Chris Johnson wrote: > Why would you be forced to answer Yes? Nobody is holding a > gun to your head. The obvious answer you should give would > be No if you don't favor that kind of government intrusion.
I think that you are missing the point here. If people refuse to answer the question because of its wording, then the results are sure to be biased because you are selectively filtering responses. If people answer it, but would have accepted an alternative if it had been offered, then the results do not accurately reflect the views of the respondents. > Yes, the wording to survey questions is critical, but I don't > think the problem you worry about in this case amounts to any > significant difference. We should all take it and the media with > a grain of salt. But I think you've gone overboard on finding fault > with this story. It's no worse than average. Being no worse than average is pretty bad and I don't think that the people who sponsored this survey are going present its results as if they are only worth a grain of salt. > By that definition, either 100% of the polls I've seen in the popular > press are unscrupulous, or your stringency test is too extreme. No > survey, sample or poll is 100% accurate. Those taken of people's > opinions are even less so than say sampling an assembly line > to detect flaws. We have to assume the research firm knows as much about > statistics, sampling error, and so forth as you and I. If there is > evidence that this particular firm doesn't do good samples > and use good math, well then, let's have it. That would certainly be > grounds for questioning this survey, and I'll do it with you. Okay, to be fair I have to admit that I'm introducing a rather novel concept: Comparison Validity. And, I have to admit that, as far as I know, it's my own term. Nevertheless, it's an important criterion for the evaluation and interpretation of experimental and survey results. I define Comparison Validity as the degree to which results measure the value between alternatives. In education it's common practice to test the significance of an intervention between an experimental group and a control group, which can yield "statistically significant" results, but it tells you nothing about how effective a particular intervention is in comparison to others. The same is true for our current discussion. The fact that 75% of voters favor a ban given no alternatives, tells you nothing about what percentage would favor a reasonable alternative. The only way to know is to include a reasonable alternative in the survey. But! That's not what this survey appears to be about. It appears as though it was intended to show support for one alternative, namely a ban. > If 55% percent said yes, I might be concerned with whether it meant > anything. But 72% is a large enough margin that most any reasonable > error is not going to matter. Further, the obvious test for any reader > to make when reading it is "does it sound reasonable given a > basic set of facts" -- like the fact that 70% to 80% of the population are > non-smokers. So it passes the smell test, if you will. The magnitude of the margin is meaningless if results are biased or if there is no comparison validity; it only reinforces our intuition of what the results would be. > One conundrum: what would make the Star Tribune writer any more > clever than the average citizen and thus capable of such > sophisticated persuasion? It's their profession? I believe that most journalists have college degrees. Shouldn't their degree programs include [good] courses that allow them to critically analysis the validity of data? If not the writer, then why not their editor? > I think it's a valid "which way the wind is blowing" > measurement for how Minneapolis residents feel about smoking. > Accurate to 3 decimal places? No. If we allowed our decisions to determined by "which way the wind is blowing" we wouldn't have a Bill of Rights. Michael Atherton Prospect Park REMINDERS: 1. Think a member has violated the rules? Email the list manager at [EMAIL PROTECTED] before continuing it on the list. 2. Don't feed the troll! Ignore obvious flame-bait. For state and national discussions see: http://e-democracy.org/discuss.html For external forums, see: http://e-democracy.org/mninteract ________________________________ Minneapolis Issues Forum - A City-focused Civic Discussion - Mn E-Democracy Post messages to: mailto:[EMAIL PROTECTED] Subscribe, Un-subscribe, etc. at: http://e-democracy.org/mpls
