Refine
Year of publication
Document Type
- Article (37)
- Working Paper (3)
- Review (1)
Is part of the Bibliography
- no (41)
The scarcity of empirical evidence surrounding the organizational challenges and successful approaches to artificial intelligence (AI) deployment has resulted in mostly theoretical conceptualizations. By analyzing policy labs and offices of data analytics across the US to understand organizational challenges of AI adoption and implementation in the public sector as well as to identify successful management strategies to address such challenges, our study moves from speculation to gathering evidence. Our findings show that most challenges are found during the implementation stage and include challenges related to skills, culture, and resistance to share the data driven by data challenges. Further, our results indicate that long term strategies and short term actions need to be put in place to address these challenges. Among the first ones, leadership and executive support and stakeholder management seem to play an important role. Data standardization, training, and data-sharing agreements also seem to be successful specific short-term actions.
Background: Instituted under the Paris Agreement, nationally determined contributions (NDCs) outline countries’ plans for mitigating and adapting to climate change. They are the primary policy instrument for protecting people’s health in the face of rising global temperatures. However, evidence on engagement with health in the NDCs is scarce. In this study, we aimed to examine how public health is incorporated in the NDCs, and how different patterns of engagement might be related to broader inequalities and tensions in global climate politics.
Methods: We analysed the NDCs in the UN Framework Convention on Climate Change registry submitted by 185 countries. Using content analysis and natural language processing (NLP) methods, we developed measures of health engagement. Multivariate regression analyses examined whether country-level factors (eg, population size, gross domestic product [GDP], and climate-related exposures) were associated with greater health engagement. Using NLP methods, we compared health engagement with other climate-related challenges (ie, economy, energy, and agriculture) and examined broader differences in the keyword terms used in countries with high and low health engagement in their NDCs.
Findings: Countries that did not mention health in their NDCs were clustered in high-income countries, whereas greater health engagement was concentrated in low-income and middle-income countries. Having a low GDP per capita and being a small island developing state were associated with higher levels of health engagement. In addition, higher levels of population exposure to temperature change and ambient air pollution were associated with more health coverage included in a country’s NDC. Variation in health engagement was greater than for other climate-related issues and reflected wider differences in countries’ approaches to the NDCs.
Interpretation: A focus on health in the NDCs follows broader patterns of global inequalities. Poorer and climate-vulnerable countries that contribute least to climate change are more likely to engage with health in their NDCs, while richer countries focus on non-health sectors in their NDCs, such as energy and the economy.
Objective: To examine country engagement with the health impacts of climate change in (1) annual statements in the UN General Debate (UNGD); and (2) the Nationally Determined Contributions (NDCs) of the Paris Agreement; and to identify what factors drive country engagement.
Methods: First, we measure engagement by the total references to the health and climate change relationship in each text, using a keyword-in-context search with relevant search terms. Second, we use machine learning models, specifically random forest models, to identify the most important country-level predictors of engagement. Our predictors are political and economic factors, health outcomes, climate change-related variables, and membership of political negotiating groups in the UN.
Findings: For both UNGD statements and NDCs, we find that that low- and middle-income countries discuss the health impacts of climate change much more than high-income countries. We find that the most important predictors of country engagement are related health outcomes (infant mortality rates, maternal death risk, life expectancy), countries’ income levels (GDP per capita), and fossil fuel consumption. Membership of political negotiating groups (e.g. G77 and SIDS) is less important predictors.
Conclusion: Our analysis indicates a North-South division in engagement, and hence there is little to suggest that a health framing of climate change overcomes existing geopolitical divisions in climate change negotiations. Countries who carry the heaviest health burdens but lack necessary resources to address the impacts of climate change are shouldering responsibility for reminding the global community of the implications of climate change for people’s health.
The world has already warmed by more than 1.2C compared with preindustrial levels, resulting in profound, immediate, and rapidly worsening health effects, and moving dangerously close to the agreed limit of maintaining temperatures “well below 2C”. These health impacts are seen on every continent, with the ongoing spread of dengue virus across South America, the cardiovascular and respiratory effects of record heatwaves and wildfires in Australia, western North America, and western Europe, and the undernutrition and mental health effects of floods and droughts in China, Bangladesh, Ethiopia, and South Africa. In the long term, climate change threatens the very foundations of human health and wellbeing, with the Global Risks Report registering climate change as one of the five most damaging or probable global risks every year for the past decade.
