Friday, August 3, 2007

Stephens Duplicity

Our last Pollster Profile of McLaughlin & Associates got me thinking. The Bill Graham campaign is consulted by the Fetzer Stephens firm, the same people who are running Elizabeth Dole's campaign. Three weeks ago Mark Stephens complained to WRAL (see the comments) about one of our polls, saying we used "questionable polling techniques."

This week Mark Stephens sent out the latest Bill Graham poll conducted by their own "traditional" pollster that showed Bill Graham leading 21% to 9% for Smith to 6% for Orr. PPP's last Republican primary poll had Graham leading by a very similar 23-9-9%.

Does Mark Stephens think our Republican primary poll was phooey? It was pretty much the exact same as his poll. Plus our numbers have been consistent in that race month after month.

Like I said three weeks ago, he was only complaining because he didn't like our results -- which indicated Elizabeth Dole was in trouble. In the case of Bill Graham he doesn't complain because he likes our results.

Oh wait a minute. He must really like our results because the Bill Graham campaign is touting our numbers on the website! He ought to think twice the next time he criticizes our "questionable polling techniques."

2 comments:

Mark Stephens said...

Justin,

The logic of your blog entry of August 3 is founded on two basic premises.

First - that because a client of mine chooses to refer to PPP polling, I must have personally encouraged the client to do so. While I wish my clients would always take my advice, that is not the case. In fact, I've never called a reporter to push a poll on them - yours or others - that employ a similar methodology utilized by PPP.

Second - that because that particular client of mine referred to a PPP poll, it somehow proves that the methodology utilized by PPP is sound and accurate. That logic is even more flawed than your polls.

I've been around this stuff for nearly 30 years. During that time, I've seen various new technologies come and go. Some good and useful - some not so good. If it's good, you keep it and use it.

The polling methodology utilized by PPP is not new. It's been around for some time. It lacks accuracy and I would never recommend it to a client that requires accuracy. I would never recommend it to anyone in the media that is attempting to analyze a political issue or campaign. The selling point for the product is that it is cheap.

I've never contended that it is totally useless. Over a period of time, it may serve to track trends and movement, but the data of any given poll is not reliable enough to make sound judgments.

I have nothing against your company. You serve the liberal political community in North Carolina and elsewhere. If they want to buy your wares - that is fine with me.

The only time I've voiced objections is when your company and clients attempt to persuade the media that your polling - or political spin based on that polling - has equal weight with traditional methodologies that have been tried and tested over decades. IMHO - that is not true and can mislead the media and their readership.

For what it's worth, that's my two cents.

Mark Stephens

Anonymous said...

Hey Justin,

Graham ended up losing when your poll showed him way ahead. Doesn't that show a problem with your polling methodology. Maybe the Dole poll is messed up too.

- Ed

 
Web Statistics