How Education Works | Education Sector Analysis & System Diagnosis — How a Country Finds What Its Education System Is Actually Doing Before It Plans the Next Reform

HEW-NODE-0117 · How Education Works · education sector analysis, system diagnosis, access, participation, learning, equity, teachers, finance, administration, external efficiency, institutional capacity, demographic context, evidence synthesis, problem diagnosis and education planning

A national education system can contain millions of learners, hundreds of thousands of teachers, thousands of institutions, several layers of government, multiple examination routes, large public budgets, private providers, training institutions, qualification bodies and years of accumulated policy—and still be surprisingly difficult to see as one system.

One ministry knows the enrolment numbers. Another agency knows the population projections. A finance ministry knows how much was appropriated. Examination authorities know results. District offices know which schools struggle to recruit. Families know where transport is difficult. Teachers know which curriculum expectations do not fit available time. Employers know where graduates arrive without the capabilities they expected. Universities know which school-leavers need bridging. None of those views is false. None, by itself, is the education system.

An education sector analysis is the disciplined act of making the whole system visible enough to decide what problem should be solved next.

This node sits beside the How Education Works hub, Education Sector Planning & Annual Operationalisation, Education Management Information Systems, School Census & Annual Education Data Collection, Education Statistics Quality Assurance & Data Validation, National Learning Assessment Systems, Education Costing, Education Budget Formulation & Medium-Term Expenditure Frameworks, School Mapping & Capacity Planning and Teacher Workforce Forecasting.

Those pages keep their jobs. EMIS owns the information infrastructure. The school census owns the recurring collection mechanism. Statistics quality assurance owns whether the resulting statistics can be trusted. National assessment owns representative evidence about learning. Costing owns the translation of policy ambitions into resource requirements. Budget formulation owns the route from priorities into funded commitments. Sector planning owns the choice, sequencing and operationalisation of a future programme of work. This page owns the layer immediately before those decisions: the integrated diagnosis of what the education system is, what it is producing, where it is losing learners or capability, why the observed problems may be occurring, which groups carry the burden, what resources are being used, and which constraints are important enough to shape the next plan.

The 60-Second Read

• Education sector analysis is diagnosis, not a reform wish list.
• It begins by defining the system being analysed: levels, providers, populations, institutions, time period and geographic scope.
• Context matters because demography, the economy, conflict, migration, language, health, labour markets and public finance change what education can and must do.
• Access analysis asks who enters education, who does not, when participation changes and where learners leave.
• Flow analysis follows cohorts through entry, repetition, promotion, transfer, completion and dropout.
• Learning analysis asks what students can actually do, not merely whether they are enrolled.
• Equity analysis disaggregates averages so disadvantage is not hidden inside national improvement.
• Teacher analysis examines supply, deployment, qualifications, workload, attrition, subject shortages and geography.
• Finance analysis asks how much is spent, on what, for whom, with what distribution and with what results.
• Management analysis examines whether the centre, middle tier and schools can execute the responsibilities assigned to them.
• External-efficiency analysis asks what happens after education: further study, work, earnings, participation and social outcomes.
• A strong analysis connects variables rather than publishing disconnected chapters.
• It distinguishes symptoms from mechanisms. Low completion may arise from cost, distance, weak learning, health, violence, language mismatch or institutional rules.
• It treats missing data as evidence about system visibility, not as permission to invent precision.
• It compares trends across time, not only one-year snapshots.
• It checks whether denominators, definitions and classifications changed before declaring progress or decline.
• It uses both quantitative and qualitative evidence where the question requires them.
• It records uncertainty, data quality and competing explanations.
• It should identify a small number of consequential system problems rather than call everything a priority.
• Its final job is to make the next planning decision more truthful.

One-Sentence Definition

Education sector analysis is a structured, evidence-based diagnosis of an education system’s context, participation, learning, equity, resources, financing, management, capacity and outcomes so that policy and planning begin from the system that actually exists rather than the system people assume exists.

The First Rule: Diagnose Before Prescribing

Education debates are often rich in proposed solutions before the problem has been specified. Build more schools. Reduce class size. Change the curriculum. Add technology. Recruit teachers. Extend the day. Increase scholarships. Introduce vocational education. Decentralise decisions. Centralise standards. Raise pay. Tighten accountability.

