To: AntiGuv; CCWoody; George W. Bush; OrthodoxPresbyterian
These results are not statistically significant. The article says so itself.
The article says there were ONLY 59 subjects in the entire study. Those subjects came under 4 groupings: Heterosexual male, heterosexual female, homosexual male, and homosexual female. I'm assuming that means there were ONLY 15 individuals in each group. (Woodster, I need your math here.)
First, unless we are dealing with a powerful trait, a small number of respondents is not good research design. Just on that basis, this study can be discarded.
Second, the fact that the numbers were not statistically significant is probably because they were all over the board for the homosexual males and females. With only 15 subjects per category and only 59 for the entire study, I'm guessing the range must have been from one end to the other for all categories to result in the cited averages for the homosexuals not being statistically significant. There was no grouping in any category against which a deviation could be compared.
Basically, this says as at point number 1: seriously flawed study to have so few subjects with a trait that is not strongly evidenced. If you're studying a "nuance," then you'd better have a huge group to make any kind of observations that are statistically valid.
59 divided into 4 groups.....sheeesh
31 posted on
10/06/2003 4:56:22 PM PDT by
xzins
(And now I will show you the most excellent way!)
To: xzins
In general, I agree with all of your observations.
35 posted on
10/06/2003 4:59:30 PM PDT by
AntiGuv
(When the countdown hits zero, something's gonna happen..)
FreeRepublic.com is powered by software copyright 2000-2008 John Robinson