Assessing the Use of Natural Language Processing NLP in Reviewing STEM Research Grant Applications

A successful grant application is a crucial component in the development of a researcher. Multiple studies have found biases in grant reviews that favour certain groups of researchers. These attitudes reduce the quality of research outcomes and contribute to a cyclic disadvantage to underrepresented groups, particularly in STEM.
Current efforts to overcome this issue have proven insufficient as bias in our judgement seems deeply ingrained in human behaviour. NLP is a promising interdisciplinary field with successful applications in analysing text strings without introducing bias.
This project proposes to evaluate the feasibility and effectiveness of NLP algorithms in assessing grant applications by detecting text-based criteria that define the success of an application and identifies bias-related trends.
By developing a proof of concept on the use of NLP to assess grant applications, this project outsets the foundation for the development of a tool that assists founding entities in having a more inclusive and fair review system.

Getting a grant is really important for researchers to do their work. Unfortunately, there's a problem: some researchers have a better chance of getting grants than others, even if their ideas are equally good. This isn't fair and it stops some really smart people from doing important research. We think we can use computers to help fix this. Computers are getting really good at understanding language, so we want to teach them to read grant applications and find out what makes a good one. By doing this, we can spot if there's any unfairness in how grant applications are judged. Our goal is to create a tool that helps people who give out grants make fairer decisions. This will give everyone a better chance to get funding for their research ideas, no matter who they are.