Hi, I am currently working on some questions in SPSS and am not sure of one of the ways I'm running the test of normality.

I have a question where I'm asked about whether statistics vary by region per year group (so I have 4 regions, across 3 years). I have been told to run a one-way ANOVA. So, in testing normality I have checked data per region per year by collating them into one column of data and one column labeling the data from the year group and region it belongs to. This then allowed me to run Kolmogorov-Smirnoff etc.

The thing is, I'm not sure if I'm supposed to test normality that way or via one column of data and one labeling it by year group (without the further breakdown of region).

The first version gives me 12 sets of sig. (4 regions, 3 times over each for each year group) and all are normally distributed (only one is close to 0.05 at 0.51). But if I just look at year and data giving me three sets of data with sig., I have a whole year group not normally distributed (0.020).

What is the correct method here?

Thank you, Vanessa