What happens in multiple logistic regression when the outcome variable
and a predictor variable contain the same construct? For instance,
using survey data, a regression model that contains an outcome
variable and 14 predictor variables was done. The outcome variable and
4 of the predictor variables were constructed from several different
questions in the survey (the other 10 predictor variables were single
questions such as age, sex, race, etc). However, one question was used
in the creation of both the outcome and one of the predictor
variables. How does this affect the results? I would like to get an
answer asap (within a day if possible). Thank
you.