I have to do a user study for a paper to see how using or not using tool X affects the quality of something (defined in my study). The study goes like this: participant 1 uses tool X to do task A and then a survey after. Then participant 1 does not use tool X to do task B then a survey after. Participant 2 does not use tool X to do task B then a survey after. Then participant 2 uses tool X to do task A. We're looking for around 50 participants.
Conditions: use tool X and not use tool X
Tasks: Task A and Task B
Task A and Task B must be done with either using tool X or not using tool X. But both tasks must be done using either mode (use tool X or not use tool X). I'm trying to ensure that the study is counterbalanced such that if my data indicates something that correlates using tool X with task B, it won't be because most participants used tool X for task B before not using tool X for task A. Again the order of using tool X then not using tool X matters. Likewise, the order of performing task A then task B matters. I hope I'm making sense.
I read around that a Latin-square generator is ideal for avoiding order effects but I'm still trying to figure out the right configuration. Any help would be appreciated.