Journal gr8scot's Journal: Did they even ask people if they trust polls?
http://lists.essential.org/pipermail/corp-focus/2008/000272.html
I see a lot of interesting topics in twitter's Journal, but I'm not sure this one is as relevant as the numbers imply. I think people knowingly waste their time replying to polls only if they feel relatively strongly about the topic, which is a type of [self-]selection bias. George F. Will has said much the same thing about voter turnout, in support of his thesis that it's no big deal that 50% +1 of eligible US voters don't, and in fact that only means they believe they will be OK whoever is elected. I'm not enthusiastic about that explanation either, but it might be right, at least for some of the ~50% of cases of not voting. Maybe I shouldn't cite him as a source in support of my speculation, especially because I can't remember the title & copyright date. Oh well, if he wants proper attribution, he can open his own account!
Anyway, 99% of people don't understand statistics, so accurate measurements of their stupid opinions by qualified statisticians really is a quantifiable but unquantified, probably enormous, waste of valuable time which has instead been expended negating value.
I see a lot of interesting topics in twitter's Journal, but I'm not sure this one is as relevant as the numbers imply. I think people knowingly waste their time replying to polls only if they feel relatively strongly about the topic, which is a type of [self-]selection bias. George F. Will has said much the same thing about voter turnout, in support of his thesis that it's no big deal that 50% +1 of eligible US voters don't, and in fact that only means they believe they will be OK whoever is elected. I'm not enthusiastic about that explanation either, but it might be right, at least for some of the ~50% of cases of not voting. Maybe I shouldn't cite him as a source in support of my speculation, especially because I can't remember the title & copyright date. Oh well, if he wants proper attribution, he can open his own account!
Anyway, 99% of people don't understand statistics, so accurate measurements of their stupid opinions by qualified statisticians really is a quantifiable but unquantified, probably enormous, waste of valuable time which has instead been expended negating value.
