Optimising adaptivity in online learning environments
De Mooij, Susanne (2021) Optimising adaptivity in online learning environments. PhD thesis, Birkbeck, University of London.
|
Text
Final_Thesis_Complete_DeMooij_0921 (1).pdf - Full Version Download (16MB) | Preview |
Abstract
Over the past two decades, the demand for online adaptive educational technology has been rising due to the promise that learning can be tailored to students’ needs. To tailor and optimise adaptivity to individual differences in learning effectively, we need to investigate what predicts students’ performance. This thesis includes experiments conducted in an online adaptive learning environment, called Learn, played daily by 180,000 primary school children. Adaptive learning is fun and challenging for children because, by definition, it adapts the difficulty of mathematical and language problems to their ability, preventing boredom or frustration. This adaptivity can be optimised in two ways: (1) by using theories and insights from cognitive research on executive function, strategy use and error monitoring to identify key individual cognitive differences when learning; (2) by inferring the involvement of specific cognitive processes or thought patterns using online measures. In Chapter 2, two online measures – eye and mouse tracking - are discussed and compared. These measures were used in two of the three studies described in this dissertation, both in a small sample in the lab and a large study in the Learn platform. Chapter 3 reports a lab study with primary school children assessing whether individual differences in cognitive abilities (working memory and inhibitory control) predict arithmetic performance. This study showed that children’s cognitive profile does have an impact on arithmetic performance, and interacts with features of the gamified environment, namely the visibility of time pressure, a key aspect of many online educational tasks. Eye fixations revealed differences in children’s attention towards the question and distracting errors between when time pressure was visible and when it was not. Both eye and mouse tracking gave some insight into children’s thinking and strategy use during their performance of the arithmetic task. Chapter 4 presents the analysis of mouse movements of children playing in a learning environment. The mouse movements reflected children’s arithmetic difficulties during problem-solving. This introduced a promising way to predict false associations children might have without the need for them to make errors, which are typically maintained at a low level in an online learning environment. Children do make mistakes while learning, and a key element of learning is for children to notice their errors and adjust their strategies to improve their performance. Chapter 5 shows evidence of adaptive behaviour in the form of post-error slowing (PES), the finding that humans slow down their performance after an error. PES was observed in a range of online mathematical and language tasks. Individual and task-related factors influencing the presence or magnitude of PES were identified. Overall, the studies identify key predictors of student’s performance as well as suggest new online measures to infer individual differences in these predictors, which is needed to fulfil the promise of tailoring the learning experience to individual children.
Metadata
| Item Type: | Thesis |
|---|---|
| Copyright Holders: | The copyright of this thesis rests with the author, who asserts his/her right to be known as such according to the Copyright Designs and Patents Act 1988. No dealing with the thesis contrary to the copyright or moral rights of the author is permitted. |
| Depositing User: | Acquisitions And Metadata |
| Date Deposited: | 21 Oct 2021 14:50 |
| Last Modified: | 01 Nov 2023 14:57 |
| URI: | https://eprints.bbk.ac.uk/id/eprint/46391 |
| DOI: | https://doi.org/10.18743/PUB.00046391 |
Statistics
Additional statistics are available via IRStats2.
Archive Staff Only (login required)
 |
Edit/View Item |