Publications & Working Papers
"Evaluation Design Report" (T4D Project Team, 2015)
Abstract: This report outlines the mixed methods evaluation design by which the T4D intervention, aimed at improving citizen empowerment and maternal and neonatal health in rural Tanzanian and Indonesian communities, will be assessed. This includes randomized control trials (RCTs), extensive case studies, focus groups, key informant interviews, surveys and ethnographic methods, which enable T4D to process-trace exactly how the interventions triggered—or failed to trigger—improvements in health care and changes in power dynamics and community relations.
"Citizen Voices, Community Solutions: Designing Better Transparency and Accountability Approaches to Improve Health" (T4D Project Team, 2017)
Ash Center Occasional Papers
Abstract: The Transparency for Development (T4D) study was designed to answer the question of whether a community-led transparency and accountability program can improve health outcomes and community empowerment, and, if so, how and in what contexts. To answer this question, researchers and civil society organization partners began by co-designing a program that would activate community participation to address myriad barriers to proper maternal and newborn care, with the ultimate goal of improving maternal and newborn health outcomes. This report presents the design of the program that was then implemented in 200 villages in Tanzania and Indonesia and studied using a mixed methods impact evaluation. In addition to detailing the program, this report outlines how the project team got there—describing a number of principles that informed some distinguishing features of the program, as well as an iterative design process that defined other features through trial and error.
"T4D Phase 2 Concept" (T4D Project Team, 2016)
Abstract: Building off of early findings from Phase 1 of the T4D project, this concept outlines the team's approach to test adapted models of the intervention in new contexts to understand whether community-led transparency and accountability interventions can be designed to better facilitate community actions targeting higher-lever government actors. This brief outlines the process through which we developed these adapted designed and what we will be testing in Phase 2 of the project.
"Transparency for Development Intervention Design" (T4D Project Team, 2016)
Abstract: This working paper details the Transparency for Development Phase I intervention, focusing on the intervention design and components. This paper will pair with upcoming briefs that we are developing in the process of co-designing the intervention with partners in Tanzania and Indonesia as well as the guidelines and tradeoffs that the team faced in designing this intervention.
"Transparency for Development: Intervention Design and Evaluation Approach" (Kosack and Fung, 2013)
Abstract: This discussion paper outlines a recently started mixed method research program to assess whether, why, and in what contexts transparency and accountability (T/A) interventions improve health outcomes. The project is intended to advance the state of knowledge about the impact of T/A interventions on service delivery in several ways. First, we will develop a new T/A intervention, based on an initial theory about how T/A affects service delivery in different contexts as well as the range of existing work in this field, that is designed to be flexible enough to work across multiple contexts. Development of this intervention will leverage the local knowledge of embedded and capable local partners through an intensive co-design process. Second, we will conduct mixed-methods evaluations of this intervention in multiple sites: our research plan combines an initial phase in which we will evaluate its impact in two carefully selected sites, Tanzania and Indonesia, with both a randomized controlled trial and an integrated and extensive qualitative evaluation to understand mechanisms and the role of context; a cost-effectiveness analysis; and a second phase, which explores the generalizability of both our first phase results and the theoretical implications of them for mechanisms and the role of context. Third, these findings will provide the basis for development of a more nuanced and empirically grounded theory of the impact of T/A — both the different mechanisms through which T/A interventions seek to affect development outcomes, such as collaborative problem solving or community pressure, and the conditions under which these mechanisms should be expected to improve service delivery. All of these elements will be used to produce a range of new tools for practitioners of T/A as well as sectoral health experts assessing whether to add T/A interventions instead of or alongside more traditional health interventions.
“Does Transparency Improve Governance” (Kosack & Fung, 2014)
Stephen Kosack and Archon Fung (2014) “Does Transparency Improve Governance,” Annual Review of Political Science, vol. 17, pp. 65-87
Abstract: In recent years, there has been increasing interest in the potential of transparency—the provision of information to the public—to improve governance in both developed and developing societies. In this article, we characterize and assess the evolution of transparency from an end in itself to a tool for resolving increasingly practical concerns of governance and government performance. After delineating four distinct varieties of transparency, we focus on the type that has received the most rigorous empirical scrutiny from social scientists—so-called “transparency and accountability” (T/A) interventions intended to improve the quality of public services and governance in developing countries. T/A interventions have yielded mixed results: some are highly successful; others appear to have little impact. We develop a rubric of five ideal-typical “worlds” facing transparency that helps to account for this variation in outcomes. Reform based on transparency can face obstacles of collective action, political resistance, and long implementation chains. T/A interventions are more likely to succeed in contextual “worlds” with fewer of these obstacles. We find that 16 experimental evaluations of T/A interventions are largely consistent with the theoretical predictions of our five-worlds rubric.