Andrew Suffield wrote: > As somebody who understands and hates statisticians and statistics, I > echo the sentiment. This is a terrible survey.
I can understand why people hate social statistics when they are often produced from bad surveys. I think that it's a bit harsh to hate statisticians, though: often the worst surveys are prepared without using a trained statistician and sometimes even without basic knowledge of statistical methodology. If there were more statisticians actively involved in collecting and producing these statistics, there would be far fewer bad surveys out there. (I would say that, though, wouldn't I?) In this particular case, there is no statistician listed on the flosspols site and I did not find the methodology documents. They might exist unpublished, so I asked. Andrew's sentiments illustrate how developers of long-running projects like debian are getting tired of poor survey design. I've picked that up anecdotally, although I've not tried to look for a quantifiable trend (yet). I think researchers would do well to recognise this irritation and address it if they want good support from the community they survey. Some other projects that sociologists find "interesting" have pages with advice for researchers. Does debian have one or would anyone be interested in (me or them) drafting one? -- MJR/slef Personal opinion http://people.debian.org/~mjr/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
