Most Scottish secondaries have more data than they know what to do with. Attainment data. Attendance data. Behaviour referrals. Exclusion logs. Pupil support records. Wellbeing surveys. Free meal entitlement. Additional support needs registers. Exam entries and results. Standardised assessment scores. Tracking spreadsheets maintained by individual departments in formats that vary from meticulous to incomprehensible. The data exists. In some schools, it exists in extraordinary quantities. What is almost always missing is the capacity to connect it — to look at a young person's attendance pattern alongside their attainment trajectory alongside their behaviour record alongside what the pastoral team knows about what is happening at home, and to ask what the picture means as a whole.
This is not a technology problem, though technology can help. It is a structural problem: the way schools organise themselves tends to produce data silos, where attainment information lives with curriculum leaders, attendance data lives with the office, behaviour records live with promoted staff, and wellbeing information lives with guidance or pupil support. Each of these data streams is maintained, analysed, and acted upon by different people with different priorities, using different systems, often on different timescales. The result is that schools typically know a great deal about their pupils in fragments, and far less about them as whole human beings moving through a school.
The silo problem
Kim Schildkamp's research programme on data use in schools — the most sustained body of work in this field internationally — has consistently identified a gap between the data schools collect and the data they actually use to inform decisions. In a series of studies across the Netherlands, the UK, and internationally, Schildkamp and her colleagues (Schildkamp, 2019; Schildkamp & Kuiper, 2010) have found that while schools generate large volumes of data, the way that data is stored, accessed, and interpreted varies enormously, and that many schools lack what Schildkamp calls 'data literacy' at the teacher and middle-leader level — the ability to formulate questions, locate relevant data, interpret it accurately, and translate findings into actionable changes in practice.
The problem is compounded in Scotland by the particular architecture of the education data landscape. Schools interact with multiple data systems that do not naturally communicate with each other. In most local authorities, the core management information system — SEEMiS in the majority of Scottish schools (or SIMS, or Bromcom, or whichever system your school uses — the problem is the same regardless of the software) — holds registration, attendance, and demographic data. Attainment data at the senior phase sits in a combination of internal tracking systems and the national Insight benchmarking tool. Achievement of Curriculum for Excellence Level (ACEL) data is collected annually for P1, P4, P7, and S3 through a separate process based on teacher professional judgement. Wellbeing information — SHANARRI indicators, child protection records, support plans — sits in pastoral systems that are often separate again. The result is that to build a picture of a single pupil that connects attainment, attendance, behaviour, and wellbeing, a school leader typically has to pull data from three or four different places, link it manually, and hope that the identifiers match.
This is not simply an inconvenience. It shapes what questions schools ask. If attainment data lives in one system and attendance data in another, the question 'is there a relationship between this pupil's attendance pattern and their attainment trajectory?' requires deliberate effort to answer. If behaviour referrals are logged separately from wellbeing records, the question 'are the pupils generating the most behaviour incidents the same pupils with the most complex pastoral needs?' — which in my experience the answer is almost always yes — requires someone to actively join those datasets. And the questions that require deliberate effort to answer are, by definition, the questions that don't get asked routinely.
The data is only as good as the instruments that produce it — and some of those instruments are more broken than schools realise.
What the research says about data use that works
Ellen Mandinach and Edith Gummer's work on data literacy for teachers (Mandinach & Gummer, 2016) identifies a crucial distinction between data and information. Data is what the system records: a number, a date, a code, a tick in a box. Information is what emerges when data is placed in context, connected to other data, and interpreted in light of professional knowledge. A pupil's attendance figure of 87% is data. The observation that this represents a decline from 94% last year, that the absences cluster on Mondays and Fridays, that the pupil is also showing a decline in attainment in practical subjects but not in others, and that the pupil support team are aware of a change in circumstances at home — that is information. And it is information, not data, that drives useful intervention.
