I have three groups in a Hierarchical Discriminant Analysis (Political Affiliation).

I want to see whether the addition of psychological variables (2 variables: beliefs of the positive effects of of XYZ, and definition of what kinds of stimuli are considered ABC) will lead to improvement in classification accuracy. After doing a crosstabs on the 3 Dependent Variables to the Independent Variables (Demographics/Background) I want to run a test of improvement using McNemar's repeated-measures chi-square.

I thought the basic outline for syntax for this was as follows, but I am not sure where I would insert my own variables (Cut and paste kind of thing) I put "politic" (political affiliation, the dependent variable in as shown below) My question is : 1) Is this the right syntax? 2) Where would I input my own variables in replace of the default variables (I got this format from someone)? 3) Is "politic" in the right location? 4) How would I run the analysis? 5) Am I missing anything?

*Create variables for correct/incorrect decisions at each step.

IF (politic eq 1 and STEP1C eq 1) STEP1D = 1 . IF (politic eq 2 and STEP1C eq 2) STEP1D = 1 .

IF (politic eq 1 and STEP1C ne 1) STEP1D = 2 . IF (politic eq 2 and STEP1C ne 2) STEP1D = 2 .

VARIABLE LABELS step1d "Classification decision - step 1" . VALUE LABELS step1d 1 "Correct" 2 "Incorrect" . FORMATS step1d (f2.0) .

IF (politic eq 1 and STEP2C eq 1) STEP2D = 1 . IF (politic eq 2 and STEP2C eq 2) STEP2D = 1 .

IF (politic eq 1 and STEP2C ne 1) STEP2D = 2 . IF (politic eq 2 and STEP2C ne 2) STEP2D = 2 .

VARIABLE LABELS step2d "Classification decision - step 2" . VALUE LABELS step2d 1 "Correct" 2 "Incorrect" . FORMATS step2d (f2.0) . EXECUTE .

Thanks,

Tamiko