Measuring R&D knowledge diffusion through large databases

stock pic of hands, contract, on table and overlay of buildings

Combining SRI International’s expertise in machine learning and science policy to answer hard questions.


Knowledge diffusion is a key metric in measuring the impact of investments in science and technology. Increased attention has been paid to the evaluation of the outcomes of government-sponsored research.

Observable milestones in the life of a technological idea include:

  1. R&D funding
  2. patents, and
  3. licensing agreements.

Each of which is a measurable step in the maturation of knowledge derived from R&D.

Recovering connections between these milestones is difficult because there is no persistent identifier that is assigned when funding is awarded that then carries through to licensing. However, professionals can make connections between the steps by using unique identifiers such as award numbers and patent numbers.


SRI processed all quarterly filings made available by the Securities and Exchange Commission extending from the present back to 1995, going back to when public electronic records were first made available. The team extracted patent references from this database of more than 12 million records and linked those patents to federal government investments, showing the economic impact of government funding.

SRI worked closely with the National Center for Science and Engineering Statistics (NCSES) to refine their seminal biennial statistical report production process by taking a holistic approach to project development. It began with scoping and developing new indicators; benchmarking the production process against other statistical report processes; and developing best practices for producing a more interactive data visualization tool.

Read more from SRI