Children facing crises

1Risks for children in crises

In emergencies, children often make up half of the affected population. The Minimum Standards for Child Protection in Humanitarian Action (CPMS) describe the main risks:

• family separation (unaccompanied and separated children);
• physical and emotional violence and harmful practices (child marriage, female genital mutilation);
• sexual and gender-based violence;
• child labour, including its worst forms;
• association with armed forces and armed groups;
• psychosocial distress and mental health disorders;
• dangers and injuries (landmines, explosive ordnance, accidents).

2The framework: children's rights and the Minimum Standards

The Convention on the Rights of the Child (CRC, 1989) protects every person under 18. Its principles guide every decision: best interests of the child, non-discrimination, the right to life, survival and development, and the right to be heard. International humanitarian law also gives children special protection, and the Paris Principles govern the prevention of recruitment and the reintegration of children associated with armed forces and armed groups.

The CPMS (Alliance for Child Protection in Humanitarian Action, 2019) set out 28 standards: quality of the response, risks, prevention and response strategies, and mainstreaming child protection in other sectors.

3Principles of action: do no harm and participation

Every action with children follows a few rules:

• Do no harm: do not expose the child, do not multiply interviews, do not promise what you cannot deliver.
• Confidentiality: share only what is necessary, with the right people, with the consent of the child or parent whenever possible.
• Participation: listen to children and take their views into account according to their age and maturity.
• Strengthen families and communities, the first protectors of children.

Never publish a photo or story that could identify a child survivor.

4Child safeguarding within organisations

Organisations themselves can pose risks. Child safeguarding requires: a signed code of conduct, reference checks during recruitment, a ban on being alone with a child outside the planned setting, rules for photos and social media, and a safe reporting mechanism.

Any suspicion of abuse by a staff member or partner is reported immediately, following the organisation's procedure, without investigating yourself and while protecting the child.

5Analysing the child protection situation step by step

A child protection response is only as good as the analysis that precedes it. In the first weeks of a crisis, the classic mistake is to launch a heavy field "assessment" with individual interviews of children, when most of the information already exists elsewhere and sensitive questions cannot be asked unless there is a service to refer to. The method recommended by the Child Protection Area of Responsibility (CP AoR) and by CPMS Standard 4 (programme cycle management) proceeds in stages and combines a secondary data review, light primary data collection and joint analysis.

Step 1: the secondary data review (days 1 to 10)

The secondary data review (SDR) brings together what is already known: pre-crisis reports (Demographic and Health Surveys, MICS surveys, reports of the Committee on the Rights of the Child, national studies on child labour or child marriage), data from the national social protection system, agency and NGO reports, multi-sector assessments, media and local networks. It is conducted by one or two analysts, preferably within the coordination group, and produces within ten days at most a 6- to 10-page note structured by risk (the seven risks of CPMS Standards 7 to 13), setting out for each risk what is known, the level of confidence and the information gaps.

Step 2: define the questions and methods for primary data collection (weeks 2 to 3)

Primary data collection covers only the gaps identified by the SDR. The preferred methods are key informant interviews (teachers, health workers, religious leaders, women leaders, site managers), focus group discussions with adults and adolescents, direct observation and, where the context allows, an inter-sectoral household survey into which a small number of non-sensitive child protection questions are inserted. An informant or a group is never asked to recount an incident involving an identifiable child; questions focus on trends ("In this community, do children work outside the household? Which children? Where?").

• Key informants: 3 to 5 per location, with diverse profiles, 30 to 45 minutes, a questionnaire of 20 to 30 questions at most.
• Focus group discussions: 6 to 10 participants, separated by sex and age group (adolescents aged 13 to 17 in groups separate from adults), 60 to 90 minutes, with a facilitator and note-taker trained in how to respond to a disclosure.
• Observation: a standardised checklist (lighting, latrines, water points, children alone, children working, presence of armed men) completed at several times of the day.

Step 3: analyse jointly (week 4)

The joint analysis brings together members of the coordination group and, where possible, the Education, GBV, Health and WASH sectors. For each risk, the analysis describes the threat (who or what causes the harm), the vulnerability (which children, and why), the capacities (families, communities, services) and the likely evolution. A severity level is then assigned by geographical area, using the scale applied in the Humanitarian Needs Overview. The final product is a 15- to 20-page report with recommendations addressed by name to each sector and to the authorities.

