Weight polls on party to reduce bias

Pew’s methods FAQ argues that, in the current era of steeply falling response rates and partisan nonresponse, adjusting survey weights to match party affiliation or past vote usually improves average poll accuracy compared with weighting only on demographics — but it is not a cure-all and can worsen some nonpolitical estimates. The report documents evidence, explains limits, and offers practical guidance for pollsters, especially those using panels. — If adopted or debated broadly, this changes how media, campaigns, and the public should read and trust polls and could shift narratives built on pre‑weighted polling numbers.