Homogeneity of regression
As I am working on my ANCOVA homework I am having trouble figuring out how to test for homogeneity of regression in SPSS. My notes are telling me to go to analyze, then the regression tab, then the linear tab, but after that I am not finding a way to regress the dependent variable onto the covariate. Can someone offer some help with this?
Thanks
Thanks
posted by Grant Goodman @ 5:08 PM
4 Comments:
At 7:16 PM,
David said…
You're almost there. Just select the DV and put it under "Dependent." Select your covariate and put it under "Independent(s)."
But to test the homogeneity of regression, you'd have to make sure the slopes (betas) are similar for each level of IV. You can do this by selecting cases and running a regression for each level. Then you can compare those betas.
At 9:45 PM,
Erika said…
Hi David,
When you say select cases, do you mean moving something over to "case labels"? I'm confused as to how to get the pretty slope graphs once I'm in the Linear Regression box.
Thanks!
At 7:51 AM,
David said…
I originally meant going to Data -> Select Cases and doing the regression twice.
Another way to do it is to to to Analyze -> Regression -> Linear. Place the appropriate variables in Dependent and Independent. Then you can place the intervention group variable in the slot "Selection Variable." Click "Rule" and it will let you select which value for group you want to look at. You would still need to run the regression twice.
For the graphs, you can use Graphs -> Chart Builder. But again, you'd have to select cases and do it for each IV level you have.
At 6:56 PM,
bfullmer said…
Hey Grant, you can also find the Beta values in the coefficient boxes when you run analyze, regression, and linear. You want them to be in the same direction (positive or negative) and similar in size. Hope that helps!
Brooke
Post a Comment
<< Home