Help us improve age disaggregated data: Technical discussion paper

7 min readApr 11, 2016

Your opportunity to take part in this process ends on 25 April 2016

This article is part of a series on age disaggregated data. Take a look and let us know what you think:

To make sure everyone benefits from the Global Goals we need to improve the quality, consistency and use of age disaggregated data. We need your help!

What is the problem?

Collectively across the globe we lack data across the full age range of people’s lives, as has been outlined in the Aging and Youth papers (links to the other open policy making papers). We are currently unable to ensure every person counts, and is counted, from the cradle to the grave.

Why the gap?

We lack basic data on the structure of the population. Civil Registration and Vital Statistics systems which can provide this data are too expensive for most countries and therefore not comprehensive. Population and housing censuses are also costly and at best only undertaken every ten years, with potential for large population changes to occur between them and not be captured — such as migration, conflict and population growth.

This all means policy making and targeting of vital services is difficult, and likely to be done based on incomplete evidence — often missing certain age groups.

A case study from Tanzania: In a 2009 Insecticide Treated Nets roll out to all children under five, with no functioning Civil Registration and Vital Statistics system it was hard to know how many nets would be needed. Population projections from the National Bureau of Statistics, based on the previous (2001) census were used. During the pilot significantly more children came to collect nets than were anticipated (42%) — some of this was no doubt due to older children or those from different districts wanting to get a free net, but work following this discovery on the projections noted that many of the assumptions based on census data were outdated — including child mortality rates which had significantly fallen in the eight years since the census.

Surveys and administrative data provide the detailed information on the lives of the people. We think that the following are the two major challenges faced in getting good data on people throughout their life in developing countries:

Key international household demographic surveys do not cover all age groups. Both the Demographic and Health Survey (DHS) and Multiple Indicator Cluster Survey (MICS) only collect individual information on those between the ages of 15 and 49.
Due to the small numbers in some age groups (particularly older people in developing countries) statistical precision or representative samples will not happen by chance for these age groups. This means that surveys will not be able to tell a representative story of these smaller age groups. This is similar to the challenges faced trying to get representative samples of any particular subgroup (e.g. those with disabilities or ethnic groups), and is further exacerbated by the intersectionality of issues. For example — disabled old people are an even smaller proportion of the population.

The population pyramid below shows Nigeria’s estimated population in 2016. Adding the age cut offs for key international surveys shows how much of the population will be missed that sit outside of the data collection of these surveys. This includes the large youth cohort below 15 years old that will be the workers, adults and families of the future, and the smaller proportion of people in the older age groups of 50 years and above compared with the rest of the population. Both these parts of the population will have their own particular planning needs.

U.S. Census Bureau-International Programs-International Data Base-9 July 2015 Version: Data:12.0625 Code:12.0321

So what can be done to address these gaps?

There are a number of things that can be done to address the problem of lack of data across the full age range of people’s lives and the fact we are currently unable to ensure every person counts, and is counted, from the cradle to the grave. Some suggestions include:

Support Civil Registration and Vital Statistics systems in all countries — to ensure that the basic facts about age and sex structure of the population are known (particularly focusing on birth and death registration, and migration). These are costly and it will take some time to get nation-wide functioning systems in place.
Population and Housing Censuses, even in developed countries, are important to gather additional data on a population and its living conditions. They aim to provide detailed information about all members of a household, all households in a community and all communities in a country — information that is used for planning and delivery of services.
Intra census population estimates are a must in all countries — ideally these would be based on Civil Registration and Vital Statistics or similar, but even in the absence of Civil Registration and Vital Statistics systems, regular updates of the assumptions used to develop the population estimates, and sensitivity testing should be undertaken.
Administrative systems should capture date of birth (or age and the date at which that age was reported) so single year age can be calculated and used in analysis of the data to establish differences required for different age cohorts.
All household surveys that ask individual questionnaires as part of the household survey should expand the age range of respondents beyond 15–49. Ideally to all those that are aged 10 and above. There will be cost and logistical implications.
To aid with understanding life experiences throughout a person’s life, more longitudinal/cohort surveys should be considered.
Support undertaking of specific surveys or booster samples in national surveys to ensure that specific sub populations are effectively captured. This includes specific age groups as well as other subgroups such as those with disabilities, or particular ethnic or migratory groups.
Collection and storage of age data should be by individual year to enable comprehensive disaggregation and facilitate any type of aggregations possible. Reporting of age should consider the needs of users, and for younger age bands reflect relevant and different groupings, and for older age groups be in five year cohorts.

These are the questions we need your help with

What experiences do you have with expanding household surveys to cover all age ranges?
What are the pros and cons of expanding data collection to all age ranges?
What experience do you have of surveys which have deliberately boosted samples for specific age groups or surveys designed to target certain specific age groups? What are the pros and cons of this approach?
Other than providing funds — how can partners be encouraged to expand household survey age ranges?

What happens next and how do I get involved?

One of our main goals is to see this problem from a range of different perspectives so that we can understand the impact that it is having on people making policy, statisticians, advocates and real life people like you and me. The good news is that we have already started. We have been reaching out to experts, academics and partners to help us to scope this work. They have helped us to tell the story of this problem from a range of different perspectives. Now it’s your turn! There are two ways for you to feed in to this process:

Want to know more about Open Policy Making?

The Policy Lab in the Cabinet Office have created lots of useful tools, guidance, case studies, tips and advice to help you take your first step toward a more open, collaborative and better policy making process. Pay them a visit over here. If you’re just getting started have a read at the Policy Lab’s introduction to design in policy making for more information on how to apply design thinking and policy making.
To follow this Open Policy Making process — follow DFID Inclusive Societies on Medium. We’re on Twitter too!