Friday, May 2, 2014

Polling in Georgia

Okay, first, read this. When finished, you're allowed to return.

You're back? Good. I assume you can be trusted, that you read Galloway's Political Insider post about two polls in Georgia. Or at least skimmed it. Some good stuff there.

The lede is simple -- we have a Tale of Two Polls.

One is a semi-traditional robo-mixed poll that shows Handel with the momentum but Perdue ahead for the GOP nomination to run as the party's rep for U.S. Senate. The other is an "internal" poll. Let's talk about that one.Here's the memo that backs up that internal poll. Released by the Kingston folks, it coincidentally shows Kingston ahead. (shock!) The brief report concludes that Kingston "leads a crowded field."

Not so fast, my methodologically-challenged friend.

The margin of error of this poll is a hair below 5 percent, and that leaves you with statistically a tie among the top four candidates. On the good side, this poll includes 80 cell phone-based interviews in its relatively small, but marginally acceptable N of 400. The results may actually be better than the other, larger poll. Except, of course, it's internal. That means you can't really trust the numbers. As if Kingston's folks would release a poll that shows him way behind. News 101: Internal polls are fascinating, but they are not to be trusted.

So which one is right? They both are, because the race is simply too close to call, at least based on polls that are not as methodologically sound as one would like. Kingston wants to stop Handel's momentum and get himself into the runoff, and nothing does that better than (1) tons of money for TV ads and (2) a sense of momentum or inevitability.
This poll release is designed to slow Handel down, or at least signal to any contributors out there that he remains the viable runoff candidate.

No comments: