"one or more cells missing data" error

[Eric Flyn]

Hi all,
I'm trying to do a simple repeated-measures ANOVA. It's
completely within-subjects - it's a perception experiment where all
subjects saw all stimuli multiple times in different treatments.
The data file is set up as follows ("ResponseNum" is the dependent
variable, it represents the subject's categorization of a sound):

Name Action ActionSound ResponseNum
subj1 treatment1 /aba/ 1
subj1 treatment1 /ava/ 0
subj1 treatment2 /aba/ 1
subj1 treatment2 /ava/ 1
subj1 treatment3 /aba/ 0
subj1 treatment3 /ava/ 1
subj2 treatment1 /aba/ 0
subj2 treatment1 /ava/ 0
subj2 treatment2 /aba/ 1
subj2 treatment2 /ava/ 0
subj2 treatment1 /aba/ 0
subj2 treatment1 /ava/ 0
subj3 treatment2 /aba/ 1
subj3 treatment2 /ava/ 1


When I type in the following, I get the error message "one or more
cells is missing data":
result <- ezANOVA(data=Main.Subs, dv=.(ResponseNum), wid=.(Name),
within=.(ActionSound, Action))

However, when I use only 1 "within" column, like in the following,
it gives me an output:
result <- ezANOVA(data=Main.Subs, dv=.(ResponseNum), wid=.(Name),
within=.(ActionSound))

Similarly, it gives me an output if I use the other column as the
"within" column:
result <- ezANOVA(data=Main.Subs, dv=.(ResponseNum), wid=.(Name),
within=.(Action))

Sorry if this is a real "newbie" question, but I can't figure out
what I'm doing wrong (I'm not very advanced with either stats or R).
Thanks for your help,
Eric

[Mike Lawrence]

ezDesign() is useful for finding missing data. Try:

ezDesign(
data = Main.Subs
, x = .(Action)
, y = .(Name)
, col = .(ActionSound)
)

Are there any empty cells? If not, try:

ezDesign(
data = Main.Subs[!is.na(Main.Subs$ResponseNum),]
, x = .(Action)
, y = .(Name)
, col = .(ActionSound)
)

Between those two, I suspect that you'll find that at least one
subject is missing data in one of the cells of the design, in which
case you'd have to drop that subject or switch from ezANOVA(), which
can't deal with missing data, to ezMixed(), which can.

[Eric Flyn]

Hi,
Thanks very much for your help. Yes, you were right, I had made a
silly error and it's working fine now.
Great package!
Thanks,
Eric