Topic models are widely used in natural language processing, allowing researchers to estimate the underlying themes in a collection of documents. Most topic models require the additional step of attaching meaningful labels to estimated topics, a process that is not scalable, suffers from human bias, and is difficult to replicate. We present a transfer topic labeling method that seeks to remedy these problems, using domain-specific codebooks as the knowledge base to automatically label estimated topics. We demonstrate our approach with a large-scale topic model analysis of the complete corpus of UK House of Commons speeches from 1935 to 2014, using the coding instructions of the Comparative Agendas Project to label topics. We evaluated our results using human expert coding and compared our approach with more current state-of-the-art neural methods. Our approach was simple to implement, compared favorably to expert judgments, and outperformed the neural networks model for a majority of the topics we estimated.
Left unabated, climate change will have catastrophic effects on the health of present and future generations. Such effects are already seen in Europe, through more frequent and severe extreme weather events, alterations to water and food systems, and changes in the environmental suitability for infectious diseases. As one of the largest current and historical contributors to greenhouse gases and the largest provider of financing for climate change mitigation and adaptation, Europe’s response is crucial, for both human health and the planet. To ensure that health and wellbeing are protected in this response it is essential to build the capacity to understand, monitor, and quantify health impacts of climate change and the health co-benefits of accelerated action. Responding to this need, the Lancet Countdown in Europe is established as a transdisciplinary research collaboration for monitoring progress on health and climate change in Europe. With the wealth of data and academic expertise available in Europe, the collaboration will develop region-specific indicators to address the main challenges and opportunities of Europe’s response to climate change for health. The indicators produced by the collaboration will provide information to health and climate policy decision making, and will also contribute to the European Observatory on Climate and Health.
Rapid growth in smart meter installations has given rise to vast collections of data at a high time-resolution and down to an individual level. However, to enable efficient policy interventions, we need to be able to appropriately segment the population of users. The aim of this paper is to consider challenges and opportunities associated with large highly-granular temporal datasets that describe residential electricity consumption. In particular, the focus is on experiments relating to aggregation of smart meter time-series data in the context of clustering and prediction tasks that are often used for customer targeting and to gain insight on energy-use about sub populations. To cluster energy use profiles, we propose a novel framework based on a set of Gaussian based models which we use to encode individuals’ energy consumption over time. The dataset consists of half hourly electricity consumption records from smart meters of households in the UK (2014–2015). The contribution of this paper comes from its investigation of how consumers or groups may be clustered according to model parameters in scenarios where additional data on consumers is not available to the researcher, or where anonymity preservation of the smart meter user is prioritised. A secondary aim is to invite greater awareness when data reduction is required to reduce the size of a large dataset for computational purposes. This may have implications for policy interventions acting at the individual or small group level, for instance, when designing incentives to encourage energy efficient behaviour or when identifying fuel poor customers.
Political science, and social science in general, have traditionally been using computational methods to study areas such as voting behavior, policy making, international conflict, and international development. More recently, increasingly available quantities of data are being combined with improved algorithms and affordable computational resources to predict, learn, and discover new insights from data that is large in volume and variety. New developments in the areas of machine learning, deep learning, natural language processing (NLP), and, more generally, artificial intelligence (AI) are opening up new opportunities for testing theories and evaluating the impact of interventions and programs in a more dynamic and effective way. Applications using large volumes of structured and unstructured data are becoming common in government and industry, and increasingly also in social science research. This chapter offers an introduction to such methods drawing examples from political science. Focusing on the areas where the strengths of the methods coincide with challenges in these fields, the chapter first presents an introduction to AI and its core technology - machine learning, with its rapidly developing subfield of deep learning. The discussion of deep neural networks is illustrated with the NLP tasks that are relevant to political science. The latest advances in deep learning methods for NLP are also reviewed, together with their potential for improving information extraction and pattern recognition from political science texts.
The 2021 report of the Lancet Countdown on health and climate change: code red for a healthy future
(2021)
The Lancet Countdown is an international collaboration
that independently monitors the health consequences of a
changing climate. Publishing updated, new, and improved
indicators each year, the Lancet Countdown represents the
consensus of leading researchers from 43 academic
institutions and UN agencies. The 44 indicators of this
report expose an unabated rise in the health impacts of
climate change and the current health consequences of the
delayed and inconsistent response of countries around the
globe—providing a clear imperative for accelerated action
that puts the health of people and planet above all else.
