October 26th, 2010 11:13 am ET
____ The (controversial) poll in question from Newsweek gives Democrats a three-point lead (48%-45%) over Republican on the generic congressional ballot. The generic ballot asks voters whether they would generally vote for a Democrat or Republican in the next election, without mentioning any specific names of candidates . . .
. . . there are some reasons for Democrats to believe the Newsweek poll may be legitimate.
First, Newsweek poll actually used a "likely voter" model, which tends to favor Republicans. Much like other polls Newsweek asked respondents questions, such as "Where is your polling place?", to determine whether they were likely to vote or not. Even after narrowing their sample using these criteria Newsweek still had likely voters favoring Democrats.
Second, the Newsweek poll included cell phone-only users. Out of the 1,000 people Newsweek polled 333 were polled on their cell phone. Some of the other pollsters, such as Rasmussen, do not sample cell phone users. Some believe the exclusion of a cell phone user sample may skew results against Democrats, since cell phone users tend to be more urban, young, Democratic voters.
Finally, the Newsweek poll included 92 people who claimed they had already voted in the election. These results, while very small, also showed Democrats winning. Polls of people who have already voted are generally considered more reliable than polls of how people will vote in the future.
read article:
http://www.examiner.com/political-buzz-in-national/midterms-recent-generic-ballot-poll-from-newsweek-has-democrats-the-leadpoll results:
http://nw-assets.s3.amazonaws.com/pdf/1006-Ftop.pdf