I want to explore independent associations with gender (male/female). Since gender is not an outcome variable, is it valid to include it as a dichotomous dependent variable for Binary Logistic Regression? I want to look for adjusted associations with gender (will be adjusting for age and a few other demographics), and was wondering whether OR (95% CI) would be a valid way of expressing these associations. Though mathematically it seems to be correct, is using gender as a dichotomous dependent variable intuitively correct (as it isn't an outcome variable)?
Thank you.
Yes you can include it as a nominal or scale variable coded 1\0 or 2/1.
In this paper, survey interviewer gender was the dependent variable in a logistic regression I conducted: https://www.researchgate.net/publication/337293480_It_Depends_on_Who%27s_Asking_Interviewer_Gender_Effects_on_Credibility_Ratings_of_Male_and_Female_Journalists_in_Six_Arab_Countries
Yes ,if your depended variable is Binary Logistic Regression , you can choose it
But, the (independent variable) causes a change in the (dependent variable). In my optioned, gender probably not so much . Usually, the gender won't make sense Your choice could lead you to perform additional experiments since it might turn out the change in independent variable .
- Badges
- Science topic