Hi, just a small query.
In M+ when running logistic regression, you define nominal covariates using the 'define' command, for example, were a (gender) var is coded 0 = male/1 = female :-
define:
female = gender == 1;
(or it could be written as)
male = gender == 0;
M+ normally assigns the highest number as the reference group. My question is (for nominal variables) in a situation like this, would defining the variable either of these ways change the reference group?, that is would the output still be interpreted as 'males when compared to females are more/less likely (at risk etc) no matter what way the var is defined, or is the reference group the group you 'define' or do not 'define', (as in any usual regression in M+ when dummy coding, the not defined group is usually the last group and would always be the reference)
Just wondering if you can manipulate the reference group in this way and get this straight in my head.
Bests