Reader David W writes the following about my article this week on Dow Constantine's lagging poll numbers:
The "problem" with the KC Executive "poll" results is that Survey USA, like most polls, for a variety of financial and legal reasons, polls only those people 1) with landline telephones, and 2) who are willing to respond to the poll (go to Survey USA's website and read the methodology link). Basically, a recorded voice asks questions to which the person being polled responds by "answering" a preset list of multiple choice outcomes.As was the case with Obama's victory, which was dramatically underestimated based on polls, these kinds of surveys do not take into account voters who either use cell phones only, or do not participate in polls using pre-recorded voices. Most of these kinds of voters typically trend Democratic, so while every poll clearly has a margin of error, landline/recorded voice polls are almost certainly less reliable or predictive than the organizations who run them (or the candidates who seemingly benefit from them) would like you to believe. Likewise, most of the voters who are likely to be contacted, and then participate, are most likely to lean Reublican/conservative—older, at home, with the time and willingness to talk to a machine.
Read the piece in question HERE.
Comments are closed.
Comments (0)