Building a District Food Security Evidence Brief in Ethiopia

A good harvest report cannot tell us whether every household can afford food. A market price series may show a rise without revealing which families have lost income or access to markets. For food security research in Ethiopia, we favour a district evidence brief that brings these different observations together and makes their limits visible. It must never disguise uncertain estimates as a confirmed assessment of hunger in a specific woreda.

Define the place, population and decision

Start with a boundary, a period and a decision. Name the woreda or other administrative area, identify the months covered and describe the groups whose circumstances the brief can reasonably represent. A household survey from one cluster should not silently stand for an entire district.

The decision could concern the timing of a market assessment, the need to investigate declining purchasing power or the quality of an early warning signal. State who will use the brief and what action the available evidence could inform. This keeps the brief focused.

Separate a description of conditions during the observation period from a forecast. A current price, an earlier household survey and a projection for the next growing season belong in different columns. Record dates before comparing them.

Triangulate agriculture, markets and households

We examine three connected kinds of information. The first describes the agricultural season: planting conditions, available production estimates and disruptions affecting livelihoods. The second concerns access, including staple prices, wages, transport interruptions and the ability to purchase food. The third covers household experience, using appropriately collected indicators of consumption, livelihood pressures and vulnerability.

None of these sources tells the entire story. A fall in an estimated harvest does not measure the food available to a household that buys most of its supplies. Stable market prices may coexist with reduced earnings. Report both observations rather than forcing them into one conclusion.

A rapid market check may cover trading centres while omitting remote settlements. Agricultural monitoring may have a different geographic boundary from a household survey. If periods or populations do not align, the brief should say so in plain language.

Our food-security research connects questions about agricultural production with households’ ability to obtain food. For a district brief, that means recording the source of every statement, including information that challenges an initial interpretation.

Label what the data can and cannot say

The World Bank’s Ethiopia Joint Food Security Monitor combines public data, transformed indicators and modelling to identify risks of deterioration. Its metadata listed version 2026-10-19 when we prepared this framework. That version matters: the series changes as new evidence becomes available. A modelled alert is a signal to investigate, not a household census or a verified count of people experiencing hunger.

The IPC acute food insecurity classification follows a defined process of evidence review, technical analysis, consensus and quality assurance. Its five phases have specified meanings. Our district worksheet cannot assign an IPC phase from a price change, one survey statistic or a private scoring rule. It can identify evidence that a qualified assessment team may need.

Likewise, an indicator of food insecurity is different from a measure of acute malnutrition. Specify what each variable captures, its sampling method and whether it describes people, households or an area.

Use a reproducible one-page evidence brief

Before filling it with data, choose one reporting period and insist on source names, version dates and clear coverage notes. The examples are placeholders for fields to populate, not findings about any Ethiopian district.

Indicator	Geography	Observation period	Source and version	Coverage gap	Policy question
Estimated crop output	Named woreda	Main season and year	Crop assessment; release date	Non-sampled fields	What needs verification?
Staple market price	Named market sites	Weeks or months	Market series; update date	Remote markets absent	Are prices changing unevenly?
Household purchasing power	Surveyed communities	Survey reference dates	Income or wage data; wave	Informal income missing	Who may face reduced access?
Food consumption indicator	Survey sample	Recall period	Household survey; methodology	Sampling uncertainty	What does the sample support?
Livelihood disruption	Affected locations	Event window	Field report; publication date	Incomplete reporting	Which changes need checking?
Early warning signal	Monitor reporting area	Forecast horizon	Model output; dataset version	Model and scale limits	Does a new assessment make sense?

If a survey excludes some households, state the exclusion where the reader will see it.

End with a transparent uncertainty note

We recommend closing the brief with three short statements: Known identifies observations supported by dated sources; Uncertain records disagreement, gaps or weak coverage; Verify next specifies the evidence and responsible team needed to resolve the question. This keeps conclusions usable.

Protect households while doing the work. Aggregate sensitive responses, remove direct identifiers and consider whether publishing small-area details could expose participants. Record when each indicator requires refresh, especially after a seasonal change or a disruption to markets.

Keep the brief ready for revision

The value of a district evidence brief lies in traceable reasoning. At EEA, we want readers to see what the data support, where uncertainty remains and what additional inquiry could change the interpretation. When newer evidence arrives, the team should be able to update the table and explain precisely why its conclusion changed.