INFORMS Miami 2001 Annual Meeting Sponsor:
Decision Analysis

Decision Analysis Computation

SA30.1 Data Mining at Microsoft

• David E. Heckerman; Microsoft Research, One Microsoft Way, 9S/1024, Redmond, WA 98052-6399; heckerma@microsoft.com

I will discuss data mining tools developed over the last decade at Microsoft Research. Some of these tools appear in SQL Server 2000 and Commerce Server 2000. Others will ship in future Microsoft products. I will illustrate the use of these tools using data from domains including web navigation, ecommerce, TV viewing and movie preferences.

SA30.2 Sequential Valuation Networks

• Riza Demirer; Kansas University Business School, Summerfield Hall, 1300 Sunnyside Ave., Lawrence, KS 66045-7585; riza@ku.edu
• Prakash P. Shenoy; University of Kansas, School of Bus., Lawrence, KS 66045; pshenoy@ku.edu,, http://htt://lark.cc.ukans.edu/~pshenoy/

We will describe a new technique for representing and solving asymmetric decision problems. Our technique is a hybrid of the best features of sequential decision diagrams and valuation networks and it succeeds in overcoming the weaknesses of these 2 techniques.

SA30.3 Principle-Agent Value Alignment

• Daniel Shapiro; Institute for the Study of Learning & Expertise, 2164 Staunton Ct., Palo Alto, CA 94306; dgs@stanford.edu
• Ross D. Shachter; Stanford University, MSE Dept., Terman Engineering Ctr., Stanford, CA 94305-4026; shachter@stanford.edu

In a principal-agent problem the principal seeks to motivate the actions of an agent through incentives. We identify necessary and sufficient conditions for establishing trust between the principal and agent, even when they perceive the world in incommensurate terms, such as when the agent is a technological artifact.

Recent Developments in Risk & Decision Analysis at Caltech/JPL

SB30.1 Programmatic Risk Analysis to Search for Life on Mars

• Elisabeth M. Pate-Cornell; Stanford University, Dept. of MSE, Stanford, CA 94305; mep@leland.stanford.edu
• Robin Dillon; Virginia Tech, Pamplin Coll. of Bus., 7054 Haycock Rd., Falls Church, VA 22041; dillon@vt.edu

NASA is developing a strategy for the future exploration of Mars. To support the planetary exploration program, we describe an analytical framework that determines the optimal budget allocation based on the maximization of a valuation function that captures tradeoffs among technical and management failure risks and the mission's scientific benefits.

SB30.2 Risk-Adjusted Mission Value for Space Missions

• Ralph Miles; EER Systems, 3608 Canon Blvd., Altadena, CA 91001; rmiles2@earthlink.net

NASA has requested their Centers and the Jet Propulsion Laboratory to incorporate probabilistic risk assessment into their risk management activities. PRA cannot make mission risk vs. mission return tradeoffs because PRA has no 'value' component. We show how to combine PRA and decision analysis to make tradeoffs for space missions.

SB30.3 Preference Reversals in Risk Management: Where They Occur & How to Avoid Them

• Matthias Hild; Caltech, Humanities & Social Sciences, MC 228-77, Pasadena, CA 91125; hild@caltech.edu

Reversals of social preferences occur when the social decision maker aggregates individuals' subjective utilities for the outcomes of a risky policy measure. The level of detail with which these outcomes are described can significantly affect the resulting policy recommendation. We discuss an alternative approach to decision-theoretically sound risk management.

Risky Decision Making

SC30.1 The Effect of Medium on Choice under Uncertainty

• Christopher Hsee; University of Chicago, Grad. Sch. of Bus., 1101 East 58th St., Chicago, IL 60637; chris.hsee@gsb.uchicago.edu

When people try to achieve outcomes, often the immediate payoff is some 'medium,' which can be traded for desired outcomes, e.g., points, stocks and even money. When the relationship between effort and the (ultimate) outcome is uncertain, the medium creates an illusion of certainty, leading people to exert more effort.

SC30.2 Learning in Detection of Regime Shifts

• George Wu; University of Chicago, Grad. Sch. of Bus., 1101 East 58th St., Chicago, IL 60637; george.wu@gab.uchicago.edu
• Cade Massey; Duke University, Fuqua Sch. of Bus., Durham, NC 27708; pbmassey@gsbpdh.uchicago.edu

In previous research, we have documented a systematic pattern of over- and under-reaction that can be explained by a system neglect hypothesis. We extend these tests to situations in which subjects may be able to learn from feedback and find evidence of learning, but continued system neglect.

