Senators Express Reservations on Bailout, Despite Pleas for Swift Passage, page 2

Originally posted by alphabetaone

However, I participated in THEIR OWN online poll which asked the IDENTICAL question that he was reporting about and the poll results are thusly:

Although, it is entirely possible that they did a random phone survey for the news report, and the online survey is a separate survey, hence the differing stats. I cant imagine anyone would use a online survey to give results on the news.

An online survey is NOT really a good sample. It is horrible methodology. One, because the demographics online are skewed, (better average education, wealth, more males) and because an online survey is not random. You are only getting a sample of the people who are reading about the topic, (thus have a higher level of interest in general in the area being surveyed) AND because those people who both read about the topic and then fill out the survey are of a more opinionated mindset.

Technically, you would expect the people doing an online survey to be more informed, and have a better idea what they thought about the issue than JoeBob sitting at home drinking a beer watching "Dancing with the stars" when the phone rings.

Lots of people say statistics are useless, and that they can be made to say anything, but it is really a problem of methodology. Self selecting polls where the surveyed chooses to participate like call in surveys and online surveys are not really very reliable.