Message from the President

The Future of the ISI

As we prepare for what promises to be a splendid Biennial Session in Sydney, April 5-12, 2005, I wish to consider the past evolution of the ISI and report on steps being taken to assure us a bright and exciting future. Over its long history, the ISI has enjoyed periods of remarkable stability separated by times of well-considered and significant change. We stand now at a threshold of potentially far-reaching change. It may be useful to review how the ISI approached such periods in the past.

Our Origin

In 1853, Adolphe Quetelet convened in Brussels the first of a series of International Statistical Congresses. There were eventually nine Congresses in all, and at their best they were energetic and productive exchanges among the leading statisticians of the day. However, the Congresses lacked an adequate continuing organization or continuing membership, and they barely outlasted the death of Quetelet in 1874. In 1885, nine years after the last of the Congresses, the ISI was founded at a meeting in London called to celebrate the Jubilee of the London Statistical Society. The 81 members, who joined in London or were invited to join by that initial group, were the elite of that era’s statisticians in government and academia, and they sought to recapture the energy of the Congresses but with a well-established continuing structure and a stable membership base. They established our first statutes, and our first half-century was a period of general stability. There was some growth in membership and publications, but major changes, such as an affiliation with the League of Nations that was proposed in 1920, were resisted.

Post-War Reorganization

The September 1938 Session in Prague was cancelled in its second day because of the threat of war, and the ISI essentially went into hibernation. We were wakened with the 1947 Session in New York, principally organized by Stuart A. Rice with an ambitious goal of adapting the ISI to a new era. New statutes were presented to the General Assembly and adopted by mail the following year, with this charge by Rice:

The ISI of the future must be regarded as more embracing than the single society of elected members that we have been in the past. … On every hand there is a new dependence upon statistics and statisticians. … There is a crying need for world leadership in this field. … Such leadership will require the revitalization of our membership. … It requires that we burst the bonds of our present statutes which restrict the capacity of our organization for evolution, growth and adaptation to the period in which we live.

The new statutes brought two fundamental changes. Before the war, the ISI had sought to influence governmental statistical agencies by facilitating collaboration and by encouraging uniformity in statistical definitions and data collection, but this role was largely taken over by the newly created United Nations. Now the ISI took on a new mission, emphasizing international communication among statisticians rather than with governments, and supporting the international promotion and dissemination of research on the theory and practice of statistics. The second major change was in the introduction of the idea of Sections of the ISI, and the formation of several Committees that in many cases lead to the creation of a Section. While the first Section was only formally adopted by the General Assembly in 1957, the fruits of this far-reaching innovation are clearly evident today with the five Sections we now have, all of which grew from this post-war reorganization. For the first time, the ISI saw itself as an umbrella organization reaching well beyond its narrow base of elected members. It was a new type of professional society, one that facilitated international communication among groups of individuals with common interests, not all of them members of the ISI. The ISI was to be the organizational key to international statistics; it did not pretend to comprise all of international statistics itself.

The Future Directions of 1979

In 1977, the General Assembly set up a Committee on Future Directions that led in 1979 to the adoption of further organizational changes, including setting up our current system of governance, with an elected Council and an Executive Committee consisting of the President, President-Elect, and the three Vice-Presidents. With the formation of the Council, the Presidents of the Sections now had a strong voice in ISI governance. The Future Directions Committee also carefully considered a series of more far-reaching changes, notably opening membership in ISI to any and all statisticians who may wish to join. However in the end, they rejected this proposal and instead reaffirmed what they called the ISI’s status as an Academy, with restricted and elected membership. They argued that the careful and conscientious election of the world’s best statisticians as members was key to the high international reputation of the ISI, and they considered that reputation in turn key to other developments they envisioned, such as founding at the ISI an International Center for Research and Development in Statistics, and actively helping with the education and training of statisticians in developing countries.

