The Open Data Institute (ODI) – Written evidence (FFF0040)
The UK’s data economy is crucial to the UK’s longer-term social and economic prosperity and influence as a global data and digital services hub and a world leader in AI, in line with the vision and ambition of the Integrated Review and the UK National AI Strategy. It is important that UK public sector data capability and data culture plays an integral role in developing trusted and trustworthy data ecosystems across sectors and borders, with public sector responsibilities ranging from 'hard' interventions such as regulation of the private sector, to 'soft' influence as such leading by example.
The UK National Data Strategy argues that 'every organisation is now a data organisation': and the Covid-19 pandemic brought data into the centre of organisational decision-making across sectors as never before, with clear needs for both access to data at speed and granularity in order to target disease transmission, monitor the emergence of new variants, and provide essential support, supplies and services to diverse communities.
But we have seen clear gaps and needs in the public sector around understanding and mapping data ecosystems in other sectors and understanding and mapping data initiatives within government; around identifying and understanding the limitations of digital tools or how to integrate them with non-digital approaches for the best outcomes; around considerations of equality and inclusion in digital public services and public sector data innovation; and around confidence and innovation for data sharing, data ethics, and data governance.
Key to this is the holistic concept of 'data literacy' as an important counterpoint to a narrow notion of 'data skills', where the latter is defined only as quantitative or numerical skills. We consider data literacy to be ‘the ability to think critically about data in different contexts and examine the impact of different approaches when collecting, using and sharing data and information’; and we consider the holistic suite of data skills needed for this sort of data literacy to include skills such as building communities, service design, data innovation and change leadership – so that data projects are impactful and lead to the best social and economic outcomes for everyone.
What are the hurdles to joint training between services? Do siloed approaches to attaining professional qualifications prevent joint training? How might better data-sharing improve joint training?
Over 2019-2020, as part of developing our 'Data and public services toolkit', we conducted applied research around public sector data use and data sharing with Devolved Administrations, local authorities, and city councils, and found that public sector teams struggled to map data ecosystems between or across services or sectors, make business cases for cross-service or cross-sector data projects, and feel confident about ethical decision-making around cross-service or cross-sector data projects. Some of the challenges appeared to be around differences in terminology, in ways of working, or in priorities; insufficient opportunities to informally discuss projects; and budget lines and processes that do not easily enable parallel training across more than one team, unit, or organisation around a shared project.
In September 2020, the UK government published the draft framework National Data Strategy (NDS) for public consultation, updating it in May 2021 in response to the views and evidence received. The NDS aims to ‘leverage existing UK strengths to boost the better use of data across businesses, government, civil society and individuals’. The strategy is structured around four core pillars of interconnected issues currently preventing the best use of data in the UK: data ‘foundations’, such as quality; data skills; data availability; and responsible use of data. It also has five missions, or priority areas for action, including:
● Mission 1: Unlocking the value of data across the economy
● Mission 3: Transforming government’s use of data to drive efficiency and improve public services
The Open Data Institute (ODI) is undertaking work to support the implementation of Mission 1; but we believe that Mission 1 can’t be achieved independently of Mission 3 and government’s own use of data. Governments are one of the largest and most powerful agents in national data ecosystems, collecting and controlling significant datasets across sectors and communities. And so the data practices of a government matters for how that country’s national data ecosystem develops overall across sectors and domains. By adopting good data practices, the government can better coordinate with private sector and civil society stakeholders involved in delivering public services, ensuring data works for the wider economy and society.
In 2021 we mapped all the data initiatives in UK government and found there are well over 100 bodies with some responsibility for ‘data’ across government. The proliferation of government bodies with responsibility for ‘data’ (and strategies, and similar topics, related to it) may create problems: for example, can we be sure that good practice in one area is being translated to others; that public sector leaders have an overarching view of what is (and isn’t) being taught and how; and that unnecessary duplication is being avoided?
We believe that there are particular risks in how government might train the public sector workforce in ‘data literacy’. The ODI defines this widely and holistically as ‘the ability to think critically about data in different contexts and examine the impact of different approaches when collecting, using and sharing data and information’. This is not about just technical data skills, but (for example) comparing how different people use numbers, graphs and infographics to convey important messages; evaluating the impact of bias and limited sampling on important decisions; and examining the ways that data is collected and the purposes of this collection (including everything from an infrequent Census to real-time sensors). We consider data literacy to be an important core competence for all civil servants, not just those working directly in data-related roles, professions and functions.
