It is vital that the statistics produced by government and other official bodies are as valuable as they can be to society. Statistics should be easy to access, relevant, and help the public understand important questions such as ‘How many people work in the UK health services?’, ‘What is the nature of the housing shortage?’ and ‘Are living standards going up or down?’
But, these are questions that can’t be answered with one statistic alone– and in some cases can’t be answered without evidence that goes beyond statistics as well. To complicate matters, in the UK system, there are multiple bodies producing statistics on each of these themes. To really understand the public value of statistics, it is important to look across a set of related statistics, as well as looking at the statistics individually. This is why we carry out systemic reviews – reviews that explore a set of official statistics in a thematic area or on a cross-cutting topic.
These reviews allow us to look at the relationships and signposting between statistics and to spot gaps and overlaps. The work involves discussions with people who use and produce statistics and a great feature is our ability to do more than publish recommendations – also intervening to support change directly, for example facilitating stakeholder meetings or other events.
To date we have reviewed health and social care statistics in England, and across the UK, statistics on income and earnings, city regions, crime and justice, housing and planning and international migration. We are currently looking at data linkage and innovations in statistics in the education area.
Each systemic review has identified examples of good practice. These include statisticians generating new statistics or analyses in response to user requests and presenting material in a user-friendly way. For example:
- ONS now has a measure of cyber crime,
- official labour market statistics can be accessed via the NOMIS platform
- the Migration Statistics Quarterly Report combines statistics on a related topic in one publication
Each systemic review has also found areas for improvement. A common one is users of statistics finding it difficult to access what they need – for example there are four main producers publishing statistics on mental health in England which causes confusion as people can’t easily access the information they need from one place.
Users also face problems understanding how definitions vary for similar statistics, for instance on housing affordability across different parts of the UK. And it can be hard for users to interpret what a set of statistics collectively show – as in the case of various data sources on income and earnings not giving a consistent message on the state of living standards. In some cases there are gaps where statistics are not available to shed sufficient light on a topical issue – such as the rents paid by those living in private rented housing, for sub-national geographies.
Often we are able to identify changes in the way the statistics system works that can generate improvements – simple things like groups bringing together different producers at strategic and working levels and user-engagement strategies can really help prioritise efforts to improve the public value of statistics. It’s great to see the English Health and Care Statistics Steering Group set up to work across health and social care statistics bodies in response to our work identifying a lack of strategic leadership; this has now delivered changes to statistical outputs – for example ‘Statistics on Smoking – England’ which combines figures from different Departments to help users. We also like prioritised plans for statistical developments, for example ONS’s Economic Statistics and Analysis Strategy.
We consistently find that statistics producers are motivated to use data to drive improvements in policy and society. But we also find resource constraints. We strongly believe that speaking to a wide range of users of statistics and understanding their questions, innovating and collaborating across the system are vital for doing the best job statisticians can – and we also recognise that this can be challenging.
We would love to hear about your good examples and any challenges you face meeting the public value expectations of the Code of Practice. We also welcome suggestions for ways we could support improvements.