I'm getting confused with the analysis of one of my experiments. Here is how the results look like. There were P participants in the experiments and they were split into two groups P1 and P2. Every student had to generate some events E for a set of steps S. Based on this I could count the quality of the answer for every student p for every step s. The quality of participant p for step s was defined as: total number (generated by all the participants P) of unique events for the given step s - number of unique events generated by participant p for the step s.
So now I have a matrix with the qualities for every participant p and step s.
What measure should I use to compare the overall quality of all the groups. I would like to know which group generated results of better overall quality. I started to count root mean squared error, but I'm not sure if this is a good way to do this.
Thanks for your help!