The Present

It is the consensus of the Executive Committee and of the Council that the time is right to again re-examine the programmatic emphases of the ISI. In light of our experiences over the two decades since our 1985 Centennial and the new challenges of the 21^st century, we ask if we should again make an evolutionary change in the ISI and if so, what change? Accordingly, in August 2004, the Council approved the appointment of a new Ad Hoc Strategic Planning Committee, charged with considering these questions and reporting to the Council in Sydney. The members of this Committee are C. Carson, D. Dawson, J. Deshpande, N. Fisher (Chair), N. Keiding, D. Moore, S. Richardson and D. Trewin. The Committee has begun its work and is considering a full range of questions, some as basic as supporting an enhanced role for the Sections and revisiting the possibility of open membership in ISI, and others as practical as how to best assure our financial future. Their recommendations are not yet ready at this writing, but it is expected that they will be circulated in advance of the Sydney Session so that we can have an initial discussion at the General Assembly. Some recommendations may be ready for adoption; others will surely require deliberation and refinement beyond Sydney, perhaps by successor committees. For my part, I am confident that whatever evolutionary changes emerge from this process will preserve what I regard as our central value: the ISI as facilitator of international communication in statistics, across nations and between sub-disciplines. We have played a uniquely successful role in this for 120 years. The keys to our success in these have been our Biennial Sessions and our flagship publications, augmented since 1948 by a network of Sections and Committees. Also, I expect that our Sessions and publications will continue to play their important roles in keeping us in touch as one organization, while we expand our efforts to reach out through Sections and Committees in new ways. Maintaining this communication across international statistics will be crucial to avoiding fragmentation as we also expand our outreach. I am convinced that we shall be equal to this, our central challenge, and that our future will be bright and prosperous.

Stephen M. Stigler
President ISI

Message from the Director

The ISI Executive Committee and Permanent Office finished the year off with a flurry of activity, culminating in a meeting of the ISI Executive Committee on December 16-17 in Voorburg, at which time a broad range of issues were discussed. As this was the last EC Meeting that will take place before the ISI Sydney Session, I would like to thank each member of the ISI Executive Committee for the considerable investment that they have made of their time and energy, as well as for their creative and insightful contributions, particularly given the difficult decisions that it has been necessary to make during this ‘transition phase’ in the ISI’s evolutionary progression. I would also like to thank all Council members for their helpful input over the course of the last 1½ years, actively helping to guide us through the various issues that have confronted the organisation.

[From left to right: President Stephen Stigler, President-Elect Niels Keiding, Vice-Presidents Pilar Martín-Guzmán, Jae C. Lee & Nicholas Fisher]

We welcome the 28 new ISI members who were elected during the second round of the 2004 ISI membership elections. The names of these new members are listed here. In 2004, a total of 135 new members were elected, a record annual amount.

As announced in the ISI Newsletter Volume 28, No. 1 (82), while we are grateful to all standing members who nominate potential new ISI members, as a token of our appreciation to the three ISI members who have submitted the largest amount of ISI membership nomination applications in 2004, we have awarded special ISI watches to Professors Jong Hoo Choi, Myoungshic Jhun and Jae Chang Lee. Many thanks for your efforts! In 2005, we would like to exceed the number of new ISI memberships that we registered in 2004 – please submit your nominations using the electronic aids available at http://isi.cbs.nl/membership.htm.

Looking forward, I would like to convey my best wishes to the membership for the New Year. There is no doubt that the highlight of 2005 will be the ISI Sydney Session (April 5-12), and I hope to meet and greet as many members as possible during that occasion. This issue of the Newsletter includes some important Sydney Session details, including the Schedule of Administrative Meetings that will take place during the Session.

In advance of previous ISI Sessions, the Report of the ISI Executive Committee to the General Assembly is typically sent to all ISI members describing ISI/Section activities during the past two years, including a brief financial overview of the organisation. As a cost and time saving measure, rather than distributing this Report to all members by post, it will be made publicly available via the ISI website at the beginning of March 2005 and can be accessed at our website via http://isi.cbs.nl/05session/05report.htm. A limited number of copies of the Report will be made available at the ISI General Assembly on April 11, 2005.

As I announced in the previous ISI Newsletter, financial considerations have made it necessary for the ISI Permanent Office to discontinue production of the publication Statistical Theory and Method Abstracts (STMA). I would like to thank STMA General Editors Constance van Eeden, Klaas van Harn and Bert van Es for their work on STMA, as well as the dedicated team of Regional Editors. I would also like to thank ISI staff members Ann Daniels and Tineke de Boer for their excellent support in the production of STMA over the years. While STMA, in its present form, will remain accessible to ISI/Section members for the remainder of 2005 via Internet (please contact Mr. Sieriel Hoesenie at @cbs.nl to obtain your personalised access password), and although STMA production via the ISI Permanent Office has now been discontinued, I am pleased to announce that the “Phoenix bird” will fly again. The ISI has recently completed an arrangement with the electronic publication Zentralblatt (see http://www.zblmath.fiz-karlsruhe.de/MATH/about/zentralblatt), allowing STMA production and dissemination to continue, albeit at a new home, now as a sub-component of the Zentralblatt database. Subscription details for 2005 are indicated here.

The ISI flagship publication International Statistical Review will witness a change of editorial direction after the Sydney Session. We are grateful to outgoing Editor Ms. Asta Manninen for her hard work during her four-year term as Co-Editor. I am delighted to announce that Professor Jarig van Sinderen will assume the role as International Statistical Review Co-Editor together with continuing Editor Professor Eugene Seneta.

Professor Jarig van Sinderen

I am pleased to announce that the second Mahalanobis Prize, established by the Indian Government through its Ministry of Statistics and Programme Implementation to commemorate P.C. Mahalanobis, will be handed out during the ISI Sydney Session General Assembly on April 11, 2005, to Professor Ben Kiregyera (Uganda) in recognition of his lifetime achievements in statistics, particularly in developing countries. We are grateful to the ISI Mahalanobis Committee selection jury, Chaired by Mr. Jean-Louis Bodin.

The Jan Tinbergen Prize selection jury, Chaired by Professor Willem Albers, has selected the following persons as winners of the 2005 Jan Tinbergen Prizes: Mr. Romain Glèlè Kakaï (Benin), Mrs. Marina Y. Ogay (Ukraine) and Ms. Kavitha Bhat (India). In addition to a cash prize, generously provided by the “Stichting International Studiefonds”, these three individuals will receive support to enable their participation at the Sydney Session and present their winning papers at a special Invited Paper Meeting (IP2) on Monday, April 11, at 13:00 hours. We are grateful to the Jan Tinbergen jury for their efforts in making this selection.

The ISI Service Certificates Committee, Chaired by Dr. Ivan Fellegi, has announced the recipients of the ISI Service Certificates, a token of the ISI's appreciation of the many years of service 'above and beyond the call of duty' that these select individuals have made to the association: Yadolah Dodge (Iran/Switzerland), Jens Ledet Jensen (Denmark), Asta Manninen (Finland), Jozef L. Teugels (Belgium), Sara van de Geer (The Netherlands).

Please join us at the ISI General Assembly in Sydney in applauding these five deserving ISI Service Certificate recipients.

ISI Service Certificate Recipients:

Professor Yadolah Dodge

Professor Jens Ledet Jensen

Ms. Asta Manninen

Professor Jozef L. Teugels

Dr. Sara van de Geer

We are proud to learn that the average number of monthly ‘hits’ to the ISI website has recently exceeded 16,459. The ISI Permanent Office is continuing its efforts to expand the amount of information available on the ISI website. We have recently added an ISI Glossary of Statistical Terms sub-site, containing a large number of statistical terms in twenty-one different languages. The Glossary is available via http://isi.cbs.nl/glossary/index.htm and has been revised with many new terms. This site will expand to include additional languages as they become available. We welcome your suggestions for new terms or descriptions of existing terms via http://isi.cbs.nl/glossary/remarks.htm. We hope that you will take advantage of this complimentary service, and access the wealth of information that the Glossary contributors have kindly submitted.

Also, please do not forget to consult the 'Career Opportunities' sub-site at http://isi.cbs.nl/AD.htm, either to obtain information about career listings, or to add a listing on behalf of your own organisation at no charge.

Interested in making a submission to the 56^th ISI Session (Lisboa, Portugal) Invited Papers Programme? This can now be done via http://isi.cbs.nl/07session/56thsessionInvitedPapers.htm.

Daniel Berze
Director ISI 

55^th ISI Session, Sydney

5 - 12 April 2005 — Register Now!

It is not every year that you will have a business reason to visit Australia. This year the 2005 Session of the International Statistical Institute (ISI) will be held in Sydney, Australia, from the 5^th to 12^th of April.

If you are intending to participate, it is highly recommended that you register now. To register, please complete the online registration form at www.tourhosts.com.au/isi2005 or return the Registration Form in Bulletin II to the Conference Managers. Even if you are undecided as to whether to attend or not, you should register your interest to ensure you are kept informed about the Session’s developments.

If you have already registered for the Session, please continue to check the Session website www.tourhosts.com.au/isi2005 on a regular basis for programme updates and conference news.

Key Dates

After 31 January 2005 – Late registration fee applies

4 March 2005 – Speakers to e-mail their presentation to the Conference Managers

4 April 2005 – Registration for the Session commences

5 April 2005 – Session Opens

12 April 2005 – Session Closes

Scientific Programme

The Scientific Programme for the ISI Session will feature leading keynote speakers from around the world and more than 100 scientific Sessions.