SC30.3 Risk Attitudes, Risk Perceptions, Revealed Preferences & the Use of Derivatives

• Joost M. E. Pennings; Wageningen University, Mktg./Consumer Behavior Group, Hollandseweg 1, Wageningen, 6706 KN , The Netherlands; joost.pennings@alg.menm.wau.nl
• Don N. Kleinmuntz; University of Illinois, Bus. Admin. Dept., 208 Wohlers, 1206 South Sixth St. MC-706, Champaign, IL 61820; dnk@uiuc.edu

We investigate and compare the risk attitudes and risk perceptions of decision makers from the financial service industries (insurance, banking) and private investors. Different elicitation methods were compared and validated against stated preferences for particular futures and options contracts, as well as actual behavior (use of derivative contracts).

SC30.4 Investment Valuation & Disclosure of Market Risk: An Experimental Investigation

• Don N. Kleinmuntz; University of Illinois, Bus. Admin. Dept., 208 Wohlers, 1206 South Sixth St. MC-706, Champaign, IL 61820; dnk@uiuc.edu
• Thomas J. Linsmeier; Michigan State University, Dept. of Acct., N253 North Business Complex, East Lansing, MI 48824-1121; tjl@pilot.msu.edu

U.S. companies must make quantitative disclosures of market risk exposures using 1 of 3 methods: tabular, sensitivity analysis or value-at-risk. We compare risk perceptions and valuations using these and several related methods for disclosure. Implications for corporate risk management and financial reporting are discussed.

Decision Theory

SC31.1 The Enterprise as a Network of Relationships: A New Approach to Employee Performance Evaluations

• Miguel S. Lobo; Duke University, Fuqua Sch. of Bus., Durham, NC 27708-0120;

As firms move away from strictly hierarchical structures, existing performance evaluation systems become limited in their information usage, and inadequate in their incentives. We present a new methodology, where evaluations from higher-ranked employees are given more confidence. We discuss probability models, applicability of network analysis tools and trial results.

SC31.2 Structural Properties of Stochastic Dynamic Programs

• James E. Smith; Duke University, Fuqua Sch. of Bus., Durham, NC 27708-0120; jes9@mail.duke.edu
• Kevin F. McCardle; UCLA, Anderson Grad. Sch. of Mgmt., Westwood Plaza, Ste. D520, Los Angeles, CA 90095-1481; kmccardl@anderson.ucla.edu

We present several fundamental results for establishing properties, i.e., monotonicity, convexity, supermodularity, etc., of value functions for stochastic dynamic programs. The results are 'meta theorems' showing that the value functions satisfy property P if the reward functions satisfy property P and the transition probabilities satisfy a stochastic version of P.

SC31.3 Uncertainty Aversion with Second-Order Probabilities & Utilities

• Robert F. Nau; Duke University, Fuqua Sch. of Bus., Durham, NC 27708-0120; robert.nau@duke.edu

The Ellsberg and Allais paradoxes can be parsimoniously explained by a simple model of smooth, partially-separable non-expected utility preferences in which the decision maker is seemingly uncertain about her decision model and is averse to the uncertainty. Equivalently, she displays different risk attitudes toward bets on 'ambiguous' and 'non-ambiguous' events.

SC31.4 withdrawn 10/25: Joint Coherence: The Continuous Case

• Kevin F. McCardle; UCLA, Anderson Grad. Sch. of Mgmt., Westwood Plaza, Ste. D520, Los Angeles, CA 90095-1481; kmccardl@anderson.ucla.edu

Practioners' Award Competition

SD30.1 Practitioners' Award Competition

• Detlof von Winterfeldt; University of Southern California, Sch. of Policy/Planning/Dev., Los Angeles, CA 90089; detlof@aol.com

Each year, the Decision Analysis Society of INFORMS presents an award for the best application of decision analysis. The finalists for the award will each present a summary of their work with the winner announced in the afternoon awards session.

SD30.2 Choosing a Development Strategy for a New Product at Amgen

• Phillip Beccue; Amgen Corp., 1 Amgen Center Dr., MS 27-5-A, Thousand Oaks, CA 91320; pbeccue@amgen.com

No abstract supplied.

