Date:         Sat, 3 May 2003 13:44:49 -0700
Reply-To:     Ellie Buteau <ebuteau@prodigy.net>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Ellie Buteau <ebuteau@prodigy.net>
Subject:      Logistic Regression Backwards LR Entry
Content-Type: text/plain; charset=us-ascii

Dear Listserv, I am having trouble interpreting spss output for Binary Logistic Regression Backwards LR entry. Bivariate analyses indicate several continuous independent variables relate to my binary outcome variable. I want to find the most parsimonious logistic regression model for my outcome, and chose the Backwards LR order of entry. But I am confused about how to interpret 1) the 'Omnibus Tests of Model Coefficients' box for the various steps in the entry, and 2) the 'Model if Term Removed' box (i.e., what does the p value for each variable in the 'significance of the change' column indicate - what does p<.05 mean in this column?). After searching google.com for any web pages that discuss spss output for backwards entry in logistic regression and finding nothing, I would appreciate any assistance you can lend! Thanks. Sincerely,Ellie Buteau, Ph.D.
Tufts University