Geoff Lee, Local Programme Committee Chair and Head of the Australian Bureau of Statistics’ Methodology Division said ''… the Invited Paper Programme is shaping up really well. The opportunity to attend a Session of the ISI in Australia is a once in a lifetime opportunity”.

The Scientific Programme will be supplemented with tutorials and short courses. Special theme days will cater for those with interests in finance and statistics, environmental statistics and genomics.

Details on the Scientific Programme, including the full list of Invited and Contributed Paper Meetings, are listed on the Session website at www.tourhosts.com.au/isi2005.

Keynote Speakers Confirmed

Renowned mathematical biologist Lord Robert May, econometrician Professor Clive Granger and Deputy Governor of the Reserve Bank of Australia Glenn Stevens have been confirmed as key speakers at the 2005 Session of the International Statistical Institute (ISI) scheduled for Sydney next April 5 - 12.

Lord May, an Australian by birth, obtained his doctorate in theoretical physics from the University of Sydney in 1959 at the age of 23. He is a world authority on mathematical biology.

In 2000, he was appointed for five years as President of the Royal Society of London, a position with a rich tradition and one of the most esteemed in the world of science. That followed a five-year appointment as the Chief Scientific Adviser to the British Government and Head of the Office of Science and Technology, playing an influential role in national scientific affairs.

Lord May holds a Royal Society Professorship jointly in the Department of Zoology, Oxford University, and at Imperial College, London, and is a Fellow of Merton College, Oxford.

Clive Granger shared the 2003 Nobel Prize in Economic Sciences with Robert Engle for their discoveries in the analysis of time series data. The work has fundamentally changed the way that economists think about financial and macro-economic data, and has led to significant breakthroughs in Statistics and Macro-economic forecasting.

Professor Granger is also noted for developing a formal statistical notion of causality based on which variables help to predict other variables. His discovery is widely used and is commonly known as "Granger causality". He is now Professor Emeritus at the University of California, San Diego (UCSD).

Glenn Stevens has been the Deputy Governor of the Reserve Bank of Australia (RBA) since December 2001 and has spent most of his professional career in RBA, joining the bank’s Research Department in 1980. He holds degrees in Economics from the University of Sydney and the University of Western Ontario, Canada. In 1990, he was a Visiting Scholar at the Federal Reserve Bank of San Francisco. He was Head of the Economic Analysis Department of the Reserve Bank from August 1992 to September 1995, and head of its International Department from September 1995 to December 1996. Since December 1996, he has been Assistant Governor (Economic), responsible for overseeing the economic analysis and research of the Bank’s staff and formulating policy advice for the Governor and the Board of the Bank. Glenn is currently a Member of Advisory Boards for the Melbourne Institute of Economic and Social Research and the Hong Kong Institute for Monetary Research.

Satellite Meetings

A number of Satellite Meetings will be held before or after the 2005 ISI Session. By combining the ISI Session with one of the Satellite Meetings, you may be able to get greater value from the time you spend travelling to and from Australia. Details and links for each Meeting are listed below:

31 March-2 April 2005
Issues for Official Statistics for Small Countries (Especially Island Nations) Nouméa, New Caledonia
 www.stat.fi/iaos/future_activities.html

29 March-1 April 2005
14^th International Workshop on Matrices and Statistics, Auckland, New Zealand http://iwms2005.massey.ac.nz

4- 5 April 2005
Statistics Education and the Communication of Statistics, Sydney, Australia
 http://www.stat.auckland.ac.nz/~iase/conferences.php?show=iase2005

13-16 April 2005
Fourth International Symposium on Business and Industry Finance, Near Cairns, Queensland, Australia
www.action-m.com/isbis4

14-15 April 2005
Measuring Small Populations, Wellington, New Zealand www.stats.govt.nz/ISIsatellitemeeting

Social Programme

The Social Programme will be a highlight of the Session and has been designed to provide participants with an opportunity to relax, experience Sydney and maximise networking opportunities. The following events are included in the registration fee for delegates and accompanying persons:

Passports & Visa Information

Delegates are reminded to check their visa arrangements are in place prior to coming to Sydney. All travellers to Australia, other than those holding New Zealand passports, require a valid entry visa prior to travel to Australia. For many nationalities, an Electronic Visa or Electronic Travel Authority (ETA) can be obtained from your travel agent or airline when making your travel arrangements. For more visa information, and to download the Visa Guide, please visit
http://www.tourhosts.com.au/isi2005/ginfo.asp.

Alternatively, please see your travel agent or airline for further details.

