History

(Also see comments by Arnold Zellner on the role of S. James Press in the history of Bayesian Statistics.)

THE FORMATION OF THE SECTION ON BAYESIAN STATISTICAL SCIENCE (SBSS) OF THE AMERICAN STATISTICAL ASSOCIATION (ASA)

by S. James Press,

Department of Statistics,

University of California,

Riverside, CA 92521-01389

jpress@ucrac1.ucr.edu

The interest in forming a Bayesian statistics Section of ASA reflected a need for statisticians and people from other disciplines who have interests in the Bayesian paradigm to formalize their common interests within the statistical community represented by ASA. Some of the people interested in forming such a section had already clustered together over the years in small special interest groups relating to Bayesian statistics. It was their hope that the ASA Section on Bayesian Statistical Science would provide a common focus for these groups, as well for the wider scientific community.

The notion of a formal Bayesian-oriented organization within the scientific community had been discussed for many years, especially in various special interest groups, such as in:

1. the Seminar on Bayesian Inference in Statistics and Econometrics (SBIES), organized and run by Professor Arnold Zellner from 1970 (for over 25 years); the group organized by and run by Professor Ward Edwards (in Michigan, and later in Los Angeles) on Bayesian methods in decision theory (with a psychology orientation);
2. in the group called the Max Ent Workshop, on Bayesian methods using Maximum Entropy (oriented towards physics and engineering), run by an informal board of advisors. The Max Ent group developed from the work of Professor Edward Jaynes at Washington University in St. Louis and was meeting annually since about 1981 (in the U.S., Canada, England, South Africa, and elsewhere);
3. and in the so-called "Valencia Conferences", run by Professor Jose Bernardo, of the University of Valencia, Spain.

These groups operated independently with various funding sources and gathered one or more times per year in a conference format to have Bayesian scholars present their latest research findings, and to discuss their research. SBIES met two times per year with funding from the National Science Foundation administered by the National Bureau For Economic Research. Papers presented at these meetings were published in various proceedings volumes. The SBIES group convened at many universities in the U.S. as well as at institutions abroad (such as Canada, Mexico, Brazil and India). This group had experience managing a budget, planning meetings, making an annual award (the Leonard J. Savage Thesis Award) for the best Bayesian Ph.D. thesis of the year (since 1977), and generating 8 books of collected papers, including a monograph series on Bayesian Econometrics and Statistics published by North-Holland and co-edited by Arnold Zellner and Joseph Kadane. Attendance at these semi-annual meetings typically ran from about 50 to over 100, depending upon where the group convened. The Valencia Conferences were held every three or four years in the vicinity of the city of Valencia, and were co-sponsored by the State of Valencia, as well as other organizations. There were discussions about forming an international Bayesian organization (eventually one was formed called the International Society for Bayesian Analysis, or ISBA), and for forming a Bayesian organization within the North American statistical community, because that community was so large all by itself.

Discussions about forming a Bayesian Section of ASA began in earnest in 1991. The lead role was taken by Professor S. James Press, Department of Statistics, University of California, Riverside. Discussions about the advisability of such a structured organization were held with a number of leading figures in the field in terms of the rationale for forming the group, what such a group would do, how it might advance the cause of Bayesian Statistical Science, etc. Professor Press created a sixteen person Steering Committee to discuss the advisability of creating such a formal Bayesian group, and to make decisions about its formation, should the group decide it would be a good idea. This committee (with their organizational affiliations at the time) included:

Professor Zellner was made an Honorary Member of the Steering Committee (involving minimal duties, but it was an advisory position nevertheless), owing to his many other concurrent duties, such as serving as President of the ASA, leading the formation of ISBA, and running the SBIES.

Once the Steering Committee agreed that it would be a good idea for a Bayesian group to have a formal voice in ASA affairs, it explored the pathways a new group had to follow to become a formally recognized Section of the ASA. The then Chair of the ASA Council of Sections was Dr. David Hoaglin who helped the Steering Committee prepare the paperwork that went with petitioning the Council of Sections for the formation of a new Section of ASA. By January, 1992, the paperwork for the formation of SBSS was basically in place.

A letter to Professor Zellner (on leave from the University of Chicago at the University of California, Irvine) from Professor Press summarizing the status of the effort at that time follows.

January 13, 1992

Professor Arnold Zellner

School of Social Sciences

University of California

Irvine, CA 92717

Dear Arnold:

We are in the final stages of submitting a request to the ASA Council on Sections that a Section of ASA on Bayesian Statistics be formed. This letter is to bring you up to date on the details of our request, and to solicit your help in the preparation of our supporting documents. As you know, you are an Honorary Member of the Steering Committee for this activity, and because of your very special status and your heavy commitments to other matters (the Presidency of ASA, and otherwise), I have not wanted to bother you with the organizational details, until now.

As I have advised you in the past, to form an ASA Section, a special interest group "may be chartered by a majority vote of the Council of Sections based upon a petition of at least 100 full members, a proposed Section Charter, and a list of officers" (a quote from the Constitution of ASA).

To accomplish these things in an efficient and democratic way I have set up a Steering Committee for the Formation of the Section. The Steering Committee currently includes:

I have written both a petition for collecting the required 100 or more signatures of ASA members, and a Charter for the proposed Section. You have seen the Petition and I am enclosing a copy of the Charter. Both the Petition and the Charter were submitted to the Steering Committee for reactions and suggested changes. After one iteration the Steering Committee approved both documents. The Steering Committee members are now out collecting 10 or more signatures each, on copies of the Petition. I will collect them and submit all documentation to ASA.

I have also formed a "Nomination's Sub-Committee, to develop a slate of officers who will be agreeable to serving for the first year, before officers can actually be formally elected by the Section membership. The Nominations Committee will contain three Steering Committee members, and currently consists of Kathryn Chaloner and Steve Fienberg. I will recruit one more member. I have also asked Steering Committee members to submit slates of candidates for officers. I have also advised the Nomination's Committee that you would be agreeable to serving as Chair.

The documentation for forming a Section must first be submitted to the ASA Committee on Section Status, chaired by David Hoaglin. This committee checks the form of all of our documents and if everything is in order, that committee recommends that the matter of Section Status for this particular group be put on the agenda for consideration by the Council on Sections, at its next meeting.

The next meeting of the Council on Sections is scheduled for Friday and Saturday, Jan. 31 and Feb. 1, 1992. I have already advised David Hoaglin of our efforts and have formally advised him that we would like the matter to be placed on the agenda of the Council on Sections meeting scheduled at that time. I have already submitted to him our Charter and the Petition. We will soon have all of our signatures and a slate of proposed officers.

And now I come to how you can be of specific help to the effort. I recently had a conversation with Tom Boardman (who is one of the members of the Council on Sections) during which I advised him about what we are hoping to accomplish. He asked me whether we had any experience running such a special interest group, whether we had run any kind of pilot test of such a group, whether we could handle a budget, and related questions. I then gave him a rundown of the history of the Bayesian Seminar, the financial support, the budget, the Savage Award, the books, the periodic meetings, etc., and he admitted that he had known nothing of this background. It then struck me that we should document all of this background as part of our proposal for Section status, to inform all Council members about this background. Moreover, if you are willing, you are the obvious person to prepare such a document for us. Could you let me know whether this is something you will have time for, with very little lead time? We will need to get it to David Hoaglin very quickly, so that his Committee will have time to react to it.

A second matter concerns the Council Meeting itself, at the end of January. Is that meeting one that you were planning to attend anyhow, as an Ex Officio member, because of your Presidential status? If so, I'm sure your presence at that meeting would raise the probability of a favorable vote by the Council Members, and you would be able to answer any questions that arise.

We are all very excited at the prospect of forming a Bayesian Statistics Section of ASA. Such a Section would mainstream our movement in North America, and would eventually provide a North American base for the World Bayesian Society.

I look forward to hearing from you as soon as possible regarding whether you will be able to prepare a background document for us.

Very best wishes,

S. James Press

Enclosure: Charter for the Bayesian Statistics Section of ASA

Professor Zellner did attend that meeting of the Council on Sections and supported the effort to form the SBSS.

Each member of the Steering Committee collected signatures that were merged to create a listing of over 150 signatures from ASA members on a petition requesting that the new Section of ASA be formed. Professor Press then submitted the Bayesian group’s petition to ASA for the formation of the new Section. That letter follows.

January, 1992

To: Dr. Barbara Bailer, Executive Director, ASA, and

The Council on Sections, ASA

From: Dr. S. James Press,

Spokesperson on Behalf of A Group Of Members Of the American Statistical Association Who Would Like To Form A New Section Called The ASA Section on Bayesian Statistical Science (SBSS)

Copies To: The Steering Committee For The Formation of SBSS

This is to formally request that the ASA form a new section to be called the Section on Bayesian Statistical Science (SBSS).

To pursue the formation of this new section some ASA members have gathered together as an informal Steering Committee to jointly develop the documentation required by the new ASA Constitution, as it appears in the 1991 Directory, for such formation. The 16 person Steering Committee includes:

James Berger--Purdue University

Kathryn Chaloner--University of Minnesota

Stephen Fienberg-York University

Ronald Christensen--University of New Mexico

Seymour Geisser--University of Minnesota

John Geweke--University of Minnesota

Prem Goel--Ohio State University

Joseph Kadane--Carnegie Mellon University

Wesley Johnson--University of California, Davis

Robert Kass--Carnegie Mellon University

John Monahan--North Carolina State University

Dale Poirier--University of Toronto

James Press--University of California, Riverside

Mike West--Duke University

Marvin Zelen--Harvard University

Arnold Zellner--University of Chicago

We have collected over 150 signatures from ASA members on a petition requesting that the new section of ASA be formed. I am including originals or photocopies of those signatures with this letter.

I am also including a copy of the proposed Charter for the new section (it is included below).

The Steering Committee has developed a proposed partial slate of officers who are agreeable to serving until the Section elects a new slate. I am submitting the partial slate now so that you can begin your evaluation. The remainder of the slate will be forthcoming within the next few days. The proposed officers are:

Chair: Arnold Zellner

Chair Elect: James Press

Program Chair: ?

Program Chair-Elect: Prem Goel

Secretary/Treasurer: ?

Council of Sections Representative: ?

Publications: ?

I am also including a small "Background" piece to acquaint ASA and The Council of Sections with some of the background surrounding the desired formation of this section (this background piece is expanded in our introduction above).

We hope you will look favorably upon our request for the formation of a new section, and we plan to work with ASA to enrich the Association through our group participation.

Sincerely,

S. James Press

Professor

On March 2, 1992, Barbara A. Bailar, the then Executive Director of ASA, informed Professor Press that the Council of Sections of ASA had approved the petition for the formation of a new Section of ASA to be called the Section on Bayesian Statistical Science, and abbreviated for ASA purposes as SBSS. There were 34 positive votes cast with no negative votes. By May, 1992, SBSS membership stood at 131, so the Section became officially recognized by ASA, could hold elections for officers, and business meetings. But the first formal Executive Committee could not begin serving until the calendar year 1993. The Section couldn’t become official until it had 100 paid members.

By late June, 1992, SBSS membership stood at 163, and by Feb. 4, 1993, SBSS membership was close to 300. The youngest member of SBSS was Graham Tierney—"Bayesian-at-Birth", born at the same time as SBSS to Kathryn Chaloner and Luke Tierney. Graham attended the first SBSS business meeting. An informal group gathered in Boston, MA at the Aug., 1992 annual ASA meetings.

At the August, 1993 annual meetings of the ASA, Professors James Press and Donald Ylvisaker of the University of California at Los Angeles discussed the idea of a contest that would award a prize for the best paper that applied Bayesian methods to solve an important real problem. Coincidentally, Dr. Toby J. Mitchell, who was an outstanding statistician, died in 1993. Dr. Mitchell had been a Senior Research Staff Member at Oak Ridge National Laboratory where he spent his entire career. He was also a dedicated user of Bayesian methods in applied problems. Following his death, his friends and colleagues raised money to support an appropriate professional cause in his honor. At the suggestion of Donald Ylvisaker, Professors Press and Ylvisaker decided to name the award for the contest they were planning, The Mitchell Prize.. The Prize included an award of $1000 and a commemorative plaque. A five person committee was appointed to judge the contestants. The evaluation committee included Persi Diaconis, Diane Lambert, Max Morris (a close friend and former Ph.D. student of Toby Mitchell), Donald Ylvisaker, and James Press, Chair. The constitution of the evaluation committee changed over time. Professor Press put the matter of having the newly formed SBSS sponsor The Mitchell Prize to the Executive Committee, and the Committee approved it. The first two Mitchell Prizes were awarded in 1994 and 1995 to Mike West, Professor of Statistics and Decision Sciences at Duke University.

The Leonard J. Savage Award of $750 is made each year for the best Ph.D thesis on Bayesian methodology in statistics and econometrics. The award had been made since 1977 by a changing committee that was associated with the Seminar on Bayesian Inference in Econometrics and Statistics (SBIES). The award was created under the leadership of Professor Arnold Zellner. In 1993, Professor Zellner proposed that SBSS co-sponsor the award, and the SBSS Executive Committee approved the proposal. Over the years the structure of the award has been modified so that currently two awards are made: one award is given for the best thesis in Theory and Methods and the other award is given for the best thesis in Application Methodology.

The initial Executive Committees were constituted as shown below.

Chair--Arnold Zellner

Chair-Elect—S. James Press

Program Chair—Don A. Berry

Program Chair-Elect—Prem K. Goel

Sect./Treasurer—D. O. Dixon

Publications Chair—J. B. Kadane

Section Representative to the Council—Kathryn B. Laskey

The Executive Committee met at the ASA Annual Meetings in San Francisco, Aug., 1993