Odds ratios and regression

BenW

New Member
#1
It's my understanding that odds ratios derived from (multiple) logistic regression is based upon a single unit increase of your variables.

If the data you are looking at is below 1, i.e a marker is measured to be 0.65, for example, then can you divide the odds ratio accordingly:

Test for marker has odds ratio of 30, for each increase of one unit.
For each increase of 0.1 unit OR is 3? (30/10)

Or do you have to involve powers eg OR^ n

30^0.1

thanks for any help.