his current position in the race with state data from a week or more ago, and things have changed since then. Obama moved up 4 in the Gallup and 5 in  the Rasmussen over that week.

Generally speaking, if you want to call the race accurately, look at the most recent polls, which are almost always national polls. Ignore the state polls.

That said, it _is_ possible for a candidate’s electoral college results to deviate significantly from what the national polls tell you – if the race is close. Enough to change the outcome.  For example, back in 2004, I noticed that Bush voters were over-concentrated – he was winning his core states by huge margins. A candidate does better if his votes are more spread out. So I concluded that if the race was close, say a 1 pt edge for Bush, Kerry would have won the electoral college: and indeed that almost happened.

Gore lost with a half a point edge in the popular vote in 2000: it’s happened farther back, too.

But with a 4-point edge or better? It’ll never happen. Even with a 3-point edge it’s not worth worrying about.
Advertisement

Comments

The American Conservative Memberships
Become a Member today for a growing stake in the conservative movement.
Join here!
Join here