Monday, 17 September 2012

ALT-C 2012: Day 1 - Eric Mazur - The Scientific Approach to Teaching

It’s been 5 years since I last got to ALT-C but thanks to a long desired reduction in our September workload I got there last Tuesday. ALT-C is a great chance to meet people and overdose on ideas, and I was quite looking forward to Eric Mazur’s keynote called "The Scientific Approach to Teaching: Research as a basis for course design" [slides]

Mazur’s work has been key in the development of Peer Instruction methods and the use of Clicker systems. In this keynote he talked about the importance of using evidence and the scientific method to guide the way we teach, rather than just anecdotal evidence. He presented three projects where data had been collected in classes, and where the data had played a role in guiding the development of the class in unexpected ways.

Firstly he looked at gender issues. In his ‘Force Concept’ class he noticed that women were getting lower marks on the section. Introduction of Peer Instruction into the class raised everyone’s grades, especially women’s, removing the gender difference in results. He didn’t suggest reasons why the women benefitted most from this method, but he argues that in education we often focus on information transfer which is the easy part of education. Assimilation of the information into students understanding of the world is the difficult part, and the part we often neglect. Using Peer Instruction type methods in the classroom is a way of focussing on the assimilation of information.

You can read more about some of his work in this area in “Reducing the gender gap in the physics classroom”.

Next he talked about using demonstrations in lectures. Researchers have found that while students enjoy watching demonstrations, tests a month after the lecture suggest that they don’t tend to learn well from those demos. Eric notes that students don’t remember facts, but models. If a fact doesn’t fit with their mental model of a concept, you need to give them time in the class to redevelop that model. Otherwise in their memories they will change the fact to fit with their mental model. It is worth noting that Eric’s classes last 3 hours which allow time for peer instruction and similar activities. If we only have one hour to present a subject, it might be more difficult to find time.

An episode of the Learning about Teaching Physics podcast, “Seeing isn't believing: Do classroom demonstrations help students learn?” talks through the research.

Finally he talked about confusion. Student surveys include questions asking if teachers provide “clear explanations”, but confusion doesn’t indicate a lack of understanding. In a research project where student results were compared with student reported levels of confusion it was found that 44% of students who said they were confused about the ‘capillary action’ topic gave correct answers about it, while only 25% of not-confused students gave correct answers. It was suggested that those who aren’t confused haven’t even began to understand it, and this might be a danger with the clear, non-chaotic lecture. We need to teach by questioning not telling and confusion is an essential step in learning.

This research reminded me about some of Derek Muller’s research where students labeled videos that helped them learn best “confusing”, and the ones which resulted in them being more confident about their own misconceptions as “clear”,”concise” and “easy to understand”. Others have linked it to the Dunning-Kruger effect.


There is an overview of Mazur’s work in Harvard Magazine which brings up some issues with the ‘flipped classroom’ and interactive approach, such as students thinking they’ve paid for you to do all the work “Not realizing that they learn precious little by that”, and classrooms generally “built with just one purpose: focusing the attention of many on the professor”.

If you are interested in using Clickers and Peer Instruction, the following resources might be of use:



No comments:

Post a Comment