The 2021 report coincides with the UN Framework
Convention on Climate Change 26th Conference of the
Parties (COP26), at which countries are facing pressure
to realise the ambition of the Paris Agreement to keep
the global average temperature rise to 1·5°C and to
mobilise the financial resources required for all
countries to have an effective climate response. These
negotiations unfold in the context of the COVID-19
pandemic—a global health crisis that has claimed
millions of lives, affected livelihoods and communities
around the globe, and exposed deep fissures and
inequities in the world’s capacity to cope with, and
respond to, health emergencies. Yet, in its response to
both crises, the world is faced with an unprecedented
opportunity to ensure a healthy future for all.
In his recent article Söderlund (2003) tests structural factors that influence the order in which the Russian regions gained a bi-lateral agreement with the federal centre emphasizing the importance of ethnicity, religion and economy. We replicate his results, and provide an extension where we argue instead that the only significant determinants of the bi-lateral process have been economic issues. Our results are substantiated by an improved methodology that addresses several debatable choices made by the author in the original article.
Following Easton’s conceptual framework discussed in the introductory chapter, a hierarchical relationship exists between three objects of support: output support, support for institutions, and support for the community. The latter two objects of support are examined in turn in two subsequent chapters on trust in European political institutions and the relationship between citizenship and identity in the European Community. This chapter focuses on the first object of support – support derived from the accrued material benefits of EU membership.
Treating Words as Data with Error: Estimating Uncertainty in Text Statements of Policy Positions
(2009)
Political text offers extraordinary potential as a source of information about the policy positions of political actors. Despite recent advances in computational text analysis, human interpretative coding of text remains an important source of text-based data, ultimately required to validate more automatic techniques. The profession’s main source of cross-national, time-series data on party policy positions comes from the human interpretative coding of party manifestos by the Comparative Manifesto Project (CMP). Despite widespread use of these data, the uncertainty associated with each point estimate has never been available, undermining the value of the dataset as a scientific resource. We propose a remedy. First, we characterize processes by which CMP data are generated. These include inherently stochastic processes of text authorship, as well as of the parsing and coding of observed text by humans. Second, we simulate these error-generating processes by bootstrapping analyses of coded quasi-sentences. This allows us to estimate precise levels of nonsystematic error for every category and scale reported by the CMP for its entire set of 3,000-plus manifestos. Using our estimates of these errors, we show how to correct biased inferences, in recent prominently published work, derived from statistical analyses of error-contaminated CMP data.
The decision to establish direct elections to the European Parliament was intended by many to establish a direct link between the individual citizen and decision making at the European level. Elections were meant to help to establish a common identity among the peoples of Europe, to legitimise policy through the normal electoral processes and provide a public space within which Europeans could exert a more direct control over their collective future. Critics disagreed, arguing that direct elections to the European Parliament would further undermine the sovereignty of member states, and may not deliver on the promise that so many were making on behalf of that process. In particular, some wondered whether elections alone could mobilise European publics to take a much greater interest in European matters, with the possibility of European elections being contested simply on national matters. Evaluating these divergent views, the subject of this article is to review the literature on direct elections to the European Parliament in the context of the role these elections play in governance of the European Union. The seminal work by Reif and Schmitt serves as the starting point of our review. These authors were the first to discuss elections to the European Parliament as second-order national elections. Results of second-order elections are influenced not only by second-order factors, but also by the situation in the first-order arena at the time of the second-order election. In the 30 years and six more sets of European Parliament elections since the publication of their work, the concept has become the dominant one in any academic discussion of European elections. In this article we review that work in order to assess the continuing value of the second-order national election concept today, and to consider some of the more fruitful areas for research which might build on the advance made by Reif and Schmitt. While the concept has proven useful in studies of a range of elections beyond just those for the European Parliament, including those for regional and local assemblies as well as referendums, this review will concentrate solely on EP elections. Concluding that Reif and Schmitt’s characterisation remains broadly valid today, the article allows that while this does not mean there is necessarily a democratic deficit within the EU, there may be changes that could be made to encourage a more effective electoral process.
Scholars estimating policy positions from political texts typically code words or sentences and then build left-right policy scales based on the relative frequencies of text units coded into different categories. Here we reexamine such scales and propose a theoretically and linguistically superior alternative based on the logarithm of odds- ratios. We contrast this scale with the current approach of the Comparative Manifesto Project (CMP), showing that our proposed logit scale avoids widely acknowledged flaws in previous approaches. We validate the new scale using independent expert surveys. Using existing CMP data, we show how to estimate more distinct policy dimensions, for more years, than has been possible before, and make this dataset publicly available. Finally, we draw some conclusions about the future design of coding schemes for political texts.
