What “People Affected” Actually Counts — the Loosest Number in Disaster Reporting

Direct answer: “affected” has no single definition, and a large share of the figures published under that word are not counts of individuals. They are the population of the administrative areas judged to be inside the disaster zone — a district is included, so everyone living in it is included. That is a defensible way to size an emergency response quickly. It is not a measurement of how many people something happened to, and the two get quoted interchangeably.

The ladder of numbers, and what each rung is asking

A disaster report usually carries several counts. They are not degrees of the same thing; they answer different questions and are produced by different processes.

• Killed — the narrowest and the slowest to settle. It has its own article, because how a death gets attributed to a disaster is a whole subject on its own.
• Injured — usually those treated by a health facility, which means it measures access to health facilities at least as much as it measures injury.
• Homeless or destroyed housing — tied to a physical assessment of buildings, so it settles slowly and is often revised upwards as surveyors reach further out.
• Displaced — people who left. This one moves in both directions, because people return, and whether a return counts as an end depends on the definition being used.
• Affected — the widest, the fastest to appear, and the least precisely defined.

Where the affected number usually comes from

In the first days of a large emergency nobody has enumerated anything. What exists is a map: which districts are inside the flood extent, the shaking intensity contour, or the storm track. The affected figure is then commonly built by taking those districts and adding up their populations from the most recent census.

This is a reasonable thing to do. Response has to be sized before it can be delivered, and the population of an affected area is exactly the right input for asking how much water, shelter and food might be needed. The problem is only what happens next: the number leaves the response document, loses the definition it was created with, and arrives in a headline as though somebody counted.

Two consequences follow, and both are counter-intuitive:

• A dense city inflates the number and a sparse region deflates it, independently of how bad the disaster was. A moderate flood across three populous districts can produce a larger affected figure than a severe one across sparsely populated terrain.
• The figure can be large while the harm is concentrated. Everyone in an included district is counted the same, whether their house was destroyed or they simply live nearby.

Why the definition varies between organisations

Different bodies need the word to mean different things, and each definition is coherent on its own terms. A humanitarian agency sizing a response is interested in people who require assistance during the emergency period. A disaster database building a comparable historical series needs criteria it can apply consistently across decades. A government declaring a disaster area works from administrative boundaries because that is what a declaration can legally cover.

None of those is wrong, and none of them is the others. This is a definitional difference rather than a disagreement about facts — the same shape as the problem described in why two agencies report the same disaster differently.

The words that get used interchangeably and should not be

• Displaced is not homeless. Someone can be displaced by an evacuation order and return to an undamaged house the next week.
• Evacuated is a measure of the warning working, not of destruction. A large evacuation figure alongside a small casualty figure is usually a success story being read as a catastrophe.
• Affected is not in need. Needs assessments come later and are almost always a smaller number, which is why the figure can appear to fall as better information arrives.

Economic losses have the same problem, twice

Damage figures carry two ambiguities that are rarely stated. The first is whether the number is insured losses or total losses — in much of the world insurance penetration is low, so an insured-loss figure can be a small fraction of the real cost, and the two are not comparable across countries. The second is whether historical comparisons have been adjusted for inflation and for how much more there now is to destroy. A coastline with four times the buildings on it will produce a larger loss from the same storm, and that is exposure rather than severity.

How to read one of these numbers honestly

1. Ask which word it is. Affected, displaced, homeless and in need are four different questions, and the difference between them is often larger than the difference between two sources.
2. Ask who published it and for what purpose. A response-planning figure and a historical-database figure are built to different specifications and should not be compared.
3. Do not compare affected counts across events. Population density does too much of the work.
4. Expect it to move. Early figures are estimates from maps; later ones come from assessment. Movement is the process working, not a correction of a lie.
5. Treat a precise-looking figure with suspicion. An affected count given to the nearest thousand in the first 48 hours is a calculation, not a census.

What this site does with them

• We publish what the source published, with the originating agency named and the original record linked, so the definition can be checked where it was written.
• We do not combine impact figures from different sources into one number. Adding a humanitarian affected count to a government one produces something that means nothing.
• We do not compute totals across events. Summing affected counts sums four or five different definitions, and the result looks authoritative in exactly the way it should not.
• Where a figure is missing we leave it out rather than substituting a zero. Zero and unknown are not the same claim, and the distinction is set out in our methodology.

The honest summary

“Affected” is a planning figure that escaped into journalism. Read it as the size of the area an emergency covered rather than as the number of lives it changed, ask which word was actually used, and be most careful with it when it is largest — because the biggest affected numbers are usually the ones drawn from the widest map, not the ones built from the most counting.

This article explains how disaster impact figures are produced in general. This site aggregates published figures and does not produce its own; for the authoritative count, use the agency named on the event.

Frequently asked questions

What does people affected actually mean in a disaster report?

There is no single definition, and that is the central problem. Broadly it means people who required assistance during the emergency period, but in practice a large share of published figures are built by identifying the administrative districts inside the disaster zone and adding up their populations from the most recent census. That makes it a measure of the size of the area covered rather than a count of individuals to whom something happened.

Why is the affected number so much larger than the number killed?

Because they answer entirely different questions. Deaths are individually attributed and verified, which is slow and narrow. Affected is a fast, wide estimate designed to size a response, and it frequently includes everyone living in an included district regardless of what happened to them. A ratio of thousands to one between the two is normal and does not indicate that either figure is wrong.

Why do affected figures sometimes go down?

Because early numbers are estimates drawn from maps and later ones come from assessment. When teams reach an area they often find that fewer people need assistance than the district population implied. A falling figure usually means better information arrived, not that anyone exaggerated.

Can I compare affected numbers between two disasters?

Not safely. Population density does too much of the work: a moderate event across dense districts can produce a larger affected count than a severe one across sparsely populated terrain. If you need to compare severity, compare physical measurements or deaths, and read the definitions attached to each figure first.

What is the difference between displaced, homeless and evacuated?

Displaced means people left their homes, for any duration and for any reason including a precautionary order. Homeless means the housing itself is not usable, which requires a physical assessment and settles more slowly. Evacuated measures the warning system working — a large evacuation with few casualties is a success being read as a catastrophe when the numbers are quoted without their labels.

Are economic damage figures comparable across countries?

Usually not. Many published loss figures are insured losses, and insurance penetration varies enormously, so the same physical damage produces a far smaller reported loss where fewer things are insured. Historical comparisons additionally need adjusting for inflation and for how much more there now is to destroy in the same place.

Does this site add up affected figures across events?

No, deliberately. Summing them would sum several different definitions produced by different organisations for different purposes, and the total would look authoritative in precisely the way it should not. We publish what each source published, name the agency, and link the original record so the definition can be checked at the point it was written.

Sources

• CRED / EM-DAT — Disaster database entry criteria and impact definitions
• UN OCHA — Humanitarian needs assessment and people in need
• IDMC — Internal Displacement Monitoring Centre: displacement definitions and methodology
• UNDRR — Sendai Framework indicators for affected populations
• IFRC — Disaster response assessment guidance
• ReliefWeb (UN OCHA) — Situation reports and how figures are revised
• WMO — Disaster loss and damage reporting

Found an error? See our corrections policy.