NHMRC Public Consultations

Skip Navigation and go to Content
Visit NHMRC website

Current and Emerging Issues for- NHMRC Fellowship Schemes submission

ID: 
41
Step 2 - Personal Details
First Name: 
Adrian
Last Name: 
Barnett
Step 3 - A. Some questions for you
1. Which of the following best explains your interest in NHMRC’s fellowship schemes: (select ONE only): 
I am currently working in the health and medical research sector
2. If you are a health and medical researcher, which of the following descriptions best classifies your research? (select ONE only): 
Public health research – please complete the next two questions
3. If you are a health and medical researcher, which of the following best describes the main source of funding that supports your salary? (select ONE only): 
institutional funds (possibly derived from multiple sources).
Step 3 - B. Consultations questions
Question 1: How should NHMRC’s funding balance between research grants and fellowships be adjusted as the total number of Project Grants available falls progressively over the next few years?: 
I believe that more money could be directed towards fellowships instead of projects. “Fund people not projects” was the title of a 2011 Nature paper by John Ioannidis in which he examined ways to reduce the time that scientists spend writing grant applications instead of doing actual research. The principle is that good researchers have good ideas and funding them directly allows them to pursue their good ideas without the need for project plans that are usually lengthy and sometimes conservative. Importantly it may be easier to reliably fund researchers than projects. Evidence for this comes from the NHMRC’s own Early Career Fellowships scheme where (with thanks to the NHMRC) we were able to examine the agreement in funding using a randomised trial (Clarke et al 2015). The agreement in funding decisions between two independent panels was 83% (95% confidence interval 73 to 92%). This was statistically significantly higher than a meta-analysis of the three previous studies of the reliability of funding which had a mean agreement of 72% (95% confidence interval 67 to 76%). Similarly we found an agreement of 74% between the official NHMRC panels and our simplified panels when reviewing projects (paper under review). Hence there is evidence that the task of selecting good researchers is more reliably completed than the task of selecting good projects. This may be because projects are based on future work, which is unpredictable, whereas fellowships are predominantly based on past achievements which are more concrete. As Donald Braben says, using peer review to fund projects should be called “peer preview” as peers are faced with the difficult task of predicting what research will be successful (Braben 2014). Five-year Project Grants The McKeon review recommended an increase in five-year Project Grants in order to create, “productivity increases both through less staff turnover and less time spent on grant application and administration matters” (McKeon et al 2013). In 2014 the NHMRC encouraged the submission of more five year grants and there was a relatively large increase in five-year grants. These grants generally cost more which reduces the overall Project Grant success rate. To keep success rates high more money would need to be directed to the Project Grant pool at the cost of the Fellowship pool. We recommend that the NHMRC uses its internal data to verify that researchers winning five-year Project Grants are actually submitting fewer applications. This could be done by comparing the subsequent application numbers of winners of five-year grants with winners of non-five-year grants. If the numbers are not decreasing then we strongly recommend that this policy is reversed. There is anecdotal evidence that researchers who win big grants are strongly encouraged by their institution to apply for more, so there is the possibility that the five-year policy is not having the desired effect of decreasing overall application times. A study from Canada found that small grants to many researchers was a better tactic for increasing overall scientific impact compared with the strategy of larger grants to fewer researchers (Fortin and Currie 2013). This provides evidence that the NHMRC should award more lower cost grants. The NHMRC could use their internal data (assembled for the 2013 “Measuring Up” report) to repeat the study of Fortin and Currie and check if this association holds in Australia. References • Braben D. Promoting the Planck Club: How Defiant Youth, Irreverent Researchers and Liberated Universities Can Foster Prosperity Indefinitely. 2014 Wiley. • Clarke P, et al. A high reliability of grant funding decisions in fellowships for early career researchers: evidence from a randomised trial. JCE 2015 in press • Fortin J-M, Currie DJ. Big Science vs. Little Science: How Scientific Impact Scales with Funding. PLoS ONE 2013;8(6):e65263. • Ioannidis JPA. Fund people not projects. Nature 2011;477:529–531.
Question 4: Noting the implications outlined in the Issues paper, should NHMRC extend the duration of Early Career Fellowships to more than 4 years and Career Development Fellowships beyond 5 years (to 7 or 10 years)?: 
Increasing the duration of the fellowships will increase their value and hence is very likely to increase the number of applications received. At the same time increasing duration will also decrease success rates. So this approach will fund fewer researchers with larger amounts. As noted in my response to question 1, one of the only scientific studies we have found that more smaller grants gave a better scientific return than fewer larger grants.
Question 8: Would this be achieved if NHMRC required institutions to commit to one or more years of ongoing support for researchers exiting from NHMRC Fellowships? : 
This would seem to be a progressive step. Research institutes benefit from their NHMRC fellows, and it therefore seems appropriate that they support their researchers in return. However, this policy would also increase the value of fellowships as researchers would win an even greater prize. This may increase the number of applications that the NHMRC receives. Our recent analysis of the impact of the streamlining of Project Grant applications found that the time spent applying actually increased after the streamlining (Barnett et al 2015). One theory for this surprising increase is that the time researchers allot to applying is based on the expected economic return rather than the length of the application form. If this theory is true then this policy is very likely to increase application numbers. Reference • Barnett AG, et al. The impact of a streamlined funding application process on application time: two cross-sectional surveys of Australian researchers.BMJ Open 2015;5:e006912

Page reviewed: 28 January, 2016