?
Beyond activity counts: object-mediated participation in project-based learning is associated with outcomes at different levels of analysis
Learning analytics of collaborative project-based learning still relies largely on activity counts and final products, which register the fact of digital activity but not the structure of collaborative participation. This study proposes an object-mediated, multi-platform approach that reconstructs joint work from digital traces as a network of ties between students and shared project artifacts, and relates the structure of digital participation to educational outcomes. Drawing on engineering project-based learning at the Moscow Institute of Electronics and Mathematics, HSE University (14,289 digital events from GitLab, Wekan and Trello, 614 student–project pairs in 251 projects), we computed volume, configuration, network-position and temporal measures of participation and estimated their associations with first-attempt project grades using rank correlations with cluster-bootstrap confidence intervals, effect sizes, multilevel decomposition and hierarchical regression. Only students with at least one observed event could be included: 838 of 1,452 eligible pairs (57.7%) left no observable trace, and observed and non-observed pairs did not differ in grade. Associations are small throughout. Activity volume is weakly associated with outcomes (ρ = +0.22), as is a student’s relative embeddedness within the team once within-project ranks are placed on a common scale (ρ = +0.16). The clearest pattern is categorical rather than graded: students who left events alongside observable teammates yet shared no artifact with any of them scored markedly lower (δ = −0.34), and this indicator is stable across pooled, mixed-effects and ordinal specifications, whereas gradations among connected students add little. A less closed structural position is associated with higher grades independently of activity volume (Burt constraint ρ = −0.190), and continued expansion of a student’s circle of shared objects is associated with higher outcomes within teams. Temporal organisation contributed no incremental variance between projects but did so within them, and its direction depends on the anchor used to close the project window. The approach offers a diagnostic vocabulary for collaboration and prompts for formative enquiry rather than an instrument for grading.