Program Development and Evaluation Program (PDE): This program seeks to support projects for the development and implementation of innovative approaches and tools to promote research integrity and prevent misconduct. This includes creating interventions based on empirical evidence, including findings from ORI-funded research, and tools to identify at-risk individuals or laboratories. This opportunity supports the creation and evaluation of interventions and assessment tools aimed at promoting research integrity. Funding Opportunity Number: IR-ORI-26-003
Institution: New Jersey Institute of Technology
PI: Cesar Bandera, PhD
Project Title: PaperBasket: a Multi-Layer Knowledge Graph Platform for Detecting Academic Misconduct
Abstract: This project expands PaperBasket.org, a system developed at the New Jersey Institute of Technology's Leir Research Institute that screens research papers before peer review by comparing them against more than 4 million published papers. The team will build a large network that maps how papers, authors, institutions, and journals connect, both through direct links such as citations and co-authorship and through shared topics and research interests. A set of AI tools will then scan this network automatically to spot unusual patterns that may signal misconduct, such as fabricated references, citation manipulation, plagiarism, or fake authors. Because unusual patterns are not always wrongdoing, the team will also create a system in which human reviewers and AI work together to examine flagged cases, rule out innocent explanations, and identify genuine problems. The project will test the system on real-world cases and publish policy recommendations on how institutions and publishers can use tools like PaperBasket to catch misconduct early.
Institution: University Of Illinois
PI: William Barley, PhD
Project Title: From Snapshot to Insight: Validating a Rapid Assessment Tool for Research Integrity Environments
Abstract: This project further develops the Rapid Environmental Adjective Descriptor (READ), a short survey that takes less than two minutes and captures how people perceive the research integrity climate where they work. Using existing survey data, the team will confirm that READ results match those of longer, established research climate surveys and test whether it can also measure how well research teams work together. The team will compare different ways of combining individual responses to see which gives the most accurate picture of a whole research group. They will then build a prototype tool that uses AI to turn READ responses into an easy-to-understand report for each team, and test it with active research teams to see how well it works in practice. The goal is a quick, low-burden early warning tool that helps research teams reflect on their culture and helps institutional leaders spot environments that need support, as well as those that can serve as models, before problems arise.
