The California recall will be very hard to accurately poll
Not all polls are created equal. The higher the turnout in an election the easier it is to poll, and conversely the lower the turnout the harder it is to poll. Why? Because the point of polling, and why polling works with in an MOE, is that a random sample, sufficiently large, will be similar to the population from which it is drawn. If you increase the size of the sample, you reduce the size of the MOE to obtain a chosen level of accuracy. But here is the issue, you draw your sample from the general population (of necessity) and thus you have to figure out a way to make your sample match the voters. People who don't vote are often different from those who do, and thus the pollster needs to find a way to make sure her respondent actually are going to vote. That is easier said than done. So, if you want to know how the recall is going, you have to hope the pollsters have done a good job and look at the votes when you can.
On edit, apparently some pollsters are way under weighting Democrats in the polls to try to account for lower Democratic turnout. But the question becomes how should the Democrats be weighted? It is hard to know but looking at returns should help us know this.
= new reply since forum marked as read