Contact Details
ISI 2005 Conference Managers
GPO Box 128
Sydney NSW 2001
Australia
Telephone: +61 02 9248 0800
Fax: +61 2 9248 0800
E-mail: isi2005@tourhosts.com.au
Website: www.tourhosts.com.au/isi2005

Have you seen Information Bulletin II? Information Bulletin II provides the latest details on the arrangements for the 2005 ISI Session and the final registration form. A copy is available on the ISI website www.tourhosts.com.au/isi2005 . To request a hard copy, please e-mail isi2005@tourhosts.com.au .

Please visit http://www.tourhosts.com.au/isi2005/programs.asp for the most recent Administrative Meetings & Scientific Programme Schedules!

Top

News of Members

	ISI elected member Professor Arjun K. Gupta, Distinguished University Professor and Professor of Mathematics and Statistics, Bowling Green State University, has been awarded the Distinguished Service Award. The Award has been presented by the Faculty Senate in recognition of his continuing quality contributions to the University.
	The Dean of the Faculty of Mathematics and Physics at Charles University in Prague has awarded Professor Carlo N. Lauro, ISI elected member and IASC Section member, the Memorial Medal of the Faculty on the occasion of the 16th Symposium of IASC COMPSTAT 2004. This Award is in recognition of Professor Lauro’s “distinguished and wide-ranging scientific work…”, as well as his “pioneering activities that helped to establish statistical computing as an important part of both theoretical and applied statistics”. Professor Lauro was President of the IASC from 1993-1995 and Vice-President of the ISI from 1997-1999.

Top

In Memoriam

Norman Lloyd Johnson, Alumni Distinguished Professor Emeritus of Statistics at the University of North Carolina at Chapel Hill and ISI elected member since 1974, died on November 18, 2004. With his death, we lost the last direct link between the founder of the British Statistical School in the early days of the twentieth century (Karl Pearson) and the maturest Statistical Sciences of the late twentieth-century.

Norman Lloyd Johnson was born on January 9, 1917, in Ilford, Blser, England. He attended the County High School, Ilford, from 1927 to 1934, and then entered (before his 18th birthday) the University College London. There, he gained a BSc degree in Mathematics and Statistics and an MSc degree in Statistics in the period 1934-38. He was appointed Assistant Lecturer in Statistics in 1938, a post to which he returned in 1945 after serving with the British Ordnance Board during World War II.

He remained on the staff of the Statistics Department of the University College London, becoming Lecturer in 1946 and Reader in 1956, with breaks in 1952-53 (Visiting Associate Professor, University of North Carolina) and 1960-61 (Visiting Professor, Case Institute of Technology), until he joined the University of North Carolina in 1962. During this period, he obtained his PhD (1948) and DSc (1962) degrees in statistics and Fellowship in the Institute of Actuaries (1949), and wrote, jointly with H. Tetley, one of the earliest statistical textbooks in England, the two-volume Statistics: An Intermediate Text Book (Cambridge University Press, 1949). In collaboration with F.C. Leone, he was subsequently joint author of (again the two-volume) Statistics and Experimental Design in Engineering and the Physical Sciences (Wiley, 1964).

He was an Associate Editor of Annals of Mathematical Statistics (1958-61), Biometrika (1962-65), and Technometrics (1967-71). Subsequently, in the seventies, he served as the Editor-in-Chief of the Journal of the American Statistical Association (JASA). More recently, he served as Associate Editor of Metron and was on the editorial board of Sequential Analysis.

During his lengthy tenure at the University of North Carolina, Chapel Hill, interspersed with brief visits to the UK, Italy, Australia, Poland, China, and other countries, Norman Johnson published over 100 papers and co-authored, co-edited and translated over twenty books.

Among his most prominent contributions are his first publications, with B.L. Welch, in 1939 and 1940 (at the age of 22) on chi-squared and noncentral t distributions, his groundbreaking papers on translation systems (1945, 1950), currently known as Johnson’s transformations, and the joint work with F.N. David (the first prominent female statistician) on various aspects of the properties of procedures when the standard ANOVA assumptions are not satisfied (robustness) and problems related to order statistics. His contributions to sequential analysis and cumulative sum control charts in the early sixties were warmly welcomed. During this period, he also made substantial contributions to sample censoring procedures and to problems related to finite population.

From the mid-sixties until 1972, Johnson was immersed in the compilation which became the four pioneering books on Distributions in Statistics, with S. Kotz, a natural continuation of the famous Elderton and Johnson volume (1969), Systems of Frequency Curves. This work was carried out simultaneously with his research on various types and aspects of statistical distributions (including quadratic forms).

