Wednesday, October 03, 2012

A Problem with Polls and a Call from Rasmussen


I’ve given certain pollsters a hard time.  But I have to admit their job is not easy.  One factor that makes it difficult is that so few cooperate with pollsters.  (And do read the interesting and detailed article at the link.)

Even though I am inclined to cooperate with what I consider legitimate pollsters, I have to admit I am one of those who have not been cooperating.

For the other day Rasmussen’s polling company called me.  I thought that was cool at first because his polls are among the few I respect.  So I eagerly cooperated even though it was a computer calling me and asking me questions.

That is I cooperated until it seemed question after question dealt in some fashion with my finances.  The computer did not ask me for bank account numbers or anything like that.  But I became suspicious that this was not Rasmussen at all but an attempt at fraud instead, and I hung up.

I was curious about the call and wanted to alert the Rasmussen people if it was fraud, so I called them.  They were very helpful, and we determined that the call was very likely indeed from Rasmussen.  And if their computer calls back again, I intend to complete the poll if I have the time.

But this episode nonetheless illustrates a problem with polls. Few answer them.  And the composition of those few who cooperate and those many who do not cooperate with polls surely affects the results.  I think it safe to say that conservatives are less likely to cooperate with a poll from NBC or CBS.  Someone, like me, who is not very trusting with financial information will be less likely to cooperate with a poll like the one from Rasmussen the other night.  And the very busy just do not have time to deal with polling calls.  Others are very prejudiced against computer calls (as I am). 

I could go on.  But surely all this self-selection has to affect polling results.

Anyway, nothing really scientific here.  I'll leave that to others.  But, as you can see, I am part of the problem with polls.

No comments: