Volume 24, No. 3, December 2005

Inside:

From the President ...... 1 DA in the UK ...... 4 Call for Papers ...... 6 Award for DAS Member ...... 3 Aloha to DA ...... 4 Perspective ...... 8 INFORMS 2006 ...... 3 DA in ...... 5

From The President year and introduce our leadership Kleinmuntz. team for the coming year. Please Greg Parnell contact anyone directly if you • Jason Merrick will continue as would like to participate in any of our Newsletter Editor. the committees or present a paper in one of our sponsored tracks at • Bob Nau will continue as our upcoming conferences. information officer and website manager. Jason Merrick and • Craig Kirkwood continues as Christina Werner will be working President-Elect/Vice President. with Bob to develop a new society Craig also represents us on the website. Subdivisions Council and Chairs our Nominating Committee. • Jim Smith will continue as our Decision Analysis Electronic • Elisabeth Paté-Cornell continues Mailing List monitor. as our Past President. From the President • Jack Kloeber will continue to • Robin Dillon-Merrill continues chair our Membership Committee. at Secretary-Treasurer. In this update, I would like to We will be doing a membership identify the new DAS leadership study this year. • Bob Bordley and Jeff Keisler team, list some of the highlights completed their service on the of the San Francisco meeting, • Kazuo Ezawa will continue council. Returning council describe the decision analysis to chair our Practice Awards members include Jim Smith, Jack events at the Society for Risk Committee. Karen Jenni will also Kloeber, Manel Baucells, and Jeff Analysis meeting, and highlight be on the committee. Stonebraker. Our two new council society membership. members are Eric Bickel and • Our Publications Committee was Karen Jenni. New DAS Leadership Team George Wu, Jim Dyer, and Jim Smith. Vicki Bier, Jim Dyer and • Continuing as the Decision I would like to thank our society Eric Bickel will be on this year’s Analysis Journal Co-Editors in leadership for their service last committee. Chief are Bob Clemen and Don continued on page 2 Page 1 Volume 24, Number 3 (December 2005) Decision Analysis Newsletter • Our Student Award Committee was Jeff Stonebraker • The Publication Award was awarded to Daniel and Eva Regnier. Next year’s committee will be Jeff Kahneman for the paper “Maps Of Bounded and Erin Baker. Rationality: Psychology For Behavioral Economics” published in American Economic Review, 2003. The following are the DAS sponsored meeting organizers: • The Student Paper Award was given to Janne Gustafsson for the paper “Contingent Portfolio • Bill Klimack is helping plan the DA session at Programming for the Management of Risky Projects” INFORMS Practice Conference Miami (30 Apr-2 with Ahti Salo. May 2006). A DA track with six presentations has been arranged.

• Jamie Jia and Shouyang Wang are our DA track chairs for Hong Kong 2006 sponsored by China OR and INFORMS (June 2006).

• Ali Abbas and Eva Regnier are organizing the Decision Analysis Sponsored Cluster at the INFORMS 2006 in Pittsburgh (November 2006).

INFORMS 2005

The San Francisco meeting was very well done, especially considering the short time the planners had to prepare for the meeting. For those who missed • The Practice Award was given to Mario Castillo, the meeting, I’ll mention some of the highlights. Our Alvaro Mendoza, and Sergio Cabrales for “Improving Decision Analysis Cluster Chairs Eric Bickel and Quality of Decision Making -- A Collaborative Jason Merrick organized a great meeting! With the Approach between Universidad de los Andes and BP challenges of the change of location, Eric and Jason Exploration Company – .” should get credit for planning two conferences!

• The Ramsey Medal, the Society’s highest honor, was given to Larry Phillips of the London School of Economics. The Ramsey Medal committee was lead by Jim Dyer. Detlof von Winderfelt introduced Larry and presented the award. Larry’s two sons also attended the ceremony.

• We also sponsored a decision analysis tutorial by Jim Matheson. Although it was the first presentation of the conference, Jim’s presentation was well attended.

continued on page 3

Page 2 Volume 24, Number 3 (December 2005) Decision Analysis Newsletter

Decision Analysis and Risk Speciality Group INFORMS 2006 (DARSG) in the Society for Risk Analysis INFORMS 2006, Pittsburgh, Nov 5th – 8th The DARSG (organized by Igor Linkov, Greg Kiker, Call for Presentations Jeff Keisler, and others) sponsors conferences, training courses, and seminars related to risk and decision It is time to start planning the DAS track for INFORMS analysis. I attended the annual SRA meeting in Orlando 2006, in Pittsburgh , while we are still inspired by the in early December which featured several DARSG- excellent DAS track in San Francisco. sponsored activities including a continuing education workshop on “Integrated Risk Communication and Please take a moment to think ahead to November Decision Analysis: Process, Methods and Tools,” 5-8th. Which topics are you particularly interested Symposia on “Multi-Criteria Decision Analysis, Risk in hearing about? Would like to organize a session? Assessment and Homeland Security Applications” Present in the decision analysis track? If so, we would and “Environmental Security in Harbors and Coastal like to hear from you. We welcome all inputs and Areas: Management using Comparative Risk suggestions to make this an enjoyable and enriching Assessment and Multi-Criteria Decision Analysis” as meeting for DAS. well as several topical sessions. We would like to thank those who have already ap- Society Membership proached us and offered to organize sessions. We have received offers for sessions on medical decision mak- DAS is now the second largest INFORMS society with ing, dependence in multiattribute utility, behavioral 936 members. We have had steady growth in regular decision making, portfolio analysis, and decision members and large growth in student members. We analysis practice. In addition we will have a featured also have 167 international members. joint session with the Homeland security track. Larry Philips will also be giving the DAS sponsored tutorial I look forward to hearing from you. on Decision Conferencing.

Remember, Decision Analysis is The Science of Better Thanks for your time, and please respond to both of Decisions! us ([email protected], [email protected]), and put “DAS - Session” somewhere in your title line. Greg Parnell Thanks again and we look forward to seeing you in Pittsburgh!!

Award for DAS Member Ali and Eva

Congratulations to Sue Goldie on receipt of a Ma- cArthur Award. Sue has made important contribu- tions to global health policy through her creative and effective application of decision analytic tools. She consistently demonstrates how decision analytic methods can be accessible and applied to real world health policy issues and policymakers. Armed with these methods, Sue’s impact on global health policy exemplifies translating decision making research into practice.

Page 3 Volume 24, Number 3 (December 2005) Decision Analysis Newsletter DA in the UK Aloha to DA

Decision Analysis in the British OR Society Con- The IFORS 2005 conference brought a pleasant gath- ference ering of speakers from the Decision Analysis Society. Thanks to John Butler for his organization of the De- The 47th conference of the UK OR Society took place cision Analysis track sessions and for ensuring that in September 2005. For the first time in the recent past, the sessions were in the first two days. The sessions the conference featured a Decision Analysis stream, included a wide variety of topics. The first day started organised by the authors of this notice. with Rakesh Sarin’s talk on money and happiness, Jim The interest in our stream exceeded our wildest Felli’s talk on graphical representations of MCDM, expectations, with a total of 22 papers, which made Tianjun’s talk on effects of reference point dependence the stream the largest in the conference. The papers on prior selection, and Kazuo Ezzawa’s talk on ad- presented covered a wide range of topics, including vances in pharmaceutical portfolio management . risk analysis, Bayesian networks, multi-criteria deci- The second session featured Bob Clemen and Canan sion analysis and multi-objective optimisation. Most Ulu’s model of probability judgement that incorporates contributors were from the UK, although presenters some empirical properties of subjective probabilities. came from as far afield as Kuwait! Next was John Butler, Jian Min Jia and Jim Dyer’s Prof Larry Phillips of the London School of Econom- model on proxy attributes in preference models, fol- ics (LSE) contributed a keynote, “Decision Analysis lowed by Bob Nau’s talk on syndicates, opinion pools in Social Context”. This stimulating paper reflected and state-dependent utility, and Sam Bodily’s talk on on years of experience of practising and teaching deci- decision and risk analysis in pharmaceuticals. sion analysis, and offered a number of frameworks for The third session featured an arcade of talks: Manel thinking about the organisational issues in undertaking Baucell’s talk on the performance of heuristics in Decision Analysis interventions. Larry also delivered binary Multiattribute choice, Eva Regnier’s talk on a popular tutorial, “Decision Analysis in 2005” in the trade-offs of costs and benefits/objectives in the pub- conference’s tutorial stream. lic sector, followed by Ali Abbas’ talk on invariant Prof Val Belton, the President of the British OR Soci- utility functions and the generalization of notions of ety and Prof Greg Parnell, the President of DAS, also additive and multiplicative delta properties, and Don presented their recent work. In Val’s case, the research Kleinmuntz’s talk on dynamic resource allocation in (with George Wright and Gilberto Montibeller) was Not-for-Profit healthcare organizations. on a comparison between the even swaps method The second day started with a session on homeland and conventional MAVT analysis. Greg reported on security led by Detlof von Winterfeldt who talked a major substantive application, his work on the US about CREATE: the first university-based center of Army’s Base Realignment Program. excellence funded by the Department of Homeland The stream seemed to be a successful undertaking, Security. Detlof’s talk was followed by Heather Ro- from the unsolicited positive feedback we received soff, Henry Willis’, and Richard Larson’s talks on dirty from many presenters and attendees. For both of us, it bomb attacks and evaluating security of the global was heartening to find that there was so much interest- containerized supply chain. ing and useful decision analytic work going on in the The second session discussed preference program- UK. We hope also that it was a way of strengthening ming with incomplete information and was presented the links between the UK and the international Deci- by Ahti Salo, Pekka Mild, Juuso Lieiso, and Antti sion Analysis communities. Punkka from the Systems Analysis Lab at the Helsinki Alec Morton university of Technology. Gilberto Montibeller The third session presented applications of risk analy- sis: Applying risk analysis to a business context by Lea Deleris and Elisabeth Pate-Cornell; matching exploration technology needs with candidate solutions

