# Information request and recommendation
Group: ____ Case: ____ Roles: ____

## Before fitting anything
Write the researcher's question in plain language: ____
Write the outcome, unit, period, main comparison and intended population: ____

| Question asked | Researcher's actual reply | Why this changes the analysis or interpretation | Known / unknown / assumed |
|---|---|---|---|
| Do researchers assign treatments to individual animals or whole pens, and how do they choose the assignments? | | | |
| Which observations share animals, pens, time or environment? | | | |
| What exactly is counted/measured and over which period? | | | |
| What contrast was planned before looking at outcomes? | | | |
| What happened to missing observations, removals or protocol changes? | | | |
| What magnitude of change would be useful and why? | | | |

Add two case-specific questions. Prioritise information needed to avoid a wrong conclusion; do not send a long generic list.

## Simulated response to the researcher (maximum 120 words)
Start with what can be answered. Recommend an outcome and analysis matched to allocation/grouping. Explain one design assumption and one diagnostic. Name the estimate and interval you would report in useful units. Separate association from a randomized-treatment comparison. Identify one missing detail and the consequence of proceeding without it. Avoid promising certainty.

## Auditor
Which recommendation follows from supplied evidence? Which assumption was silently invented? Which question mattered most? What would change the recommendation?