Unbalanced Data Designs

#1
I have four sampling groups with the following sample sizes for each N=201,120,10,40

Question (1)
I need to run an ANOVA to determine each groups preferences on how satisfied they were with our service (0-5) using a Likert scale.

How do I deal with the unbalanced data, the sample sizes aren't even close to same/normalized. Advice?

Would I use a GLM ANOVA and if so how does this test deal with sample size differences?

Question (2)

Let's say the above was categorical data and I needed to run a Chi-square analysis, how do you deal with unbalanced designs with Chi-square tests?


Thank you!
 
E

elnaz

Guest
#2
hello
about final question , you should use kruskal wallis test not chi-square test