Page 4 Volume 24, Number 3 (December 2005) Decision Analysis Newsletter by Charles Weisbin, and integrated risk assessment by Background Diane Williams. MARA is a non-profit project, organized by assistant The last session featured Steve Greidinger’s talk on professors of the Ciencias Económicas faculty of the quantifying value of information when the recommen- University of (UBA) and PhD students dation is not followed; Peter Anselmo’s talk on risk of the London School of Economics (LSE). MARA is tolerance aggregation under constrained portfolios; an official project of the LSE and is supported by the John Tindle’s talk on military utility-based architec- director of the LSE, Howard Davies, and Prof. Larry tures for government integrated planning processes, Phillips of the LSE School of Decision Analysis. On and William Tarantino and Greg Parnell’s talk on the Argentinean side, Prof. Daniel Avenburg, Prof. using decision analysis for BRAC (Base Closure and Patricia Bonatti and Prof. Alejandro Basualdo of the Realignment Commission). Ciencias Económicas faculty of the UBA support MARA. Ali Abbas MARA 2005 Conference in Argentina With the final conference “MARA 2005 – Merging DA in Argentina Methodologies” on September 23rd 2005 the results With MARA we aim to transfer decision-analytic of MARA were presented to the Argentinean scien- methodologies to allocate resources efficiently into tific community and to the wider public. The idea of the organisational practice. Over the last two de- the conference was to present the results of MARA cades, a variety of decision-analytic methods for 2005 and to contribute to raising the awareness of the budgeting decisions, project prioritisation, selection area of Decision Analysis in Argentina. Therefore all of organisational strategies and investment decisions project partners, academics and practitioners of Deci- have been developed. These approaches, however, sion Analysis in Argentina and the Argentinean press are still unknown to most practitioners. MARA helps were invited. to bridge this gap between organisational needs and The event served as a starting point to found a Centre academic methodologies. Within the framework of for Decision and Negotiation Analysis, which is the five to seven projects carried out within companies, basis for post-MARA applications for methodologies NGOs and governmental organisations in Argentina, in the area Decision Science in Argentina. The confer- MARA 2005 offers students, academics and experi- ence, moreover, served to build links to US and UK enced practitioners, the possibility of developing new communities in the area of Decision Analysis. insights into decision-analytic methodologies. These MARA 2006 Conference in Germany methods are considered to improve the resolution of ill-structured, complex problems with conflicting MARA 2006 aims to create new knowledge in the objectives and uncertain outcomes. We have chosen areas Decision and Negotiation Analysis. Designed Argentina for MARA 2005 due to the particularly as an applied, international research co-operation interesting resource allocation problems it faces fol- between universities from Germany, the UK, the US, lowing the economic crisis of 2001/2002. Moreover, Argentina, private and public sector organisations, the dealing with uncertainty is an issue of high importance project focuses on resource allocation decisions. In the in Argentina. framework of five “MARA research tracks”, interna- tional teams of young academics, young professionals Academics from the London School of Economics and and students will research current resource allocation the University of Buenos Aires, scholarship holders processes in organisations in a certain area of applica- of the Foundation of German Business and - as team tion. In addition, they will apply model-based method- leaders - practitioners with consulting experience will ologies, aiding MARA project partners to create better take part in MARA. Research focus this year is on value-for-money when allocating scarce resources. the possibility of merging different decision-analytic MARA 2006 is a follow-up project to a similar initia- methodologies. In MARA 2006, these insights will be tive of the London School of Economics, the Univer- used to transform organisational decision processes sity of Buenos Aires and the Foundation of German (“Decision Process Reengineering”) within the frame- work of five to seven projects in Germany. continued on page 7 Page 5 Volume 24, Number 3 (December 2005) Decision Analysis Newsletter

