Wednesday, October 10, 2012

All political polls are flawed…

Just a word of caution about the political polls: be suspicious, be very suspicious.

Not for the reason that's been in the news recently: the partisan argument that pollsters are surveying "too many Democrats" really doesn't hold any water, and anyway, since Romney got his media-induced "bump" in the polls last week, we haven't heard that particular argument much…

The real concern is the fatal reality that every pollster faces, from Gallup and Pew on down: it's impossible to contact anything even remotely close to a "true random sample" of the adult population by telephone. So it's impossible to get a reliable sample of opinion.

It has been impossible for at least 25 years, and the problem has been getting steadily worse every year.

I was in the market research business for 30 years, I know whereof I speak.

The fact is that many adults refuse to participate in telephone polls, period. No pollster ever has a chance to find out what those people really think.

The fact is that a large and growing number of people never answer their telephones at home, preferring to let the answering machine screen their calls. No pollster ever has a chance to find out what those people really think.

Last night I made calls at my local "Obama for President" campaign headquarters. I dialed 67 numbers, and I actually talked with 10 people. There were a couple disconnected numbers, and two slam-downs. In round numbers, I got "no answer" on 80% of the calls. You know what that meant—most of those people were home but they didn't pick up the phone.

When you do a poll and you miss 80% of the population, you're not getting a "true random sample" and so your results are not representative findings on what the general population is thinking…your results are nothing but worthless numbers.

When was the last time you gave a telephone pollster YOUR opinion?

more on political polls......

No comments:

Post a Comment