Non-negotiable ethical rule: no information is collected on a serious risk (sexual violence, recruitment, exploitation) unless the team is able to refer anyone who discloses being a victim to a functioning service. The data collection team goes out with the service mapping and the procedure for responding to disclosures.

Step 4: update

In a protracted crisis, the analysis is revised at least once a year for the Humanitarian Needs Overview cycle, and at every major change (new displacement, epidemic, offensive). Between two analyses, aggregated case management data, protection monitoring and feedback from children make it possible to track trends.

6Using the CPMS as a working tool

The Minimum Standards for Child Protection in Humanitarian Action (2019 edition) are not a reference text to be quoted in a proposal: they give concrete structure to the design, monitoring and evaluation of a response. Each standard comprises a standard statement, key actions (preparedness and response), measurements (indicators with an indicative target) and guidance notes. The handbook complements the Sphere Handbook, of which it is a companion standard, in the same way as the INEE standards for education.

The four pillars and the 28 standards

The cross-cutting principles

The CPMS take up the four general principles of the CRC (survival and development, non-discrimination, child participation, best interests) and add operational principles: safety, dignity and rights (the Sphere Protection Principles), strengthening child protection systems, strengthening local capacity, the socio-ecological approach and accountability to children. For a project manager, these translate into concrete questions at the design stage: does the activity strengthen or bypass the State's social services? Do local actors come out of it more capable? Have children been consulted?

How to use a standard in practice

1. Select the relevant standards on the basis of the situation analysis: generally three or four priority risks from Pillar 2, two or three strategies from Pillar 3, and the Pillar 4 standards corresponding to the sectors present.
2. Contextualise: the coordination group adapts the key actions and indicative targets to the context, in a short document endorsed by its members. A 100% target may become 80% in the first six months, with a justification.
3. Integrate into tools: the CPMS measurements are carried over into the logframe, the monitoring plan and reports. The indicators of the humanitarian response framework and of the AoR response plan are derived from the CPMS.
4. Review: a compliance review against the standards is carried out at mid-term, standard by standard, with a simple code (met, partially met, not met, not applicable) and an improvement plan.

Example of contextualisation

An evaluator does not ask whether you "apply the CPMS": they ask which standards you selected, why, how you contextualised them and what your documented level of compliance is. Keeping an up-to-date compliance table answers all three questions.

7Worked example: the rapid situation analysis in Kéran

The following case is fictional. It shows the full sequence of a child protection situation analysis, from the trigger to the setting of priorities.

The context

Following armed clashes in Doraliane province, some 42,000 people fled to the town of Kéran and its surroundings within three weeks. They are spread between a planned site (Kéran-Est, 14,000 people), four spontaneous sites (11,000 people) and host families in town (about 17,000 people). Site profiles indicate that 54% are under 18, i.e. about 22,700 children, of whom 19% are under 5. Before the crisis, the national survey indicated that 31% of women aged 20 to 24 in the province had been married before age 18 and 9% before age 15. The provincial social affairs department has 11 social workers for the entire province.

The timeline followed

The main findings

• Separation: 312 unaccompanied and separated children registered at the transit point; 41 unaccompanied, including 9 under the age of 5. 70% of separations occurred during the night-time flight; 18% are voluntary placements with relatives who remained in town.
• Child labour: in 15 of the 18 focus group discussions, participants mention boys aged 10 to 15 carrying loads at the market and girls aged 12 to 16 employed as domestic workers in host families.
• Child marriage: 11 of the 46 informants report knowing of marriages "to protect girls" since arrival; the trend is perceived to be rising.
• Recruitment: two informants report the presence of recruiters near the Bel-Air spontaneous site; the information is unverified and is to be handled through the channels of the monitoring and reporting mechanism.
• Distress: teachers report sleep disturbances, aggressiveness and refusal to leave their parents among many young children.
• Services: no child-friendly space in the spontaneous sites; only one health facility offering clinical management of rape.

