Just glanced at this before.
Digging a little deeper raises more skepticism.
The claimed inference is that children of deployed military personnel are more likely to engage in certain behaviors.
(p1 > p2) where (p1 = population proportion of military kids engaging in said behaviors) and (p2 = population proportion of non-military kids engaing in said behaviors)
The survey was not designed to test this hypothesis.
Some of the data was just rammed into a chi-square test (chi-square should be viewed with skepticism).
The null hypothesis of the claimed inference would be (p1 <= p2).
In order to have confidence in the inference, the null hypothesis would have to be rejected.
Not sure how they reached that conclusion.
__________________
__________________
Waiting for the perfect moment is a fruitless endeavor.
Make a decision, and then make it the right one through your actions.
"Whoever watches the wind will not plant; whoever looks at the clouds will not reap." -Ecclesiastes 11:4 (NIV)
|