XQueery Statistics

Related link: http://lists.xml.org/archives/xml-dev/200503/msg00441.html

Last week I had a little dummy-spit about the
interpretation of some survey results about XQuery
(see here)
and I have been in a statistical frame of mind all week.

So the recent Sports Illustrated survey on Homosexuality and Sports engrossed me. Only 14% of respondents didn't want to accept openly gay male players. Apart from the surprise that respondents were less tolerant of openly lesbian sportswomen, many of the other results are difficult to reconcile. Subtle wording differences in basically repeated questions resulted in very different results: I'd love to know whether this reflects nuanced views or knee jerk reactions to certain words.

The results are online, but here's my interpretation, joining the dots:

• 50 to 60% of people are positive to neutral about gays, whether in or out of sport. They probably have a gay friend, collegue or family member.
• 10 to 20% of people have misgivings, but don't support discrimination or stereotyping. They think that greater acceptence would be a good thing for the country, and that it is the private business of the person, who they would prefer to be discrete. Don't ask, don't tell, especially if you are a lesbian.
• 10 to 15% of people make some kind of distinction between "gays" and "homosexuality": I suspect they are make a distinction between tendencies and activities. Love the sinner but hate the sin. Even this group does not believe in policing the bedroom however.
• 15 to 20% of people are strongly against any open or private gayness, and are uncomfortable around gays.
  They believe that children will turn gay
  from exposure to role models.
  

SI's poll seems to have used random sampling, as distinct from a voluntary survey. When the numbers are reliable, conjectures interpreting them have some substance. When the numbers are unreliable, interpretations of the figures are mere puff.