… The question that keeps emerging is who is not on pollsters’ 2018 radar. That matters greatly because between now and November’s vote, there will be a tsunami of horse race coverage delivered in authoritative tones. However, pollsters will tell you that their analyses are more vaporous; that they are merely snapshots in time of voter sentiment, in part based on answers given (that aren’t always true) and in part based on their assumptions about how big slices of this fall’s electorate will be.
This mix of statements, assumptions and math can be useful for campaigns in ways that the public will never see — such as deciding where in their districts they are weak, need to redeploy resources, where to send their candidate and buy advertising. However, what’s good for a campaign is not necessarily good for the public. That is especially true when media organizations treat their polls as gospel; don’t discuss their poll’s assumptions, strengths and weaknesses; and gloss over their margins of error. CONT.
Steven Rosenfeld, Salon