Hi All

I looked on a post written way back in 2014 and someone was asking about a hyp. test they were doing using binomials. essentially a prtest.

It was suggested that they try switching to logistic away from their prtest wealthy, by(sex) [or whatever it was]

I decided to do this for my work... I run the hypothesis test (prtest) and the logistic regression of the exact same composition (with one IV and the DV both dummy variables) i have different results. i.e. the hypothesis tests says one is about 4.5 times more likely than the other (mean 4.5 times higher) and the logit seems to imply 5.

Why is this? both have very similar stat sig (logit is better), but I am not sure if I should continue to use one over the other

Thanks

Robbie