# Illustrative output: baseline measurement plan

**Fictional example** based on `input.md`. It is a plan, not a baseline dataset or runtime result.

## Scope and sources
- Intake: Uploaded fictional program summary.
- Framework: version 2.0, August 2026. [Evidence: `input.md#Framework`]
- Exposure: first training scheduled for October 2026; no exposure by September 2026. [Evidence: `input.md#Program`]
- Prior evidence: 2024 North district agriculture survey; method, frame, season, disaggregation, and permission unknown. [Evidence: `input.md#Existing evidence`]

## Facts, assumptions, and proposals
- **Facts:** O-1 and OC-1 wording as supplied; training start is planned for October; OC-1 target is 60% by December 2027.
- **Assumptions:** None added to create values or sample sizes.
- **Proposals:** Establish source and indicator definitions before the planned October start. Collect pre-exposure measures only where the construct is meaningful before training; an output completion count accrues during implementation and the trained-cohort outcome needs a separate starting-condition definition. The schedule's approval status is not supplied. Any adjustment remains a proposal.

## Indicator coverage matrix

| ID / wording | Level | Definition/version | Population / unit | Measurement window | Source/tool and method | Timing vs. exposure | Disaggregation | Access / permission | Prior and later-round comparability | Status |
|---|---|---|---|---|---|---|---|---|---|---|
| O-1 “Number of farmers completing training.” | Output, as labeled in the framework | Completion rule not supplied | Farmers in North and East districts stated for program; eligibility and unit/deduplication rule need confirmation | Not supplied | No source or tool supplied | This output accrues during implementation. Confirm the count's start/reset window and whether records support zero at inception; do not survey future completions before exposure | Not supplied | No collection permission supplied | Must define the completion rule and collection process for later rounds | **Decision required; no value** |
| OC-1 “Percentage of trained farmers using at least two soil-cover methods within 12 months of training.” | Outcome, based on supplied label | Denominator and survey rule not supplied | “Trained farmers”; cohort/eligibility need definition | Within 12 months of training, as stated; baseline reference window not otherwise defined | 2024 survey is candidate evidence only for North district; tool/frame/method unknown | A post-training 12-month measure is not pre-exposure baseline. Establish a pre-training measure only if the construct can validly be measured before training; otherwise document a different starting measure or no baseline for this construct | Not supplied | Prior survey permission and new collection approvals unconfirmed | Need same construct, population, season/reference window, tool/mode, and disaggregation for later comparison | **Candidate prior source, not accepted as baseline; no value** |

## Prior evidence assessment
The 2024 survey is not yet comparable to the planned two-district population. Its definition, season, frame, measurement method, disaggregation, and permission are unknown, and East district is not covered in the supplied note. Do not use a numeric value until the source and underlying data are reviewed and authorized.

## Sampling and design decisions
Sample size cannot be calculated from the supplied information. Confirm the eligible population and frame, intended estimate, design, precision or power requirements, expected response, and any qualitative component. Determine whether district or other subgroup estimates are required before selecting a design.

## Timing, quality, ethics, and permissions
- Confirm actual first-exposure date, baseline collection window, and season.
- Identify tools and maintain versioned definitions/reference periods for later rounds.
- Confirm data source permissions, ethics review, consent, safeguarding, data protection, and access roles before collection.
- Specify piloting, training, verification, completeness checks, and handling of missing values after tools and sources are selected.

## Decisions needed
Preserve O-1's supplied Output label and confirm its completion/counting rule; define OC-1 denominator and baseline construct/window; verify whether 2024 survey data can be accessed and compared; set population/frame, sampling assumptions, disaggregation, tools, permissions, responsible roles, effort and resources. No baseline value or target has been created or revised.