TO: Students in CJ 405

FROM: R. B. Taylor

DATE: 4/3/99

RE: Homework for Thursday

NOTE - SEE BEGINNING OF FILE LAB0318 FOR THE SYNTAX TO CONSTRUCT THE CONTROL2 VARIABLE, IF YOU ALREADY DO NOT HAVE THAT IN YOUR FILE. It is an average of six z scored items. You need to first get the z scores, with the descriptives procedure, then compute their average, and save the new variable. This equally weights each of the component items.

We are now doing multiple regressions. Please run at least the first two multiple regressions listed below. For each multiple regression interpret:
R squared and F test of R squared
b weight (and t test of b weight)

These are from the individual-level data file. I strongly encourage you to have your copy of the questionnaire in front of you when you are interpreting your results; remember, it is important ot interpret your results in the context of the specific regressions you are seeing. Think in terms of the units of the variables. In addition, I encourage you to be clear about each outcome: what does a high score mean?. What does a low score mean? In all cases the outcomes are (or can be treated as) averages of z scored items. Try to embed the results in your text as you go along, and try to write for a policy audience. In other words, interpret the results, presuming these folks are the representative random sample of ....

REGRESSION 1

REGRESSION

     /DESCRIPTIVES MEAN STDDEV CORR SIG N

     /MISSING LISTWISE

     /STATISTICS COEFF OUTS R ANOVA COLLIN TOL

     /CRITERIA=PIN(.05) POUT(.10)

     /NOORIGIN

     /DEPENDENT control2

     /METHOD=ENTER q1 q73 ..

REGRESSION 2

REGRESSION

    /DESCRIPTIVES MEAN STDDEV CORR SIG N

     /MISSING MEANSUB

     /STATISTICS COEFF OUTS R ANOVA COLLIN TOL

     /CRITERIA=PIN(.05) POUT(.10)

     /NOORIGIN

     /DEPENDENT confi2

     /METHOD=ENTER q14 white q79 q2_3 .

OPTIONAL REGRESSIONS - just run them out and bring them; do not need to interpret

REGRESSION

        /DESCRIPTIVES MEAN STDDEV CORR SIG N

        /MISSING LISTWISE

        /STATISTICS COEFF OUTS R ANOVA COLLIN TOL

        /CRITERIA=PIN(.05) POUT(.10)

        /NOORIGIN

        /DEPENDENT control2

        /METHOD=ENTER q1 q73 AFRAM Q2_3

==============================

REGRESSION

     /DESCRIPTIVES MEAN STDDEV CORR SIG N

     /MISSING MEANSUB

     /STATISTICS COEFF OUTS R ANOVA COLLIN TOL

     /CRITERIA=PIN(.05) POUT(.10)

     /NOORIGIN

     /DEPENDENT confi1

     /METHOD=ENTER q2_3 philly .