Should I analyze variables separately or together?

Question

I have one categorical variable (on a diet: yes or no) and two numerical variables (patients heart rate and blood pressure). Do I use a multiple logistic regression? Or should I look at each numerical variable separate with regards whether they are on a diet? That way I would only look at simple logistic regression.

Answer 1

You should analyze all of the variables you are interested in together. The relationships between your explanatory variables and the response variable may differ based on whether you are ignoring or controlling for the other variable, and the latter is certainly what you really want to know. Although written in the context of linear regression, my answer here: Is there a difference between 'controlling for' and 'ignoring' other variables in multiple regression? will help you understand this better.