Understanding education in emergencies

1Why education cannot wait

Education in emergencies (EiE) refers to the quality learning opportunities provided for all ages in crisis situations: conflicts, disasters, epidemics, forced displacement. It covers early childhood development, primary, secondary, non-formal, technical and vocational education. It is a right (article 28 of the Convention on the Rights of the Child) that does not stop when a crisis strikes.

School is also a protection mechanism. A child out of school is more exposed to recruitment by armed groups, child labour, early marriage, exploitation and sexual violence. A structured learning space provides a routine, trusted adults and an entry point to health, nutrition, protection and water services. Yet education usually receives less than 3% of humanitarian funding, while crises often last more than ten years: an entire generation can lose its schooling.

2The INEE Minimum Standards

The Inter-agency Network for Education in Emergencies (INEE) has published Minimum Standards, recognised as Sphere companion standards. They comprise 19 standards organised in five domains:

• Foundational standards: community participation, coordination and analysis (assessment, response strategies, monitoring, evaluation).
• Access and learning environment: equal access, protection and well-being, facilities and services.
• Teaching and learning: curricula, teacher training and professional support, instruction and learning processes, assessment of learning outcomes.
• Teachers and other education personnel: recruitment and selection, conditions of work, support and supervision.
• Education policy: law and policy formulation, implementation.

Each standard comes with key actions and guidance notes. The foundational standards cut across all other domains: there is no quality response without community participation, coordination and analysis.

3Barriers to access and learning

A sound analysis identifies, group by group, what prevents children from going to school, staying there and learning:

• Physical and security barriers: schools destroyed, occupied by displaced people or armed forces, distance, dangerous roads, landmines.
• Economic barriers: school fees, uniforms, supplies, child labour, loss of livelihoods.
• Social and gender barriers: early marriage, household chores, lack of separate latrines and menstrual hygiene management, shortage of female teachers.
• Disability-related barriers: inaccessible buildings, prejudice, lack of reasonable accommodation and trained teachers.
• Administrative and language barriers: lack of documents, refusal of enrolment, unfamiliar language of instruction, non-recognition of prior learning.

Data must be disaggregated by sex, age and disability; for disability, use the Washington Group questions or the Child Functioning Module (UNICEF and Washington Group).

4The Joint Education Needs Assessment (JENA)

The JENA (Joint Education Needs Assessment) is the needs assessment conducted by Education cluster partners together with the ministry. It follows a step-by-step approach:

• Secondary data review first: ministry statistics (EMIS), existing reports, multi-sector assessments. Only what is missing is collected.
• Targeted primary data collection: observation of schools and sites, interviews with head teachers, teachers and parent committees, focus group discussions with girls, boys and children with disabilities.
• Joint analysis: number of children in need (PiN), severity by area, priority barriers, existing capacities.
• Dissemination and use: findings feed into the Humanitarian Needs Overview (HNO), the Humanitarian Response Plan (HRP) and funding requests.

A JENA respects the "do no harm" principle: no interview that exposes a child, informed consent, protected personal data. A rapid assessment is possible in the first weeks, followed later by a more in-depth analysis.

5Running a JENA from start to finish: method, timeline and roles

A JENA is not a questionnaire: it is a coordination process that leads to a shared analysis that can be used for programming and funding. A poorly prepared JENA produces data nobody uses; a well-run JENA becomes the sector's reference for twelve to eighteen months. The Global Education Cluster (GEC) offers a toolkit and a guide to coordinated education needs assessments and analysis; the method below follows its logic as it is applied in the field.

Phase 0: decide and organise (week 1)

• Decision: the cluster coordination team (or the Education Working Group in a refugee response) proposes the assessment at a plenary meeting, together with the ministry. The analysis question is set out: "How many children and teachers need education assistance, where, with what severity, and why?"
• Assessment working group: 4 to 8 members (cluster coordinator, information manager, ministry or regional directorate, 2 or 3 NGOs including at least one national organisation, a child protection focal point). It validates the terms of reference, the analysis framework and the timeline.
• Intersectoral link: check with OCHA and the inter-cluster information management working group whether a multi-sector needs assessment (MSNA) is planned. It is often better to insert 10 to 15 education questions into it than to launch a separate household survey.

Phase 1: secondary data review (weeks 1 to 3)

The secondary data review (SDR) is the most cost-effective and the most neglected step. You compile: the ministry's statistical yearbook and education management information system (EMIS), census data and population projections by age, displacement data (IOM Displacement Tracking Matrix, UNHCR registration), incident reports (attacks, military use of schools), partner assessments, the previous HNO, and learning studies. Each piece of information is filed in an analysis framework (context, access, quality, protection, teachers, capacities) and rated for reliability and date. The result is a precise list of information gaps: these, and these alone, justify primary data collection.