The Comparative Manifesto Project (CMP) provides the only time series of estimated party policy positions in political science and has been extensively used in a wide variety of applications. Recent work (e.g., Benoit, Laver, and Mikhaylov 2009; Klingemann et al. 2006) focuses on nonsystematic sources of error in these estimates that arise from the text generation process. Our concern here, by contrast, is with error that arises during the text coding process since nearly all manifestos are coded only once by a single coder. First, we discuss reliability and misclassification in the context of hand-coded content analysis methods. Second, we report results of a coding experiment that used trained human coders to code sample manifestos provided by the CMP, allowing us to estimate the reliability of both coders and coding categories. Third, we compare our test codings to the published CMP “gold standard” codings of the test documents to assess accuracy and produce empirical estimates of a misclassification matrix for each coding category. Finally, we demonstrate the effect of coding misclassification on the CMP’s most widely used index, its left–right scale. Our findings indicate that misclassification is a serious and systemic problem with the current CMP data set and coding process, suggesting the CMP scheme should be significantly simplified to address reliability issues.
All methods for analyzing text require the identification of a fundamental unit of analysis. In expert-coded content analysis schemes such as the Comparative Manifesto Project, this unit is the ‘quasi-sentence’: a natural sentence or a part of a sentence judged by the coder to have an independent component of meaning. Because they are subjective constructs identified by individual coders, however, quasi-sentences make text analysis fundamentally unreliable. The justification for quasi-sentences is a supposed gain in coding validity. We show that this justification is unfounded: using quasi-sentences does not produce valuable additional information in characterizing substantive political content. Using natural sentences as text units, by contrast, delivers perfectly reliable unitization with no measurable loss in content validity of the resulting estimates.
The paper explores a question raised by the 2011 Irish election, which saw an almost unprecedented decline in support for a major governing party after an economic collapse that necessitated an ECB/IMF ‘bailout’. This seems a classic case of ‘economic voting’ in which a government is punished for incompetent performance. How did the government lose this support: gradually, as successive economic indicators appeared negative, or dramatically, following major shocks? The evidence points to losses at two critical junctures. This is consistent with an interpretation of the link between economics and politics that allows for qualitative judgements by voters in assigning credit and blame for economic performance.
Coding non-manifesto documents as if they were genuine policy platforms produced at election time clearly raises serious issues with error when these codings are used in the standard manner to estimate left-right policy positions. In addition to the long term solution of improving the document base of the Manifesto Project identified by Gemenis (2012), we argue that immediate gains in manifesto-based estimates of policy positions can be realised by using the confrontational logit scales from Lowe et al. (2011), which addresses the problems of scale content and scale construction that are exacerbated by but not unique to the problems found in proxy documents.
We address leadership emergence and the possibility that there is a partially innate predisposition to occupy a leadership role. Employing twin design methods on data from the National Longitudinal Study of Adolescent Health, we estimate the heritability of leadership role occupancy at 24%. Twin studies do not point to specific genes or neurological processes that might be involved. We therefore also conduct association analysis on the available genetic markers. The results show that leadership role occupancy is associated with rs4950, a single nucleotide polymorphism (SNP) residing on a neuronal acetylcholine receptor gene (CHRNB3). We replicate this family-based genetic association result on an independent sample in the Framingham Heart Study. This is the first study to identify a specific genotype associated with the tendency to occupy a leadership position. The results suggest that what determines whether an individual occupies a leadership position is the complex product of genetic and environmental influences, with a particular role for rs4950.
Recent literature models leadership as a process of communication in which leaders’ rhetorical signals facilitate followers’ co-ordination. While some studies have explored the effects of leadership in experimental settings, there remains a lack of empirical research on the effectiveness of informational tools in real political environments. Using quantitative text analysis of federal and sub-national legislative addresses in Russia, this article empirically demonstrates that followers react to informational signals from leaders. It further theorizes that leaders use a combination of informational and non-informational tools to solve the co-ordination problem. The findings show that a mixture of informational and non- informational tools shapes followers’ strategic calculi. Ignoring non-informational tools — and particularly the interrelationship between informational and non-informational tools - can threaten the internal validity of causal inference in the analysis of leadership effects on co-ordination.