In the next decade, his books on Urn Models and Their Application, with S. Kotz (1977), and on Survival Methods and Data Analysis, with his wife Professor R.C. Elandt-Johnson (1980), received special attention, along with the second edition of the Johnson and Leone two-volume text on statistics (which was also translated into Russian).
The eighties and early nineties are marked by N.L. Johnson’s most strenuous activities, jointly with S. Kotz and C.B. Read, related to editorship of the thirteen-volume Encyclopedia of Statistical Sciences, which is viewed as one of the most influential publications in statistics in the 20^th century. Among other favorable notices, the Encyclopedia was acclaimed in the JASA in a detailed review jointly prepared and signed by the entire Book Review Editorial Board (including an author of this obituary), under the leadership of J. Tanur. This monumental work was accompanied by the revision (and extension) of the four-volume compendium on Statistical Distributions (mentioned above), which has grown to become a five-volume set. Johnson also co-edited in the late nineties, with S. Kotz, a three-volume collection, Breakthroughs in Statistics, and a volume of essays on Leading Personalities in Statistical Sciences: From the 17^th Century to the Present.

One of his last research topics (in the late nineties and early years of the 21^st century) were intensive studies in quality control resulting in two books (with S. Kotz and X. Wu) and a comprehensive survey paper in 2002.

Among Johnson’s prominent students are such luminaries as D.J. Bartholomew, M. Ghosh, J.R. Kettenring, L. Stokes – to mention a few who provided contributions to the volume in his honor on the occasion of his 65^th birthday in 1982, edited by P.K. Sen.

During his 60 years of dedicated and brilliant toil in the fields of statistics, Professor Johnson was recognized and honored on a number of occasions for his outstanding contributions. However, so vast and wide-ranging were his contributions that these awards – which included the Shewhart Medal, given by the American Society for Quality Control, and the Wilks Memorial Award, given by the American Statistical Association – did not seem adequately to reflect his tremendous impact on statistical sciences in the 20^th century.

Norman Johnson was kind, sensitive, gentle, charitable, and broadly educated – a person who together with his beloved wife Regina was alert to the modern world’s injustices and tried to contribute towards improvements in the lives of the less fortunate. A devoted friend, he was always available in times of need.
He will be sorely missed by his colleagues and students and, indeed, by everyone who had an opportunity to cross paths with him – including those who met him only on the printed page.

Professor Dr. Guillermina Jasso (New York University)
Professor Dr. Samuel Kotz (George Washington University)

Top

ISI Membership Elections 2004

We congratulate the 28 new ISI elected members who were elected in the second round of the 2004 ISI membership elections. For those who wish to contact any of these individuals, please note that the ISI website contains a component including the names and addresses of all ISI members (see http://isi.cbs.nl/ISImembers/isimembers.htm), and these new members will be added to this list in the coming weeks.

Top

Historical Anniversaries: Optimality of the Sample Mean

We have all learned at some point in our careers, usually fairly early on, about the optimality properties of the sample mean as an estimator, best linear unbiased under certain model frameworks being one common example. The long road to achieving these now well-known results was first embarked upon 250 years ago.

The sample mean was not always the estimator used for the centre of location. One common estimator was the midpoint between the maximum and the minimum, or the two extreme values. This estimator has ancient origins going back to Aristotle (384 BCE-322 BCE) who used this mean not in a statistical sense but in ethics in his Nicomachean Ethics where he argued that virtue is a kind of moderation that aims at the mean or moderate amount among the possible actions. The mean is that which is equidistant from both extremes of the range of actions under consideration. Other estimators were also considered. For example, there is an early use of trimmed means predating robust statistics by several hundred years. Leon Batista Alberti (1404-1472) in his De statua described for artists what he thought were the human proportions that would best reflect beauty. He chose several subjects that he thought to be the most beautiful and made various measurements on them. To obtain the representative measurement in each dimension of the human body that he considered, he omitted the extremes and took the mean of the remaining measurements.

The first step in obtaining some optimality properties about the sample mean were obtained in 1755 by Thomas Simpson (1710-1761) in an article in Philosophical Transactions of the Royal Society. Simpson, who was made a fellow of the Royal Society in 1745, was a self-taught mathematician. He wrote books on probability theory and annuities that are mainly derivative from books on the same topics by Abraham De Moivre (1667-1754). It is his work on the theory of errors that provides the first interesting insights into the theory around the sample mean. Simpson’s work on the mean was motivated by a problem in astronomy. Simpson began his article with:

“It is well known … that the method practised by astronomers, in order to diminish the errors arising from the imperfections of instruments and the organs of sense, by taking the Mean of several observations, has not been so generally received, but that some persons, of considerable note, have been of opinion, and even publickly maintained, that one single observation taken with great care, was as much to be relied on as the Mean of a great number.”

