Help us improve age disaggregated data

This article is the concluding part of a series on age disaggregated data.

A huge THANK YOU! to all of you who participated in DFID Inclusive Societies experiment in open policy making. We posted four technical blogs in April and asked for your feedback, in advance of two meetings on age disaggregation which were held in DFID on 3rd May. Your comments formed the basis of the technical discussions at the meetings, and together with the discussion from the day, will form the basis of a DFID Data Disaggregation Action Plan and DFID’s future engagement with partners on data disaggregation.

We were thrilled by the level of response. We received approximately 120 comments in response to questions under seven headings, plus eight longer email responses covering all questions. Comments came from a wide range of countries including: Germany, Burkina Faso, Serbia, Rwanda, Nepal, Ghana, Kenya, Russia, Kyrgyzstan, India, Tajikistan and the UK.

The comments were overwhelmingly supportive, containing lots of support for improving data on older or younger populations, and clear examples of the limitations that the current situation poses for developing appropriate policies for these groups.

There were some powerful examples of the importance of collecting data on all age groups:

“In Kenya and South Africa the HIV survey which expanded to include ages above 60 the statistics were quite revealing. It was found that the prevalence rates were much higher than the national rates. The pros were that it opened up the eyes of decision makers and they started planning for those 49 and above in terms of prevention, treatment and care.”

On the question as to whether household surveys should be expanded to cover all age ranges, respondents agreed that it will lead to:

“..advancing on the principle of leave no one behind.”

“..inclusion; better information from the household, which can help to target better the social public expenditure.”

But raised concerns around:

“..additional resources needed, time, money and interviewers.”

“[A] large quantity of data will be difficult to collect, collate and process.”

“Sample sizes [of specific sub groups] are often relatively small meaning it is hard to undertake in depth analysis of older people.”

On additions to national surveys:

“The suggestion of oversampling population groups, like older persons, is sensible but needs to be based on agreed international guidelines. This is essential to generate comparative datasets.”

It was highlighted that a Demographic and Health Survey (DHS) in Maldives had a specific module on older people.

“In many contexts, it is likely that carefully designed smaller scale surveys or qualitative studies designed to inform effective interventions to support specific groups such as young LGBT people [or] on specific health conditions facing older people would be more appropriate than attempting to collect large scale data. But they are limited as an all-purpose tool to assess socio-economic factors.”

Suggestions were made on how to improve data collection across all the suggested approaches:

“Questions could often be improved to be more age-sensitive. For example, questions around why individuals are not in employment often include “old age” or “retired”. This means it can be hard to unpack the specific reasons for not working in old age (e.g. ill health, choice, discouragement, receipt of other income sources etc.).”

A number of responses highlighted that care needs to be taken about who answers on behalf of an older or younger person — if the head of the household answers on their behalf, the answers may not be robust.

You also commented that we must be aware that:

“Cultural codes and expectations for younger [or] elder people may cause misunderstandings or conflict when data about taboo topics like, for example, sexual and reproductive health and rights are being collected.”

Further to these suggestions, many respondents highlighted their concerns that data is often collected, but not analysed effectively:

“An important principle in terms of research practice is to only collect data which will be used.”

“Initiatives to collect more data on populations in low and middle-income countries are welcomed, but in my experience the main issue currently is access to existing datasets for researchers.“

“Where data exists, it is often not used.”

“The level of age-disaggregation in published reports varies from country to country, and is often limited.”

And finally we asked what partners like DFID can do, you said that we can help by:

“Promoting use of open data/collaboration… The growth of social protection programmes also creates opportunities to better exploit surveys with an expanded age range, for example on coverage to elderly people.”

“collaborat[ing] with national statistical offices to develop systems for data collection and data analysis. The more nationals analyse their data, the higher the likelihood that they will demand more and better data from their national institutions.”

“Explain[ing] the importance of including older persons… emphasise the SDGs and how the goal is to leave no one behind.”

“What cannot be stressed enough is the need to strengthen national data collection systems in developing countries, be it statistical systems or especially CRVS. They build the cornerstone of evidence based decision making for the whole population, sending more or less valid projections and additional specific surveys to the second line of a complimentary tool.”

Thanks once again for your inputs! These inputs together with comments and discussion from the 3rd May day will all feed into development of DFID’s Disaggregation Action Plan. This action plan will set an agenda both for DFID internally to look at how it can disaggregate more in its own programme and results work, but also externally to discuss with partners and stakeholders as to what they can do to support great data disaggregation.

The ultimate aim of this work is to begin to address some of the data challenges of Leave No One Behind in the Global Goals.

This blog post was written by Emily Poskett — DFID’s Head of Profession for Statistics. You can follow Emily on Twitter.