The government analytics handbook: Leveraging data to strengthen public administration. By DanielRogger, ChristianSchuster (Eds.), Washington, DC: World Bank. 2023. pp. 760. ISBN (paper): 978‐1‐4648‐1957‐5; ISBN (electronic): 978‐1‐4648‐1981‐0

At a whopping 760 pages, the Government Analytics Handbook, edited by Daniel Rogger and Christian Schuster, is a monumental effort to provide an overview of the many sources of data from and on government and how analyzing these data can help strengthen public sector management. It is an invaluable resource not only for government decision makers and reformers who want to capitalize on the data analytics revolution but also for public administration researchers, teachers, and students in academia, think tanks, and beyond. The chapters can also be read individually, depending on interest, and the book offers a handy “how to use the handbook” section for its different audiences (pp. 53–55). It is accompanied by a website 1 as well as appendixes with various tools, checklists, and indicators for government analytics. In short, it offers a wealth of information for a diverse set of potential users. The Government Analytics Handbook has 5 parts and 30 chapters. Part 1 (chapters 1–3) provides an overview of the book, including summaries and lessons learned from all chapters and suggestions for moving the government analytics agenda forward. Part 2 (chapters 4–8) deals with foundational, cross-cutting issues in government analytics, such as “measuring what matters” (chapter 4), ensuring transparency and reproducibility of analysis (chapter 5), being cognizant of the ethics of data collection in public administration, especially about public sector staff (chapter 6), understanding and encouraging the use of performance information in government (chapter 7), and using multiple data sources to identify and diagnose corruption in the public sector (chapter 8). Part 3 of the book gives a comprehensive overview of the various sources of administrative data and examples of how they can inform government decision-making. They include human resources and payroll data (chapter 10), expenditure data (chapter 11), procurement data (chapter 12), data on the quality of administrative processes (chapter 13), customs data (chapter 14), administrative case data (chapter 15), and data on task and project completion (chapter 17). Part 3 also includes a chapter on using machine learning techniques for government analytics (chapter 16). Importantly, Part 3 starts with a chapter on how to build the infrastructure for collecting, processing, and using data in the public sector, based on the example of Human Resources Management Information Systems (HRMIS) (chapter 9). The chapter highlights the importance of a sequential approach where the basic modules are developed first, and the system is scaled up gradually, based on testing, technical support, and building in-house capacity to maintain and use Management Information Systems (MIS). Part 4 of the Handbook deals with a particularly familiar yet challenging topic for public administration researchers: public servant surveys. An increasing number of governments regularly conduct such surveys, covering public sector management issues such as employee motivation, engagement, job satisfaction, management practices, and perceptions of leadership (chapter 18). Part 4 discusses extensively the methodological choices and best practices in designing civil servant surveys, including different modes of administration (chapter 19), sample sizes (chapter 20), effective questionnaire design to ensure variability and completeness of responses (chapters 21, 22, and 23), and comparability of results across countries and organizations (chapter 24). One of the strengths of this section is that the chapters are not just theoretical discussions. The authors leverage the wealth of data available under the banner of the Global Survey of Public Servants 2 to empirically test and develop concrete recommendations for each of these issues. The last two chapters of Part 4 illustrate how governments use public servant surveys and offer recommendations for action drawing on the experiences of Australia, Canada, Ireland, the United Kingdom, and the United States (chapters 25 and 26). Recommendations include, for example, publishing results and setting up easy-to-use dashboards, breaking down data to the agency level, and building in-house capacity for interpreting survey results linked to senior management and embedded in a culture of transparency and responsiveness. Part 5 turns to external data sources for government analytics, including household surveys and labor market surveys (chapter 27), citizen surveys on specific topics such as trust in government (chapter 28), service delivery surveys and assessments along the service delivery chain (chapter 29), and anthropological methods (chapter 30). For example, labor market surveys can inform comparisons between public and private sector employment trends and conditions (Baig et al., 2023). Alternatively, anthropological approaches can illuminate aspects of the inner working of the bureaucracy, such as informal practices and their relational embeddedness, which remain hidden in structured and standardized modes of data collection such as surveys (Asdal et al., 2023). The Government Analytics Handbook has many strong suits beyond its comprehensiveness. First, despite being a book about data and analytics, it does not reify these. Rather, it acknowledges the importance of balancing them with practical and tacit knowledge to inform decision-making. As the authors put it, “What government analytics can do is strengthen the quality of conversations about how to improve public administration, rather than dictating managerial responses to specific analytics findings” (Rogger & Schuster, 2023, p. 11). It also engages openly and directly with ethical issues inherent in collecting and using public administration data throughout various chapters as well as more in-depth in chapter 6, which discusses how to protect individual dignity and privacy, allow for voice and dissent, and