Phase 2: targeted primary data collection (weeks 3 to 6)

• Units of data collection: the school (school form: condition of the building, functionality, enrolment by sex and grade, teachers present, water and latrines, incidents); the key informant (head teacher, chair of the school management committee, teacher, community leader, IDP focal point); focus group discussions separated by sex and age (adolescent girls, adolescent boys, mothers, fathers, female teachers).
• Sampling: in an emergency, purposive sampling is used, covering every administrative unit and every population type (IDPs in sites, IDPs in host families, host community, returnees). The results are therefore indicative, not representative, and the report must say so. If statistical representativeness is required, rely on the MSNA household survey.
• Teams: mixed-sex pairs, trained for two days (tools, ethics, protection, mobile data entry), with one supervisor for every 4 to 6 pairs and daily quality control of the forms.
• Children: children are not interviewed individually about experiences of violence. Consultations with children are held in groups, using participatory methods, with trained facilitators, parental consent and the child's assent, and a referral pathway available on the same day.

Phase 3: joint analysis (weeks 6 to 8)

A one- to two-day analysis workshop brings together the working group and the main partners. Raw percentages are not presented: they are interpreted. For each area, five questions are answered: how large is the need, who is most affected, why, what already exists, and what will happen without intervention. Disagreements over interpretation are recorded, together with the level of confidence in each conclusion.

Phase 4: dissemination and use (weeks 8 to 10)

Expected outputs: a short report (15 to 25 pages), a dashboard by area, a cleaned and anonymised database shared with partners, and the figures fed into the HNO. The JENA has succeeded if its figures appear in the response plan, partners' proposals and pooled fund priorities.

Rule of thumb: if a question has no planned use in the analysis or programming, it does not go into the tool. Every question costs respondents time and increases the risk of errors.

6Estimating children in need and severity: method and full worked example

The number of people in need (PiN) for education and the severity by area are the two figures that structure the HNO, the HRP and a large share of allocations. Since the adoption of the Joint Intersectoral Analysis Framework (JIAF), severity has been expressed on a scale from 1 (minimal) to 5 (catastrophic) by geographic unit and by population group. The GEC publishes annual guidance on PiN calculation; the principles below follow its logic.

Step 1: establish the school-age population

Start from the school-age population by population group (IDPs, returnees, host community, refugees) and by area. Age bands follow the national system (for example 3 to 5 years for pre-primary, 6 to 11 years for primary, 12 to 17 years for secondary). Where age-disaggregated data are lacking, apply the demographic structure from the census or demographic surveys: the 6 to 17 age group often represents 28 to 35% of the total population in high-fertility countries.

Step 2: define who is "in need"

A child is in need if they meet at least one condition: they have no access to education (out of school or school closed); they are enrolled in an environment that does not meet minimum standards (overcrowded classroom, damaged or occupied building, no water or latrines, no trained teacher); they are exposed to school-related protection risks (attacks, dangerous journeys to school, violence). Teachers may also be counted in the PiN where the cluster so decides (unpaid, displaced or untrained teachers).

Step 3: calculate without double counting

The main pitfall is adding indicators together. An out-of-school child whose school is also damaged counts only once. The usual method: PiN = out-of-school children + children enrolled in inadequate conditions (taking, for the latter, the highest indicator or a non-overlapping combination).

Full worked example: the fictional district of Kessala

Kessala district has 412,000 inhabitants, including 96,000 internally displaced persons who arrived seven months ago after clashes. The demographic structure shows 31% aged 6 to 17. The pre-crisis enrolment rate (EMIS) was 68% for the host community. The JENA establishes the following data.

Total district PiN: 84,665 children, or 66% of children aged 6 to 17. Disaggregation to apply next: sex (according to the observed proportion, here 51% girls among IDPs), age group and disability. For disability, in the absence of local data, the cluster often applies the WHO estimate that around 15% of the population lives with a disability, bearing in mind that prevalence among children measured with the Child Functioning Module varies widely from one country to another; the assumption chosen must be written down.

Assigning severity

Severity combines several indicators, each with thresholds. Example of a grid adopted by a fictional national cluster:

In Kessala, among IDPs: 73% out of school (4), 22% non-functional schools in the hosting areas (3), 6 incidents (4), 78 pupils per qualified teacher (4). With an aggregation rule setting final severity at the level reached or exceeded by at least 50% of indicators, severity is 4. The aggregation rule must be decided before seeing the results; otherwise the temptation to adjust thresholds to obtain the desired severity is strong.

A credible PiN can be defended in two sentences: which population it is calculated on and which conditions cause a child to be counted. If the team cannot explain it, the figure will be challenged during the HNO review.

7Analysing barriers group by group: tools and reading the data

The general list of barriers (physical, economic, social, disability-related, administrative) is well known. Professional work consists of prioritising them for each group and each area, and linking them to responses. A barrier that is not quantified, or at least ranked, guides no decision.

