The dataset incorporates (N = 160) 9th-grade secondary education students across 11 classes in three different schools. The students participated in a digital writing task within their regular classroom setting as part of the CELLA research (Centre for Learning and Living with AI) funded by the Jacobs Foundation. Throughout three separate sessions, the students engaged in one pre-test examination, two 45-minute essay-writing tasks based on biology (Azevedo et al., 2022) and artificial intelligence (AI), and those students who volunteered (n = 50) participated in a 15-minute post-test interview. The data collection involved a range of methodological approaches, including a background questionnaire, a junior metacognitive awareness inventory questionnaire (Sperling et al., 2002), a self-reported metacognitive strategy questionnaire (Karlen et al., unpublished), a self-reported metacognitive experiences questionnaire (Dindar et al., 2022), digital trace data, students' essays, screen recordings, pre- and post-domain knowledge tests, and post-test interviews. The task was conducted in a custom Moodle-based digital learning environment equipped with five support tools: a timer, planner, annotation tool, search tool, and essay writing interface. Before the main activity, students received tool training and completed questionnaires on demographics, prior domain knowledge, and metacognitive strategy knowledge. Multimodal data collection included time-stamped trace log data of tool use, typing behaviour, and navigation actions, enabling the modelling of SRL sequences. Rule-based AI library was used to label the sequences of learning actions into different cognitive and metacognitive processes (writing, first-reading, re-reading, orientation, planning, monitoring, and evaluation; Fan et al., 2022).

2025