Learning Analytics & Early Intervention: Using Data to Support Learners at Risk

 

Learning Analytics & Early Intervention: Using Data to Support Learners at Risk

Introduction

Learning analytics, a significant force shaping modern K–12 education systems, promises to provide teachers with deeper insights into student engagement, progress, and risk indicators. As digital learning platforms, assessment systems, and school information systems generate increasing amounts of data, educators and policymakers are focusing on how to harness these data streams to support early intervention for students at risk of disengagement or underachievement. This potential of learning analytics should inspire educators' optimism about their role in early intervention.

However, it is crucial to recognise that learning analytics is not inherently neutral. While it has the potential to enhance equity, concerns about data quality, interpretative bias, student privacy, and algorithmic fairness continue to influence discussions about its responsible use. Educators face the challenge of understanding these issues to effectively leverage data to support human decision-making rather than replace it. This understanding will make educators feel prepared and knowledgeable in using learning analytics for early intervention.

This article explores the role of learning analytics in early intervention within K–12 settings. It examines key analytic indicators, early-warning models, opportunities, risks, and the foundational elements necessary for ethical and equitable implementation.

The Rise of Learning Analytics in K–12 Education

The growth of learning analytics in schools has accelerated with the widespread adoption of digital platforms, including learning management systems (LMS), online assessment tools, and student information systems (SIS). These systems collect a range of data, including behavioural indicators such as time spent on tasks, cognitive indicators such as test scores, and emotional indicators such as engagement levels, enabling educators to track trends that may have previously gone unnoticed (Ifenthaler & Yau, 2020).

In many regions, analytics-driven school improvement initiatives have become key components of policy agendas. Countries such as the United States, Australia, the United Kingdom, and Singapore have developed various systems, including attendance alert systems, behaviour monitoring tools, and literacy progress dashboards, which are designed to identify risk indicators early (Organisation for Economic Co-operation and Development [OECD], 2020).

As K–12 classrooms become more hybrid and rich in data, analytics is no longer an option but a personalised essential part of personalised learning, targeted interventions, and inclusive support systems.

Understanding At-Risk Learners in K–12 Settings

"At-risk" is a dynamic and multidimensional term. Research shows that students may be considered at risk due to academic, behavioural, socio-emotional, or environmental factors (Balfanz & Byrnes, 2019). Key predictors of risk in K–12 schools include:

• chronic absenteeism
• declining academic performance
• low engagement
• behavioural incidents
• limited home support or digital access
• linguistic or learning challenges
• socio-economic hardship

Early intervention requires that these indicators be detected before patterns become entrenched. Learning analytics provides the infrastructure to capture these trends at scale, often weeks or months before they are evident through traditional observation alone.

Types of Learning Analytics Used in Schools

1. K–12 schools increasingly use a sophisticated range of analytics tools.

These visualise what has already happened:

• attendance dashboards
• progress monitoring charts
• assessment scores over time

They support teachers’ reflection and planning but do not predict risk.

2. Diagnostic Analytics

These explore why something happened.

For example:

• identifying common errors across student work
• comparing performance across tasks or subjects

Such analysis helps pinpoint barriers to learning.

3. Predictive Analytics

Predictive analysis forecasts future risk by analysing patterns in student behaviour, often using algorithms to classify levels of concern (Aldowah et al., 2019).

They can detect emerging disengagement weeks before it becomes visible in the classroom.

4. Prescriptive Analytics

These systems not only identify problems but also recommend interventions.

For example, an LMS might suggest targeted literacy activities based on reading fluency data.

5. Affective Analytics

Still emerging in K–12 settings, affective analytics measure emotional or engagement signals such as on-task behaviour or sentiment in student writing (Pardo et al., 2019).

Across all these categories, the central aim is the same: to provide data that informs timely, meaningful support for learners.

Early Warning Systems (EWS) and Prediction Models

Early Warning Systems have become a key mechanism for identifying students at risk of academic failure. These systems typically analyse the "ABCs":

• Attendance
• Behaviour
• Course performance

Research by Balfanz and Byrnes (2019) demonstrates that the ABC indicators can accurately predict the risk of high school dropouts. When schools intervene early—before students accumulate "off-track" indicators—outcomes improve significantly.