Call for Papers Don Kleinmutz & Bob Clemen Special Issue on free to contact either Bob Clemen or Bob Winkler Bayesian Statistics and Decision Analysis ([email protected]).

Decision Analysis, a journal of the Institute for Op- Editorial Objectives erations Research and the Management Sciences, is Decision Analysis is dedicated to advancing the soliciting submissions for a special issue on Bayesian theory, application, and teaching of all aspects of statistics and decision analysis. Ideally, papers should decision analysis. The primary focus of the journal emphasize connections between Bayesian statistics is to develop and study operational decision-making and decision analysis, focusing on how Bayes-ian methods, drawing on all aspects of decision theory methods can play an important role in decision analy- and decision analysis, with the ultimate objective of sis. Preference will be given to papers emphasiz-ing providing practical guidance for decision makers. As Bayesian methodology typically not encountered such, the journal aims to bridge the theory and prac- in the decision analysis literature, discussing how tice of de-cision analysis, facilitating communication Bayesian methods can be utilized more effectively and the exchange of knowledge among practitioners in any aspect of decision analysis, or presenting and re-searchers in academia, business, industry, and actual applications involving novel use of Bayesian government. Articles will contribute to these goals methods. Articles focusing on theory, methods, and in many ways, using a wide variety of methods and applica-tions are all welcome, as long as there is a approaches. For example, articles might discuss new clear connection with decision analysis. or exist-ing algorithms, procedures, or processes for implementing decision analysis; develop new theory This special issue will be jointly edited by Professor or em-pirical studies related to cognitive, organization- Robert Winkler of Duke University, serving as guest al, or social issues that have implications for decision editor, and Professor Robert Clemen, also of Duke and analysis; propose and test innovative uses of informa- currently co-editor-in-chief of the journal. The special tion technology to perform decision analysis; or raise issue is tentatively scheduled for publication in 2007. issues related to the application of decision analysis in Manuscripts will be considered as they are submitted, real-world situations. Articles should generally remain but for full consideration should be submitted no later faithful to the intellectual foundations of decision than March 31, 2006. Articles not ac-cepted for the theory and decision analysis. However, the journal special issue may be considered for publication in welcomes original contributions that genuinely chal- subsequent issues of the journal. lenge the field, for example by showing how concepts, ideas, and methods from other fields can improve the How to Submit a Paper theory or practice of decision analysis. The journal Please submit your paper electronically by email at- also publishes articles that review and summarize tachment to co-editor Bob Clemen (clemen@duke. important topics or advances of interest to decision edu). Include in your cover letter a statement that analysts or that provide original historical, scholarly, you wish your paper to be considered for the special or practical perspectives on the field. In addition, the issue. All submissions will be peer-reviewed. For journal encourages articles that support the teaching information about the journal, including instructions of best practices, such as state-of-the-art applications, to authors, please visit http://da.pubs.informs.org. We case studies, and tutorial articles on decision-analysis also encourage authors to review the journal’s editorial methods. objectives below and to ensure that submissions are suitable for the journal in both style and substance. If you have questions about whether your submission fits the objectives of the special issue, please feel

