Hi.

i have been given some advice by someone to carry out two different analyses:

1. Use transformed data in a non-linear model so that the effect of my explanatory variable can be monotone decreasing.

2. Use raw data in a logistic regression model

i have a couple of questions about this as I'm quite a novice when it comes to regression.
- When I'm doing the non-linear regression, is it possible to also include a number of other variables and do a backward stepwise deletion to get to the best model?
- When I'm doing the logistic regression can I include my explanatory variable as a non-linear function...or do i just include it as a variable in the normal logistic regression model?

Thanks,
Nicola