1
$\begingroup$

SPSS Help provides the following help for testing group differences in a correlation, using the GLM approach. See link.

It first requires that X and Y are transformed into z scores.

Question: When is this transformation necessary?

$\endgroup$

1 Answer 1

2
$\begingroup$

The GLM approach is pretty painful here, but your GLM is really doing a regression analysis, so if the variables have a different scale, you would reject equality just because of that. By standardizing the variables, the coefficients are equivalent to correlations, so you are doing the test on equality of correlations, and scale differences don't count.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.