User talk:Nwash/Player-Avatar Gender Questionnaire

Okay, I'm definitely inviting comment on this, but before the floodgates are opened, let's get the obvious stuff out of the way. The best hope for overcoming the weaknesses above as much as possible is a large sample size. The larger it gets, the more the law of large numbers should mitigate any bias. The other part is that, as currently written, it only addresses the what, but assessing the why is more complicated. Do we developed separate categories for that, as in males-playing-male characters, males-playing-female-characters, and so on with a list of typical reasons, or do we want to go with an unsigned talk page approach and allow more detailed comments? (Or both?) Nwash 08:41, 10 June 2009 (UTC)
 * It's self-selected. I don't think we have a good way to get a random sample, which would be far more ideal for our purposes. We can only hope that the lack of a random sample doesn't skew the results in any meaningful way.
 * We're choosing from a subset. The ideal would be to choose our sample from all Guild Wars players. This will be limited to those who use this wiki, and we'd never be able to rule out that we have someone different qualities from average players that might skew the results.
 * It relies on self-reporting. We can only hope that people will be honest with this being anonymous. That's never a guarantee, and since we don't have telepathy at our disposal (nor would using it for this purpose be ethical, anyway), we'll have to hope people will be honest.
 * It doesn't have internal checks. I thought about using such tricks as having the same question listed but phrased in different ways as a kind of sanity check for the results, but there are a few reasons I've opted against it for now:
 * It's better to make the questionnaire both look and be short if we want to get maximum participation.
 * We really can't make it a scientific poll without a well-chosen random sample anyway.
 * Internal checks would be pointless unless we're going to use them, which would mean either manually checking every submission or coming up with some automatic method of rejecting inconsistent results.