While the cumulative nature of knowledge is recognized as central to economic growth, the microeconomic and institutional foundations of cumulativeness are less understood. Although "Open Science" is widely recognized to play a central role in the production and diffusion of fundamental knowledge, few formal analyses support this understanding of the impact of policies and practices on the rate and direction of scientific progress. This study focuses on the development and implementation of novel tools for quantitative analysis of the impact of science policy interventions on the process of cumulative scientific discovery. This analysis extends prior research by exploiting the recent availability of detailed citation data with frontier methods from the program evaluation literature. The approach moves beyond traditional cross-sectional comparisons of citations associated with knowledge in different institutional or policy environments; instead, "natural experiments" are utilized, where the conditions governing access, diffusion or follow-on research funding associated with a given piece of knowledge are changing over time. This approach allows the role of institutions and policy to be disentangled in shaping scientific progress from the intrinsic variation in scientific importance across discoveries. Specifically, the study outlines three types of tools: a differences-in-differences approach to citation analysis, the explicit comparison of changes in citation behavior in different subpopulations, and the development of an approach to recover the "distribution" of the impact of policy interventions. These tools can be fruitfully applied to provide novel policy analysis for a range of science policy interventions, from choices about the level of (and restrictions on) public funding, rules governing access to scientific research materials and data, and policies regarding intellectual property rights for discoveries resulting from the scientific process. In particular, the tools allow for the evaluation of science policy intervention on the rate and direction of scientific progress, and allows for the evaluation of the distributional consequences of policy initiatives. There are three potential applications of the tools in some detail, including (a) the impact of intellectual property rights on the diffusion and use of academic science, (b) the impact of national science policies on the geography and distribution of stem cell research, and (c) the impact of institutions that facilitate the sharing of research resources on the dynamics of knowledge accumulation in life sciences research. In sum, the outputs of the research will include papers that focus on the development of the tools per se, and papers that apply the tools in the context of applications focused on important science policy challenges. This study will inform: (a) science policy analysis, by developing and implementing tools that can assess the impact of specific policies and institutions on the rate, direction, and composition of scientific activities; (b) the study of the economics of science, by elaborating the microeconomic and institutional foundations of knowledge accumulation, which supports economic growth; and (c) the sociology of science, by highlighting the roles of preexisting relationships, status, and networks in the expansion of the scientific community and explaining interactions between the features of the scientific system and its growth. More generally, the broader impact of this study is to provide a set of tools for a range of science policy questions that have so far resisted quantitative analysis, and to offer a novel domain for the application and adaptation of frontier methods for program evaluation. In particular, the results of this research will have a specific impact in the economics and sociology of science, as well as in science policy analysis per se. In addition, the study provides a bridge between the explosion of quantitative data about science as a potential new area for the application of tools in the program evaluation literature.
While the cumulative nature of knowledge has long been recognized as central to economic growth, the microeconomic and institutional foundations of cumulativeness have been less-well understood. This project was designed to address the dearth of systematic, large scale, quantitative analyses of the impact of policies and practices on the rate and direction of scientific progress. We aimed both to develop tools that could be used to analyze the impact of scientific institutions and policies on the rate and direction of progress and to apply these tools in a number of important contexts. The tools we develop rely on a number of features of the scientific landscape: (1) academic publications and citations help trace the extent and importance of scientific progress, (2) in a number of important situations, science policies and institutional environments change in ways that were unanticipated by the scientists affected by these policies, and (3) we can use these data and these changes to estimate the impact of science policies and institutions on the rate and direction of scientific progress. An additional feature of our research is that our projects are undertaken with an eye towards understanding the nature and drivers of knowledge accumulation – the long-term accretion of knowledge on which economic growth depends. Specifically, our tools rely on the assumption that academic publications and the citations to academic publications, known as "bibliometric measures," can provide useful insights into the rate and direction of knowledge creation and accumulation. Bibliometric data are particularly helpful measures, as they contain detailed information on individual scholars, their institutions, the geographic locations in which knowledge is created, and specific information on the academic fields and sub-fields to which they are relevant. We exploit these details in our analyses. We combine these measures with research designs and recently-advanced statistical (econometric) techniques that enable identification of causal effects. Rather than investigating correlations in the data, these techniques exploit unexpected "shocks" that act similarly to controlled experiments to enable us to see the impact (rather than just the association) between specific policies or institutions and the outcomes associated with them. For example, by carefully choosing a control sample and controlling for a number of key factors, we are able to identify the impact of the U.S. Federal Human Embryonic Stem Cell (hESC) policy on the competitiveness of U.S. hESC research relative to that of the rest-of-the-world. In this grant, we have developed and applied these techniques in a number of contexts and to address a number of questions, including: the role of Biological Resource Centers in affecting the rate and direction of follow-on research in the life sciences the role of Openness in the rate and direction of research involving transgenic mice the drivers of mistakes and fraud (retracted articles) in scientific literature the way in which changes in the enforcement of intellectual property rights shape patterns of knowledge exchange and accumulation within knowledge communities the impact of the U.S. administration policy with regards to funding human embryonic stem cell research on the rate and direction of follow-on research the impact of IP restrictions on follow-on research associated with the human genome project the conditions supporting the disclosure of privately funded new knowledge through scientific publication, patenting, or both in the human genome Each of these lines of inquiry has borne fruit and we have published more than 10 academic articles on these topics, delivered more than two dozen academic and public policy presentations on these issues, and have even testified before Congress on these and related themes. Taken together, this research has clarified the importance of open access to scientific materials and intermediate for supporting knowledge accumulation, demonstrated that lifting intellectual property protections over scientific research inputs leads to a profound acceleration in knowledge generation, clarified the impact of a nuanced Federal policy governing Human Embryonic Stem Cell (hESC) on the nature of U.S. competitiveness in this scientific arena, and examined the effectiveness of a key institution governing scientific publication (the system of retractions) on the progress of knowledge. Our findings have specific implications for each of the policy-oriented issues and institutions we examine. Further, our analyses point the way for future examinations of science policies and institutions and our techniques have been adopted by numerous researchers who are making progress in these areas.