The most productive frameworks for school-level data use tend to share a common structure: they begin with a question rather than with the data. Robert Balfanz's work at Johns Hopkins on early warning systems — the ABC framework, tracking Attendance, Behaviour, and Course performance as leading indicators of disengagement — is built on this logic. Balfanz and colleagues (Balfanz, Herzog, & Mac Iver, 2007) demonstrated that a relatively small number of readily available indicators, when tracked longitudinally and combined, could identify pupils at risk of dropping out with a high degree of accuracy — often years before the outcome became visible in attainment data alone. The power of the ABC framework is not its sophistication — the indicators are simple — but its integration: it asks schools to look at attendance, behaviour, and course performance together, as a composite picture, rather than treating each as a separate data stream managed by a separate team.
Viviane Robinson's research on student-centred leadership (Robinson, 2011) makes a related point about the relationship between data and improvement. Robinson's meta-analysis of school leadership effects found that the leadership dimension with the largest effect size on student outcomes was 'promoting and participating in teacher learning and development' — but that this effect depended on leaders' ability to use evidence about student learning to inform the focus of that professional development. In other words, it is not enough to be a data-rich school. You have to be a school that uses data to ask the right questions about teaching and learning, and then acts on the answers.
The pastoral-data connection
There is a persistent assumption in many schools that data and pastoral care are separate functions — that data is the province of attainment-focused DHTs, assessment coordinators, and the people who run the tracking spreadsheets, while pastoral care is the province of guidance staff, pupil support leaders, and the people who have the difficult conversations with families. This division is understandable organisationally — the skills and sensibilities involved are genuinely different — but it produces a blind spot that costs schools dearly.
The blind spot is this: the pupils who are struggling most are almost always struggling across multiple domains simultaneously. The pupil whose attainment is declining is often the pupil whose attendance is deteriorating, whose behaviour is escalating, and whose wellbeing is under pressure. These patterns do not exist in isolation, and they do not respect the organisational boundaries between the attainment team and the pastoral team. But because the data about each domain is collected, stored, and analysed separately, the composite picture — which is the only picture that makes sense of the young person's experience — often does not exist in any single place.
Scotland's wellbeing framework, SHANARRI (Safe, Healthy, Achieving, Nurtured, Active, Respected, Responsible, Included), provides a holistic language for thinking about young people's needs. And HGIOS4's Quality Indicator 3.1 (Ensuring wellbeing, equality and inclusion) explicitly asks schools to demonstrate that they know their pupils' wellbeing needs and are addressing them. But frameworks and quality indicators do not, by themselves, produce the data infrastructure needed to make this work in practice. A school can have excellent pastoral relationships — teachers who know their pupils deeply, who notice when something is wrong, who follow up with families — and still lack the systematic capacity to identify patterns across the cohort, to track the impact of interventions over time, or to allocate limited support resources to the pupils who need them most rather than to the pupils who shout loudest.
The most effective pastoral practice I have encountered is not less data-driven than the most effective attainment tracking. It is more data-driven — but the data it draws on is broader, more varied, and harder to aggregate. It includes attendance patterns, behaviour records, and attainment trajectories, but also wellbeing assessments, support plan records, involvement of external agencies, engagement with extracurricular activity, and the soft intelligence that pastoral staff carry in their heads but rarely record in any system. The challenge is not whether this information exists — it always does, in some form — but whether it is connected, visible, and used to inform decisions rather than simply filed.
Observation is one of the richest data sources schools have — when it's designed to capture structured evidence rather than anecdote.
From data collection to data use: the practical gap
There is a version of data-driven school improvement that involves investing in dashboards, platforms, and analytics tools — and this can be genuinely useful, provided the tools are designed around the questions schools need to ask rather than the questions that are easiest to visualise. But the bigger barrier to effective data use is not technical. It is cultural and structural.
The cultural barrier is that data in schools is overwhelmingly associated with accountability. Teachers experience data as something that is done to them — a tracking cycle that asks them to predict grades, a performance management conversation anchored to exam results, a local authority request for attainment figures. This association between data and judgement makes teachers wary of data, which in turn makes them less likely to use it formatively — to genuinely interrogate what the data is telling them about their pupils and their practice. Schildkamp's work confirms this: in schools where data is used primarily for accountability, teachers are less likely to engage with data for improvement purposes. The two uses are not merely different; they are, in many school cultures, actively in tension.
