0
$\begingroup$

I have conducted a hierarchical regression with 2 categorical variables. One of which I am controlling for (ethnicity-dummy coded).

I need to check the assumptions of normality, linearity and homoscedasticity of residuals. My question is, do I check the distribution of the residuals at each of the 2 levels of my independent variable, or do I need to check the distribution of the residuals all together (simultaneously)?.

Improve this question
$\endgroup$
0
$\begingroup$

You should check the distribution of the residuals. You can either look at the qq-plot, at the histogram of the residuals or make a formal test for normality (e.g. the Shapiro-Wilk test).

Improve this answer
$\endgroup$
2
  • $\begingroup$ Hi Anna. Thank you for your assistance. Should I check the distribution of the residuals for each level of my IV (0= no medication, 1= medication) separately, or should I check my residuals all together (E.g.,. Medication and no medication combined). I ask this because when I run some additional analyses, and check the normality of the residuals seperately for each group, one level of the Iv is normally distributed and the other is not. I hope this makes sense, sorry. $\endgroup$
    – RosettaJo
    Sep 14 '20 at 5:16
  • $\begingroup$ I don't know if I understood it well. You should simply run your regression with all the data on all the IVs you have and then check for normality of the residuals. Maybe this tutorial with dummy variables can help you also from the practical point of view rpubs.com/Loy/dummyvariables $\endgroup$
    – Anna
    Sep 14 '20 at 12:32

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.