SD30.3 Risk Assessment & Management for Chemical Weapons Disposal

• Gary J. Boyd; SAIC, 1809 Fashion Court, Ste. 107, Joppa, MD 21085; gary.j.boyd@saic.com
• Greg St. Pierre; US Army Program for Chemical Demilitarization, Risk Mgmt. & Quality Assurance; gregory.stpierre@pmc.apgea.army.mil

No abstract supplied.

SD30.4 Asset Strategy Analysis for an Oncology Drug

• Eric Johnson; Pharsight Corp., 800 West Camino Real, Mt. View, CA 94002; ejohnson@pharsight.com

This case study describes how we helped our client reach a well-reasoned consensus regarding how to continue development of their successful oncology drug. By doing so, we added $100 million of value. We describe the project management strategy that generated the consensus and some of the analytic techniques that made it well-reasoned.

Decision Making: Applications & Empirical Studies

MA30.1 Multiple Stakeholders' Perspectives on Management of Water Systems & Pollution

• L. Robin Keller; University of California, 350 Grad. Sch. of Mgmt., Irvine, CA 92697-3125; lrkeller@uci.edu,, http://www.gsm.uci.edu/~keller
• Jeffery L. Guyse; Cal Poly, Pomona; jlguyse@csupomona.edu
• Dipayan Biswas; University of California, Irvine; dbiswas00@gsm.uci.edu

We construct and compare different stakeholder groups' hierarchies of multiple objectives for management of water pollution in the Talbert Marshland and near-shore ocean in Huntington Beach. We trace the evolution of the pollution management problem from early detection through beach closures, to expansion of scope with the recent energy crisis.

MA30.2 Challenges in Measuring Financial Investors' Risk Aversion

• Philippe Delquie; INSEAD, Blvd. de Constance, Fontainebleau, , France; philippe.delquie@insead.fr
• Gillaume Cousin; INSEAD, Blvd. de Constance, Fontainebleau, 77305 , France; gcousin@be-partner.com

We review some difficulties encountered in eliciting the lay investor's preferences for return distributions: tendency to let attention slip to loss side only, valuation of a risk highly relative to the extremes of the range of risks, difficulty with continuous distributions, etc. From this, we try to design an ergonomic elicitation method.

MA30.3 An Experimental Study of Risk-Value Theory

• John Butler; Ohio State University, Dept. of Acct. & MIS, 2100 Neil Ave., Columbus, OH 43210;
• Jim Dyer; University of Texas; jim.dyer@bus.utexas.edu
• Jianmin Jia; Chinese University of Hong Kong, Fac. of Bus. Admin., Shatin, NT, , Hong Kong; jjia@cuhk.edu.hk

There are 2 basic assumptions in our risk-value theory: risk and preference are conversely consistent for lotteries with zero-means and preference order between 2 lotteries with the same mean will not change when the common mean changes. We will report the results of an experimental test of these assumptions.

Markets in Uncertainty

MA31.1 Decision Markets

• Robin D. Hanson; George Mason University, Dept. oc Economics, MSN 1D3, Carow Hall, Fairfax, VA 22030-4444; rhanson@gmu.edu

Speculative markets are consistently more accurate than competing institutions, recently forecasting elections, product sales, and scientific progress. Decision markets harness this power to forecast which decisions produce better consequences. They can thus tell a company when to dump its CEO and can even enable a new form of government: futarchy.

MA31.2 Election Forecasts from a Futures Market

• Forrest Nelson; University of Iowa, Dept. of Economics, Tippie Coll. of Bus., Iowa City, IA 52242; forrest-nelson@uiowa.edu
• Joyce Berg; University of Iowa, Dept. of Acct., Tippie Coll. of Bus., Iowa City, IA 52242; joyce-berg@uiowa.edu
• Thomas A. Rietz; University of Iowa, Dept. of Finance, Tippie Coll. of Bus., Iowa City, IA 52242; thomas-rietz@uiowa.edu

Futures markets aggregate information of traders about future values. The political stock markets conducted by the Iowa Electronic Markets capitalize on this price discovery mechanism to predict election outcomes. We examine the accuracy of such markets in 4 presidential elections and derive and compare alternative estimates of forecast errors.

MA31.3 Compact Securities Markets

• Michael P. Wellman; University of Michigan AI Laboratory, 1101 Beal Ave., Ann Arbor, MI 48109-2110; wellman@umich.edu
• David M. Pennock; NEC Research Institute, 4 Independence Way, Princeton, NJ 08540; dpennock@research.nj.nec.com

Complete securities markets, comprising one security for every state of nature, support Pareto optimal allocations but require a number of markets exponential in the uncertain events. By applying representational ideas from Bayesian networks, we identify conditions where one can construct 'operationally complete' securities markets with a relatively compact market structure.

