Date: Thu, 31 Oct 2002 08:31:22 -0800
Reply-To: "Parise, Carol A." <PariseC@sutterhealth.org>
Sender: "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From: "Parise, Carol A." <PariseC@sutterhealth.org>
Subject: logit linearity in ordinal logistic
Content-Type: text/plain; charset=us-ascii
One of the assumptions of logistic regression when using a continuous
predictor is that the variable is liner in the log odd of the DV. The way I
have been doing this in SPSS for 2-category logistic regression is by taking
the contiunous variable and making quartiles or quintiles out of it and then
sticking it a logistic model with the DV. if the ORs increase or decrease
linearly for each category of the DV then I say this is linear in the log
odds and the assumption is met. Is this the same approach that needs to be
taken for ordinal logistic regression or is there a more streamlined method
of testing this assumption that i'm missing?
Thanks for any ideas.