0
$\begingroup$

I have a variable I do not know how I should handle my logistic regression. The variable is the number of registered students each semester. If I plot it against my binary outcome, I get the following plot: enter image description here

what kind of explanatory variables should I use? Linear, polynomial, categorical? I feel myself confused when it looks like this and would therefore need some tips.

$\endgroup$
1
  • $\begingroup$ Are you showing us the entire sample? Even a linear relationship would be a stretch with just twelve data points, so higher order terms are pretty much out of the question. $\endgroup$ Commented Apr 29, 2015 at 12:32

1 Answer 1

1
$\begingroup$

Without any other information, a linear term is the obvious choice, since the number of students registered is on a ratio scale. This may or may not make sense given all the missing context of the purpose of the study, the nature of the outcome, the other variables available, the sample size, etc.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.