Artifical grammar learning is supposed to tap into processes so basic to human cognition that many seem to assume that they are "universal", the same for every human. As a result researchers focus too much on group averages and do not look at individual differences in performance. If you do you find out that even healthy individuals do very different things within the same experimental condition. My life would be easier if this were not the case.
First I thought that this had something to do with the fact that in these experiments participants have to make conscious decision, so even if grammatical processes are universal, participants may override their outcome. But if this is the case, why do adult birds of the same species show individual differences in a modified artificial grammar learning experiment?
I came up with my own type of analysis for this paper. Since I was a PhD student at that time I was very proud. With that said, the analysis was a bit clumsy and today I would treat the data differently (probably using logistic regressions).