I want to simulate a data set for a dependent variable that takes on two levels (0 and 1). I want one of the IV to be linearly related to the DV and another that is non-linearly related. I can do this (if I know how) in either SAS or SPSS (meaning I have both - I don't know how to do it).
Any suggestions would be appreciated. One problem for me is that logistic regression (where this is really aimed) assumes that the logit of the DV is linearly related not the raw value of the DV (unlike linear regression). I am not certain if a variable that is linearly related to a binomial DV (or non-linearly related) will always be related the same way to the DV's logit.
Any suggestions would be appreciated. One problem for me is that logistic regression (where this is really aimed) assumes that the logit of the DV is linearly related not the raw value of the DV (unlike linear regression). I am not certain if a variable that is linearly related to a binomial DV (or non-linearly related) will always be related the same way to the DV's logit.