Hello,
I have trouble deciding on whether these scatterplots indicate homoscedasticity for my 3 samples (a different one for each hypothesis).
If I understand correctly, I should be checking, if the data points vary the same for each standardized predicted value. So it does not matter, if there are for example 3 data points at (x = -3) and 100 data points at (x = 1). It just does matter, if they are distributed the same way?
I just do not understand how this can be estimated just by looking at these scatterplots. My question is: Would you say, these indicate homoscedasticity and why or why not?
I am using a hierarchical regression with 2 steps. In the first step, I included the same 2 demographic variables for each hypothesis. In the 2nd step, I included 1 independent variable for each hypothesis.
Kind regards,
Max
Hello Max Müller, I'm not really sure about my answer or interpretation, but you say you look at the data. Don't you perform a statistical test for homoscedasticity?
Take a look at the last case in this paper
Which shows heteroscedasticity and suggests a way to deal with the problem. Best wishes David Booth
Your plots show homoscedasticity as the variance of residuals appears to remain stable across the range of predicted values.
However: If the goal of this exercise is to verify the assumptions of ANOVA (or a linear model), it fails for two other reasons:
Perhaps, these three chapters of my book help you to find a better model for your data. It looks like Likert scale items:
https://schmettow.github.io/New_Stats/wwm.html#model-criticism
https://schmettow.github.io/New_Stats/glm.html#elements-glm
https://schmettow.github.io/New_Stats/glm.html#rating-scales
- Badges
- Science topic
- Similar topics
- Analytical Chemistry
- Extraction