Wednesday we wrote about polling difficulties, today brings another article about this phenomenon. It reports the Washington Post and the George Mason U. polling outfit set out to look at Kansas voters’ views about the senate race in their state.
They found Democrats willing to talk to their interviewers, Republicans not very willing. 32% of Democrats were willing to be interviewed, only 12% of Republicans would do so.
WaPo interpreted this disparity as reflecting GOP voter apathy, but then left-leaning WaPo would spin it that way. Red State, where the report appeared, chose to see it as GOP voters’ negative view of biased pollsters. This time next week we should know who was right and who was wrong.
One thing we know now, before the votes are counted. A 12% sample is much less likely to be representative of the whole than a 32% sample, much more likely to be an unrepresentative outlier subgroup.
