In my experiment, I’ll like to present 4 practice trials to participants, and obtain 100% correct responses before they move on to the actual trials. If accuracy <100%, a further 2 trials should be presented. If both trials are answered correctly, then participants will move on. If not, another 2 trials should be presented.
This can be done. It is already (mostly) explained in the draft manual of SuperLab 4.0, which is nearly complete.
I will post (or email you) a PDF of the manual next Wednesday. Sorry for not being able to post it sooner due to the 4th of July holiday and me being out of town on Monday.