The 2011 election in Ireland was one of the most dramatic elections in European post-war history in terms of net electoral volatility. In some respects the election overturned the traditional party system. Yet it was a conservative revolution, one in which the main players remained the same, and the switch in the major government party was merely one in which one centre right party replaced another. Comparing voting behaviour over the last three elections we show that the 2011 election looks much like that of 2002 and 2007. The crisis did not result in the redefinition of the electoral landscape. While we find clear evidence of economic voting at the 2011 election, issue voting remained week. We believe that this is due to the fact that parties have not offered clear policy alternatives to the electorate in the recent past and did not do so in 2011.
In the absence of public information on the inner workings of the Russian political regime, especially during Medvedev’s presidency, outside observers often have to rely on politicians’ unguarded comments or subjective analysis. Instead, we turn to quantitative text analysis of political rhetoric. Treating governors as a quasi-expert panel, we argue that policy positions revealed in regional legislative addresses explain how elites perceived the distribution of power between Putin and Medvedev. We find that governors moved from a neutral position in 2009 to a clearly pro-Putin position in 2011, and that policy initiatives advocated by Medvedev all but evaporated from the rhetoric of governors in 2012.
Examining the careers of democratic heads of state and government from 1960–2010, we find that one in every seven turns to the private sector after office. Distinguishing between the factors that attract leaders to business and those that render leaders attractive, we find that the global CEO compensation rates, cultural norms, having served in office in Anglo-Saxon countries as well as their personal background, matter. We also find that certain economic outcomes and policies in office such as economic growth and reduction in state spending are often associated with post-tenure business careers. We do not find evidence, however, that leaders are able to implement policies with future careers in mind, which would in turn raise concerns over accountability.
Empirical social science often relies on data that are not observed in the field, but are transformed into quantitative variables by expert researchers who analyze and interpret qualitative raw sources. While generally considered the most valid way to produce data, this expert-driven process is inherently difficult to replicate or to assess on grounds of reliability. Using crowd-sourcing to distribute text for reading and interpretation by massive numbers of non-experts, we generate results comparable to those using experts to read and interpret the same texts, but do so far more quickly and flexibly. Crucially, the data we collect can be reproduced and extended transparently, making crowd-sourced datasets intrinsically reproducible. This focuses researchers’ attention on the fundamental scientific objective of specifying reliable and replicable methods for collecting the data needed, rather than on the content of any particular dataset. We also show that our approach works straightforwardly with different types of political text, written in different languages. While findings reported here concern text analysis, they have far-reaching implications for expert-generated data in the social sciences.
Every year at the United Nations (UN), member states deliver statements during the General Debate (GD) discussing major issues in world politics. These speeches provide invaluable information on governments’ perspectives and preferences on a wide range of issues, but have largely been overlooked in the study of international politics. This paper introduces a new dataset consisting of over 7300 country statements from 1970–2014. We demonstrate how the UN GD corpus (UNGDC) can be used as a resource from which country positions on different policy dimensions can be derived using text analytic methods. The article provides applications of these estimates, demonstrating the contribution the UNGDC can make to the study of international politics.
We present a database of parliamentary debates that contains the complete record of parliamentary speeches from Dáil Éireann, the lower house and principal chamber of the Irish parliament, from 1919 to 2013. In addition, the database contains background information on all TDs (Teachta Dála, members of parliament), such as their party affiliations, constituencies and office positions. The current version of the database includes close to 4.5 million speeches from 1,178 TDs. The speeches were downloaded from the official parliament website and further processed and parsed. Background information on TDs was collected from the member database of the parliament website. Data on cabinet positions (ministers and junior ministers) was collected from the official website of the government. A record linkage algorithm and human coders were used to match TDs and ministers.
Detecting Policy Preferences and Dynamics in the UN General Debate with Neural Word Embeddings
(2017)
Foreign policy analysis has been struggling to find ways to measure policy preferences and paradigm shifts in international political systems. This paper presents a novel, potential solution to this challenge, through the application of a neural word embedding (Word2vec) model on a dataset featuring speeches by heads of state or government in the United Nations General Debate. The paper provides three key contributions based on the output of the Word2vec model. First, it presents a set of policy attention indices, synthesizing the semantic proximity of political speeches to specific policy themes. Second, it introduces country-specific semantic centrality indices, based on topological analyses of countries’ semantic positions with respect to each other. Third, it tests the hypothesis that there exists a statistical relation between the semantic content of political speeches and UN voting behavior, falsifying it and suggesting that political speeches contain information of different nature then the one behind voting outcomes. The paper concludes with a discussion of the practical use of its results and consequences for foreign policy analysis, public accountability, and transparency.