The barriers-by-group matrix

Build a matrix whose rows are the groups (girls aged 6 to 11, boys aged 6 to 11, adolescent girls, adolescent boys, children with disabilities, unaccompanied children, children associated with armed groups, children from a linguistic minority) and whose columns are the barriers. Each cell receives a score from 0 to 3 derived from triangulation (household data, key informants, focus group discussions), together with a source.

Reading: for adolescent girls, four barriers score 3. A response that addresses only cost (school kits) will have only a marginal effect; it must combine economic support, safe journeys to school, flexible schedules and work with families on early marriage. For children with disabilities, distance and accessibility dominate: a ramp in a school 4 km away solves nothing without a transport solution or a nearby school.

Measuring disability correctly

The Child Functioning Module (UNICEF and the Washington Group) exists in two versions: 2 to 4 years and 5 to 17 years. It asks about functional difficulties (seeing, hearing, walking, communicating, learning, concentrating, making friends, anxiety, depression) with four response categories. A child is generally classified as having a functional difficulty if they have "a lot of difficulty" or "cannot do at all" in at least one domain (and, for anxiety and depression, a daily frequency). Common mistakes: asking "is there a person with a disability in the household?", which massively underestimates prevalence; translating the questions without cognitive testing; having the module administered by an untrained enumerator.

Reading access data

• Enrolment, attendance, retention, completion: four different indicators. A high enrolment rate with 55% attendance signals a daily barrier (work, hunger, safety) rather than a barrier to entry.
• Age profile: a sharp drop at age 12 or 13 among girls points to marriage, puberty, menstrual hygiene and the transition to secondary school; among boys, to work or recruitment.
• Over-age children: a high proportion of 12-year-olds in second grade justifies an accelerated education offer rather than simple re-enrolment.
• Comparing IDPs and host community: a small gap indicates a systemic problem (insufficient school provision for everyone); a large gap indicates displacement-specific barriers (documents, cost, language, discrimination).

The role of consultations with children

Focus group discussions with children and adolescents often provide information missing from household surveys: fear of a specific route, harassment by a teacher, shame about clothing, hunger. Suitable methods: mapping safe and unsafe places around the school, ranking picture cards showing reasons for not going to school, "a day in my life". Groups of 6 to 10 participants of the same sex and age group, a facilitator of the same sex, a maximum duration of 60 to 90 minutes, no questions about personal experiences of violence, and a referral protocol in case of disclosure.

A barrier is only useful to the analysis if it leads to a decision: who is affected, how many, where, and which response removes it. A list of barriers without prioritisation is a description, not an analysis.

8The INEE Minimum Standards as a diagnostic and accountability tool

The INEE Minimum Standards are often cited in proposals and rarely used. They become a working tool when turned into a self-assessment grid: for each standard, you check what is in place, what is missing, and who is responsible for closing the gap. They are linked to the Core Humanitarian Standard (CHS), whose nine commitments apply to every response, and to the Minimum Standards for Child Protection in Humanitarian Action (CPMS), one of which is dedicated to education.

Self-assessment grid for an education programme

How to use it in practice

1. At start-up: complete the grid with the team and one or two community representatives. It forms the basis of the quality action plan.
2. Every six months: update it, ideally with an outside perspective (another cluster partner, a regional technical adviser).
3. During evaluation: evaluators use the OECD DAC criteria (relevance, coherence, effectiveness, efficiency, impact, sustainability); the INEE grid provides material for the relevance, coherence and sustainability criteria.

What donors and evaluators look at

• An explicit link between the needs analysis (JENA, HNO) and the strategy: why these areas, these groups, these activities.
• Data disaggregated by sex, age and disability from the baseline onwards, not only in the final report.
• Learning and well-being indicators, not just access indicators (number of enrolled children, kits, classrooms).
• Compliance with cluster standards and coherence with national policy.
• Accountability mechanisms towards learners and parents (complaints, feedback) that actually work, with logs and responses.
• The organisation's child safeguarding policy and its practical application: background checks, signed code of conduct, training, reporting.

Common mistakes

• Citing the INEE Minimum Standards in the introduction without any activity or indicator deriving from them.
• Confusing standards and indicators: a standard describes a qualitative outcome; indicators are defined locally, with the cluster.
• Mechanically applying figures (one square metre per pupil, 40 pupils per class) without testing them against national standards and the context.
• Neglecting the "Education Policy" domain on the grounds that it concerns the ministry: it is what determines the recognition of learning and sustainability.

A standard that is not verified is an intention. The difference between a compliant programme and a programme that cites the standards can be seen in the evidence: registers, minutes, visit reports, assessment results.