Although ‘data literacy’ is often highlighted as being important, it is seldom defined. When it is introduced in an important strategy document, such as the National Data Strategy, it is often conflated immediately with more quantitative data skills and not clearly defined in its own right. An additional challenge is that there are other terms with which it overlaps and can be conflated: for example, data skills, AI skills, AI literacy, media literacy, and digital literacy. There appears to be no coherent or consistent taxonomy in the public sector across these different concepts, nor between the taxonomy used in relation to the public sector workforce and the taxonomy used in relation to the wider population. We are currently researching the government’s approach to data literacy and expect to publish our findings in the first half of 2022.
How might the public sector become more attractive as an employer, particularly in comparison with the private sector? How might it become attractive enough to retain workers throughout their careers while maintaining a level of turnover that brings fresh ideas to organisations?
In our response to the UK National Data Strategy consultation 2020 we recognised that 'every organisation is now a data organisation': with this in mind, we believe it is important for the public sector to be - and to be seen to be - data literate, in order to influence the private sector and civil society by the high example that it sets. Furthermore, we believe that if the public sector is seen to use data in an effective and trustworthy way, this will increase the standing of the public sector in the wider data economy, which will also help make it an attractive employer relative to the private sector. In the Royal Society’s 2019 report 'Dynamics of data science skills', data scientists described how the public sector and civil society can effectively compete with the private sector for data science talent by offering opportunities to work with rich datasets, opportunities to work on challenge and complex problems with high social impact, and opportunities to be part of a rewarding and collaborative team environment that supported ongoing skills development.
At the ODI, we map the holistic breadth of data skills around ODI Data Skills framework, which breaks down the complex landscape of data skills into the sets of skills required by different people in an organisation. The Data Skills Framework is a tool that anyone can use to understand the skills needed to work with data. It helps users analyse their current approaches to data literacy, identify imbalances of skills, and address gaps in data literacy across an organisation. Research shows that if organisations focus solely on technology and technical skills, they are less likely to unlock the full value of data. The Data Skills Framework illustrates how technical data skills must be balanced with other skills – such as building communities, service design, data innovation and change leadership – to help ensure data projects are impactful and lead to the best social and economic outcomes for everyone. The Data Skills Framework can be used to identify and understand the requirement for this balance of skills – by analysing current skills and existing training programmes within an organisation. This helps ensure that skills and capability development plans align with organisational goals and that silos are not created. We believe that for the public sector to be an attractive employer, it should offer training and opportunities across the data skills framework, making the civil service a rewarding place to pursue a data literate career.
What role can digital tools play in increasing the accessibility of public services workers to service users, and in improving the quality of their work? How might we anticipate and mitigate any inequalities of access to public services that may arise from the expansion of such technologies?
In our project 'Open Cities', we explored data sharing and data openness in civic infrastructure as an alternative to the more tech-solutionist approach of 'Smart Cities'. A core theme of this project was exploring and understanding how digital tools typically only work for the intended outcome if they are well-integrated within a community or ecosystem: poor community engagement, for example, may not be easily 'fixed' by making a new digital tool and instead may require more substantive exploration around the purpose and drivers of community engagement. For these reasons, a focus on digital skills might be necessary but not sufficient: other skills such as service design, agile working, and even openness itself, may be needed for modernising public sector teams.
At the ODI we view data as a kind of national infrastructure, and so we believe that major national data assets, such as the government’s official statistics, must be representative of our society and accessible by all as a public good. In our response to the UK Statistics Authority consultation on the Inclusive Data Taskforce, we argued that protected characteristics are complex, multidimensional concepts that intersect with other features. Data collection should take account of this complexity and be specific about what characteristics are actually being measured and why. Consideration also needs to be given to other characteristics such as socio-economic or regional characteristics that may be relevant. We also argued that to make data more inclusive, it is important that those involved in collecting data, designing services and analysing results take a holistic, interdisciplinary approach to ‘data literacy’ that includes social awareness and emotional intelligence.
