Repeated Measures Anova question

#1
Thanks so much for taking a minute to read this!

I am planning a dissertation and IF could get some feedback as to if I am on the right track, that would be great!

I am planning on conducting a study that compares the effect of student type (ELL, students with LD and typical students) on orthographic development as measured by 3 scores (feature points, spelling stage and words correct). Test scores have been measured at 3 different times (Fall, Winter and Spring). I will be conducting repeated measures ANOVAs for students in grades K, 1, and 2 separately.

Would I run these as three separate repeated measures ANOVA for each of the scores?

And one more question: The one measure is spelling stage. This test assigns a sequential stage as a descriptor to a cluster of developmental skills. (i.e the first stage is Emergent Writing- Early, then comes Emergent Writing- Middle etc). I was planning on assigning a numerical value to each of the stages- the first stage=1, the second stage = 2, etc. But I am wondering if that is flawed statistically? What is your opinion?

Thank you Thank you Thank you!!!
 
#2
Yes, I would be inclined to say that you would run at least 3 separate analyses; one for each of the outcomes.

If you are scoring writing on levels (i.e., 1/2/3) I would be particular about analyzing it; when you arbitrarily score things you cannot get statistical differences easily. Instead you are left comparing groups (low/mid/high abilty) or ranking them as ordinal (no set scale between ranks, but there is a definitive increase between group 1 and group 2, group 2 and group 3, etc.). It can be a bit awkward to analyze.

Hope this helps.