The worthlessness of polls confirmed

I have been complaining for years that polls are completely worthless (and that goes for surveys and most marketing data collected in online forms and all that crap). I know about the statisticians and the sample size and the correlations, etc, but I have always simply said that you cannot rely on what people tell you they might do in the future (people lie intentionally). Even if polling companies were 100% accurate in recording what they were told (which they likely arent) and if polling questions were 100% neutral (leaving out all Republican and Democratic polls, \"Should John McCain and his illegitimate black daughter be chosen as the Presidential candidate to run against the Socialist Flip Flopper that shot young men in loincloths in the back!\") so as to garner a useful result, even then, you still have the human factor (and I am aware statisticians think they can account for this but I dont buy it).

And above and beyond just the normal fallibility of polls we also get reports that popular polling companies just make shit up.

Comments

RE: The worthlessness of polls confirmed

One of the comments on TechDirt:

"I worked (on the phones, as a supervisor, then as a manager) in two different market research offices (one of which was global and i think they were the 5th biggest MR firm in the world...) I can tell you, not just "polls are worthless." BUT ALL of market research is worthless!

First off the collected information is skewed, the respondents who actually talk to us make up 1/2 the stuff they say. Secondly the need a quick buck minimum wage high school interviewer is falsifying information left and right. After the interviewer gets done, the supervisors edit the responses, then the coders edit the responses into catigories which most the time don't accurately represent what the respondent said in the first place.

I could go on for hours..."

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.