ensure that “analytics enable public servants to deliver on public sector values: accountability, productivity, and innovation” (Wittels, 2023, p. 95). Considering, for example, the dangerous implications of collecting granular data on employees in repressive regimes, such questions on the ethics of government data collection and use warrant more attention in the public administration community. The most impressive feat of the book is that it is both a technical manual and a manifesto for action. The latter is what makes it more than a compendium and makes it even more relevant to public administration researchers and practitioners. Some of the Handbook's most interesting parts are the case studies and examples that illustrate the utilization of government analytics. Given the costs and capacity required, documenting and understanding what drives the actual use and benefits of such efforts is essential. The public administration literature has engaged extensively with the use of performance information in the public sector—its upsides, downsides, and ambiguities (e.g., Van Dooren & Van De Walle, 2008). The Government Analytics Handbook draws to some extent on insights gained from this literature, including through a chapter specifically dedicated to “Measuring and Encouraging Performance Information Use in Government” (Moynihan, 2023). Nevertheless, this is also where the reader might want more, especially on how data analytics systems are or can be built and used outside the Global North. In the United States, the use of performance information and employee surveys has a comparatively long history as part of broader reform efforts—from the Government Performance and Results Act (GPRA) of 1993 (Moynihan & Kroll, 2016) to the introduction of the Federal Employee Viewpoint Survey in 2002 to the current Biden-Harris management-focused learning agenda. The takeaway is that not only do systems supportive of data analytics in government take decades and substantial investments to build, but so does generating demand and ensuring the impact of these systems. The latter, especially, remains an ongoing battle, even in the best of cases. To its credit, the book does a lot in terms of laying the foundation for data analytics outside the Global North. In addition to examples and case studies, for example, from Brazil, Nigeria, Colombia, Ghana, or Pakistan, to name just a few, the book also includes a chapter specifically assessing the cross-country comparability of responses for certain civil servant survey questions, including on culturally specific concepts such as transformational leadership (Lipinski et al., 2023). The authors find suggestive evidence that such comparisons are possible, and that comparability is highest for countries in the same region and at similar income levels. This is good news for comparative public administration scholars, who have advocated for increased data collection outside the Western world to enable more cross-national public administration research (Bertelli et al., 2020). Yet, some questions remain about how the political and administrative context influence not only the use of government analytics, but also the development of data and analytics infrastructure, data collection options and choices, and even the results of the data collection efforts. For example, whether online or in-person modes of data collection yield better data for civil servant surveys is not only a function of response rate but also of social desirability bias. Using an experiment where the survey mode was randomized, Han et al. (2023) find that differences between the two modes of administration were most pronounced among managers and for ethical questions. The authors rightfully point out that understanding what drives these differences is an important area for further research. Taking this a step further, we can argue that context matters even for survey administration. Civil servant expectations and fears about how the results will be used can induce social desirability bias, especially in contexts with little experience with such surveys and no history of using them constructively to improve working conditions and management practices in the public sector. Acknowledging the need to understand such fears and expectations could reinforce another message of the Handbook: that analytics needs to be accompanied by strong public and political communication to build confidence in its value and showcase its benefits for decision makers, public employees, and citizens alike. Given the comprehensiveness of the Government Analytics Handbook, these are not as much limitations as they are suggestions for a way forward. As the book's website is being refined and updated, it offers opportunities to keep building on this foundation to expand not just the kinds of data and methods used but to accumulate practical insights about how government analytics systems can be built and what determines their success across countries. The World Bank is uniquely positioned to both draw on and speak to decision makers and researchers on a truly global scale. The Government Analytics Handbook reflects this. The hope is that further iterations will delve deeper into country experiences from the Global South in building and using data analytics systems. This can help the public administration community not only improve their methodologies and sources of data but also understand the politico-administrative drivers of the data revolution and how they can be leveraged to serve citizens better. Sabina Schnell is Assistant Professor and Senior Research Associate at the Program for Advancement of Research on Conflict and Collaboration (PARCC), Maxwell School of Citizenship and Public Affairs, Syracuse University. Her research interests include transparency, open government, public values, HRM, and comparative and international public administration. She has also worked as an advisor on public sector reform projects with international organizations and national governments, including the World Bank. Email: [email protected]

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.