Another Statistics Question

The other statistics thread caused me to ask this question. I hear about polls and statistical research all the time. However, I have never been polled not have I ever stopped to take a survey. Survey takers are at the mall all the time, and the only people who stop to talk to them are old people who have nothing else to do and no one to talk to.

Doesn't that mean that all statistics are skewed? They are not getting a true random sample or cross section. All polls should state "of the 1000 people who would talk to us surveyed, a sample that did not include busy people, parents with kids, people with a life, shy people and people in a hurry, the majority would vote for mandatory adult diapers".

I find that that most polls are taken over the phone, you can normaly tell from caller ID. My wife refuses to take them. I will take them if I have time in the evening, especialy in Election years. It doesn't mean I'm a lazy unemployed bum, it means they happend to call just before or after dinner, I can take the poll at the same time as stacking the dishwasher no problem

I dated a gal whose family was chosen to be a Neilsen TV rating family. There was a box that monitored what you watched and who was watching (based on whether or not you pressed a button when you watched.)

I hope you like your programming based on their German Shepherd's preferences because they programmed the dog as one of the viewers.

Yep, you're 100% correct. This is a well known problem called sampling bias that pops up in ALL fields of science where data is collected in some sort of overview format, but especially the social sciences with surveys and the like. Now, a good, careful researcher can work around the bias or find ways to neutralize it - there are many techniques for doing so. Those people who run polls for the local TV news, or politicians pushing an agenda, or marketers... they are not good, careful researchers. (Often they know full well that their sample is biased, understand the problem completely, and don't care one bit.)

I once answered a phone survey on "freedom vs. security issues" out of boredom. It was during an election period but it was supposedly an independent survey. The questions were horribly designed and it was clearly intended to put people with opinions like mine off from even answering. I got really mad at more than a few of them and thought about hanging up a bunch times, but once I was far enough in I felt like I was duty bound to finish the thing just so my 2c would be counted. (SPOILER: they didn't matter E36 M3.) The next week all the papers ran something like "80% of Canadians are afraid of everything and want the goobermint to give them moar security!!!!1" Gee, there's no WAY that "poll" could have had an agenda behind it...

Ask any statistician (like my mum, who is one), and the first thing they will tell you is that polls and surveys are a crock of horse turds. You can rig them to say whatever you want them to. Yay mass media.

In reply to Jay:

http://en.wikipedia.org/wiki/Push_poll

EDIT: This is a gem-

Perhaps the most famous use of push polls is in the 2000 United States Republican Party primaries, when it was alleged that George W. Bush's campaign used push polling to torpedo the campaign of Senator John McCain. Voters in South Carolina reportedly were asked "Would you be more likely or less likely to vote for John McCain for president if you knew he had fathered an illegitimate black child?"