2
$\begingroup$

In my paper, in the section methodology, I have a subsection called dependent variable. But since I am performing a logistic regression, there seem to be 2 (or even 3) things that could be named the dependent variable:

First you have your Y, which is a measured variable.

Second there is the predictedlog-odds of the probability.

Third there is the predicted probability (easily calculated from the log-odds).

I am wondering which one to describe in my paper in this section? I prefer not to talk about the log-odds, since it makes no sense to make things more complicated at that stage of the paper. I wanted to say the actual Y, cause that's what people are used to in linear regression, but it seems the Y is not dependent of the independent variables, its dependent of the probability. Could I safely talk about the probability, without breaking any rules?

Does anyone have any experience with this? Maybe an example of a research paper/article.

$\endgroup$

1 Answer 1

7
$\begingroup$

I don't think one needs a paper. The dependent variable is the Y variable (these are used as synonyms.)

The log odds, and the predicted probabilities, are what are modelled from the logistic regression -- these are results, rather than model inputs.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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