Do polls account for people who systematically refuse to respond?

Question

I never participate in polls. There must be other people who don't either, especially phone polls. If this trait is common, wouldn't it skew poll data. For example, if 40% of all people never respond to phone calls, isn't that an important polling trait?

Answer 1

The question of people who do not respond to polls has been a serious one for many years. While statisticians attempt to get around it in many ways, it is, quite simply, an enormous issue.

Suppose you're doing an exit poll at the elections, and you wear a blue shirt. Blue being the "democratic" color, a lot of republicans might completely refuse to talk to you. Those that do talk to you might lie, call you names, or whatever. Meanwhile, the democrats might assume you're really cool, and try to hang out with you. The result: the people you could talk to probably voted democratic a lot more than the general population of voters.

You don't answer phone polls. You probably also wouldn't talk to people doing polls at the mall, or at the ballot. Statisticians can make a note of the fact that you don't talk to them if they ask. They might even be able to go to the IRS, or some other organization, and find out various things about you to figure out why you don't do polls, but since you don't do polls, they cannot KNOW why you aren't talking to them.

Presumably, there's some section of the population that doesn't answer polls, but what that section is, what their beliefs and attitudes are, and how they live their life may be quite difficult to find out. Is there one specific type of person that doesn't answer polls? are they rich? poor? white? black? Green? Are there a few different groups that won't answer polls? or are the people who don't answer polls about the same as the average folks that do answer polls? probably not, because the simple fact that one group answers polls, and another group does not suggests that the people who don't answer polls have different values than the people who do answer polls. How different s a serious question, and the only way to even guess at it is to get sneaky, and search all kinds of databases to find out about people before polling them on... anything.

Answer 2

In a telephone survey, you can expect to make 40 phone calls per completed interview. A study of telephone surveying found that the rate of Initial Refusal (as opposed to No Answer/Machine or Breakoff) was 12.6%. It's possible to collect a mathematically valid sample that can be generalized to the population, even with this low rate of completion, simply by volume. It depends on stratification (what traits are desired in the sample) and incidence (how frequently those traits occur in the general population). It's time-consuming and expensive, which is why there's a niche industry that specializes in conducting phone interviews.

Depending on how the data was gathered, there are simple statistics than can be run that will give you the sampling error. This is usually in the fine print if it's in a mainstream publication, but will be part of the main body of the research if it's in a scholarly journal. It'll say something like: "We are 95% confident that our finding - that 65% of respondents voted against the incumbent due to perceptions of corruption - is within +/- 5% of the true population percentage."

That's why Zogby and Gallup polls are so money - they've got it down to a science and keep reliable, long-term trend information, and regularly collect reliable and valid social science data.

Answer 3

Presumably, there's some section of the population that doesn't answer polls, but what that section is, what their beliefs and attitudes are, and how they live their life may be quite difficult to find out. Is there one specific type of person that doesn't answer polls? are they rich? poor? white? black? Green? Are there a few different groups that won't answer polls? or are the people who don't answer polls about the same as the average folks that do answer polls? probably not, because the simple fact that one group answers polls, and another group does not suggests that the people who don't answer polls have different values than the people who do answer polls. How different s a serious question, and the only way to even guess at it is to get sneaky, and search all kinds of databases to find out about people before polling them on... anything.

Answer 4

Polls say more about the person asking and the people that made up the poll than what is being asked. It has been shown that how you ask changes how people will respond. You can make a poll that will get any answer you want.

Polls are a joke on us.

Answer 5

Most polls account for nothing they are tainted for both sides they
have your name and affiliation before they call you they also word questions for a specific result then spin it publicly by wording it again to slant it their way.

Answer 6

no polles are people that do responed