Fishing in the Bay

If I am reading the results of the ARC Discovery grants correctly, the profession of statistics has been absolutely decimated in this year’s round of grants. I have gone through the outcomes for Victoria and New South Wales, and searched on the primary RFCD code for “STATISTICS”. Some sobering facts follow.

For Victoria: I looked at all seven listed institutions. I could only find ONE project that we would ordinarily call statistics, Peter Hall’s. There are three others with the Statistics RFCD code but they really are not statistics. One is about stochastic differential equations and another is about stochastic modelling of gas flows - no doubt worthy projects but just not statistics as we normally use the term.

For NSW: There is not a single statistics project funded at the University of Sydney or Macquarie. At UTS there is a project on “boundary crossing analysis for random processes” which is stochastic processes, not statistics. The only statistics project I could find at UNSW concerns Bayesian methods for assessing climate change. So congratulations are very much due to Scott Sisson and Yanan Fan (thanks to David Warton for the correction).

This lack of success for research support for statistics as a subject stands in stark contrast to our recent success in having university funding for statistics students increased by 25%. It is also a fact that the Maths, Information and Communication Science (MICS) group had the same success rate as other subject areas. So our low numbers imply either (a) a low number of quality applications or (b) projects going to maths and info science instead of statistics within the MICS grouping.

There is clearly a success quota within each subject group. The success rates are way to close to each other (unless there are many thousands of applications in each areas). If you read through the results you will find that the vast bulk of grants in the MICS group have indeed been given to maths. For instance, at UNSW there are 6 Maths projects to 2 Statistics (one of which is probability). At University of Melbourne there are 9 Math projects to 2 Statistics (one of which is stochastic processes). At University of Sydney there are 5 Math projects to 0 statistics. This is all the more remarkable as University of Sydney are politically powerful at DEST - just look at the way they dominate the RQF panels.

It may be that there were very few statistics applications this year - it is not possible to geta list of all applications to check this. In any case, I really do not have the time to do this, nor to look at all institutions across the nation and compare results this year with the past few years. What is clear though is that this year’s lack of ARC success is incommensurate with the overall success and impact of academic statisticians in Australia.

The result is a debacle for the profession, especially for younger academics who need some research grants to kick start their career. It is clear that advances in statistical theory over the last 30 years have had almost immediate impact on the practice of quantitative analysis under uncertainty. Perhaps we need someone on the ARC panel who can make this case, and argue strongly for Statistics (by which I mean methodology, modelling and inferential theory) as opposed to mathematics and stochastic processes. Thoughts anyone?

You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.