The Lancet Countdown tracks progress on health and climate change and provides an independent assessment of the health effects of climate change, the implementation of the Paris Agreement, and the health implications of these actions. It follows on from the work of the 2015 Lancet Commission on Health and Climate Change, which concluded that anthropogenic climate change threatens to undermine the past 50 years of gains in public health, and conversely, that a comprehensive response to climate change could be “the greatest global health opportunity of the 21st century”.
Big Data and Artificial Intelligence will have a profound transformational impact on governments around the world. Thus, it is important for scholars to provide a useful analysis on the topic to public managers and policymakers. This study offers an in-depth review of the Policy and Administration literature on the role of Big Data and advanced analytics in the public sector. It provides an overview of the key themes in the research field, namely the application and benefits of Big Data throughout the policy process, and challenges to its adoption and the resulting implications for the public sector. It is argued that research on the subject is still nascent and more should be done to ensure that the theory adds real value to practitioners. A critical assessment of the strengths and limitations of the existing literature is developed, and a future research agenda to address these gaps and enrich our understanding on the topic is proposed.
Public sector organizations are increasingly interested in using data science and artificial intelligence capabilities to deliver policy and generate efficiencies in high uncertainty environments. The long-term success of data science and artificial intelligence (AI) in the public sector relies on effectively embedding it into delivery solutions for policy implementation. However, governments cannot do this integration of AI into public service delivery on their own. The UK Government Industrial Strategy is clear that delivering on the AI grand challenge requires collaboration between universities and public and private sectors. This cross-sectoral collaborative approach is the norm in applied AI centres of excellence around the world. Despite their popularity, cross-sector collaborations entail serious management challenges that hinder their success. In this article we discuss the opportunities and challenges from AI for public sector. Finally, we propose a series of strategies to successfully manage these cross-sectoral collaborations.
The Lancet Countdown: tracking progress on health and climate change was established to provide an independent, global monitoring system dedicated to tracking the health dimensions of the impacts of, and the response to, climate change. The Lancet Countdown tracks indicators across five domains: climate change impacts, exposures, and vulnerability; adaptation, planning, and resilience for health; mitigation actions and health co-benefits; finance and economics; and public and political engagement.
This report is the product of a collaboration of 27 leading academic institutions, the UN, and intergovernmental agencies from every continent. The report draws on world-class expertise from climate scientists, ecologists, mathematicians, geographers, engineers, energy, food, livestock, and transport experts, economists, social and political scientists, public health professionals, and doctors.
The Lancet Countdown’s work builds on decades of research in this field, and was first proposed in the 2015 Lancet Commission on health and climate change,1 which documented the human impacts of climate change and provided ten global recommendations to respond to this public health emergency and secure the public health benefits available.
There has been substantial focus on China’s influence in Africa in recent years. Some argue that China’s growing economic ties with African states have increased its political influence across the continent. This article examines whether trade with China leads African states to adopt more similar foreign policy preferences to China in the United Nations. We examine foreign policy similarity using voting patterns in the United Nations General Assembly and country statements in the United Nations General Debate. The analysis demonstrates that more trade with China has paradoxical effects on foreign policy positions of African states—it leads them to align more closely with US foreign policy positions in the United Nations, except on human rights votes. Our findings suggest that African states are engaged in balancing behavior with external powers whereby African elites seek to play off rival powers against one another in order to strengthen their own autonomy and maximize trade.
Advances in community detection reveal new insights into multiplex and multilayer networks. Less work, however, investigates the relationship between these communities and outcomes in social systems. We leverage these advances to shed light on the relationship between the cooperative mesostructure of the international system and the onset of interstate conflict. We detect communities based upon weaker signals of affinity expressed in United Nations votes and speeches, as well as stronger signals observed across multiple layers of bilateral cooperation. Communities of diplomatic affinity display an expected negative relationship with conflict onset. Ties in communities based upon observed cooperation, however, display no effect under a standard model specification and a positive relationship with conflict under an alternative specification. These results align with some extant hypotheses but also point to a paucity in our understanding of the relationship between community structure and behavioral outcomes in networks.
