Identifying “Driver Grants” to Accurately Analyse Time-Lag from Grant Funding to Public Knowledge

Academic research relies heavily on grants from various funding mechanisms, which are typically awarded with the expectation that researchers will disseminate their findings. However, the “time-lag” between receiving funding and the publication of research output remains unclear. This ambiguity makes it challenging to objectively evaluate the effectiveness of individual grants beyond the self-assessments provided by the investigators involved. A major complicating factor is that many researchers work on multiple projects simultaneously, making it difficult to establish clear causal links between specific grants and research outputs. In this study, we propose a novel, field-agnostic method to predict “driver grant” relationships to determine which grants were pivotal to the publication of specific papers or patents. We validate this approach using a qualitative survey completed by 482 researchers in Japan, achieving moderate overall accuracies of 0.755 for publications and 0.713 for patents. We applied this method to a large dataset encompassing nearly 250,000 awarded grants and over 1.76 million publications and 34,000 patents produced by researchers at prominent Japanese research universities. Through this analysis, we examine “time-lag” and the factors contributing to its variability. Our findings demonstrate that filtering for “driver grant”-output relationships improves time-lag estimation and may enable more objective evaluations of research grant effectiveness.