Mapping from proximity traces to socio‐spatial behaviours and student progression at the school

Abstract

Identifying students facing difficulties and providing them with timely support is one of the educator's key responsibilities. Yet, this task is becoming increasingly challenging as the complexity of physical learning spaces grows, along with the emergence of novel educational technologies and classroom designs. There has been substantial research and development work focused on identifying student social behaviours in digital platforms (eg, the learning management system) as predictors of academic progression. However, little work has investigated such relationships in physical learning spaces. This study explores the potential of using wearable trackers for the early detection of low‐progress students based on their social and spatial (socio‐spatial) behaviours at the school. Positioning data from 98 primary school students and six teachers were automatically captured over a period of eight weeks. Fourteen socio‐spatial behavioural features were extracted and processed using a set of machine learning classifiers to model students’ learning progression. Results illustrate the potential of prospectively identifying low‐progress students from these features and the importance of adapting classroom learning analytics to differences in pedagogical designs. What is already known about this topic Learning analytics research on predicting students’ academic progression is emerging in both digital and physical learning spaces. Students’ social behaviours in learning activities is a key factor in predicting their academic progression. Emerging sensing technologies can provide opportunities to study students’ real‐time social behaviours in physical learning spaces. What this paper adds Fourteen progression‐related socio‐spatial behavioural features are extracted from students’ physical (x‐y) positioning traces. Predictive learning analytics that achieved 81% accuracy in prospectively identifying low‐progress students from their real‐time socio‐spatial behaviours. Empirical evidence to support the need for classroom learning analytics to have instructional sensitivity (ie, be calibrated according to the learning design). Implications for practice and/or policy Sensing technologies and machine learning algorithms can be used to capture and generate valuable insights about higher‐order learning constructs (eg, performance and collaboration) from students' physical positioning traces in classrooms. Researchers and practitioners should be cautious with generalised classification algorithms and predictive learning analytics that do not account for the pedagogical differences between different subjects or learning designs. Researchers and practitioners should consider the potentially unforeseen ethical issues that can emerge in using sensing technologies and predictive learning analytics in authentic, physical classroom settings.

The role of indoor positioning analytics in assessment of simulation‐based learning

“…For the first research question (RQ1), students' spatial‐procedural behaviours extracted from indoor positioning traces demonstrated significant and large correlations with teachers' assessment on task and collaboration performances in Scenario A but not in Scenario B of the simulation. The significant findings in Scenario A resonate with prior research on the relationships between spatial behaviours and student performances (Yan, Martinez‐Maldonado, et al, 2022). This finding illustrates the potential of indoor positioning analytics in understanding assessment reliability by uncovering the behavioural consistency in teachers' assessment and scoring criteria.…”

“…The current study used data anonymisation to remove individual identities from the dataset, thus protecting students' privacy. This approach has also been applied in several recent MMLA studies (Bassiou et al, 2016; Praharaj et al, 2018; Yan, Martinez‐Maldonado, et al, 2022; Yan, Martinez‐Maldonado, et al, 2022). Likewise, focusing on group‐level analytics also removes individual elements from any results used in future debriefs.…”

“…In particular, while previous work on indoor positioning analytics has investigated students' movement within a team (Echeverria et al, 2018; Fernandez‐Nieto, Martinez‐Maldonado, Kitto, & Shum, 2021), presence in spaces of interest (Fernandez‐Nieto, Martinez‐Maldonado, Echeverria, et al, 2021), and interaction with other team members (Fernandez‐Nieto, Martinez‐Maldonado, Echeverria, et al, 2021; Zhao et al, 2022), the current study focused primarily on formative assessment and group/team‐level feedback. This work may be similar to our previous works that used indoor positioning analytics to model students' progression of math and reading abilities in open learning spaces (Yan et al, 2021; Yan, Martinez‐Maldonado, et al, 2022), but we focus specifically on the assessment of task and teamwork performance during simulation‐based learning in this study. To our knowledge, this study is the first that explored the potential of indoor positioning analytics in supporting teachers' formative assessment and immediate feedback during simulation‐based learning.…”

“…A potential instance of collaboration was identified when two students were within one meter proximity for more than 10 seconds. This method of modelling collaboration from positioning data has been used in several prior studies (Martinez‐Maldonado, Echeverria, Schulte, et al, 2020; Yan et al, 2021; Yan, Martinez‐Maldonado, et al, 2022) and grounded in the theoretical foundations of proxemics (Hall, 1966). The 10 second threshold was adopted to reduce the likelihood of misidentifying unintended collocation as meaningful collaboration, for example, when students were merely passing by each other (Greenberg et al, 2014).…”

“…Predictive analytics ( validity of measurement ) have also been frequently used in MMLA research for the automatic identification of students' task performance (see reviews Sharma & Giannakos, 2020; Yan, Zhao, et al, 2022). A previous study has illustrated the potential of position‐based predictive analytics in supporting teachers to pinpoint low‐progress students (Yan, Martinez‐Maldonado, et al, 2022). Likewise, predictive analytics may also provide the opportunity to support inexperienced teachers by automating the distinction between low‐performing and high‐performing teams in simulation‐based learning.…”