Monday, April 19, 2010

The Problem With Polling

Okay, here's the thing about polls: they might be good tools to gauge socio-political trends…or, they might not. I was reminded of this by reading this sentence in a recent Rasmussen Reports poll claiming to show a majority favor repeal of the recently-enacted health care law. The article stated: "The latest Rasmussen Reports national telephone survey finds that 56% of likely voters nationwide favor repeal…"

Except, that's not really 56% of likely voters. That's 56% of likely voters…who answered the phone and took the survey. Not the same thing at all.

Because, I never take the fucking things. I don't listen to someone cold-call me and ask for money. I always hang up on robo-calls. ALWAYS. So, the numbers are there, yes. But those numbers alone don't tell the whole story. It's not 56% of voters…it's 56% of phone-answerers.

I'm not required to answer the phone, or the door, for that matter. I'm certainly not required to take part in a survey. So, the better question is: who is taking the time to do these things?


Blogger Strangeite said...


I have worked for enough campaigns over the years to know how damn influential polls are to the key players. We know and they know that they are very ineffective tools, but, they are one of the only tools in the chest.

This is why I politely agree to take the poll when the phone rings and the pollster says it will only take 5 minutes, even though I know it will be more like 20.

Dig around and look at the raw data for one of these polls sometime. It is amazing and scary at the same time. While most pollsters won't give the methodology of their likely voter models, you can usually infer the components.

The real 800lb gorilla is cell phones. Most of the major firms have added cell phones to their call lists but it is incomplete and very poor. Cell phone numbers are not as readily available as landline numbers.

This has become a huge problem and is only going to get worse.

5:28 PM  

Post a Comment

<< Home