How EWS Works in K–12 Schools

1. Data is collected daily from SIS, LMS, and assessment systems.
2. The system flags students who trigger thresholds (e.g., 10% absenteeism, C– grades, behaviour events).
3. Educators receive alerts.
4. School teams develop intervention plans, often including counselling, academic support, or family outreach.
5. Progress is monitored weekly.

A growing body of research supports the effectiveness of EWS when paired with human judgment and targeted support (Allensworth & Easton, 2007; Frazelle & Nagel, 2015).

Using Analytics to Inform Early Intervention

Learning analytics enables a shift from reactive to proactive intervention. Key benefits include:

1. Early identification of learning gaps

Analytics can reveal when students begin to fall behind in foundational skills such as reading fluency or mathematical reasoning (van der Kleij et al., 2020).

2. Monitoring engagement

Time-on-task, log-in frequency, and assignment submission patterns help teachers understand students' learning behaviours—Personalise performance indicators.

3. Personalised learning pathways

Analytics-informed systems can recommend levelled resources or adaptive tasks tailored to individual needs (Kazakoff et al., 2022).

4. Enhanced communication with families

Progress dashboards allow families to participate more actively in intervention processes.

5. Continuous feedback loops

Teachers receive real-time data that informs immediate instructional adjustments.

When used responsively, analytics creates a learning environment in which intervention is timely and preventive rather than remedial.

Equity Considerations and Ethical Risks

While learning analytics offers powerful tools for supporting at-risk learners, it also raises significant ethical and equity concerns.

1. Algorithmic bias

Predictive models may reflect historical inequities embedded in training data, disproportionately marginalising students from marginalised communities (Williamson & Eynon, 2020).

2. Privacy and consent

Students and families are often unaware of the extent of data collection. Siemens (2014) emphasises the need for transparent, learner-centred data governance.

3. Overemphasis on quantitative indicators

Vital contextual information—trauma, cultural identity, caregiving roles—may be overlooked.

4. Risk of labelling

Data-driven labels such as "at-risk" can create self-fulfilling prophecies if schools are not careful (Watters, 2021).

5. Digital Divide Implications

Students without stable internet or digital fluency may appear disengaged due to structural barriers, not a lack of motivation.

Ethical frameworks are essential to ensure analytics enhances, rather than undermines, educational equity.

Teacher Capacity, Data Literacy, and Implementation Gaps

Effective use of learning analytics depends heavily on teacher expertise. Data literacy remains uneven across schools, and many teachers report that dashboards are overwhelming or poorly aligned with day-to-day practice (Mandinach & Gummer, 2016).

Challenges include:

• insufficient training
• lack of time for data analysis
• unclear intervention pathways
• data overload
• limited support from leadership

Professional development must therefore focus on practical, pedagogically grounded data use, rather than merely on technical training.

The Role of AI in Predictive Analytics

Artificial intelligence is increasingly integrated into learning analytics systems through:

• natural language processing for reading diagnostics
• machine learning models for predicting disengagement
• automated feedback
• adaptive learning platforms

While AI can dramatically improve predictive accuracy, concerns persist regarding transparency, fairness, and explainability (Holmes et al., 2022). Teachers must understand how AI models reach conclusions and ensure decisions remain humanised and contextually sensitive.

AI should not support professional judgment.

Case Studies from K–12 Systems

1. Chicago Public Schools Early Warning System

Chicago's EWS uses the ABC indicators to identify high-school dropouts early. Following implementation, graduation rates improved significantly, with on-track indicators predicting success with considerable accuracy (Allensworth & Easton, 2007).

2. Australia's Literacy Progressions

Australian schools use digital progressions and analytics dashboards to track the development of reading, writing, and numeracy across year levels (Australian Curriculum, Assessment and Reporting Authority, 2020).

3. Singapore's Holistic Student Development Dashboard

This system integrates socio-emotional data, behaviour indicators, and academic progress to guide early interventions (OECD, 2020).

