Repeated measures ANOVA with 2 variables with different levels

Hi, I am trying to run a 2 way RMANOVA on this data called RMANOVA_Table_Analysis. I have also attached this dataset.

I am trying to get the main effect of background, color and the interaction between them. This is what I tried, looks like I am making an error in creating the repeated measures model.

Any pointers on how to do this correctly, would be much appreciated!

[Scott MacKenzie]

What do you mean by "looks like I am making an error"? What error are you seeing?

[MM]

Answer 1

[Scott MacKenzie]

Change

rm = fitrm(dv, 'uniR - texB ~ 1 ', 'WithinDesign', withinDesign);

to

rm = fitrm(dv, 'uniR-texB ~ 1 ', 'WithinDesign', withinDesign);

Apparently, the fitrm function doesn't like the spaces on each side of the dash.

[MM]

Thank you, Scott!! Removing the space made it work. Rookie mistake.

Another question:

This is the code I use to create the ANOVA table.

[file,path] = uigetfile('*.xlsx');
Q = readtable(file);

dv = array2table(reshape(RMANOVA_Table_Analysis.mean_final_offset, num_subj, []));

% codes for column names: uni = uniform, tex = textured, R=red, G=green B=blue
dv.Properties.VariableNames = {'uniR','uniG','uniB', 'texR', 'texG', 'texB'};

% create the within-subjects design
withinDesign = table([1 1 1 2 2 2]',[1 2 3 1 2 3]','VariableNames',{'Background','Color'});
withinDesign.Background = categorical(withinDesign.Background);
withinDesign.Color = categorical(withinDesign.Color);

% create repeated measures model
rm = fitrm(dv, 'uniR-texB ~ 1 ', 'WithinDesign', withinDesign);


% perform anova (remove semicolon to view table generated by ranova)
AT = ranova(rm, 'WithinModel', 'Background*Color');

If I replace

rm = fitrm(dv, 'uniR-texB ~ 1 ', 'WithinDesign', withinDesign); 

with

rm = fitrm(dv, 'uniR-texB ~ Background*Color+Background+Color ', 'WithinDesign', withinDesign)

I'd think I would get a similar response. But instead I get an error instead (below):

rm = fitrm(dv, 'uniR-texB ~ Background*Color+Background+Color ', 'WithinDesign', withinDesign)
Error using classreg.regr.FormulaProcessor (line 385)
The model formula contains names not recognized as predictor or response names.

Error in classreg.regr.MultivariateLinearFormula (line 47)
f = [e-mail](varargin{:});

Error in RepeatedMeasuresModel.fit (line 1312)
formula = classreg.regr.MultivariateLinearFormula(model,varNames);

Error in fitrm (line 77)
s = RepeatedMeasuresModel.fit(ds,model,varargin{:});

Would you know why this is the case?

[Scott MacKenzie]

@MM, you're welcome. In specifying the modelspec for fitrm, the between-subjects variables are placed on the right-hand side of the tilde symbol. Your design does not have any between-subjects variables, so "1" is used on the right-hand side of tilde in the modelspec. In your modified code, you are trying to position the within-subjects variables (Background, Color) as between-subjects variables. That's the reason for the error.

In your original code (with the correction I noted), you will get three effects: the main effects for Background and Color and the Background x Color interaction effect. From your question, I think that's your main objective, right? So I'm not sure what you are trying to do in the second version of fitrm.

BTW, the effect of Color on mean_final_offset was statistically significant, F(2,22) = 4.006, p = .0328. The other two effects were not statistically significant.

[MM]

Thank you for explaining! Makes sense why I get the error.