Page 6 Volume 24, Number 3 (December 2005) Decision Analysis Newsletter

Business, which was organised in DAS Officers 2005 in Argentina. During MARA DAS Council 2005, our/five research teams helped BASF Argentina, Siemens President: Manel Baucells Argentina, Schering Argentina, Greg Parnell IESE Business School Volkswagen Argentina and a local Department of Systems Engineering Universidad De Navarra NGO to optimise portfolios in the US Militray Academy Barcelona 08034 areas marketing, Corporate Social West Point, NY 10996-1779 Spain Responsibility, sales strategy and Tel: 845.938.4374 [email protected] project development. With MARA Fax: 845.938.5919 [email protected] 2006, organisations and research partners are invited once again President-Elect: Eric Bickel to benefit from collaborating at Craig Kirkwood Industrial and Systems Engineering the forefront of applied academic BA 422, Main Campus Texas A & M University PO BOX 874706 College Station, TX 77843 research. Tempe, AZ 85287-4706 [email protected] Excerpts taken from http://www. Tel: 480.965.6354 projectmara.com/ Fax: 480.965.8629 [email protected] Karen Jenni Past President: Geomatrix Consultants, Inc. M. Elisabeth Paté-Cornell 1401 17th Street, Suite 600 Chair, Department of Management Denver, Colorado 80202 Stanford University [email protected] Stanford CA 94305-4024 Tel: 650.723.3823 Fax: 650.725.8799 Jack Kloeber [email protected] Director, Portfolio Management Treasurer: Johnson and Johnson Robin Dillon-Merrill JJPS Bldg 700 Georgetown University Rte 202 South 418 Old North Raritan, NJ 08869 McDonough School of Business [email protected] Washington, DC 20057 Tel.: 202.687.5398 Fax: 202.687.4031 James E. Smith [email protected] Fuqua School of Business Newsletter Editor: Duke University Jason Merrick Durham, NC 27708-0120 Statistical Sciences & Operations Research [email protected] Virginia Commonwealth University PO Box 843083 Richmond, VA 23284-3083 Tel: 804.828.1301 x. 136 Jeff Stonebraker Fax: 804.828.8785 GlaxoSmithKline [email protected] Five Moore Drive Web Editor: PO Box 13398 Robert F. Nau Research Triangle Park, NC 27709 Fuqua School of Business [email protected]. Duke University Durham, NC 27706 Tel: 919.660.7763 Fax: 919.684.2818 [email protected]

Page 7 Volume 24, Number 3 (December 2005) Decision Analysis Newsletter fective plural evaluation will, in fact, usually include Perspective one or more NDT models. If not, we decision analysts may need to explore other lines of work—or at least define decision analysis more broadly! How Prescriptive is Decision Theory? Rex Brown The role NDT (normative decision theory) plays in improving rationality is less obvious than I used to think. In particular, substituting the output of an NDT model for my judgment may not make me a more rational decider.

NDT produces tests of coherence, not prescriptions (at least not directly), much like arithmetic. If I think there are there are 8 men and 12 women in my class, and 18 in total, arithmetic alerts me that I am incon- sistent; it does not tell me what to change. So it is with NDT and its derivative, “Bayesian” inference. If I see black clouds and assess 80% chance of rain, but Bayes’ theory produces 50% from my prior and likelihoods, NDT alerts me that I am inconsistent; but it does not oblige me to change my probability of rain from 80% to 50%. Similarly, if I propose to bet $100 on a horse, but give only 5% chance of win- ning $1000, NDT alerts me that I am inconsistent . However, it does not tell me not to bet.

These alerts may help me make my judgment sounder, but not in any unambiguous way. There may be some unique ideal judgment that I would adopt if I could apply NDT perfectly to all my initial judgments--and not just those I choose to model . NDT could in prin- ciple get me there, if I had enough skill and resources. However, that does not mean that any NDT model that is practically feasible, i.e. DA (decision analysis), will get me closer to my ideal than direct judgment, because the model may tap less well into my total knowledge. (This may help explain why seasoned deciders often back their own judgment their DA.) Addressing my judgment several different ways (in- cluding intuition, non-DA techniques and canvassing other opinions) may use more of what I know than any single NDT model, however sound technically. Reconciling inconsistencies among conflicting results somehow (say, by changing least what I am most confident about), should close in on my ideal judg- ment--but I cannot prove it.

Getting my judgment closer to the ideal may not involve NDT at all. However, I like to think that ef-

Page 8