The analysis matrix

The reasoning

The sub-group ranked separation as priority 1 for two reasons: the potential harm is immediate (young unaccompanied children) and the chances of reunification decline rapidly with time and the dispersal of families. Child labour and child marriage share economic drivers: they were addressed together, with a recommendation to the cash working group to include child protection-related vulnerability criteria. Recruitment was not ranked for lack of verified data, but enhanced monitoring was put in place and a confidential alert was transmitted through the designated channels. The report made 14 recommendations, 9 of them addressed to other sectors (for example: lighting for 40 latrines in Kéran-Est, priority registration of the spontaneous sites, inclusion of 12 child protection questions in the next multi-sector assessment).

What the example illustrates: figures on scale are presented with their estimation method and confidence level. A report that states "600 separated children" without explaining where the figure comes from loses its credibility at a donor's first question.

8Involving children: methods, consent and ethics

The child's right to be heard (Article 12 of the CRC) is a CPMS principle, but it often translates into tokenistic consultation or, conversely, into intrusive data collection. Quality participation is voluntary, age-appropriate, safe and inclusive, and it has visible consequences for the programme. The nine basic requirements for ethical and meaningful participation, set out by the Committee on the Rights of the Child in its General Comment No. 12 (2009), serve as a checklist: transparent and informative, voluntary, respectful, relevant, child-friendly, inclusive, supported by training, safe and sensitive to risk, and accountable.

Consent and assent

This table follows the logic of the Inter-Agency Guidelines for Case Management and Child Protection and of the guidelines Caring for Child Survivors of Sexual Abuse (IRC and UNICEF, 2012). Consent is informed when the child understands what will happen, who will have access to the information, what the limits of confidentiality are, and that they can refuse or stop without any consequence for the assistance they receive.

Age-appropriate methods

• Ages 5 to 9: drawings ("draw a place where you feel safe and a place where you are afraid"), simple body mapping, role play with puppets, sessions of 30 to 40 minutes, groups of 5 to 8 children.
• Ages 10 to 13: participatory risk mapping of the site, pairwise ranking of concerns, story completion with a fictional character, sessions of 45 to 60 minutes.
• Ages 14 to 17: focus group discussions separated by sex, accompanied "safety walks", adolescent committees consulted on activity design, short self-administered surveys with no questions about personal incidents.

The fictional character technique ("Imagine a girl your age who lives here. What bad things could happen to her? What can she do?") makes it possible to talk about risks without leading the child to tell their own story in front of the group.

Organising a consultation session

1. Prepare: a precise objective, decision-oriented questions, parental agreement, a venue that is visible but out of earshot, at least two facilitators (including one of the same sex as the participants), the disclosure procedure and the service mapping.
2. Open: introductions, explanation of the objective, group rules (no personal stories, respect, the right to stay silent), assent of each child.
3. Facilitate: main activity, paraphrasing, attention to children who withdraw.
4. Close: a calming activity, a reminder of whom to contact if needed, thanks without financial compensation (a snack is acceptable).
5. Feed back: report back to the children within four to six weeks on what has changed thanks to their contributions. This is the step most often forgotten.

If a child discloses abuse during a group activity, the facilitator gently interrupts ("What you are saying is important; let's talk about it together just afterwards"), continues the activity, then applies the procedure: listen, no detailed questions, no promise of absolute secrecy, immediate referral to a caseworker.

Including the least visible children

Consultations naturally reach children in school, boys and children in planned sites. Active efforts are needed to reach married adolescent girls, children with disabilities (using the Washington Group questions adapted for children, the Child Functioning Module developed with UNICEF), working children (sessions at compatible times, at the workplace if it is safe), children living on the street and minority children. A tracking table of participant diversity (sex, age, disability, status) makes it possible to show evaluators who has been heard and who is missing.

9Setting up an operational safeguarding system

Child safeguarding is not just a signed policy. Donors and auditors now check that it works: the UK FCDO conducts safeguarding due diligence before funding, the European Commission checks policies on protection from sexual exploitation and abuse under its agreements, USAID's BHA includes clauses on protection from sexual exploitation and abuse and on child safeguarding in its awards, and United Nations agencies assess their partners' PSEA capacity under the common protocol adopted in 2018 and its assessment tools. The Core Humanitarian Standard (CHS) makes the prevention of sexual exploitation, abuse and harassment a cross-cutting requirement.