Any of those interventions can be sensible in the right system state. Any can also consume years of attention while leaving the binding constraint untouched. If low participation is caused mainly by travel time, changing the curriculum will not create a route to school. If students attend but cannot read the instructional language, building more places does not solve the learning problem. If teachers have been recruited nationally but are concentrated in already well-served areas, another intake may enlarge the workforce while preserving the geographic shortage. Diagnosis exists to stop the solution from outrunning the evidence.

Sector Analysis Is Not the Same as Sector Planning

The distinction matters. Analysis asks: What is happening? For whom? Where? Since when? Compared with what? Through which mechanism? With what level of confidence?

Planning asks: What should we change? In what sequence? At what cost? Under whose responsibility? By when? With which implementation and monitoring arrangements?

The Education Sector Planning & Annual Operationalisation page owns the second problem. This node protects the first. Mixing them too early creates confirmation bias: analysts begin searching for evidence that justifies a preferred programme instead of testing what the system most needs.

Define the System Boundary Before Measuring It

A sector analysis can quietly become incoherent if different chapters are analysing different systems. Does “education” include early childhood? Non-formal education? Technical and vocational education and training? Tertiary education? Private schools? Faith-based providers? Community schools? Adult learning? Specialised institutions? Alternative pathways? Refugee education? Apprenticeships?

The answer can vary by purpose, but the boundary should be explicit. The analysis must say which institutions and learners are inside the denominator, which are outside, which are poorly observed and which cross institutional boundaries.

Time Is a System Boundary Too

A one-year photograph can be useful for operations and misleading for diagnosis. Enrolment may look high because a temporary cohort is unusually large. Spending may jump because a capital programme peaked. Examination performance may change because the test framework changed. Teacher vacancies may fall because posts were frozen rather than filled.

Sector analysis should therefore separate structural trends from short-run shocks. Five- or ten-year sequences often reveal what a single year hides.

Start With Context, Because Education Does Not Float Above Society

The same school system behaves differently under population growth, population decline, urbanisation, rural dispersion, high migration, fiscal contraction, economic expansion, conflict, epidemic risk or climate exposure. These are not decorative background sections. They are constraints and demand drivers.

A country with a fast-growing school-age population may need to add classrooms and teachers merely to keep participation constant. A country with sustained fertility decline may instead face shrinking cohorts, excess capacity in some areas, school consolidation questions and a chance to improve resource intensity without increasing total spending. The education system can look unchanged institutionally while the population underneath it has changed completely.

Demography Changes the Denominator

Enrolment counts alone do not reveal access. Ten million primary pupils may represent near-universal participation in one country and a large exclusion gap in another. Analysts need the relevant population by age, sex, geography and often other characteristics before participation ratios acquire meaning.

This is why population estimates and projections belong upstream of capacity planning. The related School Capacity Planning node translates future demand into enough real places; sector analysis first establishes the demographic shape of that demand and how well the present network serves it.

Access Is More Than Enrolment

A learner can be administratively enrolled and educationally absent. A child may have a school place but travel two hours each way. A student may attend intermittently because transport is unreliable. A teenager may remain on the roll while working enough hours that learning time disappears. A student with a disability may be formally admitted but unable to use parts of the campus or curriculum.

Access analysis therefore asks whether a place is reachable, affordable, usable and sustained. The School Attendance, Journey to School, School Fees & Household Education Costs and School Accessibility & Reasonable Accommodation pages own those operating mechanics in detail.

The Cohort Is One of the Most Useful Units of Analysis

Imagine 100 children who should enter Grade 1 in the same year. How many actually enter? How many enter late? How many repeat? How many reach the end of primary? How many transition to lower secondary? How many complete? How long does the journey take? Which groups leave disproportionately?

Cohort flow turns a static system into a moving one. It reveals where learners accumulate, where they disappear and where policies create delays or bottlenecks.

Repetition Can Look Like Participation While Consuming Capacity

If many learners repeat a grade, classrooms can remain full even as progression is weak. Repetition raises the number of pupil-years needed to produce a completer, occupies places that new entrants need and can increase age-grade distortion. But the interpretation requires care: automatic promotion without learning can produce a different failure. The Student Promotion, Progression & Grade Repetition page owns the policy mechanics. Sector analysis measures their system consequences.

Completion Is Not the Same as Learning

A system can improve enrolment, attendance and completion while leaving foundational learning weak. That is why participation indicators and learning indicators must be read together. A seat is necessary. It is not sufficient.

