Why You Have Not Been Polled
by Tom Riley and Howard Fienberg October 23, 2000
There seems to be at least a half-dozen opinion polls in the media every day. You have opinions and want to have your say. It makes you want to shout, "Hey, pollsters! I have an opinion! Poll me!" People routinely question how anyone can poll 500 or 800 selected individuals and say that the results represent what America thinks and feels.
Some people see darker forces than mere chance at work. Everyone they know has the same opinions they do, yet opinion polls consistently come up with differing results. For instance, everyone you know feels that candidate A is a weasel and could never vote for him or her. But a majority of people in the opinion polls still intends to vote for candidate A. What's going on here?
The "I've never been polled" canard is so common that when you put a bunch of like-minded people together, charges of fabricated numbers and rigged polls quickly develop.
It's actually not so complicated.
There are three explanations for why so many of us go "un-polled." First, there are only about 500 major news polls conducted each year. Even assuming that all the polls surveyed at least a thousand people (and many don't), that still works out to rough odds of over 500 to 1 that a given individual would be polled in a year.
"But," counter the unpolled, "I asked some of my friends, and not one of them has been polled either." What these people fail to realize is that pollsters attempt to survey representative slices or strata of the overall population. People who are demographically "like" you (co-workers, friends, relatives, etc) should be exactly as likely -- or unlikely -- to be polled as you are. This also explains why some people are polled numerous times: their pertinent demographic information has been stored in a database, and they can be used to fill in needed niches.
You see, you are not as individual as you might like to think. You do not need to drink the entire ocean to note how much salt there is in it - just a sip will do. Similarly, a pollster can sample a small segment of the population to represent the whole. Although you are an individual, you share enough characteristics with other people that you are roughly alike in the eye of sampling.
Finally, you may have been called for a poll (perhaps many times) without realizing it. Few of you think you have been called by a pollster. How many of you, though, have been called by long distance companies or telemarketers at dinnertime? Probably most of you. Oftentimes, that person mispronouncing your name on the phone is actually conducting a poll rather than trying to make a sale.
But given the choice between "letting your voice be heard" and the tedium of another pitch for "fabulous long distance savings," many people might just opt not to be polled after all.
- Tom Riley is director of research for the Philanthropy Roundtable and was formerly senior analyst at the Statistical Assessment Service (STATS). Howard Fienberg is research analyst for STATS.
return to Howard Fienberg's page