The posting about the 62% thing is useless without also posting either the raw numbers or the overall rate. Those particular numbers, 62%, 34%, and 4% no-response, could have been achieved with as few as 47 respondents overall: 29 female, 16 male, and 2 no-response.
In other words, as few as 13 additional responses by women could have tipped the scales drastically. Yet if it was only 47 respondents out of many thousands, the overall rate would be negligible, far less than 1%.
-- and this is very
likely to be the case -- the numbers for people feeling unwelcome in a religious setting could be far
higher than in the non-religious setting, and still yield a 'lower' percentage of female versus male, by the mere fact that larger sample sizes have less variance in their means. For example, it could be that there is a ten times higher rate of feeling uncomfortable in religious settings, but if the numbers happen to be 225 female and 225 male, then it would show up as 0.5 vs. 0.5, which appears to show that women feel less comfortable in the secular movement, but meanwhile 225/29 shows a 7.75 times better climate for women in secularism than in religion. This is all just speculation, without the actual data, no one
can make an accurate assessment.
This is an example of the Base Rate Fallacy
, and closely related to the Prosecutor's Fallacy
and Simpson's Paradox
, all of which are examples of misunderstanding conditional probability. See also http://en.wikipedia.org/wiki/Conditiona ... _fallacies
Myers has no excuse for this. He is pushing propaganda, pure and simple. I highly recommend folks here read a bit on conditional probability, if they are unfamiliar with it. It's actually a pretty interesting concept (that we already use in our daily lives, usually intuitively and without realizing it), and learning the theory behind it can really help clarify one's thinking. A great tool to add to your baloney detection kits.