Did the switch to using partial randomisation at The British Academy change the characteristics of applicants?
Rhys Thomas
(University of Oxford)
Adrian Barnett
(Queensland University of Technology)
Ken Emond
(The British Academy)
Short abstract
This paper examines the impact of the British Academy's switch to a conditional lottery for allocating research funding in its Small Grant Scheme. Findings show a significant overall increase in applicants, with a disproportionately higher rise among Asian and British Asian applicants.
Long abstract
In 2022, the British Academy (BA) began trialling a conditional lottery to allocate research funding through its Small Grant Scheme. The aim was to determine whether using a lottery-based allocation system could reduce biases and, in turn, increase the diversity of grant recipients.
This paper examines the impact of replacing a traditional peer review system for awarding small grants with a system in which grants are allocated randomly, provided they meet a minimum quality threshold. Specifically, it analyses how partial randomisation influences applicants' characteristics.
Using data from the BA on applicants to the Small Grant Scheme, Postdoctoral Fellowship, and Mid-Career Fellowship—both before and after the introduction of partial randomisation—this paper employs a difference-in-differences design to assess whether applicant characteristics changed following the shift to a lottery-based allocation. We examine the number and proportion of applicants across various demographic and institutional characteristics.
Our findings indicate a substantial overall increase in the number of applicants, with approximately 500 additional individuals applying to the scheme each year. The number of applicants increased across all demographic categories analysed and rose in both Russell Group and non-Russell Group universities. However, there was a disproportionate rise in the number of Asian and British Asian applicants, while the number of white applicants declined disproportionately. These results have important implications for the role of partial randomisation in promoting ethnic diversity in research funding. They also highlight the potential for broadening the "diversity of ideas" in research by expanding the applicant pool.
Accepted Paper
Short abstract
Long abstract
In 2022, the British Academy (BA) began trialling a conditional lottery to allocate research funding through its Small Grant Scheme. The aim was to determine whether using a lottery-based allocation system could reduce biases and, in turn, increase the diversity of grant recipients.
This paper examines the impact of replacing a traditional peer review system for awarding small grants with a system in which grants are allocated randomly, provided they meet a minimum quality threshold. Specifically, it analyses how partial randomisation influences applicants' characteristics.
Using data from the BA on applicants to the Small Grant Scheme, Postdoctoral Fellowship, and Mid-Career Fellowship—both before and after the introduction of partial randomisation—this paper employs a difference-in-differences design to assess whether applicant characteristics changed following the shift to a lottery-based allocation. We examine the number and proportion of applicants across various demographic and institutional characteristics.
Our findings indicate a substantial overall increase in the number of applicants, with approximately 500 additional individuals applying to the scheme each year. The number of applicants increased across all demographic categories analysed and rose in both Russell Group and non-Russell Group universities. However, there was a disproportionate rise in the number of Asian and British Asian applicants, while the number of white applicants declined disproportionately. These results have important implications for the role of partial randomisation in promoting ethnic diversity in research funding. They also highlight the potential for broadening the "diversity of ideas" in research by expanding the applicant pool.
Metascience Lab (II): brokering experiments
Session 1 Tuesday 1 July, 2025, -