Simpson considered two error distributions,
P(X=x)_∞r^x
and P(X=x)_∞(v-|x|+1)r^x both defined for x = –v, –v + 1, –v +2,...,–2,–1,0,1,2,..., v –2, v –1, v

Since the motivation was astronomical measurement, in the examples, Simpson considered x is measured in seconds of an arc. In a numerical example, Simpson considered only the second error distribution under the special case of v = 5 and r = 1. This error distribution is shown graphically here. Under this distribution, Simpson calculated P( |X| ≤ k ) based on a sample size n=6 and P( |X| ≤ k ) for a single observation when k = 1 and 2. In both situations (k = 1 and 2), Simpson found that
P( | X | ≤ K ) > P( | X | ≥ k ) (in particular 0.725 compared to 0.444 when k = 1 and 0.967 compared to 0.667 when k = 2). He concluded that:
“Upon the whole of which it appears, that of the taking of the Mean of a number of observations, greatly diminishes the chances for all the smaller errors, and cuts off almost all possibility of any great ones: which last consideration, alone, seems sufficient to recommend the use of the method, not only to astronomers, but to all others concerned in making experiments of any kind …”

Someone had to throw a little water on this fire. That someone turned out to be Thomas Bayes (ca. 1701-1761), who pointed out in modern jargon that the mean of biased measurements will still be biased. In his words, “… the more observations you make with an imperfect instrument the more certain it seems to be that the error in your conclusion will be proportional to the imperfection of the instrument made use of …”

Those Bayesians have been making life difficult for the rest of us from the very beginning it seems.

David Bellhouse
Chairman
Christiaan Huygens Committee on the History of Statistics

Top

ISI Honorary Member Interviews: Edmond Malinvaud

Who were the three people who influenced your career the most, and why?

The post war period “was a time of great idealism among young people in [France]; we felt a deep desire to build a better world after the depression of the 1930’s and the horrors of the war” (Professor James Durbin, ISI Honorary Member Interviews, ISI Newsletter, Volume 28, N°. 2, 2004). Such was the reason why, in 1945 already, I had decided to become an economist and in 1946 I chose the newly created INSEE as a natural site for my career, following my earlier study of mathematics. During the two years 1946-48, in the INSEE School, which was later to become ENSAE, I was particularly influenced by two professors.

Georges Darmois, a shrewd and charming probabilist and mathematical statistician, who became ISI President in 1953-55, attracted me to the circle of his preferred students. In the early 50’s, he asked me to give a small course in econometrics, which grew larger over the years up to the publication of my Statistical Methods of Econometrics.

Maurice Allais, 1988 Nobel Prize winner in Economics, considerably helped me to put order in my knowledge of economics, which in 1947 was mainly the result of my unorganised readings. He also attracted me to the group of his selected students and obtained for me a Rockefeller Fellowship, which was spent in 1950-51 at the Cowles Commission. This exciting research group was the third main influence on my career.

If you could relive any part of your career, which part would it be? If you would prefer to omit any part, which would it be?

To these strange questions, my answer is simply that I am happy to have had the opportunity of serving in various capacities under conditions which made each one of these services relevant: taking part in the building of the French system of public statistics, teaching and writing textbooks in economics, advising in economic policymaking (starting from French planning, going on to macroeconomic policies, finishing with European economic unification), and on the side pursuing research of my own liking.

INSEE is recognised as a highly advanced NSO in the international official statistics community and it is common knowledge that in your previous capacity as Director General you had played a major leadership role in obtaining this status. What advice would you give to someone who had assumed the position as Director General of a NSO?

In order to be good, any society must know itself: where does it come from? Where does it want to go, considering its current state? A proper answer to this second question much depends on the existence of a corpus of good socio-economic statistics and on people’s confidence in their public statistics. Building such a corpus and such a confidence is a long run affair, in which official statisticians must not only be competent but also share a common vision of their role. A NSO Director General, who ideally should remain in office for about a decade, is a successor of its predecessors and should think beyond his or her term in office. He or she should thus mainly care about the long run destiny of the NSO.

What was the most memorable part of your term as President of the ISI?

