Exploring the instructional efficiency of representation and engagement in online learning materials

Using two 2 x 3 in-situ experiments in an introductory programming course, we study the effect of representation and engagement on the instructional efficiency of learning materials. In the experiments, we controlled for used representation and the level of engagement and accounted for prior experience and prior cognitive effort. We observe that analogical representations with little engagement are more beneficial for those already familiar with the topic. No significant effect from engagement or prior experience was observed when students studied using traditional representations. Low cognitive effort before studying was related to studying being less cognitively demanding, regardless of the condition. No single way of presenting information seems to work better or worse universally for all participants.