Pollster.com

The actual voters and the likely voters gives an indication about the intensity of opinions. People who don't bother to vote typically have very wishy-washy views about politics anyway. They don't have convictions strong enough that they will take 15 minutes out of a single day per year to cast a vote.

I think it gives even a better indication than asking if the respondent strongly agrees with a position or just agrees with it. Someone who says they strongly agree with something and yet also says they are not likely to vote leaves you to doubt that their idea of strongly agreeing is less than that of a person who also plans to follow up on that disagreement with a vote.

There's a lot of factors in likeliness to vote other than strength of opinion, like whether you think it'd matter, whether you can afford to take the time away from the job (either monetarily or because of bad bosses), and whether you're away from home and bothered to have your absentee stuff set up.

I should clarify that bit about "whether you think it'd matter" - there was an article a while ago about the game theory of voting. If you're acting selfishly, it's so unlikely that your vote is the one that makes a difference that voting isn't really worth it. But it is worth it if you're altruistically considering the difference it makes for the whole country (and if you think there's a fairly significant difference between the candidates).

I still believe that Rasmussen and other pollsters underpoll non-white Voters and this could be a factor in parts of the US where Hispanics, Blacks and some Asians are underrepresented. I remember that was a factor in VA this last Fall. Minority voter turnout was extremely low. I think in Arizona and some in southern California could have a huge turnout of hispanic voters.

I think this new ant-immigration legislation in Arizona that borders on gestapo tactics. "papers please" is the new word that dark skinned foreigners will hear from cops.