Why disaggregate?
A total hides who is missing. Imagine a food program reaches 1,000 people. The report looks fine. Split by sex and age and it may show that few older people or women-headed households came. That is a finding you can act on.
Disaggregation makes inequality visible. It lets you ask: are we reaching the people we meant to reach?
Which breakdowns should you use?
Sex, age and disability are the usual core. Add others only if they help you decide something, for example displacement status, location or household type.
| Breakdown | What it shows | Typical use |
|---|---|---|
| Sex | Differences between women and men | Almost every person-level indicator |
| Age | Differences between children, adults, older people | Health, education, protection, food |
| Disability | Whether people with disabilities are reached | Access and inclusion in all sectors |
Choose breakdowns when you write the indicator reference sheet. See how to write an indicator reference sheet. Adding a breakdown after collection is often impossible, because the data was never gathered.
How do you collect sex and age data?
- Ask at the source. Record sex and age (or date of birth) when you first register or interview the person.
- Use fixed categories. Agree the age bands before collection and use them everywhere.
- Record, do not guess. Enumerators should not decide a person’s sex or age by looking. If the person cannot say, record “not known”.
- Check for consistency. The groups should add up to the total. If they do not, find out why before you report.
Local law and context vary. Check what is safe and legal to ask in each place, and what your donor requires.
How should you ask about disability?
Avoid asking “Are you disabled?”. Many people do not use the word, and the answer varies by culture. A better method is to ask short, standard questions about difficulty doing everyday activities, such as seeing, hearing, walking, remembering, caring for oneself, and communicating. Answers on a scale from no difficulty to cannot do it at all let you group people consistently.
Use an established question set from a recognized source, and train enumerators to ask it respectfully and in private. Do not make up your own questions, because they will not compare with other data.
How do you protect people in small groups?
If you report “one woman aged 60 or more with a disability in Village B”, people may know who she is. Disaggregation can expose individuals when the numbers are small.
- Combine categories. Merge age bands, or report by district instead of village.
- Suppress small cells. Set a minimum, such as not showing any cell under a number you choose, and apply it the same way every time.
- Control who sees the detail. Keep the raw, person-level data to a small group of staff.
Our article on keeping a beneficiary registry that protects people covers the person-level records behind these figures.
What mistakes are common?
- Collecting too many breakdowns. Long forms tire people and lower the quality.
- Different age bands across projects. Results cannot be compared or combined.
- Counting people twice. If the same person appears in several records, every group total is inflated. See why reach totals may count the same person twice.
- Treating the breakdown as a target. A gap is a signal to investigate, not a number to adjust.
Here is a simple example of reading disaggregated data. A project reports that 600 people joined a livelihood training. Split by sex, the numbers are 420 men and 180 women. Split by age, very few are over 60. Split by disability, almost nobody reports difficulty with walking.
The total of 600 hid three questions worth asking:
- Is the training held at a time or place that women find hard to reach?
- Do older people know about it, and can they get there?
- Is the venue accessible to someone who cannot climb stairs?
Each question leads to an action, such as a second session time, a home visit to share information, or a ground-floor venue. Then you measure again next period. That is the real value of disaggregation: it gives you something to change.
Set targets for the groups that matter, not only the total. A target of 50 percent women is easier to manage than a general wish to reach women.
Where EMPOBase fits
EMPOBase’s indicator library includes disaggregation guidance for its entries, which you can adapt to your donor’s requirements. The Beneficiary registry can hold the person-level records your breakdowns come from. EMPOBase totals are counts of records, not unique people, and nothing deduplicates beneficiaries, so check for repeats before you report any group total. Choose categories and the questions you ask yourself, and check local rules. See the indicators page for the library.