Logistic Regression: A Self-learning Text, Third Edition (Statistics in the Health Sciences)

GRADE is the response variable. GRADE has three levels, coded 0 for well differen-
tiated, 1 for moderately differentiated, and 2 for poorly differentiated.

The model is stated as follows:

ln

PðGRADEg*jXÞ

PðGRADE>g*jXÞ



¼a*g*b* 1 RACEb* 2 ESTROGEN forg*¼ 0 ; 1

Note that this is the alternative formulation of the ordinal model discussed in Chap.

13. In contrast to the formulation presented in the SAS section of the appendix, SPSS
    models the odds that the outcome is in a category less than or equal to categoryg.
    The other difference in the alternative formulation of the model is that there are
    negative signs before the beta coefficients. These two differences “cancel out” for the
    beta coefficients so thatbi¼bihowever, for the intercepts,ag¼ag, whereagandbi,
    respectively, denote the intercept andith regression coefficient in the model run
    using SAS.

To perform an ordinal regression in SPSS, select Analyze!Regression!Ordinal
from the drop-down menus to reach the dialog box to specify the logistic model.
Select GRADE from the variable list and enter it into the Dependent Variable box,
then select and enter the covariates (RACE and ESTROGEN) into the Covariate(s)
box. Click on the Output button to request a “Test of Parallel Lines,” which is a
statistical test that SPSS provides that performs a similar function as the Score test
of the proportional odds assumption in SAS. In the main dialog box, click on OK to
run the model with the default settings.

The command syntax for the ordinal regression model is as follows:

PLUM grade WITH race estrogen

/CRITERIA¼CIN(95) DELTA(0) LCONVERGE(0) MXITER(100) MXSTEP(5)

PCONVERGE(1.0E-6) SINGULAR(1.OE-8)

/LINK¼LOGIT

/PRINT¼FIT PARAMETER SUMMARY TPARALLEL.

The output generated by this code follows:

PLUM – Ordinal Regression

Test of Parallel Lines

Model 2 Log likelihood Chi-square df Sig.

Null hypothesis 34.743

General 33.846 .897 2 .638

The null hypothesis states that the location parameters (slope coefficients) are the
same across response categories.

SPSS 645