Peter Hall is Professor of Statistics, The Australian National University and Chair of the National Committee for the Mathematical Sciences (Australia)

The following article, invited by a US professional newsmagazine, was initially written for an American audience.  It is published by TFW with permission of the author.

I wrote the first draft of this contribution during a short, extremely enjoyable visit to UC Davis, revelling in the intellectual riches that exemplify a Statistics Department in a strong US research university. I appreciate that in US universities the situation is not quite ideal; for example, recruitment of students and junior faculty is becoming increasingly challenging, and the level of funding is not as high as it might be. But the position of US academic Statistics today is pre-eminent, relative to that in the rest of the world, to a degree that could hardly have been anticipated twenty years ago.

Do not take this for granted. In the Australian experience, and also that of several other countries, the position of Statistics is unexpectedly fragile. In Australia, a serious shortage of both staff and funding has seriously curtailed teaching and research in statistics. Let me tell you a cautionary tale...

There is no dispute that Statistics is in decline in Australian universities. We are experiencing very significant difficulty filling vacant positions, and we shall continue to face this problem well into the future. Relatively few PhD students are attracted to the discipline, partly because of the attraction of better-paying opportunities in industry and in state and federal public service departments. (Nevertheless, all these areas report problems attracting staff.) These difficulties come at a time when increasing numbers of retirements are reducing senior staff numbers.

The challenges facing Statistics in Australia reflect, but are more serious than, those experienced across the mathematical sciences. Data gathered by the Australian Mathematical Society show that the number of mathematicians and statisticians working in Australian universities is today between 60 and 70% of what it
was in the mid 1990s. The number of BS (with Honours) mathematics and statistics graduates in the five-year period from 1997 to 2001 was only three-quarters of what it had been in the previous five years (Johnston, 2003). The number of Departments of Statistics remaining in Australia is currently only three; fifteen years ago there were three times that number, when there were fewer than half as many universities and when the demand by employers for trained statisticians was far less than at present. Today, the majority of university statisticians in Australia work in cognate or merged departments, such as Departments of Mathematics & Statistics.

The shortage of Statistics graduates in Australia is so acute that it inhibits foreign investment. The Bulletin, Australia's counterpart of Time magazine (and incorporating the version of Newsweek published in Australia) reported in 2002 that "a pharmaceutical...company's US once was considering shelving plans for growth in Australia because it couldn't recruit the senior statisticians that it needed to run clinical trials" (Bagnall, 2002).

The reasons for these difficulties are multifaceted and interacting, and as a result their combined impact is greater than one would expect from the sum of the individual contributions. However, a major cause is the decline in university funding. Expressed as a percentage of GDP, Australian federal government expenditure on higher education has fallen each year for the last eight, from 0.72% in 1996-7 to 0.53% in 2003-4. (Of Australia's 42 universities, 38 are federally funded.)

These contractions have contributed significantly to the substantial fragility of Australia's mathematics and statistics communities. Those in North America and Europe are more robust. Nations with stronger mathematical-science cultures than Australia often successfully supplement their depleting ranks of mathematicians and statisticians by drawing away our strongest (Thomas 2001, 2002). Funding arrangements for Australian universities make it difficult for them to accept foreign PhD students, and so Australia has little access to the new blood (much of it from Asia and Eastern Europe) which will shape US science for the next generation.

Of course, issues outside the university sector play a role in producing these outcomes. They include a widely-publicised shortfall in trained, skilled high-school mathematics teachers (see e.g. Bagnall, 2002). The Statistical Society of Australia is actively calling for better statistics education in schools. The similarity of Australian culture to that in the USA, Canada and the UK, where performance in mathematics and statistics is (on the whole) more highly valued, has made it relatively attractive for Australian statistical scientists to emigrate.

However, the catalyst that has multiplied the impact of these difficulties has been the indifference of the Australian government's science and education policies to the consequences of lack of investment in our profession. Along with a decline in the real value of total university budgets, noted above, mathematics and statistics continue to suffer a debilitating shortfall of revenue relative to other areas of scholarship. In 2005, under the Australian government's "Relative Funding Model," government financial support for a Statistics student will be only two-thirds of that for a student of computer science, although the costs of teaching (including staff salaries) will be virtually identical. The result will be a continuation of the decline in statistics teaching in Australian universities. Courses are increasingly directed towards a low common denominator of students, in order to maximise class sizes and recover expenses. Funding for research in Statistics is likewise very unsatisfactory.

