There has, and continues to be, a lot of discussion about the (statistical) challenges in defining the indicators that measure the sustainable development goals (SDGs). But what about the statistical challenges that are faced in helping countries reach the SDGs?
This was the topic of the talk I was invited to give at the Royal Statistical Society’s Annual Conference in a session organised by their International Development Working Group on the SDGs. The first speaker Kenny Bambrick of DfID introduced the SDGs and their wide ranging set of goals, targets and indicators The second speaker Geoffrey Greenwell of PARIS21 discussed how indicators based on administrative data (ie routinely collected data by, for example government departments such as health) could help countries focus on improving their basic information systems and service delivery.
An underlying motivation of my talk was to show that statistical issues around international development and the SDGs don’t relate only to national statistics, capacity building and the creation of indicators. Although these are clearly very important issues I worry that this is all that the most statisticians perceive as the statistical challenges are in international development.
I wanted to show that to help provide better evidence to inform policy decisions around international development there are challenging statistical problems that need solving. These statistical problems are related to several areas of current statistical research. The talk expanded on many of the points I have raised in previous posts (see here and here) and a copy of the slides can be found here.
One topic we all discussed was the SDG aim to “leave no one behind”. Many of the poorest and most vulnerable communities, for example street children, pastoralists, or slaves are difficult to find, let alone count or consider and investigate how they are affected by different policies. Special sampling and analytical tools, such as mark-recapture ideas or adaptive sampling techniques may be are required to help study, understand and estimate these populations (at local, regional or national level).
I also noted that we need to be careful about the excitement around the big data revolution. Although this leads to exciting opportunities for acquiring different types of data and engaging with many different communities the most vulnerable populations are the least likely to have access to these technologies and are in danger of falling even further behind.
Overall, my main message was that there are many interesting statistical research problems in international development. We need to encourage communication between statisticians and those working in international development. This will be something I will be working on so watch this space.
Any thoughts and contributions on how to do this would be welcome in the comments below.