My Take on "Next to Useless"
Mark Blumenthal | November 23, 2009
Topics: Issue polls , Joe Klein , Measurement , National Journal column
Last week, I highlighted a post by Time's Joe Klein concluding that "polling on issues," as reported by most news media, "is next to useless--especially on issues as emotionally complicated as wars and as technically complicated as health care reform." I share my thoughts on in my column, posted this morning on NationalJournal.com. Short version: He's right that no one poll question captures all of public opinion on an issue like health care, but we get closer to the truth, including some on the same CNN poll the provoked Klein's criticism.
By Mark Blumenthal | November 23, 2009 11:57 AM | Permalink | Comments (1) | TrackBacks (0)
Comments
Mark,
The Journal piece is good. Certainly the argument that one needs to examine a large number of questions using a large number of different wordings is one that I am very sympathetic to.
But there is another question: are all "issue questions" alike? Or is there some way to think of (classify) issue debates that should be considered by pollsters and students of polls?
Some while ago, Edward Carmines and Jim Stimson did a nice piece on this (
"The Two Faces of Issue Voting"
Edward G. Carmines and James A. Stimson
The American Political Science Review, Vol. 74, No. 1 (Mar., 1980), pp. 78-91).
One of the distinctions they make is between the "ends" of policy versus the "means". So asking about the goal (end) of universal coverage is not the same as asking a technical question about means (financing, administration, cost control), etc. Since the current bill contains many ends and many means, it represents a particular polling challenge.
In any case, the Carmines/Stimson article is useful for students of issue polling.
Posted on November 23, 2009 12:58 PM
Post a comment