1
$\begingroup$

Suppose I'm conducting a multiple regression with two predictor variables (X1 and X2) and have found a significant interaction effect.

Let's say X1 = aggression, X2 = gender (men/women), and Y = health. So I want to know how aggression predicts health differently for men and for women. Because the interaction (X1*X2) is significant in the regression model, I run simple slopes and plot the slopes for men and women at 1 standard deviation below and 1 standard deviation above the mean of aggression: https://i.sstatic.net/GIXno.jpg

Great, so I know the slopes are significantly different from one another.

But what I want to know is: how do I determine if point "Ym1" is signficantly different than point "Yw1"? Also how do I determine if point "Ym2" is significantly different than point "Yw2"?

Does anyone know what the standard method of doing this is? And what the method is called? Any help would be really appreciated!

Improve this question
$\endgroup$
2
  • $\begingroup$ Have you looked at Post-hoc tests? I think this is what you are looking for. Have a look here too. $\endgroup$
    – Stefan
    Commented Feb 25, 2017 at 3:47
  • 1
    $\begingroup$ May be you should recraft your hypothesis and objective of your study. This will help users to respond correctly. $\endgroup$
    – user10619
    Commented Feb 25, 2017 at 5:08

0

Reset to default

Your Answer

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

Browse other questions tagged or ask your own question.