Thursday, June 22, 2006

More on Polling

A couple of weeks ago I was looking forward to a poll to be released. In retrospect, I may need to take that back. Let me offer a lengthy explanation.

It is my personal belief that no one should be allowed to gush over any poll result, unless they have spent a serious amount of time actually administering a poll. I don't mean writing the questions or commissioning the poll, but actually calling people on the phone to get their opinion. I had the misfortune of doing this for a temporary job while doing more serious career hunting and it really opened my eyes.

I discovered that the people that will: 1. See your number on caller ID and pick up the phone. 2. Will take the time to listen to your questions, and 3. Answer you honestly in a thoughtful manner, are simply not an indicative sampling of society as a whole. I believe this would have gotten worse since I did polling and not better. If you look at the past political polling for the last few national elections it has really been all over the place and is only truly useful in showing candidates with a big lead as eventual victors. I would submit that those polling places that came close to the actual results were more lucky than good. If I remember correctly, Rassmussen who was the hero of 2004 was a goat in 2000, and vice versa for Zogby. Now did each of them drastically change the way they polled? Probably not. And don't even get me started about the exit polling in 2004 (and I don't want to hear "Diebold," either.)

Truth is, polling national contests should be a lot more accurate as there are recognizable names and a lot larger sample generally. Now imaging polling for the Arizona Eighth race. You have a smaller population base to sample from, will want to poll people who are likely primary voters, they will need to identify what party they belong to and are eligible to vote for, and must actually be contacted and agree to be polled.

There are about a dozen candidates to track, at varying degrees of name recognition and no incumbent. So it should be no problem accurately predicting the support of each candidate both in the primary and in the general 12 weeks out, right?. I'm certainly glad I am not administering that poll.

And unless one or two candidates are simply blowing the others away, I am not sure that I am going to trust much of what comes out in a poll at this point. I would argue that the signature collections for filing would be more accurate, and I don't really trust them for a benchmark either.

What this means is that we are going to experience one of the most interesting political horse races possible, probably the most interesting we will see in this district in decades. It is good to be in the blog/pundit business with something such as this taking place.

P.S. I will still be happy when a poll comes out.


Anonymous said...

You are still a little biased in your reporting. You only have 4 Republican candidates on your site and 5 have filed.

Framer said...

That was certainly not an intentional ommission. I must have overwritten Mike when I was updating the links a few weeks ago. I apologize to Mike. I even have some time scheduled to to do a "State of the Campaign" fo him.

I will make sure that he is put back on the list, and I will fire my editor :)