When you run a logistic regression in SPSS, keep in mind what constitutes an event. I am speaking of the dependent variable where a value of ‘1’ is under study. You are calculating the odds of obtaining a ‘1’ . And it’s usually based off a secondary variables that has an exposed and unexposed group. Know what the value ‘1’ represents because the interpretation depends on it. If the value represents the occurrence of something NOT happening you want to realize that BEFORE you do the interpretation. Becausr the software is not going to tell correct you.

Your logistic model is dependent on your definitions. If you start to interpret your odds and it does not make sense, revisit the definitions. The problem is not the odds ratio. For example, if your odds suggests that men get breast cancer more than women, we know to double check the dependent variable. There may be a chance that our definition of an event (equal to one) is reversed making ‘0’ the event. The message here is to just check your work.

-Moore to follow Amy

### Like this:

Like Loading...

*Related*