Case in point, some of the comments in response to my post on the General Social Survey (GSS) conducted by the National Opinion Research Center (NORC) at the University of Chicago. Here are two, related comments:
First up, our resident troll, 'Anonymous:'
Do you really expect gunloons to tell someone over the phone that they own guns?Next up, frequent commenter 'TS,' who should know better:
Do you think that would skew the results?
Some people may not be getting on the internet or answering phone surveys. Call them secretive or paranoid- whichever you want.First, it should be noted the GSS is one of the most respected and utilized surveys in this country. In fact, only one other survey is used more--the US Census. So, it's not as if this is some rinky-dink survey conducted by a junior high student.
As such, professional surveys such as the GSS, understand--and account--for respondents who may lie or answer erroneously for some reason such as not understanding the question or some other reason. That's why all surveys or polls of this kind note a margin of error (MoE). Our resident crackpot, Anon, would have us believe three things: 1. ) most or all gunloon respondents lie about their ownership of guns; 2.) no non-gunowners lie about their non-ownership; and 3.) gunloons choose to lie rather than refusing to respond. This is the only way respondents can skew a survey of this size; there would have to be a massive and concerted effort by one segment of the population to both elect to participate and lie and mislead, while the other segment answers completely truthfully.
As an aside, isn't it interesting gunloons believe Kleck's 2.5M DGUs annually survey is perfectly valid--a survey in which gunloons are brandishing and sometimes shooting their firearms--while a much larger survey involving gun ownership is completely bogus?
WRT TS's comment--if one chooses not to participate in a survey, one is a non-respondent. But it's important to understand that such surveys are extremely accurate. It's called statistical inference where we can make accurate inferences about a large population from a smaller sample.