Good magazine takes a look at the historical track record of different media polls in terms of forecasting the general election. ABC looks like the winner to me, with Washington Post and Harris also worth taking seriously.
« O Canada | Main | The Zero Factor »
Polling Accuracy
25 Apr 2008 09:08 am
Comments (3)
Not to mention, judging a polling firm based just on the last day of polling isn't entirely fair. Most of the polls conducted in the months before the election have to deal with a large number of undecided and leaners, and how they push them is important.
To many people, ARG is an outlier when Obama is not winning. Survey USA says ARG was one of the best in Pennsylvania and ARG nailed Wisconsin, California, Texas, and Ohio when other more pro-Obama polls were way off.
Comments closed May 09, 2008.
| 
Heads in the Sand
I think the sample size is too small to really make predictions on how they'll do in 2008. Using election AND primary data would be better. For example if you look at this year's primaries, it seems that Survey USA (not even mentioned) has been doing pretty well while American Research Group is regularly an outlier.
Posted by LFC | April 25, 2008 10:56 AM