National examinations may provide one signal, but they often cover only selected grades, pathways or candidates. A representative national learning assessment can answer a different system question: what can the population of learners at a defined stage actually do? The National Learning Assessment Systems node owns that measurement architecture.

Learning Data Need Construct Boundaries

A reading score is not “quality of education.” A mathematics score is not “school effectiveness.” Examination results are shaped by who reaches the examination, what the test measures, curriculum alignment, language, opportunity to learn, socioeconomic conditions and many other factors.

A disciplined analysis avoids turning one convenient measure into a total judgement about the system.

Equity Begins Where the Average Ends

A national completion rate of 90 per cent can coexist with much lower completion for rural learners, children with disabilities, displaced learners, linguistic minorities, poor households or particular regions. An average can rise because already advantaged groups improve while a marginalised group remains unchanged.

Sector analysis should therefore ask not only “how much?” but “for whom?” and “where?” Disaggregation should follow policy-relevant and ethically appropriate categories for which data quality is adequate. Small groups require privacy protection and cautious interpretation.

Intersection Matters

Disadvantage does not always arrive one variable at a time. A rural girl with a disability may face barriers that cannot be understood by looking separately at rurality, gender and disability averages. A migrant student in a low-income household may experience language, documentation, transport and cost constraints simultaneously.

The goal is not to produce endless cross-tabulations. It is to find combinations that reveal a mechanism the system can act upon.

Teacher Numbers Are a Starting Point, Not a Workforce Diagnosis

A country can have enough teachers nationally and severe shortages locally. It can have enough teachers overall and too few mathematics, science, special education or language teachers. It can recruit enough new teachers while losing experienced teachers faster than they can be replaced. It can have favourable pupil-teacher ratios on paper while absence, non-teaching duties or split shifts reduce real instructional capacity.

Useful workforce analysis therefore connects headcount, full-time equivalent capacity, subject, qualification, deployment, age, attrition, workload, absence and geography. See Teacher Workforce Forecasting, Teacher Deployment, Teacher Retention and Teacher Absence & Substitute Coverage.

Infrastructure Analysis Should Measure Usability

Counting classrooms is not the same as counting usable learning spaces. Buildings may lack electricity, water, toilets, ventilation, accessibility, furniture or safe structural condition. A laboratory may exist but be unavailable because equipment is broken. A computer room may exist without reliable connectivity. A school may have nominal capacity that cannot be used during extreme heat.

Sector diagnosis should distinguish assets from functioning services. The relevant operating nodes include School Infrastructure Maintenance, WASH in Schools, School Electrification & Energy Reliability and School Connectivity.

Finance Analysis Starts With the Resource Envelope

How much public money is available? How has it changed in real terms? How large is education relative to total public expenditure and national income? What is the mix between recurrent and capital spending? How much comes from households or external partners? What fiscal pressures limit future growth?

Those questions establish the envelope, but sector analysis must go further. Money should be connected to levels of education, inputs, geography, learner groups and results. A spending total without distribution can hide major inequality.

Unit Costs Reveal the Cost of Producing Educational Opportunity

Per-student expenditure can be compared across levels, institutions or regions, but interpretation needs care. A remote school may legitimately cost more per learner because fixed costs are spread across a small enrolment. A technical programme may require equipment that general education does not. Specialised supports may deliberately raise costs to secure equal access.

The goal is not “lowest unit cost.” It is understanding what cost structure is buying and whether resources match intended service standards and learner needs.

Efficiency Has More Than One Meaning

Internal efficiency examines how smoothly learners move through the system: repetition, dropout, completion and the number of years of schooling consumed per completer. Technical efficiency asks whether a given result could plausibly be achieved with fewer resources or whether more result could be achieved with the same resources. Allocative efficiency asks whether resources are distributed toward the priorities that matter most. None should be reduced to “spend less.”

External Efficiency Looks Beyond the School Gate

Education is not complete when a student receives a certificate. Systems should ask what happens afterward: entry into further education, employment, occupational match, earnings, civic participation, health, family outcomes and other socially valued capabilities.

The Graduate Tracer Studies & Labour-Market Feedback node owns one important evidence stream. Sector analysis places such outcomes beside access, learning, equity and cost rather than treating employment alone as the purpose of education.

Administration Is Part of the Education System

A technically excellent policy can fail because nobody has the authority, staffing, systems or routines required to implement it. Sector analysis therefore needs to examine educational administration itself: role clarity, staffing, skills, information flow, procurement, financial management, supervision, coordination, decision rights and the relationship between central, regional, district and school levels.

