Assessing the validity of the cognitive load scale in a problem-based learning setting