|
|
Dear all !
I have three groups (1,2,3). I have three states at two points in time - pre- (1a,2a and 3a) and post treatment (1b,2b,3b). Groups are binary (membership: 0/1).
The "states" I have in two forms - binary and continous both for pre- and post treatment measures. I am told to do a logistic regression in SPSS, in order to obtain odds/ratio measures for groups as well as specificity and sensitivity.
However - performing a logistic regression does not give my groups ANY predictive power at all - DESPITE they differ when crosstabulating them - either with pre- or post treatment measures. So the groups are different and have different outcomes. What is wrong? Can you advice - please !!!
/ Anders
|
Locked