This is where the analysis crosses from “what policy says” to “what the organisation can actually do.”

Capacity Is Specific, Not Generic

Saying “capacity is weak” is almost useless. Which function? Which organisation? Which role? Is the constraint staffing, expertise, authority, workload, systems, funding, leadership, incentives, data or coordination?

A ministry may have strong curriculum expertise and weak capital-project management. A district may monitor attendance well but lack school-improvement coaching capacity. A school may have capable teachers and no administrative support. Capacity diagnosis should be attached to a task.

Political Economy Is Not an Excuse to Avoid Evidence

Education systems contain interests. Teacher organisations, private providers, universities, examination bodies, communities, local governments, ministries and political leaders can have different incentives and different control over implementation. Understanding those incentives can explain why technically straightforward reforms stall.

But “political economy” should not become a vague label for anything difficult. A useful analysis identifies actors, formal authority, resources, incentives, likely gains or losses, implementation dependencies and the specific mechanism through which a decision may be supported, altered or blocked.

Stakeholder Voice Adds Information the Database Does Not Contain

Administrative data can tell analysts that absenteeism is higher in a district. Interviews or fieldwork may reveal that a bridge floods during the rainy season, that students are caring for siblings, that the timetable conflicts with transport schedules or that families no longer trust a particular route.

Qualitative evidence is not a substitute for measurement when measurement is possible. It is often the route from pattern to mechanism.

The Evidence Inventory Comes Before the Dashboard

A sector analysis should map its sources before combining them: school census, learner records, teacher records, finance systems, payroll, assessments, examinations, household surveys, labour-force surveys, population censuses, demographic projections, health data, geospatial data, infrastructure inventories, administrative records and qualitative studies.

For each source, record coverage, reference period, definitions, missingness, known biases, revision status and whether the data can be linked safely to other sources. This prevents the attractive dashboard from hiding incompatible foundations.

Definitions Are Part of the Data

“Teacher,” “school,” “dropout,” “completion,” “disability,” “rural,” “trained teacher” and “public spending” can all be defined differently across systems or over time. A change in definition can create an apparent trend with no underlying behavioural change.

The Education Statistics Quality Assurance & Data Validation node owns the detailed integrity process. Sector analysis should consume only statistics whose definitions and limitations it understands.

Missing Data Are a Finding

If the system cannot determine how many teachers work in each subject, cannot connect spending to schools, cannot identify which children with disabilities are out of school, or cannot track students who transfer between providers, that is not merely an analyst inconvenience. It is evidence that the system lacks visibility over a function it may be expected to manage.

A responsible report distinguishes “unknown” from zero and “not measured” from “no problem.”

Triangulation Is a Protection Against One-System Blindness

If administrative enrolment rises sharply but household survey participation does not, investigate. If payroll reports 10,000 teachers but the school census finds 9,200 in post, investigate. If examination pass rates rise while a stable national assessment shows no corresponding improvement, investigate. If budget execution is high but schools report chronic shortages of materials, investigate.

Conflicting evidence is not automatically a problem to average away. It may be the most informative signal in the analysis.

Comparators Need a Reason

International comparison can help, but the nearest country alphabetically is not necessarily the most useful comparator. Analysts may choose peers by income, demographic structure, language, geography, institutional model, regional context or educational starting point. Different comparisons answer different questions.

The comparison should reveal a plausible range or illuminate a mechanism, not simply rank countries.

Benchmarking Is Not Copying

If another system achieves better learning with similar resources, that is a question worth asking. It is not proof that importing one visible policy will recreate the outcome. Institutions, teacher labour markets, curriculum coherence, family conditions and implementation histories differ.

Strong sector analysis uses international examples to sharpen questions, not to bypass local diagnosis.

Move From Indicators to Mechanisms

An indicator says something changed. A mechanism explains how. Suppose lower-secondary completion falls in remote districts. Possible mechanisms include distance, insufficient school places, teacher shortages, transport cost, language transition, household labour demand, safety, examination rules or poor learning in primary school that makes later work inaccessible.

The analysis should progressively eliminate implausible explanations and strengthen plausible ones using converging evidence.

Build a Problem Tree Carefully

Problem trees are useful when they prevent premature solutions. Start with a clearly observable problem. Identify direct causes, deeper causes and consequences. Then test each link against evidence. Do not put “lack of awareness” or “lack of capacity” everywhere simply because those phrases are easy to write.

A problem tree should become more precise as evidence enters, not more elaborate for presentation.

Distinguish Constraint, Cause and Correlate

Three variables can move together without one causing the other. Schools with more books may have better results because they are better funded generally. Teacher experience may correlate with outcomes because experienced teachers are more likely to be assigned to advantaged schools. Household income may correlate with attendance through transport, nutrition, housing stability and expectations.

Sector analysis need not turn every chapter into a causal study, but it should be disciplined about what the evidence can and cannot establish.

Quantify the Size of the Problem

A problem can be severe for individuals and small in system scale, or modest for each learner and enormous in aggregate. Decision-makers need both dimensions. How many learners are affected? How large is the outcome gap? How concentrated is it? What is the trajectory? How much system capacity would repair require?

Locate the Problem in the System

National averages can conceal geographic concentration. Mapping access, teacher shortages, infrastructure conditions, learning outcomes or expenditure can reveal clusters that change the policy response. A national programme may be inefficient if the problem is highly localised. A local programme may be inadequate if the problem reflects a national rule.

Geography is therefore both a descriptive and a governance variable.

Find the Level at Which the Constraint Can Be Changed

A school cannot fix a national salary scale. A ministry cannot directly manage every classroom routine. A district may control teacher deployment but not the number of funded positions. A family cannot repair a road. Diagnosis should identify not only the mechanism but the authority level capable of acting on it.

This links analysis to the Education Decentralisation & Delegated Authority and District & Regional Education Offices nodes.

Look for Bottlenecks, Not Just Weak Indicators

If teacher training capacity can produce only 2,000 new teachers annually but workforce need is 5,000, recruitment advertising is not the binding constraint. If the ministry can fund new classrooms but land acquisition takes six years, construction budget is not the first bottleneck. If textbooks reach district warehouses but not schools, central procurement is not the whole supply problem.

Sector analysis becomes operationally useful when it identifies where throughput is constrained.

Look for Feedback Loops

Some problems reinforce themselves. Weak learning increases repetition. Repetition increases class size. Higher class size reduces individual support. Frustration increases dropout. Lower completion weakens the supply of future teachers in remote areas. Those loops mean a problem can accelerate unless interrupted.

Conversely, improvement can compound: better early literacy makes later curriculum more accessible, reducing repetition and freeing capacity.

Separate Stocks From Flows

A stock is what exists at a point in time: teachers employed, classrooms available, students enrolled, debt outstanding. A flow changes the stock: recruitment, attrition, construction, graduation, dropout, transfers. Systems can look stable until the flows become visible.

For example, a teacher workforce may still be large while retirements are accelerating and the training pipeline is shrinking. Analysis that sees only the stock notices the crisis late.

Separate Capacity From Utilisation

A school may have twelve classrooms but use only eight because four are unsafe. A training college may have 1,000 seats but graduate 600 teachers because practicum placements are scarce. A budget line may exist but execute slowly because procurement starts late. Nominal capacity and usable capacity are different variables.

Separate Inputs, Processes, Outputs and Outcomes

Inputs include teachers, buildings, materials and money. Processes include teaching time, attendance, supervision and management. Outputs include lessons delivered, learners completing or qualifications issued. Outcomes include learning, progression, employment, wellbeing and other capabilities.

A system can increase inputs without changing processes; improve processes without immediate long-run outcomes; or produce credentials whose external value is weak. Keeping the layers distinct prevents causal shortcuts.

Use a Theory of the System Before a Theory of Change

Before designing how an intervention is supposed to work, understand how the existing system works. Who allocates teachers? Who approves textbooks? How does money move? What determines school places? Who controls examinations? How are complaints escalated? Which data arrive too late to influence decisions?

A theory of change built on the wrong institutional map will fail elegantly.

Risk Analysis Belongs Inside Sector Diagnosis

Climate hazards, conflict, pandemics, cyber incidents, fiscal shocks, energy interruptions and demographic change can alter the future trajectory of education. A sector analysis should identify major risks, exposure, vulnerability and existing resilience rather than assuming the next planning period will resemble the last.

The point is not to predict every shock. It is to avoid building a plan whose success depends on the absence of foreseeable disruption.

Scenario Analysis Helps When the Future Is Not One Number

Population growth may be uncertain. Economic growth may vary. Migration can change local demand quickly. Teacher attrition may respond to labour-market conditions. Rather than hiding uncertainty inside one “best estimate,” analysts can build plausible scenarios and test which decisions remain robust across them.

Current Authoritative Baseline: IIEP-UNESCO Sector Analysis

The International Institute for Educational Planning at UNESCO maintains a formal education sector analysis methodology developed over two decades with partners including the World Bank, UNICEF and the Global Partnership for Education. Its published Education Sector Analysis Methodological Guidelines, Volume 1 organise sector-wide diagnosis around context, access, costs and financing, quality, system capacity and management, external efficiency and equity. A later Volume 3 extends the methodology to areas including disability inclusion, risk and resilience, educational administration and stakeholder/problem-driven analysis.

The important lesson is architectural. A serious sector analysis is not a collection of unrelated indicators. It intentionally joins participation, quality, finance, management, equity and outcomes so that one chapter can explain another.

Current Global Data Baseline: UNESCO UIS

The UNESCO Institute for Statistics remains the official international source for comparable education statistics. Its World Education Statistics 2025 organises evidence across the SDG 4 agenda, while the 2026 Education Data Refresh, updated in June 2026, expanded the published global education database to more than 8.26 million country datapoints. The scale is useful; the analytical lesson is more important: national diagnosis should know which international indicators can support comparison and which questions still require national administrative, survey or qualitative evidence.

A Strong Analysis Has a Data Quality Ledger

• source owner;
• reference year;
• coverage;
• unit of observation;
• definition;
• revision status;
• missingness;
• known bias;
• comparability across years;
• comparability across regions;
• disaggregation available;
• privacy constraints;
• linkage feasibility;
• confidence in use;
• questions the source cannot answer.

A Strong Analysis Has an Assumption Ledger Too

Not every claim rests on directly observed data. Analysts interpolate missing years, estimate private enrolment, project populations, classify expenditures, infer teacher full-time equivalents or assume transition rates. Those assumptions should be recorded so readers can see where the diagnosis is measured and where it is modelled.

Do Not Bury the Base Population

Rates require denominators. Analysts should state whether population figures come from a census, intercensal estimate, household survey or demographic projection and whether age definitions align with official school ages. Small denominator changes can materially alter participation rates.

Do Not Compare Nominal Money Across Time

A budget can rise in nominal currency while purchasing less after inflation. Exchange-rate conversion can distort trends. Capital spending can be lumpy. Analysts should choose real-price measures and purchasing comparisons appropriate to the question and state them explicitly.

Do Not Confuse Budget With Expenditure

An approved budget is permission to spend, not proof that resources arrived. Sector analysis should distinguish approved allocations, releases, commitments, payments and actual service delivery. See Budget Execution & Public Expenditure Tracking and Education Treasury & Cash Management.

Do Not Compare Examination Cohorts Without Selection Context

If fewer weak students reach an examination, average scores can rise without learning improving for the original cohort. If participation expands to include previously excluded students, average scores can fall while the system becomes more inclusive. Sector analysis should examine who is in the tested population before interpreting the score.

Do Not Treat Private Education as Invisible

Where non-state providers educate a significant share of learners, excluding them can distort enrolment, teacher, spending and equity analysis. At the same time, data quality may differ from public systems. The Non-State Education Provider Regulation node owns the governance layer; sector analysis should make provider coverage explicit.

Do Not Treat Household Spending as Free Money

Household spending can supplement public provision and also create exclusion. Fees, transport, uniforms, tutoring, devices and opportunity costs can shift the effective price of education. If public spending analysis ignores household burden, apparent system affordability may be misleading.

Do Not Turn Every Correlation Into a Priority

A large dataset can produce many statistically significant relationships. The test for planning relevance is stronger: is the relationship substantively important, plausibly causal, actionable by an identifiable actor, large enough to matter and supported by other evidence?

Prioritisation Begins After Diagnosis, but Analysis Should Prepare It

A good sector analysis ends with a manageable set of problems described in a way that planning can use. For each, state the scale, affected groups, geography, trend, likely mechanisms, uncertainty, responsible system level, related resource implications and evidence gaps.

That is very different from ending every chapter with “more investment and capacity building are needed.”

A Practical Problem Statement

Weak: “Rural education quality is poor.”

Stronger: “In 42 remote lower-secondary schools, Grade 8 mathematics learning is substantially below the national distribution while teacher vacancy rates, out-of-field teaching and annual teacher turnover are materially higher; attendance is similar to comparable schools, and the pattern has persisted for four cohorts. The evidence suggests that subject-specific staffing instability is a major constraint requiring further causal and operational validation.”

The stronger statement does not pretend certainty. It makes the next investigation and planning step possible.

Case Study: The Country With “Too Few Teachers”

Invented example: a country reports a national pupil-teacher ratio of 27:1, apparently close to its policy target. Yet district data show ratios below 20:1 in large cities and above 50:1 in several remote areas. Mathematics and science vacancies are concentrated even more sharply. Teacher-training output is adequate nationally, but new teachers can transfer after one year and housing is scarce in remote districts.

The diagnosis changes from “train more teachers” to “repair deployment, subject matching and retention in specific geographies.” The relevant operational owners then become Teacher Deployment, Hard-to-Staff School Teacher Incentives and Teacher Retention.

Case Study: The Country With Rising Completion and Flat Learning

Invented example: primary completion rises from 78 to 91 per cent over eight years. This is a real access achievement. A stable learning assessment, however, shows little improvement in foundational reading. Class sizes grew, textbooks arrived late in two regions and instructional time fell because of recurrent closures.

The analysis refuses two bad conclusions: “the reform failed” and “the system succeeded.” It records improved completion and a persistent learning bottleneck, then asks what changed in the learning production process.

Case Study: The Country Spending More but Reaching Schools Late

Invented example: education expenditure rises substantially and budget execution appears high. School surveys show that operating grants arrive near the end of the academic year. Schools delay repairs, materials and local services, then rush to spend before year-end.

The sector problem is not simply “insufficient finance.” It includes cash timing and execution quality. The diagnosis routes to Education Treasury & Cash Management and School Grants & Direct-to-School Funding.

Case Study: The Shrinking Cohort Nobody Planned For

Invented example: births have fallen for twelve years, but school staffing and capital formulas still assume growth. Some rural schools now have very small cohorts while cities continue to receive internal migrants. National averages suggest spare capacity; local maps show simultaneous surplus and shortage.

The diagnosis requires demographic projection, school mapping, teacher allocation and transport analysis together. No single indicator owns the problem.

Failure Mode 1: Start With the Minister’s Preferred Reform

Repair: define diagnostic questions before selecting interventions and maintain a visible separation between evidence synthesis and policy choice.

Failure Mode 2: Produce 300 Indicators and No Diagnosis

Repair: organise indicators around system questions, mechanisms and consequential bottlenecks rather than data availability.

Failure Mode 3: Use Only National Averages

Repair: disaggregate by geography and relevant learner groups, then test whether apparent national progress is broadly shared.

Failure Mode 4: Use Only Administrative Data

Repair: triangulate with surveys, assessments, geospatial evidence and qualitative work where administrative systems cannot observe the mechanism.

Failure Mode 5: Use Only Surveys

Repair: use administrative systems for operational detail and recurring monitoring, while respecting their quality limitations.

Failure Mode 6: Hide Missing Data With False Precision

Repair: show ranges, confidence levels, assumptions and unresolved gaps; treat the visibility failure as a system finding.

Failure Mode 7: Compare Money Without Inflation or Context

Repair: use real-value and per-learner measures appropriate to the question and separate recurrent from capital effects.

Failure Mode 8: Interpret Test Scores Without Participation

Repair: examine who was eligible, who participated and whether cohort composition changed.

Failure Mode 9: Call Every Problem “Capacity”

Repair: specify the function, actor, missing capability and operational consequence.

Failure Mode 10: End With a Long List of Priorities

Repair: identify a small set of high-consequence problems, dependencies and decision-relevant uncertainties for the planning stage.

The Education Sector Analysis Operating Chain

1. Define the sector boundary.
2. Define the time period and geographic scope.
3. Map institutions, providers and governance levels.
4. Inventory administrative, survey, assessment, finance, demographic and qualitative evidence.
5. Record data definitions, coverage and quality.
6. Describe demographic, economic, social, fiscal and risk context.
7. Analyse entry and participation.
8. Analyse cohort flow, repetition, transfer, dropout and completion.
9. Analyse attendance and usable learning time.
10. Analyse learning evidence and its construct limits.
11. Analyse equity and intersectional patterns.
12. Analyse teacher supply, deployment, qualification, workload and attrition.
13. Analyse infrastructure and service usability.
14. Analyse financing, expenditure composition and distribution.
15. Analyse internal efficiency.
16. Analyse external outcomes and transitions.
17. Analyse administrative and implementation capacity.
18. Map authority and stakeholder incentives for major functions.
19. Triangulate conflicting evidence.
20. Build trend and geographic views.
21. Distinguish stocks from flows.
22. Distinguish capacity from utilisation.
23. Identify bottlenecks and feedback loops.
24. Test plausible mechanisms.
25. Quantify problem scale and affected groups.
26. Record uncertainty and missing evidence.
27. Identify the governance level capable of acting.
28. Run plausible future scenarios where uncertainty matters.
29. Write decision-ready problem statements.
30. Hand the diagnosis to planning without pre-empting the plan.

