Automating the analysis of social interactions in educational settings: a scoping review

Social interactions are crucial for children’s development, attracting significant interest from educators and researchers. Traditional methods of data collection in educational settings have limitations, prompting the exploration of ICT devices for more accurate and efficient quantitative data collection. This systematic review, following the PRISMA framework, analysed 21 studies that used sensor devices to collect data on social interactions among children aged 0-12 in educational environments. The studies investigated various aspects such as interaction mapping, disease transmission, homophily, and play types, predominantly using observational or descriptive approaches. They were categorized into three levels of ecological complexity: (1) validation of sensor devices, (2) interaction analysis as a predictor of disease spread and vocabulary growth, and (3) examination of children’s social dynamics. Findings indicate that sensor devices are particularly effective when combined with Social Network Analysis (SNA), which facilitates comprehensive analysis and graphical representation of social networks. Quantitative data on social interactions could help identify and support children facing exclusion or marginalization, allowing for targeted educational interventions. However, there is a notable gap in the literature regarding the use of sensor devices in educational interventions, underscoring the need for further research to evaluate their effectiveness and facilitate their application in education.