How should I organise data for unbalanced two factor anova analysis (anovan)

Hi All,

I have two groups (control; n=13 and interven; n=12). For each group, I have collected the peak amplitudes of each subject at four different stimulus strengths. I would like to conduct a two way anova analysis: group and intensity. I am aware that for an unbalanced multifactorial anova I could use the function 'anovan' in MATLAB. However, I am a bit confused as to how to organise my data so that I can conduct the analysis correctly.

Currently I have created two tables (per group), with each column representing a different intensity.

peakCON = table(idx0p4CON,idx0p6CON,idx0p8CON,idx1p2CON); %13x4; control group - 4 different intensities
peakFA = table(idx0p4FA,idx0p6FA,idx0p8FA,idx1p2FA); %12x4; intervention group - 4 different intensities
STIM = [peakCON(:,1),peakFA(:,1),peakCON(:,2),peakFA(:,2),peakCON(:,3),peakFA(:,3),peakCON(:,4),peakFA(:,4)]; %concatenate the CON and INTERV group
GRP = [1 2 1 2 1 2 1 2]; %Grouping variables

%P=anovan(STIM,GRP);

When I concatenate my tables at 'STIM' I get an error saying that the rows are unequal (obviously). I am unsure how to proceed from this point, and would appreciate any help.

Answer 1

[Jeff Miller]

To start with, you have to make a table with 25 lines, one for each member of each group, and 4 variables for intensity plus one categorical variable for group. The code should look something like this:

peakCON = table(idx0p4CON,idx0p6CON,idx0p8CON,idx1p2CON); %13x4; control group - 4 different intensities
peakCON.Properties.VariableNames = {'idx0p4', 'idx0p6', 'idx0p8', 'idx1p2'};
peakCON.Group = ones(height(peakCON),1);
peakFA = table(idx0p4FA,idx0p6FA,idx0p8FA,idx1p2FA); %12x4; intervention group - 4 different intensities
peakFA.Properties.VariableNames = {'idx0p4', 'idx0p6', 'idx0p8', 'idx1p2'};
peakFA.Group = 2 * ones(height(peakFA),1);
STIM = [peakCON; peakFA];
STIM.Group = categorical(STIM.Group);

[NA]

Thanks Jeff; appreciate it.