SIMULATED FEEDBACK EVIDENCE WORKSHEET
BLANK TEMPLATE — contains no persona, participant, model run, study, or result

Question:
System under test:
System revision:
Record date:
Reviewer:

INPUT COVERAGE
Input ID:
Exact hand-authored or generated input:
Input provenance:
Why this case belongs in the set:

SYSTEM BEHAVIOR
Observation method:
Deterministic output or event:
Expected behavior:
Observed difference:
Reproduction record:

SIMULATED INTERPRETATION — COMPLETE ONLY IF A MODEL ACTUALLY RAN
Run ID:
Model and immutable version:
Prompt/template version:
Persona definition provenance:
Generated interpretation:
Known sampling and model limits:

REAL-USER EVIDENCE — COMPLETE ONLY FOR AN APPROPRIATELY CONSENTED STUDY
Study ID:
Consent/ethics record:
Participant sampling boundary:
Observed participant behavior:
Participant-reported feedback:
Measurement limitations:

CONCLUSION
Supported conclusion:
Unsupported claims to avoid:
Next evidence needed:

Do not copy generated interpretation into the real-user evidence section.
Synthetic frequency is not population prevalence.