Research organizations, data repositories, and universities are beginning to assign persistent web-accessible identifiers to scientific resources. Assigning these persistent identifiers serves two crucial purposes. First, they are assumed to increase the traceability and reusability of scientific resources by allowing data sets, software, and other resources to be cited within the scientific literature. Second, without such traceability, it is difficult, if not impossible, to develop tools to measure the impact such resources have within the communities they belong to, or to understand the spread of that impact. Significant progress has been made in the past few years in the development of recommendations, policies, and procedures for creating and promoting these citable identifiers. While these efforts have raised the awareness of citation principles, which primarily address solutions to this issue moving forward, little progress has been made in tracking how data sets and other digital resources have actually been identified and cited in the past. This project focuses on developing efficient and extensible computational approaches to such attribution tracking. This effort will inform broader efforts to understand how scientific communities benefit from the use of scientific resources, facilities, and services.
Using a multifaceted experimental design, the project examines whether resources with persistent identifiers receive more attribution (through citations or acknowledgements) than those that do not. The research questions focus on the benefits of assigning persistent identifiers to scientific resources, namely, whether the overall citations of such resources increase, and whether the traceability of such citations of resources increased with resource persistent identifiers. The outcomes and contribution of this project center on providing insight into 1) whether human or computational approaches compare favorably in terms of their accuracy and efficiency in finding relevant references to scientific resources, 2) whether persistent identifiers provide quantifiable benefit in terms of increasing the amount a resource is referenced by users, and 3) whether persistent identifiers are being included in significant numbers in references to scientific resources.
