To: Tenchusatsu who wrote (1058383 ) 3/7/2018 2:08:37 PM From: RetiredNow Read Replies (1) | Respond to Well, there are inherent problems with all polls. So let's say you do the right thing, like Rasmussen, and look at likely voters who are also registered voters. Then you construct your sample to be broken down into Democrats, Republicans, Independents, etc. to reflect the percentages among those likely registered voters. This is most likely to minimize sampling bias, but what if someone like Obama comes along who is a smooth talker and gets all the Independents like me to swing to vote for him? How would Rasmussen's poll predict that? All polls can do is take very good guesses and put the appropriate margin of error around it.