The strategic importance of statistical science to Australia's future can hardly be overstated. Virtually all contemporary technologies, especially those in which Australia is endeavouring to gain a foothold (e.g. bioinformatics, and information & communications technology) have Statistics at their heart. Internationally, the need to increase quantitative skills across all scientific disciplines is being repeatedly endorsed. We frequently make reference to this point, in the hope of reversing current trends in Australia.

We draw encouragement, and considerable grist for our mill of persuasive argument, from the many reports that have appeared in the US in recent years. For example, the Australian government has been reminded more than once of the recommendations of the US National Research Council (Anon, 2003) on undergraduate training of future research biologists. Four of the eight recommendations call for increased involvement with the mathematical, physical and information sciences. Likewise, the reports and arguments that have led the NSF to increase its support of mathematics and statistics by an annual rate of 20% (see e.g. Tondeur 2001, 2003) have been pressed into service to show how inadequate is the view that some Australian authorities take of the need for training and research in Statistics.

The recent NSF report on the health, size and nature of the US Statistics profession (Kettenring et al., 2003) will find many uses of this type. As a resource to help mathematicians and statisticians prepare submissions to the Australian government, and to help them defend their own departments against further budget cuts, we have established a web page of useful links to documents of this type: http://wwwmaths.anu.edu.au/other/ncms/mathsdocs/

Tondeur (2001) observed that "public literacy in mathematics and statistics, and appreciation of their role in modern day life, are critical for societal progress." For all of us in the ASA, these words are hardly necessary. But for Australian politicians, bureaucrats and journalists the words require both emphasis and repetition. We are endeavouring to convince Australian authorities of the relatively high degree of acceptance which statistical science enjoys in the US, as part of a persuasive case that is being mounted, by many, for a better deal for Statistics in Australian universities.
 

Acknowledgements. I am grateful to Iain Johnstone, Hans-Georg Müller, Alex Reisner and Jane-Ling Wang for helpful comments.

REFERENCES

ANON (2003). BIO2010 | Transforming Undergraduate Education for Future Research Biologists. Committee on Undergraduate Biology Education to Prepare Research Scientists for the 21st Century. (Board of Life Sciences, Division of Earth and Life Studies, National Research Council of the National Academies.) National Academies Press, Washington, DC. Obtainable from: http://www.nap.edu/books/0309085357/html/

BAGNALL, D. (2002). The not-so-clever country. The Bulletin, 9 October. Obtainable from: http://bulletin.ninemsn.com.au/bulletin/EdDesk.nsf/All/156ED8FF8A71D681CA256C400025F9E0 All/156ED8FF8A71D681CA256C400025F9E0

JOHNSTON, P. (2003). Honours degrees in mathematics at Australian Universities, 1959-2001. Obtainable from: http://www.austms.org.au/AustMath/allhons.html

KETTENRING, J., LINDSAY, B. AND SIEGMUND, D. (2003). Statistics: Challenges and Opportunities for the Twenty-First Century. NSF report. Obtainable from: http://www.pnl.gov/scales/docs/nsf_report.pdf

THOMAS, J. (2001). Brain drain issue. Briefing paper, Aust. Math. Soc. Obtainable from: http://www.austms.org.au/AustMath/SMP_brain_drain_briefing.pdf

THOMAS, J. (2002). Mathematical Sciences in Australia: Still Looking for a Future. Briefing paper, Aust. Math. Soc. Obtainable from: http://www.austms.org.au/AustMath/braindrain_2002.pdf

TONDEUR, P. (2001). Opinion: NSF Mathematics Initiative. Not. Amer. Math. Soc. 48, 293. Obtainable from: http://www.ams.org/notices/200103/commentary.ps

TONDEUR, P. (2003). A view of the future of the mathematical sciences. European Math. Soc. Newsletter, September 2003, No. 49 p. 21. Obtainable from: http://www.emis.de/newsletter/current/current9.pdf