9Adapting the analysis to the context: sudden-onset emergency, protracted crisis, urban setting, restricted access

The analytical approach is the same everywhere, but its modalities vary greatly depending on the type of crisis, the population and access conditions. An experienced education manager adapts the timeline, the tools and the level of ambition rather than applying a single model.

Working with local partners with limited capacity

Local organisations have access, language and the trust of communities, but often little experience of structured assessments. Rather than reducing them to the role of enumerators, give them real responsibilities with support: co-facilitating team training, taking part in the analysis workshop, co-authoring the report. Adapt the tools: short forms, simple mobile data entry, an interview guide translated into the local language. Budget for their participation (staff time, transport, communication) rather than assuming it is free.

Insecure contexts

• Minimise sensitive data: no precise GPS coordinates of schools in public documents where attacks have occurred; no teachers' names.
• Protect enumerators: enumerators from the area, but not in their own village if this exposes them; routes cleared by the security manager; the option to suspend data collection.
• Data on attacks: serious incidents are passed on through the channels agreed with the actors of the Monitoring and Reporting Mechanism (MRM); the JENA is not a tool for investigating violations.

The "do no harm" principle applied to analysis

An assessment can cause harm: it raises expectations that no funding will meet, it exposes respondents, it takes teachers away from their classrooms. Practical responses: explain clearly to communities what the assessment can and cannot deliver; share the results with them; coordinate so that three organisations do not assess the same school in the same month; respect personal data protection principles (minimisation, consent, secure storage, limited retention).

Where access is limited, a cautious analysis based on secondary data and key informants, with an explicit level of confidence, is better than an ambitious survey that puts teams at risk or is never completed.

10Interview questions and key references

Recruitment for programme officer, education coordinator or cluster coordinator positions almost always includes questions on needs analysis and standards. Here are the most frequent questions and what a panel expects.

Questions and elements of a good answer

1. "How would you organise a JENA in a district where you do not yet have a team?" A good answer starts with coordination (cluster, ministry, OCHA, existing MSNA), moves on to the secondary data review and the list of gaps, proposes targeted data collection with local partners, and ends with the joint analysis workshop and integration into the HNO. It mentions ethics (consent, child safeguarding, data protection) and a realistic timeline.
2. "How do you calculate the number of children in need of education?" School-age population by group and by area; an explicit definition of need (out of school, enrolled in inadequate conditions, exposed to risks); no double counting; disaggregation by sex, age and disability; written assumptions.
3. "What are the barriers to education for displaced adolescent girls and how would you measure them?" Cost, safety of journeys, domestic work, early marriage and pregnancy, menstrual hygiene, lack of female teachers, transition to secondary school. Measurement: attendance and dropout data by age and sex, separate focus group discussions, interviews with female teachers. The candidate links each barrier to a response.
4. "How do you identify children with disabilities in an assessment?" The Child Functioning Module (UNICEF and the Washington Group), trained enumerators, questions tested in the local language, involvement of organisations of persons with disabilities. The candidate explains why a direct question about disability underestimates prevalence.
5. "Name the domains of the INEE Minimum Standards and give an example of practical application." The five domains, then a specific example: for instance, applying the standard on teachers' conditions of work through an incentive scale harmonised at cluster level and written letters of engagement.
6. "The available data are contradictory: the ministry reports 85% enrolment, your assessment 52%. What do you do?" Check the definitions (enrolment versus attendance), the dates and the populations covered (IDPs included or not), present both sources with their limitations, discuss with the ministry before any publication, and avoid public confrontation.
7. "How do you make sure your assessment does not harm communities?" Coordination to avoid duplication, managing expectations, supervised consultations with children, data protection, feeding back the results.

Key references

• INEE, Minimum Standards for Education: Preparedness, Response, Recovery, and the revised edition published by INEE.
• INEE, Guidance Note on Inclusive Education and Pocket Guide to Supporting Learners with Disabilities.
• INEE, Guidance Note on Gender (education in emergencies).
• Global Education Cluster, Joint Education Needs Assessment (JENA) toolkit and guide to coordinated education needs assessments and analysis.
• Global Education Cluster, Education Cluster Coordinator Handbook and annual guidance on PiN and severity calculation.
• OCHA and partners, Joint Intersectoral Analysis Framework (JIAF) and guidance on the Humanitarian Programme Cycle (HNO and HRP).
• Washington Group on Disability Statistics and UNICEF, Child Functioning Module.
• Sphere Association, The Sphere Handbook (2018 edition), including the WASH standards applied to schools.
• CHS Alliance, Groupe URD and the Sphere Project, Core Humanitarian Standard on Quality and Accountability (CHS).
• Alliance for Child Protection in Humanitarian Action, Minimum Standards for Child Protection in Humanitarian Action (2019 edition).
• Convention on the Rights of the Child (Articles 28 and 29) and Convention on the Rights of Persons with Disabilities (Article 24).