interpreting logistic regression

#1
I am having a tough time interpreting output in terms of what the coefficients and odds ratio say about the estimated effect. For example, say you are comparing a dichotomous dependent variable x with an independent continuous variable y. You get a constant coefficient of -5.0 and an x variable coefficient of 0.05, with an odds ratio of 1.05. p value indicates significance.

What is the effect of y on x? With every 10 unit increase in y, what happens to x? How do you state this in typical speak found in the literature? If using the odds ratio, is it:

"for every one unit increase in y, there is a 5% increase in the odds of x"?

or is it "for every one unit increase in y, there is a 5% increase in the log odds of x?"

Using the coefficient, is it "a one unit increase in y results in a ??? increase in x?"

Thanks.....
 

WeeG

TS Contributor
#2
this is exactly how it's done...

you say: for a 10 unit increase in y, the odds of having the event x, are 1.648 times bigger.
( 1.648 = exp(10*0.05))