Reply #83: People do tend not to question polls results they like, that's true [View All]

[daleo]

But exit polls should be highly reliable, because the proper universe of people (actual voters) are available to the pollster, and the respondents are chosen with simple random sampling (every nth voter). This should be a very reliable methodology. The only real possibility of sample bias relates to the idea that Republicans just don't like to fill out polls while Democrats do. This is nothing more than an assumption in this case, so suspicions about the variance between the exit polls and the reported election results are well founded.