How do I know if a particular poll is meaningful? We have the answer to this problem

Rules for polling analysis:

Understand that the polls this editor sees as predictive include

a) issue related polls that, on issues, mirror the overriding concerns of the people and,

b) opinion polling as to the approval percentage of the current president's (Obama, in this case) job related effectiveness. In other words, Obama can score 55% on an "approval poll" but, only 42% on "job effectiveness" in the same poll. I ignore anything that smacks of a popularity contest.

c) A third factor in polling analysis is whether or not the poll was taken of "most likely voters." If not, I disregard that poll and most polls are "random called" polls. If you use "random calling" and your target audience is on a university campus, or the District of Columbia or Jacksonville, Mississippi, your results will be screwed in favor of that demographic and may not be representative of the larger population.

d) Forth, I consider two things together: the bias of the polling entity (all polls have a bias) and the demographics of that particular poll.

e) Fifth, I concern myself with polling trends, not a one time, result oriented poll designed to make the evening news. Polls that rate at the bottom of the scale, in my book, include anything from the "major" media (NBC and its mentions, ABC and CBS) and, USA Today, Five Thirty Eight, the Public Policy Poll, the New York Times and the Washington Post (which includes the Public Policy Poll).

f) Finally, I make note of a poll whose results run counter to what one would expect when bias and demographics are considered. For example, if a poll taken of college students by USA Today found that the majority of those polled were leaning away from federally funded abortions, well, I would think this to be important, especially if I could determine that this was a representative trend.
Post #2887

No comments:

Post a Comment