The widespread implementation of digital devices in educational environments has facilitated educational innovation based on learner data. Particularly catalyzed by the COVID-19 pandemic, metropolitan and provincial offices of education across the nat...
The widespread implementation of digital devices in educational environments has facilitated educational innovation based on learner data. Particularly catalyzed by the COVID-19 pandemic, metropolitan and provincial offices of education across the nation have implemented large-scale "one device per student" initiatives. Building upon this digital infrastructure, the Ministry of Education plans to gradually introduce "AI Digital Textbook" from 2025. AI Digital Textbook is a textbook that utilizes intelligent information technologies, such as artificial intelligence, to support personalized instruction for each student. The introduction of AI Digital Textbook will enable individualized learning tailored to each student's pace and level, potentially overcoming current classroom limitations. However, this transformation means that students will spend considerable time operating digital devices. The expansion of digital device-based learning raises concerns about decreased attention spans and non-educational device usage. Previous studies have shown that increased digital device usage correlates with lower attention levels and negatively impacts academic achievement in the long term. Additionally, these devices can easily distract students by inducing multitasking behavior. Given that digital device-based instruction offers clear advantages, students need to develop meta-attention—the ability to properly use their digital devices and self-regulate their attention. To cultivate this meta-attention, students require assistance in objectively recognizing their situational awareness. However, the current curriculum lacks systematic support for improving students' meta-attention capabilities, and student data are primarily limited to learning processes and outcomes, making it difficult to assess students' attention states.
Therefore, this study developed and implemented meta-attention support strategies that enable learners to recognize and regulate their attention based on data in digital education, collaborating with field experts and theoretical experts. The research questions are as follows: First, what are the learner data-based meta-attention support strategies and detailed guidelines? Second, how is the instructional process structured when applying learner data-based meta-attention support strategies? Third, what are the effects and areas for improvement in classes implementing learner data-based meta-attention support strategies? Fourth, what learner characteristics influence attention improvement? To address these research questions, the study employed Design-based Research (DBR) methodology. Through an expert council comprising six field experts and three theoretical experts, the study identified practical problems and needs, and developed and continuously improved support strategies by building theoretical foundations through literature review. The field implementation involved conducting strategy-based instruction with one homeroom teacher and 68 first-year high school students, and examining both instructor and learner responses. Learner attention data were collected using edutech, which was used to create an attention dashboard. Students completed advanced inquiry tasks in integrated science following the meta-attention support strategy phases, and the FAIR attention test was administered to verify effectiveness. Additionally, qualitative data were collected through in-depth interviews with one teacher and 13 students. The collected data were analyzed using descriptive statistics and paired t-tests for quantitative data, while thematic analysis was applied to qualitative data. This analysis revealed the effectiveness and areas for improvement of the support strategies, detailed instructional processes, and learner characteristics that influence attention improvement.
The study resulted in the development of learner data-based meta-attention support strategies comprising 12 support strategies and 35 detailed guidelines across four categories: meta-attention guidance, attention planning, attention monitoring, and attention evaluation, with three strategies per category. The meta-attention guidance phase includes motivating meta-attention activities through the presentation of successful attention improvement cases, providing detailed guidance on meta-attention concepts and procedures, and incorporating practice exercises and comprehension checks for meta-attention. The attention planning phase involves a process in which students initially examine their attention status through a learner data-based attention dashboard, subsequently analyze the causes of their behaviors, and ultimately establish specific attention goals and improvement strategies for the current session. The attention monitoring phase provides guidance on recognizing and regulating situations of decreased attention during class, encouraging students to self-check their status when their attention wanes, and implementing attention improvement strategies to regain focus. Individual feedback can be provided through teacher circulation during this phase. In the attention evaluation phase, students assess their attention goal achievement and strategy effectiveness after completing activities. The strategies used are cumulatively recorded, allowing students to reflect on these records and identify the most effective strategies for themselves. The class concludes with students sharing their meta-attention experiences with peers, providing peer feedback, and establishing foundations for future attention planning.
The detailed instructional process analysis revealed changes in both teacher strategy implementation and student participation patterns. A heatmap analysis of the teacher's strategy usage frequency by session demonstrated temporal changes in strategy implementation. Specifically, the first session showed the highest strategy usage rates across all phases, with meta-attention guidance and attention planning decreasing over subsequent sessions, while attention monitoring and evaluation maintained consistent levels. Regarding specific strategies, those with low or no usage frequency appeared to transition naturally into implicit learning through class participation, primarily due to time constraints. Subsequently, time-series data was used to verify the sequence of strategy implementation. The instruction process generally proceeded according to the developed strategy sequence. Among the various strategies, only the strategy of identifying students needing assistance through student data review was implemented first, before class. Changes in student participation processes were thoroughly examined based on survey data collected each session and interview results, providing detailed insights into students' dashboard checking methods and reactions, categorization of established attention goals and improvement strategies, strategy effectiveness evaluation and modification processes, and peer interaction content. The FAIR attention test results demonstrated statistically significant improvements in students' attention levels. All subcategories—selective attention, sustained attention, and self-control—demonstrated substantial improvement. Student interviews revealed positive effects including development of self-directed learning strategies, enhanced intrinsic motivation and self-efficacy, improved attention, and voluntary transfer of meta-attention skills. Limitations included individual activity-centered instruction and technical constraints in data collection. Instructor interviews identified key achievements: providing different meta-attention experience opportunities for learners of varying abilities, maintaining continuous motivation through the dashboard, and meta-attention activity transfer. However, limitations were noted, including insufficient individualized support for different learner levels, technical limitations in attention data collection, and limited peer interaction opportunities. Finally, analysis of learner behavior patterns and perceptions identified factors influencing attention improvement. Based on classroom activity data comparing high and low improvers, high improvers tended to set specific, quantified goals, while low improvers set abstract, ambiguous goals. Strategy utilization was higher among high improvers, with their strategy use increasing over time, while low improvers exhibited stagnant or declining strategy use. Regarding data reliability, high improvers demonstrated strong trust in the data, viewing the attention dashboard as a tool for discovering previously unrecognized information. In contrast, low improvers displayed skepticism about data accuracy and demonstrated off-task behavior after identifying data limitations.
This study presents a novel educational approach for attention improvement in the digital transformation era. It systematically developed strategies and detailed guidelines for instructors to design learner data-based meta-attention instruction to effectively enhance learner attention in digital education. The study empirically verified the instruction's effectiveness and conducted detailed analysis of the instructional process. Its significance lies in presenting a new paradigm that supports learners’ self-directed attention development through the integration of learner attention data and meta-attention.