You are a senior MEAL specialist with expertise in field data collection management. Your task is to create a comprehensive Field Supervision Checklist and Data Quality Spot-Check Protocol for "data collection exercise" (from the material, if given).
Context:
- Data collection exercise: "name" (from the material, if given)
- Number of data collection teams: "number" (from the material, if given)
- Data collection method: "method" (from the material, if given)
- Duration: "duration" (from the material, if given)
- Supervisor-to-enumerator ratio: "ratio" (from the material, if given)
Deliverables:
1. Daily Supervision Checklist Design a checklist that supervisors complete each day. Organize by timing:
Before fieldwork (morning):
- Team briefing conducted (attendance, daily targets, route plan)
- All tablets charged and functional
- Consent forms and paper backups available
- Enumerator appearance and identification (ID badges, branded vests)
- Safety check (security situation, communication devices working)
During fieldwork (observation):
- Informed consent properly administered (observed at least once per enumerator per day)
- Correct respondent selected per sampling protocol
- Interview conducted in appropriate language
- Appropriate probing techniques used
- Sensitive questions handled with care and privacy ensured
- Skip logic followed correctly
- Interview duration within expected range (flag interviews under 50% or over 150% of average)
- No leading questions or answer suggestion observed
- GPS coordinates captured at correct location
After fieldwork (evening):
- All forms submitted/synced to server
- Daily data review completed (completeness, outliers, duplicates)
- Replacement interviews documented if needed
- Enumerator performance discussed individually
- Incident log updated
- Daily summary report submitted to MEAL coordinator
Format as a printable checklist with Yes/No/NA columns and a comments field.
2. Data Quality Spot-Check Protocol Design a systematic spot-check process:
Back-check interviews (verification visits):
- Sample: revisit 5-10% of completed interviews within 48 hours
- Selection: random selection from each enumerator's completed forms
- Verification questions: 5-8 key factual questions that should not change between visits (e.g., household size, roof material, number of school-age children)
- Acceptable discrepancy threshold: less than 10% disagreement rate per enumerator
- Escalation: if discrepancy exceeds 15%, flag all of that enumerator's data for review
Direct observation scoring: Create a standardized observation rubric for supervisors:
| Competency | Score 1 (Poor) | Score 2 (Adequate) | Score 3 (Good) | Score 4 (Excellent) |
|---|
Competencies to score:
- Consent administration
- Rapport building
- Question reading accuracy
- Probing technique
- Skip logic compliance
- Sensitive question handling
- Data entry accuracy
- Time management
Minimum acceptable score: 2.5 average across all competencies.
Automated data checks (server-side):
- Forms per enumerator per day (flag if below minimum or above maximum)
- Interview start/end times (flag implausible durations)
- GPS clustering (flag if all interviews from same coordinates)
- Duplicate detection
- Outlier detection for numeric fields
- Completeness rates by enumerator
3. Performance Dashboard Template Design a simple daily tracking table:
| Enumerator | Forms Today | Cumulative | Avg Duration | Completeness % | Spot-Check Score | Issues |
|---|
4. Corrective Action Protocols Define escalation procedures:
- Level 1 (coaching): minor issues, resolved through on-the-spot feedback
- Level 2 (formal warning): repeated errors, documented in writing
- Level 3 (retraining): systematic quality issues, half-day refresher required
- Level 4 (removal): fabrication, ethical violations, or persistent poor quality after retraining
5. Incident Reporting Template Provide a template for documenting field incidents: security events, respondent complaints, equipment failures, sampling protocol deviations, and safeguarding concerns.
Where the material is silent, write "Not in the material" rather than guessing.
Source material: Attached, or pasted below. Base everything above on it and do not invent details that are not in it.