Analysing students‘ response times with classroom response systems

Classroom response systems ('clickers‘ etc.) are well documented as having potential for developing knowledge and engagement in the classroom, so much so that they are ubiquitous, almost passé. The wealth of data gathered from such systems is beginning to be analysed and recent work in the peer instruction community (e.g. Miller et. al, 2014) suggests that in some contexts students answering correctly tend to have the shortest (quickest) response times. Recent work at Kingston University has deployed and evaluated clickers in various-sized groups of students across STEM subjects (and elsewhere) which has confirmed previous findings of their benefits in active learning scenarios. In a small pilot study we took advantage of the data gathered from these trials to investigate if there was a relationship between students‘ speed and consistency of responses to in-class clicker questions and their summative assessment grades. Linear Algebra was introduced in a flipped learning environment with preparatory quizzes before class and clickers-based peer instruction in class, and the response times and accuracy of question responses was compared with preparatory formative quiz scores and final summative results. Preliminary results (Orwell et al., 2016) suggest that quiz preparation is more strongly associated with summative correctness than in-class accuracy and that the students responding correctly tend to do-so more slowly than other students, the opposite of Miller et al. To follow-up this apparent contradiction in response timings we are enlarging the set of response timings for analysis beyond the relatively controlled scenario of the peer instruction Linear Algebra class to a broader set of mathematics classes and will report on these with a view to identifying indicators that may be useful in data analytics.