A presentation from the STEM Annual Conference 2014.
As part of the ‘flipped class’ pedagogic approach the first year physics courses at the University of Edinburgh employ Peer Instruction (PI). In this pedagogy the class is posed a question and students vote individually on what they think the answer is. If a non-trivial proportion of the class answer incorrectly the students then discuss the question in small groups explaining their reasoning and resolving any disagreements before voting again. In most cases the proportion of students selecting the correct answer increases significantly following this peer discussion.
Individual PI episodes can be characterised as successful or unsuccessful depending on what proportion of students who initially answered incorrectly subsequently adopt the correct answer. We have been able to observe the discussions students have during these episodes using smart pen equipment which provides some insight into why episodes may or may not have been successful. In some cases this is due to unanticipated flaws in the characteristics of the original question.
Based on the information gleaned from these observations we have reformulated some of the PI questions and have found them to be transformed from instigators of unsuccessful episodes into ones which function very well. Using a case study approach we will highlight some important factors for effective PI questions and demonstrate their consequent improvements in PI episode outcomes; while our case study is drawn from a physics context the findings are more widely applicable to all STEM disciplines.
