Comment by UweSchmidt

If you get an answer outside of what you expected, reevaluate your approach, fix your study and redo it all, probably with a new set of participants.

If you can't do science, don't call it science.