Multi-way Regression with Prevalence as the Outcome

Return to Multi-way Page 2

This exercise repeats the previous exercise for running model 7, but instead of using the outcome variable waz, the outcome variable is switched to a dummy for underweight (<-2 SD weight /age) yes/no called waprev, which yields a prevalence outcome.  Follow these steps to run the regression model:

1.  Open keast4j.sav

2.  Click on Statistics, Regression, Linear...

3.  Enter the variable waprev from the variable list into the box labeled Dependent using the arrow key.

4.  Enter these variables, one-by-one into the Independent(s) box using the arrow key:  dlowedn, dbadro, notoilet, dpiped, dwell, vincome.

5.  Click on the box to select Enter Model.

6.  Click on OK.

wpe7.jpg (9740 bytes)

wpe8.jpg (13748 bytes)

wpe9.jpg (21227 bytes)


In this model, as before when waz was used, the results have similar significance levels.  Still, education and income are the strongest influences on the outcome (p=0.000).  The previous model using waz showed a -0.321 point change in waz score for each unit decrease in education, and in this model that corresponds to a 15% change in the prevalence of malnutrition.

Return to Top of Page