A Sector-Diagnosis Dashboard Should Connect, Not Merely Display

• school-age population and demographic trend;
• entry rate by age;
• gross and net participation where appropriate;
• out-of-school population estimates;
• attendance and chronic absence;
• age-grade distortion;
• repetition and promotion;
• transition between levels;
• completion;
• learning distributions, not only averages;
• equity gaps;
• teacher headcount and full-time equivalent capacity;
• teacher deployment by subject and geography;
• vacancy and attrition;
• usable classrooms and facility services;
• instructional materials availability;
• public expenditure in real terms;
• expenditure per learner;
• spending by level and function;
• household education costs where available;
• budget execution and cash timing;
• administrative staffing and function capacity;
• graduate transitions;
• major system risks;
• data gaps that prevent management.

Quality Check 1: Can the Report Explain a Contradiction?

If enrolment rises while completion falls, can the analysis explain why? If spending rises while material availability declines, can it trace the route? If teacher numbers increase while class sizes remain high, can it locate where teachers went? Contradictions are tests of whether the analysis has connected the system.

Quality Check 2: Can the Report Name Who Is Missing?

A sector analysis that describes only students already visible in schools can miss the population most in need of policy attention. Compare administrative records with population and household evidence to ask who never entered, who left and who is poorly represented in the data.

Quality Check 3: Can the Report Distinguish Money Problems?

Insufficient allocation, inequitable allocation, late release, procurement failure, weak execution, leakage and inefficient design are different finance problems. A useful diagnosis does not call all of them “lack of funding.”

Quality Check 4: Can the Report Separate Policy From Delivery?

A country may have a strong policy on paper and weak service in practice. The analysis should test implementation: staffing, timing, coverage, compliance, support and local variation.

Quality Check 5: Can the Report State What It Does Not Know?

Credibility rises when uncertainty is visible. The strongest analysis is not the one with an answer to every question. It is the one that knows which answers are well supported and which remain hypotheses.

Canonical Owner Boundaries

• Education Management Information Systems owns the recurring information infrastructure.
• School Census & Annual Education Data Collection owns the annual system return.
• Education Statistics Quality Assurance & Data Validation owns statistical integrity and validation.
• National Learning Assessment Systems owns population-level learning measurement.
• Education Costing owns resource requirement models for proposed policy.
• Education Budget Formulation & MTEFs owns the budget-preparation route.
• Education Sector Planning & Annual Operationalisation owns policy choice, sequencing, implementation planning and annualisation.
• School Improvement Planning owns the school-level improvement plan.

This node owns the integrated system diagnosis that sits upstream of those jobs: assembling credible evidence across domains, locating problems, testing mechanisms, understanding distribution and capacity, documenting uncertainty and producing decision-ready statements about the education system’s present condition.

The Return Path

Return to the ministerial conference room before a new five-year plan is written.

On one screen is an enrolment chart. On another is a learning assessment. The finance team has a spending table. The teacher directorate has vacancies. The infrastructure unit has a list of projects. A regional director says the averages are hiding remote schools. A parent organisation says transport and cost are keeping adolescents out. A university says incoming students need remedial mathematics. The planning team has three weeks before the strategy workshop.

The easiest move is to stack the presentations together and call the result a sector analysis.

The harder and more useful move is to ask how the evidence connects. Which cohorts are being lost? Which problems are concentrated? What changed first? Which constraints are upstream? Which numbers are trustworthy? Where is the system blind? Which problem can be changed by a school, a district, a ministry or the treasury? Which apparent success is being purchased by an unsustainable mechanism? Which apparent failure is actually the side effect of broader inclusion?

That is the work of diagnosis.

A good education plan begins one step earlier than ambition: with a truthful account of the system that must carry it.

Return to the How Education Works hub.