I have conducted an experiment where the subject is first presented with a set of 7 stimuli. Thereafter, the subject is presented with a set of 5 stimuli. Finally, the set of 7 stimuli is presented again. The set of 7 stimuli is divided into two subgroups A and B with 5 and 2 stimuli, respectively. This kind of experiments was repeated 15 times with different sets of stimuli. The purpose is to study on the basis of the subject’s responses how the presentation of the set of 5 stimuli effects on the perception of the subgroups A and B.

I have planned to calculate average values over the 20 experiments and include the results in the 2*2 (groups, presentation order) repeated measurements ANOVA table as follows:

A1=first presentation of group A

A2=second presentation of group A

B1=first presentation of group B

B2=second presentation of group B

C=set of 5 stimuli

The ANOVA design would then be:

A1C A2C B1C B2C

Since the sample size is different in the groups A, B and C, I would select randomly a subset of 30 samples from groups A and C to make them comparable with group B (group B includes 30 stimuli in total, i.e. 2 stimuli in each experiment and the total number of experiments is 15). Additionally, I have planned to conduct a separate ANOVA where I only include groups A and C (i.e. A1C A2C). These groups have both 75 stimuli in total (20 experiments and 5 stimuli per experiment).

I wonder if someone could please let me know if this is statistically correct way to compare these groups.

Many thanks already in advance!

Best Regards,

mh