I have a problem and I don´t know whether is due to SPSS software or just because of statistics in general.
I hava tried to make binary logistic regression model with one covariate being binary coded end point (1= yes, 0= no case). In other group in this analysis no such end point exist (so all cases in other dependent variable group there´s zero cases = zero data = 0%). Therefore the SPSS binary stepwise logistic regression model cannot calculate any meaningful result for this variable. It only has very high B, Wald 0.000 and Exp(B) has a crazy non-meaningful value.
Does anyone has some good ideas or methods how to avoid this problem?