Do any expert fools ever think that just maybe someone would give a wrong answer just for fun? Maybe that's why polls are not correct.

...or they were honest but the polls have a margin of error. If the exit polls showed Obama up 5%, a margin of error of +/- 3% (probably optimistic), and he lost by 2%, that discrepancy is only a couple standard deviations. Predictions that are off by two standard deviations happen all the time. We don't need to conclude a poll is inaccurate just because, by pure chance in the sampling process, it turns out to be a bit off.

It's fixed.

I always lie to pollsters.

So do I. A lot of people I know do the same