MA31.4 The Power of Play: Efficiency, Information Aggregation & Forecast Accuracy in Market Games

• David M. Pennock; NEC Research Institute, 4 Independence Way, Princeton, NJ 08540; dpennock@research.nj.nec.com
• Steve Lawrence; NEC Research Institute, 4 Independence Way, Princeton, NJ 08540; lawrence@research.nj.nec.com
• C. Lee Giles; Pennsylvania State University, CSE Dept., 504 Rider Bldg., 120 South Burrowes St., University Park, PA 16801; giles@ist.psu.edu

Information aggregation in securities markets yield prices that constitute accurate forecasts. I show that this property extends even to Internet market games run with play money. Despite their lack of grounding in tangible assets, these market games show clear signs of economic efficiency, informational efficiency, and forecast accuracy.

Portfolio Decision Analysis

MB30.1 The Asset Allocation Decision: Portfolio Optimization for Investors

• Dane Odell; US Military Academy, Dept. of Systems Eng., West Point, NY 10996;
• John B. Willis; US Military Academy, Dept. of Systems Eng., West Point, NY 10996; john-willis@usma.edu

The US Government's Thrift Savings Plan is a retirement savings and investment program designed for government employees. Participants can invest in a choice of 5 funds, each with a historical risk/return profile. We address a method to optimally allocate money into the funds based on an individual's risk tolerance.

MB30.2 A Framework for Combining Multiple Objective Decision Analysis & Optimization

• Brian Stokes; US Military Academy, Dept. of Systems Eng., Mahan Hall, Thayer Rd., West Point, NY 10996; fb8654@usma.edu
• Gregory S. Parnell; US Military Academy, Dept. of Systems Eng., West Point, NY 10996-1779; gparnell@usma.edu
• Brad Foster; ;

In several recent DoD decision analysis applications, multiple objective decision analysis is used with optimization to develop the optimal resource allocation plan. We believe that fundamental modeling issues have not been adequately addressed. We develop a framework for alternative multiple objective value models.

MB30.3 Modeling the Cost Objective in a Decision Maker's Value Structure

• Frank Snyder; US Military Academy, Dept. of Systems Eng., West Point, NY 10996; frank-snyder@usma.edu

We examine 2 approaches for representing cost. First, minimizing cost is a fundamental objective in the value structure. Second, minimizing cost is treated separately. We plot value (utility) against cost for the decision maker. We analyze and compare these 2 cost modeling approaches. We conclude a survey of the audience.

MB30.4 Air Force Portfolio Analysis Application: Selecting F-16 Engine Modifications

• Stephen P. Chambal; AFIT/ENS, Bldg. 640, 2950 P St., WPAFB, OH 45433-7765; stephen.chambal@afit.edu
• Cihangir Yay; ;

We demonstrate a multi-attribute decision making process for the F-16 program office. An initial value hierarchy is developed for selecting engine modifications based on safety, cost and combat capability. Sample alternatives are ranked and evaluated based on cost. We select portfolios based on changing budget constraints.

Marketing Decision Analysis

MB46.1 Bayesian Updating of Customer Expectations

• Roland Rust; University of Maryland;

Some common beliefs about customer-perceived quality are wrong, once customer expectations are viewed as distributions, rather than simple point expectations. This may help explain why current measures of customer satisfaction only partially predict future behavior. We experimentally confirm most of the predictions of this model.

MB46.2 Integrating Gap Analysis & Utility Theory in Service Research

• Robert F. Bordley; General Motors, Renaissance Center, MC 482-D20-B24, PO Box 100, Detroit, MI 48265-1000; robert.bordley@gm.com

The value of a service is generally a function of the gap between how that product performs and customer expectations about how it should perform. As we show, the target-based approach to utility allows us to easily model this behavior. We show that while information and past experience generally increases the desirability of a product expected to perform above expectations...

MB46.3 Incorporating Endogenous Demand into the Newsvendor Problem

• David Bell; Harvard Business School;

In the classic newsvendor problem, lost sales have only a financial impact perhaps including some effect on goodwill. A reputation for lost sales lowers future demand. Surprisingly, the optimal inventory solution is still of the classic style. We extend the analysis to optimal inventories for multiple products.

