Hello! I am conducting a 2-way factorial ANOVA. Normality was violated (Shapiro-Wilk Sig.=.000) and homogeneity was also violated (Levene's Sig.< 0.05). I understand that here isn't a common nonparametric test for Factorial ANOVA. I am using Excel and SPSS. Do you have any recommendations on how I should proceed with this?
I must mention that I am somewhat inexperienced in statistics and will appreciated all of your feedback.
Thank you.
How about the normality and homoscedasticity of residuals? If they are OK you can go.
There are robust ANOVA methods, but you may not need them.
As Mehmet said: it's about the residuals, not about the data itself.
You haven't mentioned what your response variable is. If it is a continuous measure for some kind of concentration or intensity, you should analyze the logarithms, what likely solves the issues with the violations of assumptions. If you have counts, proportions, binary data, ordinal data, you should consider a negative binomial, binomial, logistic or ordered logistic model instead.
You might calculate the variance of each group, both factors, and all interactions.
Thereby, maximum var / minimum var should not exceed 3.5
That's a rule of thumb.
If higher, assumptions are severely violated.
Does the data possess a balanced design?
Also check histogram of response variable.
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA