Wednesday, July 17, 2013

More Bad Poll Journalism

The other day I wrote about some of the awful polls being used by news media to gauge -- in no real way whatsoever -- public reaction to the Zimmerman verdict.

Don't think we're done yet with the bad poll stories.

My favorite of the day is this one in which 74 percent agreed with the verdict.  This poll has the advantage of a big N (3,557 respondents) but the disadvantage of being complete bullshit as it does not use a random sample but rather relies on people who happened to use this web site and happened to participate.  I discussed this briefly in my post the other day, the weaknesses of a SLOP (self-selected opinion poll).  And don't be fooled by the large number of respondents.  As we know from our history (see Literary Digest debacle), when it comes to polls, size doesn't matter.

Simply put, when a news org uses one of these polls, it needs to slap a big NON-SCIENTIFIC FOR ENTERTAINMENT USE ONLY label on the poll itself.

No comments: