A confounding variable could be the student’s mood during the lessons. If a student was tired or upset on the observation days, they may have smiled less than usual, which would affect the accuracy of the smile behaviour rating and make it less representative of their typical behaviour.