MB46.4 Applying Principles of Decision Quality & Brand Management in Marketing Niche Vehicles

• Jennifer L. Meyer; Strategic Decisions Group, 2440 Sand Hill Rd., Menlo Park, CA 94025; jmeyer@sdg.net,, jmeyer@sdg.com
• Nazir Ahmad; Strategic Decisions Group, 2440 Sand Hill Rd., Menlo Park, CA 94025; nahmad@sdg.com

A stylized case study will demonstrate how a company proceeded with its problematic marketing of several niche brand vehicles in an industrialized market. Dialogues with management dealt with thorny analytical questions and with navigating through complex organizational and cultural issues about a company, its people and their identity in the marketplace.

Panel: Pre-College Decision Analysis Education

MC30.1 Panel: Pre-College Decision Analysis Education

• Ronald A. Howard; Stanford University, Dept. of MSE, Terman Eng. Ctr., Sch. of Eng., Stanford, CA 94305-4026;

For many years, there have been endeavors to introduce some form of decision analysis education in grade school and high school classes. A distinguished panel with relevant experience will discuss the benefits and challenges of such enterprises. Attendees will be invited to share their views.

Decision Analysis Awards

MD30.1 Decision Analysis Awards

• L. Robin Keller; University of California, 350 Grad. Sch. of Mgmt., Irvine, CA 92697-3125; lrkeller@uci.edu,, http://www.gsm.uci.edu/~keller

The Decision Analysis Society of INFORMS will announce the recipients of the 2001 Ramsey Medal for lifetime contributions to decision analysis, 2001 publications award for best publication in the year 1999 and the 2001 student paper award. Each winner will be invited to speak briefly. The winner of the Decision Analysis Practice award will also be announced. The practice award competitors will make their presentations in an earlier session.

Environmental Applications of Decision Analysis

TA30.1 Ecological Risk Ranking: Evaluation of a Method to Improve the Quality of Public Participation in Environmental Decision-Making

• Henry Willis; Carnegie Mellon University, Dept. of Eng. & Public Policy, Pittsburgh, PA 15213; hhw@andrew.cmu.edu
• Michael DeKay; ;
• Paul Fischbeck; ;
• Baruch Fischhoff; ;
• Keith Florig; ;
• Granger Morgan; ;
• Claire Palmgren; ;

Our research group has developed a method for ranking risk to health, safety and the environment. The method elicits ranking results using both holistic judgment and a multi-attribute utility model. We describe how the risk attributes were selected and report preliminary ranking results using our method and these attributes.

TA30.2 A Framework for Estimating the Value of Improving Toxics Release Inventory Data

• Charles D. Linville; American University, Dept. of CS & IS, Washington, DC; linvill@email.clark.american.edu
• Robert R. Bruno; American University, Dept. of CS & IS, Washington, DC;

Databases of geographically-linked environmental information are of considerable value in environmental protection. We describe a framework for estimating the expected value of improving the information in one such database, the US EPA's Toxics Release Inventory, established by the Emergency Planning and Community Right-To-Know Act and the Pollution Prevention Act.

TA30.3 Evaluating the EPA's Toxicological Testing Program for Children's Health: A Decision Analysis Approach

• Fumie Yokota; Harvard School of Public Health; yokota@fas.harvard.edu

A Bayesian decision theory framework is used to characterize the expected value of information from the Voluntary Children's Chemical Evaluation Program. The prior distribution is modeled using a hierarchical Bayesian approach threshold analysis of the posterior distribution will determine the conditions under which VOI exceeds the cost of the program.

TA30.4 Connecting Data Quality to Decision Making: Developing a Framework

• Kara M. Morgan; Research Triangle Institute, 1516 M St. NW, Ste. 740, Washington, DC 20036; kmorgan@rti.org
• Malcolm J. Bertoni; Research Triangle Institute, 1615 M St. NW, Ste. 740, Washington, DC 20036;

The US Environmental Protection Agency generates a large amount of environmental measurement data. The relationship between data quality and decision making is complex and not well understood by data generators or data users, i.e., decision makers. We investigate methods for using data quality information to improve decision making.

