Review Vulnerable Population Protections

You are an expert in research ethics and safeguarding for evaluations involving vulnerable populations reviewing the protections in a deliverable. The protections section may be embedded in a methodology, ethics chapter, inception report, or standalone safeguarding plan. The goal is to assess whether protections are operational rather than rhetorical. **VULNERABLE-POPULATION PROTECTIONS SECTION TO REVIEW:** [paste the protections section here] **Review Requirements:** 1. **Population identification.** Assess whether the specific vulnerable populations involved are named with their distinct risks (children, refugees, GBV survivors, people with disabilities, etc.) rather than grouped under a generic 'vulnerable' label. 2. **Specialized consent.** Check whether consent and assent processes are adapted for each population, including parental consent plus child assent, low-literacy adaptations, sign language, plain-language alternatives, and language matching. 3. **Trained personnel.** Verify whether personnel skills and training requirements are specified for working with each population, including child safeguarding, trauma-informed practice, sign language proficiency, or matched-gender enumerators. 4. **Referral pathways.** Assess whether referral pathways are named for cases requiring follow-up (GBV disclosure, child protection concerns, suicidality, urgent health or legal needs), with named services, contact procedures, and handover responsibilities. 5. **Data protection.** Confirm whether heightened data protection measures are specified for sensitive populations, including stricter access controls, separated identifiers, secure transmission, retention limits, and protections against re-identification. **Output Format:** Produce: 1. A 1-paragraph overall assessment of whether protections are operational or rhetorical. 2. A scored review table: dimension, score (1-5), evidence from the document, recommended action. 3. A prioritized revision list (must-fix vs. should-fix), with must-fix items flagged for resolution before any data collection from vulnerable populations begins. 4. A short note on the single highest-risk gap that would most likely cause harm if data collection began as designed.