(Community Matters) If I’m reading Nate Silver correctly, the increased sampling of Republicans in this week’s polls may be simply the tally of those polled rather than a deliberate weighting. Certainly would be the case with Gallup and SurveyUSA which do not force the samples. Rasmussen weighs samples for party id according to a three week average.
More disturbing is imagining that average 9-ish point increase in sample would be a bounce or a shift from convention.
Community Matters 