10-25-2013 04:49 AM
After running a logistic regression i get the output containing the estimates for the scoring variables.
My question is the following, how do i interpret the estimates of the continuous variables that i have used to someone? How can i show him the likelihood of an event based on that estimate?
What does a 0.42 estimate for No Accounts tell me for a persons likelihood to be an event int he future?
Thanx ina advance
10-25-2013 09:54 AM
Well, exponentiating the estimate would give an odds ratio, based on a one unit change. So for each change in No Accounts of 1 unit, the odds of <insert response here> are roughly 1.5 times higher. Is that what you are seeking?
10-30-2013 12:33 PM
Here is link to SAS STAT user guide that has some details on how to interpret proc logistic. For more depth, check books written by Paul Allison. If you visit www.lexjansen.com, you can find many past proceedings covering the subject.