At 05:45 PM 12/10/2014, Shin wrote:
>Question 1
>How can one do a 1:2 or 1:many case control matching? I'm using
>trial version 22. I was able to do 1:1 but want to get a bigger control group.
It depends on whether you have some exact key on which to match your
cases and controls; AND that key is a unique identifier in the cases
file, but not necessarily unique in the controls file.
If you have this, then the operation is a MATCH FILES with the case
file specified as /TABLE=, and the controls file specified as /FILE=.
That will give you ALL controls that match each case. If you want no
more than, say, 2, you should then select 2 of each, at random.
Two other points:
. If you have only approximate matches of cases to controls on
relevant characteristics, it's more complicated, and you'll have to
tell us more about your data.
. I've never used the Complex Samples option to SPSS, but if you have
it, and some understanding of it, see if it can be used to solve your problem.
>Question 2
>Is there any way to perform case control matching using version 17?
Everything I've written will work that far back, and farther. Of
course, if you use Complex Samples, that option will have to have
been installed on the particular copy of SPSS you are using.
=====================
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