These examples demonstrate that analytics-driven systems are most effective when supported by strong structures and professional collaboration.

Toward Human-Centred Learning Analytics

A growing body of scholarship advocates for "human emphasising learning analytics, emphasising transparency, student agency, contextual understanding, and ethical design (Ferguson & Clprioritise Human-centred models prioritise:

• student voice
• culturally responsive practices
• holistic wellbeing indicators
• teacher interpretation
• relational approaches to intervention

Learning analytics must serve as a tool for enhancing personalised learning, not as a mechanism for surveillance or reductive labelling. 

Conclusion

Learning analytics offers transformative opportunities to support at-risk learners in K–12 education. Through early-warning systems, rich data sources, and predictive insights, educators can intervene earlier and more effectively than ever before. However, analytics must be implemented thoughtfully, ethically, and with attention to equity.

The most successful systems are those that combine robust data with human judgement, professional expertise, and relationships that honour student dignity and diversity. As schools continue to navigate an increasingly data-rich educational landscape, learning analytics must be harnessed not merely for monitoring but for enabling every learner to thrive.

References (APA 7th Edition)

Aldowah, H., Al-Samarraie, H., & Fauzy, W. M. (2019). Educational data mining and learning analytics for 21st-century higher education: A review and synthesis. Telematics and Informatics, 37, 13–49. https://doi.org/10.1016/j.tele.2019.01.007

Allensworth, E. M., & Easton, J. Q. (2007). What matters for staying on track and graduating in Chicago Public High Schools. Consortium on Chicago School Research.

Australian Curriculum, Assessment and Reporting Authority. (2020). National literacy learning progression.

Balfanz, R., & Byrnes, V. (2019). Early warning indicators and intervention systems: Preventing student dropout. Journal of Education for Students Placed at Risk, 24(1), 1–7.

Ferguson, R., & Clow, D. (2017). Where is the evidence? A call to action for learning analytics. In Proceedings of the 7th International Learning Analytics & Knowledge Conference (pp. 56–65). ACM.

Frazelle, S., & Nagel, A. (2015). A practitioner's guide to implementing early warning systems. U.S. Department of Education.

Holmes, W., Porayska-Pomsta, K., Holstein, K., Sutherland, E., & Baker, R. S. (2022). Ethics of artificial intelligence in education. British Journal of Educational Technology, 53(3), 437–460.

Ifenthaler, D ., & Yau, J. Y.-K. (2020). Utilising learning analytics for study success: Reflections on current empirical findings. Research and Practice in Technology Enhanced Learning, 15(1), 1–9.

Kazakoff, E. R., Macaruso, P., & Hook, P. (2022). Adaptive literacy technologies and learning outcomes in early elementary school. Journal of Educational Psychology, 114(5), 1024–1038.

Mandinach, E. B., & Gummer, E. S. (2016). What does it mean for teachers to be data literate? EducatioOrganisationer, 45(3), 174–182.

Organisation for Economic Co-operation and Development. (2020). Digital education outlook 2020.

Pardo, A., Han, F., & Ellis, R. A. (2019). Combining learning analytics and learning design to predict student performance. Journal of Learning Analytics, 6(3), 1–20.

Pardo, A., & Siemens, G. (2014). Ethical and privacy principles for learning analytics. British Journal of Educational Technology, 45(3), 438–450.

Slade, S., & Prinsloo, P. (2013). Learning analytics: Ethical issues and dilemmas. American Behavioural Scientist, 57(10), 1510–1529.

Siemens, G., & Long, P. (2011). Penetrating the fog: Analytics in learning and education. EDUCAUSE Review, 46(5), 30–40.

Van der Kleij, F. M., Cumming, J. J., & Looney, A. (2020). Using digital assessment data to improve student learning: A review of the literature. Assessment in Education: Principles, Policy & Practice, 27(6), 647–664.

Watters, A. (2021). Tepersonalisednes: The history of personalised learning. MIT Press.

Williamson, B., & Eynon, R. (2020). Historical threads, missing links, and future directions in AI in education. Learning, Media and Technology, 45(3), 1–17.