Capturing engagement in early science learning: triangulating observational, psychophysiological, and self-report measures
Revueltas Roux, Alexia Michelle
This thesis examined engagement in early science learning in the context of a science centre, particularly how, and to what extent, the process of engagement can be captured and measured in early years children. To achieve it, this thesis explored the potential of using a multimodal approach which captures the three components of engagement (cognitive, behavioural, and emotional) by triangulating simultaneous data from different tools, proposed theoretically by Azevedo (2015). Firstly, an observational study of children naturally interacting in a science centre showed that early-years children gravitate more and for longer to a hands-on exhibit compared to a planned-discovery exhibit. Secondly, a feasibility study showed it was feasible to use and triangulate the following tools in children from 3-7 years in a real-world context: engagement scales for behaviour coding using videorecording, a head-mounted camera, an electrodermal sensor (EDA), and a self-report questionnaire. Thirdly, the main triangulation study involved 28 children interacting with a sand exhibit at a Science Centre. Findings showed a relationship between children's cognitive-behavioural observations and their emotional arousal (EDA markers), but not with their self-report measures. Specifically, results showed that children's emotional arousal peaks were more likely to increase if they were doing cognitive-demanding behaviours such as strategic decision-making or looking away from the exhibit whilst searching for their parents. There was an increase in this effect the longer its duration lasted, but no effect was found for specific timepoints when a behaviour happened.. Children also engaged more when using new or previously used strategic behaviours rather than when adapting or immediately repeating them, as well as when persevering on a goal until fulfilled. A final study evaluated expert science practitioners' perception of engagement when they judged engagement as an outcome compared with as a process. A slider tool was developed to capture practitioners’ continuous perception of the process of engagement while they watched a video of the interaction, as well as their overall perception of engagement by giving a single value. Their agreement amongst the three different videos used was also examined, and how much they aligned with results from the previous triangulation study. Results showed no differences comparing between dynamic and discrete single scoring, however, the continuous dynamic score showed more nuances behind practitioner’s rating of engagement throughout the interaction. These dynamic ratings also showed more agreement between the practitioners, and although practitioners’ perceptions aligned to identify the video classified as the highest level of engagement, when levels were lower, evaluation of the level of engagement was challenging and practitioners did not agree on the intensity of the perceived engagement. Overall, this body of research highlights how multiple sources of data can provide a richer picture of what could be understood as engagement, particularly when engagement is conceptualised as a continuous process, which may inform improvements in both facilitation and exhibit design through deeper understanding of engagement processes and tailoring them specifically to different age groups. However, the findings also highlight some of limitations some of the different tools used here have for specific situations. This research presents a theoretical advancement by conceptually examining engagement as a process as well as an outcome along with contributing a thorough examination of early years' informal science learning engagement and relevant tools to capture it. This is an area which has been greatly understudied compared to other age groups and contexts. This research can improve informal learning experiences in key developmental stages, particularly for populations with limited access to informal learning contexts.