Max Jasper wrote:
> 1:1 & 1:n matched case-control conditional logistic data can be analyzed
> using Cox in SPSS but NOT m:n matched case-control data. Why not?
>
It is a good idea to search the list before asking questions. It is
quite likely that some questions might have already been asked (and
answered) before. This is an excerpt from a message David Nichols sent
to this list in 1998:
"Conditional logistic regression models are also used for 1:N and M:N
matched situations, where there are respectively N controls for each
case, or sets of M cases matched with N controls. The use of COXREG for
conditional logistic regression discussed here will work as long as
there is only one case per set. If there are multiple cases in a set,
the method used in COXREG for handling tied survival times, known as
Breslow's method, isn't going to produce very good results, so I don't
recommend using COXREG in such a situation."
(Full message:
http://www.listserv.uga.edu/cgi-bin/wa?A2=ind9801&L=spssx-l&P=4731)
HTH,
Marta García-Granero
--
For miscellaneous statistical stuff, visit:
http://gjyp.nl/marta/=====================
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
