A variety of schtuff

[stevencwatts]

I've been neglecting this thing for quite a while, and I figured it was time to shape it up again. Most of the reason I've neglected it is because my mind has been a bit occupied. That would happen to you, too, if you were served a subpoena. So for a few things on my mind

Comments

[wattsu]

The thing about polls, though, is that while they are designed to be somewhat representative of the larger population, they simply can't be. That's one of the reasons I don't really pay attention to polls, because different polling organizations use different samples, different questionnaires, and may even have different agendas. So how are we to know the prevailing opinion? Sure, 80% of people in a survey might think George W. Bush is a douchebag, but where was the survey taken? How was it administered?

Basically, I think polls, unless they're really well thought out and have enormous sample groups, are not exactly the most reliable sources of information.

[stevencwatts]

Agreed. However, Gallup is the most trusted name in polling due to its enormous sample groups and experience in asking fairly well-rounded and bipartisan questions. Moreso, polls are inaccurate, but when administered by the same place they tend to be inaccurate to the same degree.

To illustrate: When I was speaking to radio sales staff about their ratings books, they basically explained that it is done by random sample polling. Most people, they say, forget to fill out their sheets and do it on Thursdays, the day that the poll booklets are due. This is why a lot of radio stations, especially during the times of the polling, will make special offers on their promotions on Thursdays ("Double Cash Payday!" or some such incentive for a contest). The thing is, all radio stations do this. So the system is unavoidably flawed, but it's equally flawed to all radio stations. So you may not get a pinpoint accurate number, but you probably do get fairly accurate proportional results