Hello,
I want to perform a single mediation (model 4, hayes) with Process in SPSS.
Before I can do that, I have to check the assumption for linearity. I created a matrix scatter-plot, but because my independent variable is binary it doesn't look right. I have 1 dependent variable, which is metric.
Does anyone know, how I can check for the linearity assumption, if my independent variable is binary and single level (Intervention group vs. control group)?
There is no need to check for linearity with a binary independent variable. You can always draw a straight line through two points (the two group means on your binary variable). More relevant would be homoskedasticity (the assumption of equal variances/error variances in the two groups in the population).
@christian geiser
thank you very much for the answer!
Do you have a source, so I can cite this or is it just logical?
Do I have to check for homoskedasticity even if I use bootstrapping with the Makro Process from Hayes. I have read that Process is robust against most assumptions.
Sorry I'm late to the party but regression.they're 2 possible answers depending on what your question is about. Christian's answer is correct assuming the standard regression interpretation of your question. However in logistic regression linearity often means linearity of the logit transformation on the DV side. In that case in that case it is usually true. See the literature on logistic regression
- Badges
- Science topic
- Similar topics
- Indexes