Why do people cite surveys?

[tame_deuces]

Basically you just have to disprove of survey research which does not come from a reputable source or isn't peer referenced to similar studies and where it isn't possible to read about the sampling techniques.

[hitch1978]

[ QUOTE ]

For example, in one survey, 60% of respondents claimed they were self conscious about their dandruff.


[/ QUOTE ]

98% of students at my middle school were very conscious of my science teacher's dandruff.

Hello Mr. Beard.

[Johnny Drama]

I've been seeing this commercial on TV that claims something like, "1 in 4 people thinks they sweat more than average"

[Jamougha]

[ QUOTE ]
I've been seeing this commercial on TV that claims something like, "1 in 4 people thinks they sweat more than average"

[/ QUOTE ]

If you ask men to estimate their IQ, polls say that the average answer is 137. [img]/images/graemlins/smile.gif[/img]

[xSCWx]

[ QUOTE ]
I've been seeing this commercial on TV that claims something like, "1 in 4 people thinks they sweat more than average"

[/ QUOTE ]

There used to be one a while back that said 25% of people sweat more than average. It wasn't a joke.

[Siegmund]

It might even be true. If there are a lot of people who sweat very little, but a few people who sweat very profusely, there will indeed be only a few people who sweat more than the mean amount of moisture per day. (There will always be 50% who sweat more than the median amount per day.) A fine example of the sort of thing that gets changed when a copywriter tries to summarize a scientific study.

[ QUOTE ]

Ummm, actually the point is that we can know for sure the distribution of various ailments (e.g., dandruff) in the population.

[/ QUOTE ]

And how could we possibly know that? We'd have to estimate it, either by collecting a sample of people and seeing which ones have dandruff, or by doing something fancy with the numbers of people who see doctors for it or buy shampoos for it. Surveying, in some fashion.

It's certainly possible that two surveys on the same topic will yield different results according to how the questions are worded. There really is no way to avoid that except to test the survey on a pilot group and get some feedback on how their interpreted the questions. There are some standard rules about neutral phrasing of questions, which are a starting point.

[mbillie1]

It is obviously the case that people quite often simply do not know accurately the answer to some questions about themselves despite the fact that they are sure... but it's also the case that you can get statistically legitimate results (though the interpretation might be more difficult) from surveys. Look at the MMPI, the actual meanings to the questions don't exactly matter... but maybe that's sort of a different thing...

also lol @ the unintentionally borrowed Nietzscheism of sarcastically arguing the uselessness of supporting something with such-and-such and concluding by supporting your argument with precisely such-and-such (for as Homer says, "many lies tell the poets")