FEATURE PUTTING DATA TO WORK
Putting data to work W
HEN THE WORD ‘statistics’ was first used in 1787, it referred to the numerical information needed to run the state. At that time, the amount of
such material was modest. But in the modern world, a vast range of government departments gather data that can inform both policy and academic research. In response, the ESRC is collaborating with the
Office for National Statistics (ONS), and centres in the devolved UK administrations, to establish the Administrative Data Research Partnership. This new structure is designed to support both government and academic objectives through better access to data. ‘Administrative Data’ is the term for statistical
material gathered by any arm of government. The census is the most obvious example, but data on income and spending, educational achievement, housing, health and employment are among many other vital datasets generated by government. Anna Vignoles, professor of education at Cambridge University and an ESRC council member, has been advising on the creation of ADRP. She says: “The aim of this partnership is to ensure that data is used for research, and especially for research which is policy-relevant and practice- relevant. We want to find out what questions we can answer by linking data from a range of government sources.” The data available via the ADRP will be
physically in the hands of the ONS, which Vignoles says will guarantee its security and anonymisation. “ONS has a high level of public trust but we still expect debate on issues such as risk and privacy. This means we have to make the case to government and the public about how important this initiative is. We need to convince government departments of the real benefits in sharing their data through ADRP.” Dr Emma Gordon is the ESRC director
responsible for ADRP. On the basis of her previous experience at the UK Treasury and elsewhere, she says: “Government departments are good at using their own data. But they can find it hard to get at other departments’ data or to link it to their own.” Part of the problem is the legal uncertainty of sharing and linking, and this is where the ADRP’s modus operandi is intended to help. “In the past, individual departments have had to sort this out: and someone could go to jail if they got it wrong.” In addition, the 2017 Digital Economy Act makes it simpler to justify the use of administrative data,
26 SOCIETY NOW WINTER 2018 “
The new Administrative Data Research Partnership is a collaboration between the ESRC, the Office for National Statistics and centres in the devolved UK administrations, designed to support government and academic objectives through better access to data. By Martin Ince
except on health, in cases where there is a clear public benefit. As the main aim of ADRP is to meet the policy needs of government departments, this ought to be feasible in most cases. She points out: “ESRC has invested heavily in ADRP data infrastructure at the ONS. It will allow us to store a lot of data, to provide linkage to it and to keep it anonymous. In the past there was no real reason for government departments to give us data.” Now, she hopes they may start to see the point.
Most deeper policy questions
span more than one department, and our approach should too
Anna Vignoles adds: “There is an awful lot of
good data in government departments. The insights we can get are enhanced, for instance, if we can link educational data to data held by the Department for Work and Pensions on employment and by the Ministry of Justice on crime. Most of the deeper policy questions span more than one department and our approach to data should do so too.” As well as holding rich resources of data, the
”
intention is for the ADRP to help develop new ways of linking and analysing it. Vignoles says: “The Administrative Data Research Network [ADRP’s predecessor body] did extensive work on methodology, and we have a budget to do more. For academics to provide value to users in government departments, we need to have the right methods, the right data and of course the right questions.” Gordon agrees that the development of novel
methodology will be “a big part” of the ADRP mission. “The devolved administrations [of the UK] realise the importance of training in the use of linked datasets, and this will happen in our partnership with the ONS as well. As more people get access to these data, we will need more how-to knowledge on methodology for linking data. There is already a growing realisation of the power of these methods.” In Vignoles’ view, the UK is already a world leader in some forms of data, and is especially good at linking longitudinal data collected through people’s lives to administrative data. But she adds that there are creative uses of linked data in Scandinavian nations, and that in the US more use is made of commercial data, for example on spending transactions. Gordon agrees that Scandinavia has some advantages over the UK when it comes to linking
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36