The structural barrier is time. Interpreting data — connecting the attendance pattern to the attainment trajectory, investigating whether the behaviour incidents cluster in particular lessons or at particular times, following up with the pastoral team to understand what lies behind the numbers — takes time that most teachers and middle leaders do not have. Data drops happen termly. Pastoral meetings happen weekly. The gap between the frequency of data collection and the frequency of data interpretation is where useful information goes to die.
The schools that use data most effectively tend to have solved both of these problems simultaneously. They have built cultures where data is understood as evidence — morally neutral information about how things are, not a verdict on how well someone is performing. And they have created structures — regular data conversations with a specific focus, cross-functional teams that bring together attainment and pastoral perspectives, accessible systems that allow patterns to be identified quickly rather than requiring hours of manual cross-referencing — that make data interpretation part of routine practice rather than a periodic burden.
What schools need
The gap is not more data. Schools already have more than they can use. What they need is connection: systems and practices that bring different data streams together around the pupils they serve, that make patterns visible without requiring heroic manual effort, and that support the professional conversations through which data becomes insight and insight becomes action.
This is, at its core, a design problem. The data architecture of most schools was not designed to support the kind of integrated, question-driven analysis that the research identifies as effective. It was designed to meet reporting requirements — to return ACEL data to the local authority, to submit SQA entries, to log attendance for legal compliance, to record behaviour incidents for the sake of the record. These are necessary functions, but they are not the same as improvement functions, and building an improvement culture on top of a compliance architecture is harder than it needs to be.
What good looks like is surprisingly simple, at least in principle. It means being able to answer questions like: Which pupils are showing declining attendance and declining attainment simultaneously? Or: Are the pupils we identified as at risk in September still at risk in January, and has what we've done made a difference? Or: What does the pattern of observation evidence tell us about the classroom experience of our most vulnerable learners? These are not exotic questions. They are the questions that any thoughtful school leader would want to ask. The problem is that, in most schools, answering them requires a level of manual data integration that makes them impractical to ask routinely — and questions that cannot be asked routinely do not shape practice.
The ambition should be modest in scope but radical in implication: to make it as easy for a school to see the whole picture of a young person's experience as it currently is to see their latest attainment data. That is not where most schools are. But it is where the evidence suggests they need to be.
Having the data is one thing. Building it into a school improvement process that changes practice is another.
If you're thinking about how observation and analytics could join the rest of your evidence in one place, Learning Lens is available for Scottish schools — you can book a conversation from the home page.
References
Balfanz, R., Herzog, L. & Mac Iver, D.J. (2007). 'Preventing Student Disengagement and Keeping Students on the Graduation Track in Urban Middle-Grades Schools: Early Identification and Effective Interventions.' Educational Psychologist, 42(4), 223–235.
Education Scotland (2015). How Good Is Our School? (4th edition). Livingston: Education Scotland.
Mandinach, E.B. & Gummer, E.S. (2016). Data Literacy for Educators: Making It Count in Teacher Preparation and Practice. New York: Teachers College Press.
Robinson, V.M.J. (2011). Student-Centered Leadership. San Francisco: Jossey-Bass.
Schildkamp, K. (2019). 'Data-based Decision-making for School Improvement: Research Insights and Gaps.' Educational Research, 61(3), 257–273.
Schildkamp, K. & Kuiper, W. (2010). 'Data-informed Curriculum Reform: Which Data, What Purposes, and Promoting and Hindering Factors.' Teaching and Teacher Education, 26(3), 482–496.
Jamie Scobie writes from extensive experience in Scottish secondary education, including pastoral care, data for improvement, and school self-evaluation. This blog is an independent publication: he writes in a personal capacity as the creator of Learning Lens, writing about classroom observation, teaching evidence, and education policy. He speaks here only for himself and for Learning Lens, not for any employer or other organisation. He holds an MSt from Cambridge (Distinction) and a Masters from Stirling.