Topics in Decision Analysis

TB15.1 What Question to Ask Next to Aid Multi-Attribute Alternative Selection

• Hillary A. Holloway; University of Michigan, 1205 Beal Ave., IOE Bldg., Ann Arbor, MI 48109-2117; hhollowa@engin.umich.edu
• Chelsea C. White, III; University of Michigan, Dept. of IOE, 1205 Beal Ave., Ann Arbor, MI 48109-2117; ccwiii@umich.edu

We examine the problem of what question to ask next when the decision model imprecisely specified multi-attribute value theory is applied in an interactive decision-aiding environment. We formulate the problem as a dynamic program and illustrate how it can be applied in a computationally efficient manner.

TB15.2 Real Options Valuation using Decision Trees & Net Present Value

• Zeger Degraeve; London Business School, Sussex Place, Regent's Park, London, NW1 4SA , UK; zdegraeve@london.edu
• Bert De Reyck; London Business School, Regent's Park, London, NW1 4SA , UK; bdereyck@london.edu

We reject the claim that flexibility in investments necessarily has to be evaluated using real option methodology and show that the net present value framework can still be used for a sound investment analysis. We extend decision tree analysis by determining appropriate discount rates at different chance nodes, reflecting the project's unique and market risk at that stage.

TB15.3 Scored Portfolios, Efficient Frontiers & Dominant Policies

• Zvi Covaliu; Fair, Isaac & Co., Inc., Baltimore, MD; zvicovaliu@fairisaac.com
• Peter Beling; University of Virginia; beling@virginia.edu
• Robert M. Oliver; Fair, Isaac & Co., Inc., Baltimore, MD; boboliver@fairisaac.com

Account acquisition decisions in scored credit or loan portfolios have added additional business objectives to risk management, e.g., profit and market share. We establish equivalence between ROC dominance, maximizing expected profit and efficient frontiers dominance in multiple objective space. These interesting results guide policy choice when multiple scorecards are available.

TB15.4 Stimulating the Creation of Design Alternatives using Customer Values

• Ralph L. Keeney; University of Southern California, 101 Lombard St., Ste. 704W, San Francisco, CA 94111; keeneyr@aol.com

Knowing your customers' desires, you can create alternatives better for them that also increase your organization's profits and market share. We elicited objectives and desired characteristics from knowledgeable individuals and customers for cell phone and cellular phone plans. Techniques using these objectives and characteristics illustrate how to identify new alternatives.

DA Arcade I

TB30.1 A Simple Mixture Model for Approximating Binary Dependence

• Donald L. Keefer; Arizona State University, Dept. of Supply Chain Mgmt., Tempe, AZ 85287-4706; don.keefer@asu.edu

We propose a new simple model for approximating positive probabilistic dependence among binary events such as project success/failures. This model is a mixture between high and low dependence models and requires assessment of just one conditional probability in addition to the n marginals.

TB30.2 A Strategy for Regulating Environmental Risk

• Rex V. Brown; George Mason University, Sch. of Public Policy, 2018 Lake Breeze Way, Reston, VA 20191; rbrown@gmu.edu

We present an approach to formulating environmental regulations that balances public protection, cost and effective enforcement, and builds on existing judgment of responsible regulators. They set a hierarchy of 'redundant' requirements that decrease in stringency. A personalist extension of conventional Probabilistic Risk Assessment tests whether hazardous activites comply with them.

TB30.3 Telephone Utility Assessment of Cancer Pain Health States

• Scott B. Cantor; University of Texas, Anderson Cancer Ctr., 1515 Holcombe Blvd., Box 196, Houston, TX 77030-4009; sbcantor@mdanderson.org
• Guadalupe R. Palos; University of Texas, Anderson Cancer Ctr., 1515 Holcombe Blvd., Box 221, Houston, TX 77030-4009; gpalos@ndanderson.org
• LuAnn Aday; University of Texas, Sch. of Public Health, 1200 Herman Pressler St., Houston, TX 77225; laday@sph.uth.tmc.edu
• Tito R. Mendoza; University of Texas, Anderson Cancer Ctr., 1515 Holcombe St., Box 221, Houston, TX 77030-4009; tmendoza@mdanderson.org
• Charles S. Cleeland; University of Texas, Anderson Cancer Ctr., 1515 Holcombe St., Box 221, Houston, TX 77030-4009; ccleeland@mdanderson.org

We evaluated the feasibility of performing utility assessment by telephone for health state related to cancer pain and its treatment. We elicited preferences using the numerical rating scale method for 302 adults. We evaluated the validity of the methodology and analyzed the trade-offs between pain severity and adverse side effects.