In 2019, with funding from The Legal Education Foundation, our project Monitoring equality in digital public services explored adherence to legal requirements around the Public Sector Equality Duty in digital public services. We’ve explored how the protected characteristics of people using the digital services are being collected, to make it possible to tell how they might be affecting excluded communities. We have found during this project that those providing digital public services don’t know the demographic make-up of who uses them. Data about the protected characteristics of people using these services isn’t currently collected and statistics aren’t published in a consistent or collective way. This means it is harder to find out who is excluded from using these services and why. Responsible collection and publication of data and statistics on protected characteristics would enable the monitoring of digital public services to determine whether everyone is being treated equitably by the system.
In our response to the 2021 NHSX draft data strategy consultation, we argued for the importance of open models and open innovation for widening participation in data innovation and trust in AI. In our response to the 2020 DCMS draft National Data Strategy consultation, we argued that government should look beyond the possible harmful impacts of data and AI on marginalised groups and additionally consider the ways in which data and AI might help ameliorate existing biases and injustices in society; and in our project Experimentalism and the Fourth Industrial Revolution we argue for a reappraisal of the 'deficit model' of marginalised groups, and instead consider the ways in which their data practices and AI innovations can enrich and expand the data practices of others such as the public sector.
What have been the effects of the COVID-19 pandemic and Brexit on the public services workforce? Have these events created opportunities for workforce reform?
In March 2021, the ODI published a report, 'Data on teachers’ lives during the pandemic', as UK schools were beginning to reopen fully for the second time during the coronavirus (Covid-19) pandemic. There was also very little data about the experiences of teachers during the pandemic; an under-researched area during compulsory school closures, but one that is particularly important to understand. About 10% of the teaching workforce leaves the profession each year and this ‘teacher wastage’ is likely to be particularly problematic at a time when we need a stable, motivated workforce as children recover from the effects of the pandemic. For this report, one of the major UK teachers’ unions, the NASUWT, made data from a large longitudinal survey of its members available to the ODI; survey data was combined with other open and publicly available data by Mime, an educational data consultancy, before being anonymised, aggregated and made available for exploration through an interactive tool. We found that about half of teachers were required to take on a dual teaching role: providing classroom supervision to those children still attending school, and providing remote learning to others, and most teachers reported that their workload had increased significantly in the previous 12 months; but despite the reported increase in workload and levels of stress, fewer teachers were considering leaving the profession when compared with surveys from previous years. This project serves both as an evidence point for the effects of the COVID-19 pandemic on the public services workforce; and also as an example of how the pandemic created an opportunity for reform and innovation in the use of workforce data, by catalysing organisations that steward data on the public sector workforce to make that data more available in order to produce timely insights.
Online learning in higher education exponentially accelerated during the Covid-19 pandemic, with students, staff and academic institutions having to rely on technologies and educational platforms to deliver lessons, record classes and introduce new interactive methods of learning. From reports of student and staff surveillance, to concerns around online learning harassment and ‘Zoom-bombing’ (where a person joins a Zoom meeting without invitation and aims to disrupt the session), the digitisation of online learning not only affects the academic experience, but also the relationship between educational institutions and how they collect, use and process personal data related to students and staff. Our 2021 project on ODI project 'Data governance for online learning' identified some of the new challenges that have emerged as a result of adapting to online learning through Covid-19, as well as recognising the efforts that students, staff and institutions have all made in transitioning to the digital classroom at short notice. Our findings were that the digitisation of education and increase of online learning data has also impacted stakeholder relationships with education and their institutions; and that although there might have been disruption in the distribution of resources within the institution, there have typically also been new opportunities for innovation, governance and policy.
In autumn 2021 as part of our engagement with the DCMS consultation on the future of UK data protection, we convened an expert roundtable in partnership with the Institute for Government (IfG) on data for public services. Key findings from the discussion were that the most salient challenges are around organisational culture, incentives, and data infrastructure, and around the availability of guidance, support and relevant tools. A related theme was that the law might not be the main opportunity for supporting data sharing in government, but might be a useful tool for strengthening redress mechanisms. Some believed that the problems with government data sharing are often simply problems of government; and while there was support for learning from pandemic experiences of data sharing in government, in terms of informing legislative change this was caveated by the importance of learning carefully because ‘hard cases make bad law’.
February 2022
8