3
$\begingroup$

What are the assumptions that need to be met when using hierarchical regression, and subsequently simple slopes analysis to test and probe interaction effects and how can I test for them in SPSS?

I have a continuous DV and IV, with a categorical moderator (3 categories).

Thanks in advance.

$\endgroup$
3
$\begingroup$

The assumptions are the same as those that are made for hierarchical regression analysis without interactions, including the following:

  1. Variables are approximately normally distributed. With large sample sizes (n > 300) it is best to chek distributions with the SPSS EXAMINE command. You should also attend to bivariate and multivariate outliers using Mahalanobis distance.
  2. The relationship is linear in the parameters. It is fine to have curvilinear or higher ordfer effects, modeled by raising the IV to a power and then weighting that term by a parameter - you still have an additive model. But you should not have a model in which a parameter is raised to the power of the IV - that model would not be suitable for OLS regression.
  3. Variables are reliably measured
  4. Homoscedacity of error variance. Check for lack of fit due to curvilinear and higher order effects - including interactions - by examining the plot of the residuals.
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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