The four pillars of the Keeping Children Safe standards

Safer recruitment, step by step

1. A job advertisement stating the organisation's commitment to safeguarding and that checks will be carried out.
2. A mandatory safeguarding question at interview (for example: "A child asks you for your personal phone number. What do you do?").
3. At least two references, including the most recent employer, contacted by telephone with an explicit question about any concerns regarding behaviour towards children or vulnerable people.
4. For member organisations, a check through the inter-agency Misconduct Disclosure Scheme, which shows whether a candidate has been sanctioned for sexual exploitation, abuse or harassment.
5. A criminal record check where national law allows it and the system is reliable.
6. Signature of the code of conduct and training before any contact with children, not "within the first three months".

Assessing the risks of an activity

Risks scored 6 or more (likelihood multiplied by impact) require a documented mitigation measure before the activity starts.

Responding to a report

1. Within the hour: the child's safety, medical care if needed, referral to a caseworker or a GBV service, report to the safeguarding focal point.
2. Within 24 hours: notification of country management and headquarters according to the procedure, risk assessment for the child and witnesses, precautionary measures (temporary removal of the person concerned without prejudging the outcome).
3. As per agreements: notification of the donor within the contractual deadlines, often 24 to 72 hours for a credible allegation, with non-identifying information.
4. Investigation: by trained investigators independent of the line management concerned, using a survivor-centred approach; the child is interviewed again only if essential and by a qualified person.
5. Closure: documented decision, follow-up of the child until their own case is closed, root cause analysis and improvement of procedures.

Auditors do not treat the number of reports as an indicator of failure. An organisation of 300 people working with children that has received no reports in three years probably has a mechanism nobody uses.

10Common mistakes, contextual variations, interview questions and references

Common mistakes and how to avoid them

Variations by context

• Sudden-onset crisis: the SDR is limited to 72 hours and primary data collection to a few sentinel sites; low confidence is accepted and a revision is planned at 6 weeks. The immediate priorities are almost always preventing and responding to separation and integrating protection messages into distributions.
• Protracted crisis: the analysis draws on trends in case management data and on representative surveys; the emphasis is on strengthening the national system (government social workers, legal framework, justice for children).
• Camp: risks can be located (water points, latrines, distributions); coordination with camp management is central.
• Urban setting: dispersed population, children invisible in host families, domestic work and street life; the response relies on schools, health centres, municipal structures and neighbourhood associations.
• Remote management and insecurity: telephone data collection limited to non-sensitive questions, networks of trained and protected community focal points, no identifying data transmitted via messaging apps.
• Local partners with limited capacity: safeguarding is built with them (template policy, training, monthly accompaniment) rather than demanded as an impossible precondition; a capacity-strengthening plan with deadlines is annexed to the partnership agreement.

What donors and evaluators look for

• A recent, joint and sourced situation analysis, and an explicit link between priority risks and the proposed activities.
• Reference to the CPMS and to the indicators of the AoR response plan.
• The safeguarding policy, the actual training of staff and partners, and a complaints mechanism accessible to children.
• Documented participation of children in design and monitoring.
• Disaggregation of all data by sex, age and disability.

Interview questions

Key references

• Alliance for Child Protection in Humanitarian Action, Minimum Standards for Child Protection in Humanitarian Action, 2019 edition.
• United Nations, Convention on the Rights of the Child, 1989, and its Optional Protocols; Committee on the Rights of the Child, General Comment No. 12 on the right of the child to be heard, 2009.
• Child Protection AoR and the Alliance, child protection needs and situation analysis tools, including the guidance on secondary data review.
• Sphere Project, The Sphere Handbook, 2018 edition, Protection Principles.
• CHS Alliance, Groupe URD and the Sphere Project, Core Humanitarian Standard on Quality and Accountability, 2024 edition.
• Keeping Children Safe, International Child Safeguarding Standards.
• IASC, Six Core Principles Relating to Sexual Exploitation and Abuse, revised 2019; United Nations, Protocol on Allegations of Sexual Exploitation and Abuse Involving Implementing Partners, 2018.
• SCHR, Inter-Agency Misconduct Disclosure Scheme.
• Washington Group and UNICEF, Child Functioning Module.