Data analysis

Vasileia_chr

Hello everyone,

I know that this question doesn't belong in this forum but the truth is that in the past you have helped me a lot with my OpenSesame related questions, so I hope that this time it will be no exception😊. I am actually writing about the SPSS analysis of my data.

I have run the Shapiro-Wilk normality test and I' ve found some of the variables not being normally distributed, yet my Levene's test for homogeneity was confirming its null hypothesis. All in all, the question is can I still run a repeated measures within factors ANOVA (with a sample of 34 people), even if I don't have normally distributed data but there is homogeneity in them?

Thank you all in advance!

Yours faithfully,

Vasileia

EJ

Dear Vasileia,

The ANOVA is usually robust to violations of normality. Specifically, this paper (https://link.springer.com/article/10.3758/s13428-017-0918-2) suggests that when the min vs max variance ratio < 1.5, the results should be robust. Of course you might run a Welch ANOVA (which does not assume variance homogeneity) to confirm that you reach the same conclusion.

Cheers,

E.J.