We are having a controversy here on how to enter "duration of treatment"
in a design with 2 comparison groups which unfortunately are not control
groups.
In examining whether the relation between X and Y is different when a
treatment is applied to a group, we simply,
1) enter X, flag1, flag2
2) enter flag1*xcentered flag2*xcentered.
comparison group 1 is represented by 1 and 0 for flag1 and flag2
comparison group 2 is represented by 0 and 1 for flag1 and flag2
treatment group is represented by 0 and 0 for flag1 and flag2.
However, the treatment is not of the same duration.
We have a variable that shows the duration of treatment, so that it
would be 0 for group1 and 0 for group2, but a number of months for the
treatment group.
Art Kendall
Social Research Consultants
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Locked