AI for SDG-16 on Peace, Justice, and Strong Institutions: Tracking Progress and Assessing Impact
(2019)
The transition from the Millennium Development Goals (MDGs) to the Sustainable Development Goals (SDGs) brought with it significant changes in the process of creating the goals and with the actual content of the SDGs. One of the most important developments was the inclusion of SDG 16, which recognises the central role of effective, accountable and inclusive political institutions in promoting sustainable development. Yet, a significant shortcoming is the difficulty in measuring progress on this SDG 16. In addition to general issues linked with data availability across the various indicators, a key challenge is aggregating trends across these wide-ranging indicators to track overall progress on SDG 16. A second issue that follows, is that despite claims regarding the centrality of SDG 16 for achieving the other SDGs, little is known about the causal pathways from the different indicators in SDG 16 to the other SDGs and associated indicators. In other words, questions remain over how changes in SDG 16 indicators impact a country’s progress towards indicators linked to health, gender equality, water and sanitation, and climate change.
Research on user satisfaction has increased substantially in recent years. Studies to date tend to test for significance of pre-defined factors thought to have an influence with no scalable means to verify the validity of the assumptions made. Digital technology has enabled new methods to collect user feedback, for example through online forums where service users post comments. Topic models can help analyze large volumes of such feedback and are proposed as a feasible solution to aggregate user opinions for use in the public sector. Insights can contribute to a more inclusive decision-making process in public services. This novel approach is applied to process reviews of publicly-funded primary care practices in England. Findings from the analysis of over 200,000 reviews indicate that the quality of interactions with staff and bureaucratic exigencies are the key drivers of user satisfaction. Moreover, patient satisfaction is strongly influenced by factors not considered in state-of-the-art patient surveys. These results highlight the potential benefits that text mining and machine learning for the public administration field.
In the context of recent economic and financial crisis in Europe, questions about the power of the core executive to control fiscal outcomes are more important than ever. Why are some governments more effective in controlling spending while others fall prey to excessive overspending by individual cabinet ministers? We approach this question by lifting the veil of collective cabinet responsibility and focusing on intra-cabinet decision-making around budgetary allocation. Using the contributions of individual cabinet members during budget debates in Ireland, we estimate their positions on a latent dimension that represents their relative levels of support or opposition to the cabinet leadership. We find some evidence that ministers who are close to the finance minister receive a larger budget share, but under worsening macro-economic conditions closeness to the prime minister is a better predictor for budget allocations. Our results highlight potential fragility of the fiscal authority delegation mechanism in adverse economic environment.
Improving Public Services by Mining Citizen Feedback: An Application of Natural Language Processing
(2020)
Research on user satisfaction has increased substantially in recent years. To date, most studies have tested the significance of pre‐defined factors thought to influence user satisfaction, with no scalable means of verifying the validity of their assumptions. Digital technology has created new methods of collecting user feedback where service users post comments. As topic models can analyze large volumes of feedback, they have been proposed as a feasible approach to aggregating user opinions. This novel approach has been applied to process reviews of primary‐care practices in England. Findings from an analysis of more than 200,000 reviews show that the quality of interactions with staff and bureaucratic exigencies are the key drivers of user satisfaction. In addition, patient satisfaction is strongly influenced by factors that are not measured by state‐of‐the‐art patient surveys. These results highlight the potential benefits of text mining and machine learning for public administration.
Rapid growth in smart meter installations has given rise to vast collections of data at a high time-resolution and down to an individual level. However, to enable efficient policy interventions, we need to be able to appropriately segment the population of users. The aim of this paper is to consider challenges and opportunities associated with large highly-granular temporal datasets that describe residential electricity consumption. In particular, the focus is on experiments relating to aggregation of smart meter time-series data in the context of clustering and prediction tasks that are often used for customer targeting and to gain insight on energy-use about sub populations. To cluster energy use profiles, we propose a novel framework based on a set of Gaussian based models which we use to encode individuals’ energy consumption over time. The dataset consists of half hourly electricity consumption records from smart meters of households in the UK (2014–2015). The contribution of this paper comes from its investigation of how consumers or groups may be clustered according to model parameters in scenarios where additional data on consumers is not available to the researcher, or where anonymity preservation of the smart meter user is prioritised. A secondary aim is to invite greater awareness when data reduction is required to reduce the size of a large dataset for computational purposes. This may have implications for policy interventions acting at the individual or small group level, for instance, when designing incentives to encourage energy efficient behaviour or when identifying fuel poor customers.