Certainly the World Fertility Survey which was run by the Institute, under the leadership of Maurice Kendall, in order to bring good information on human fertility in less developed countries. In 1979-1981, it had reached the stage of publication of results. It was most exciting to be associated with that great achievement (except for the fact that fertility surveys could not be made in China, India, Brazil and Nigeria!). However, there is also quite an unpleasant spot in my memory; namely, to have had to organise the Session in Buenos Aires with the regime which was responsible for the “disappearance” of the Argentinean Chief Statistician, Carlos Noriega.

Do you have any suggestions as to how the contacts and exchange of knowledge between academic and official statisticians can be increased?

We must realise that, over hardly more than a century, statistics as a discipline has been blooming in many directions, required as it is for many fields of application. Statistics as data have developed in all these various fields, to the point that they become part of modern culture. Looking for exchanges between academic and official statisticians is now too narrow a way of posing the question of the division of labour and exchanges between statisticians. When the ISI was created and up to the first decade or so after I entered the profession, the Institute was an appropriate international centre of exchanges between all statisticians on all main subjects of interest. I submit that for this century the vocation of the ISI has to be reconsidered and focused on matters of common interest; particularly, on the development of a solid statistical culture in our societies: how to foster the teaching of statistics within mathematics in primary and secondary schools; what should the deontology be for the collection, diffusion and use of quantitative information; how to best pattern forums in which specialised international statistical associations would jointly discuss on specific subjects; and so on.

In your impression, what new statistical challenges are awaiting the profession in the forthcoming years?

Perhaps the wise attitude would be to decline answering such a difficult question. I shall, however, venture to say that those challenges are intertwined with challenges facing our rich societies: difficult public finances which endanger the allocation of resources to research or to official statistics; schizophrenic attitudes in the uncertain choice of our ethics or of priorities for our investigations; and so on.

What would your advice be to young statisticians today?

Here really, I do not think young statisticians could benefit from my advice. They should know by themselves what to do, depending on their capabilities and inclinations.

Top

ISI Committee Matters: Women in Statistics

The ISI Committee on Women in Statistics (CWS) was organised to enhance the role for women within the ISI and the statistics profession. The Committee was organised in 1997 under the leadership of Mary Regier, who served until 2001 as Chair. Beverley Carlson served from 2001 to 2003 and I recently succeeded her. The Committee has identified five objectives:

• To promote and strengthen the representation of women statisticians in the ISI and its Sections;
• To help in providing opportunities for women members to assume active and visible roles in the ISI and other statistical associations;
• To collect information on women in statistical professions in different countries and to facilitate the flow of information among women statisticians;
• To stimulate interest in statistics among women and encourage women in schools and colleges to study statistics;
• To support the compilation of statistics on women, with a view to generating relevant studies concerning women’s roles in the various activities in their countries.

One of the vehicles through which the Committee functions is its website. The site can be accessed at http://www.nass.usda.gov/cws/ . Activities and articles of interest are put onto the site for access to individuals throughout the world. The Committee has a new webmaster, Martha Farrar (mfarrar@nass.usda.gov) who is also a Committee member.

A recent objective has been to identify individuals that are interested in being country and regional correspondents. These correspondents would periodically contribute brief notes on statistics or statisticians in their country several times a year to be published on the CWS website. The country correspondents would also take on the role of identifying potential ISI members in their country or region and support their nomination for membership.

The CWS will be sponsoring an Invited Paper Meeting (IPM) at the 2005 ISI Sydney Session. IPM 74 is entitled Surveying Women’s Issues. It will be held on Saturday, the 9^th of April, from 9:00 a.m. until 11:15 a.m. Five papers will be presented during the session addressing issues of health, abortion, sexual assault, literacy, and voting. The papers come from both developed and developing countries – Canada, the United States, India, Egypt and New Zealand.

A recent activity that might be of interest to women statisticians working in the drug or pharmaceutical industry is a survey conducted by Dr. Sheela Talwalker, former Director of Clinical Statistics at Pharmacia. The focus of the survey was to evaluate the opportunities that women statisticians working in the drug industry get for career growth and their satisfaction with management help in achieving it. Dr. Talwalker presented her analysis of the survey at the 2003 ISI Session in Berlin. The results of the survey are on the CWS website.

If you are interested in becoming involved in the activities of the ISI Committee on Women in Statistics, please contact Cynthia Clark (Cynthia.clark@ons.gov.uk). In particular, the Committee is soliciting individuals interested in being country or regional representatives.

Cynthia Clark
Chair
ISI Committee on Women in Statistics

Top

Back to Home Page

News from ISI sections Volume 29, No. 1 (85) 2005