TB30.4 A Clinical Decision Model for Breast Cancer Prevention Decisions

• Elissa Ozanne; Stanford University, Dept. of MSE, Terman Engineering Ctr., Stanford, CA 94305-4023; elissa@stanford.edu
• Ronald A. Howard; Stanford University, Dept. of MSE, Terman Eng. Ctr., Sch. of Eng., Stanford, CA 94305-4026;
• Laura Esserman; UCSF Carol Franc Buck Breast Care Center, 1600 Divisadero St., 2nd Floor, San Francisco, CA 94115; laura.esserman@ucsfmedctr.org

Using a structured consultation plan focused on patient education, individual risk estimation, and preference assessment we are integrating decision analysis into real-time constraints of a clinic, toward improving decision quality for patients at high risk for breast cancer. We present an overview of the method and results of pilot tests.

Analysis of Management & Organizational Factors

TC30.1 Using Strategic Reference Points in Developing Models of the Enterprise

• Robert F. Bordley; General Motors, Renaissance Center, MC 482-D20-B24, PO Box 100, Detroit, MI 48265-1000; robert.bordley@gm.com

No abstract supplied.

TC30.2 A Dynamic Approach to Organizational Fit & Performance

• Richard M. Burton; Duke University, Fuqua Sch. of Bus., Durham, NC 27708-0120; rmb2@duke.edu
• Borge Obel; Odense University, Dept. of Organization & Mgmt., Odense, DK-5230 , Denmark; boe@sam.sdu.dk
• Michael Roach; Duke University, Fuqua Sch. of Bus., Durham, NC 27708-0120; michael.roach@duke.edu

Quantitative and qualitative public data on US bank holding companies are used as input for an expert system to identify organizational misfits. Longitudinal techniques are applied to explore the dynamic relationship between misfits and performance. We argue that organizations should reduce all misfits to realize improvements in performance.

TC30.3 Influences of Management & Organizational Factors on Risk

• Vicki Bier; University of Wisconsin, 1513 University Ave., Madison, WI 53706; bier@engr.wisc.edu

Management and organization have significant effects on risk. There is no one 'correct' management style or structure. Rather, key aspects of an organization should match the environment and be internally consistent. Understanding this issue is a challenging research agenda at the interface of engineering and social science.

TC30.4 withdrawn 10/10: A Framework for Organizational Risk Analysis

• Gregg Courand; Synergia LLC, 2400 Broadway, Ste. 203, Reswood City, CA 94063; courand@synergia.com
• Michael R. Fehling; Synergia LLC, 2400 Broadway, Ste. 203, Redwood City, CA 94063; fehling@synergia.com

DA Arcade II

TD30.1 withdrawn 10/12: Paradox-Free Decision Modeling

• Fatma Mili; Oakland University; mili@oakland.edu

TD30.2 A Multi-Objective Decision Analysis Approach to Watershed Management

• Jason Merrick; Virginia Commonwealth University, Dept. of Stats Science & OR, 1015 West Main, PO Box 842014, Richmond, VA 23284; jrmerrick@mail1.vcu.edu
• Gregory S. Parnell; US Military Academy, Dept. of Systems Eng., West Point, NY 10996-1779; gparnell@usma.edu

In an NSF funded project, we integrated the views of diverse watershed stakeholders using multi-objective decision analysis. Each of the stakeholders' beliefs, strategies and goals were brought together under a watershed management framework to meet the overall objective of the project committee: to maximize the quality of the watershed.

TD30.3 Using a Bayesian Approach to Quantify Scale Compatibility Bias

• Richard M. Anderson; JHU, 3400 North Charles St., 313 Ames Hall, Baltimore, MD 21218; anderson@titan.me.jhu.edu
• Benjamin F. Hobbs; JHU, Sch. of Engineering, Geography & Environ. Eng., Baltimore, MD 21218; bhobbs@jhu.edu

We use a Bayesian method to compute probability distributions of attribute weights, assuming that the ratios produced by each respondent's judgments are subject to random error and scale compatibility bias. The bias is found to be significant and in the predicted direction. A heuristic is effective in eliminating the bias.

TD30.4 Exploiting Supply & Demand Curves for Decision Analysis

• Jeffrey Keisler; University of Massachusetts, 100 Morrissey Blvd., M/5-230, Boston, MA 02125; jeff.keisler@umb.edu

Business strategy decisions can be efficiently modeled using standard microeconomic structures combined with traditional DA. I will briefly discuss motivating examples from practice and specific modeling techniques and challenges.