C O V E R F E A T U R E

0018-9162/06/$20.00 © 2006 IEEE66 Computer P u b l i s h e d b y t h e I E E E C o m p u t e r S o c i e t y

sible. Sophisticated urban simulation models can sup-port such analyses by predicting the long-term effects ofalternative policies.

But data is not the only element of effective democ-ratic decision making: In a democratic society, publicdeliberation by citizens and their elected representativesmust precede such major decisions. Thus, in addition toproviding accurate information, a design goal for anurban simulation system should also be to facilitate pub-lic understanding and citizen engagement.

We present a snapshot of ongoing research into the design of user interactions around the results from one such simulation system. UrbanSim1 (www.urbansim.org), a large-scale urban simulation systemdeveloped at the University of Washington, projects pat-terns of land and transportation use and the environ-mental impact of various policies and investments over20 years or more.

Results from UrbanSim take the form of indicators2—variables that represent key aspects of the simulationresults, such as population and employment density. Tobetter support the use of urban simulation systems inpublic deliberation, we are designing tools to supporturban planners, citizens, and other stakeholders in theirinteraction with UrbanSim indicators. The developmentof these tools is guided by Value Sensitive Design,3 a the-oretically grounded approach to technology design that

Sophisticated simulation systems such as UrbanSim model the long-term impacts of

transportation and land-use alternatives. Accounting for human values throughout the

design process helps in designing interactions that engage both planners and citizens in

the decision-making process.

Janet DavisGrinnell College

Peyina Lin, Alan Borning, Batya Friedman, Peter H. Kahn Jr., and Paul A. Waddell University of Washington

R egional officials, urban planners, and citizensmust grapple with issues such as traffic jams,resource consumption, and urban sprawl. Deci-sions about new freeways, transit service expan-sion, or land-use regulations are often contro-

versial and expensive, with long-term consequences. In many cases, neither planners nor citizens have suf-

ficient information about how the various alternativeswill play out over the long term. This represents a gapin informed decision making because long-term assess-ment is essential.

A metropolitan region that is considering a new lightrail system or a new freeway—both major invest-ments—must look beyond the immediate effects on traf-fic congestion. Decision-makers must attempt to under-stand how different alternatives might affect land use,transportation, and environmental impacts over the nextseveral decades. Without carefully considering long-term effects, the chosen alternative might lead to theexact opposite of the original goal: A new freeway’s con-venient access, for example, could result in more devel-opment of new houses and businesses far away fromexisting employment and population centers, which inturn could increase traffic congestion.

Because these decisions will affect the entire regionfor many years, it makes sense to inform decision mak-ing with long-term analyses that are as accurate as pos-

Simulations for UrbanPlanning: Designing forHuman Values

September 2006 67

accounts for human values throughout thedesign process. The “What Is Value SensitiveDesign?” sidebar gives more details aboutthis approach.

Our application of this method has led tofive interaction design goals:

• Improve the system’s functionality bydeveloping new tools for stakeholders tolearn about, select, and visualize indica-tors to use in decision making.

• Support citizens and other stakeholdersin evaluating alternatives with respect totheir own values.

• Enhance the system’s transparency withrespect to its design, assumptions, andlimitations—so it is not a black box.

• Contribute to the system’s legitimacy byproviding information that is credible andappropriate to the use context.

• Foster citizen engagement in the decisionprocess by providing tailored informa-tion and opportunities for involvement.

To meet these goals, we developed and arerefining three tools to help a variety of stake-holders—planners, modelers, citizens—understand UrbanSim’s indicators: technicaldocumentation designed to make informa-tion about indicators readily accessible; indi-cator perspectives that provide a platformfor organizations to advocate for the use ofparticular indicators in decision making; andhousehold indicators that let citizens look atsimulation results from the viewpoint oftheir own household within the region.

INDICATORS AND URBANSIMULATION

UrbanSim is a complex software systemthat models a region’s urban processes overthe next several decades. The system takeshours—sometimes days—to run, resultingin a massive database that contains detailedinformation about the region’s households,jobs, travel routes, and real estate in eachsimulated future year. Our indicator toolsaim to help stakeholders extract useful information fromthis very large database.

UrbanSim currently supports 55 indicators for extract-ing information from simulation results, and one of ourgoals in the indicator tools is to support urban plannersin developing new indicators for their particular region.Presenting simulation results with a consistent set ofindicators for all the candidate policy alternatives cangreatly enhance scenario assessment and comparison.

For example, suppose that stakeholders want to fos-ter compact urban neighborhoods that make it easy andpleasant to walk and at the same time keep outlyingareas rural by containing sprawl. In urban planning lit-erature, population density is a key indicator of devel-opment character (in this case, walkable neighborhoodsversus urban sprawl). Planning agencies can monitorpopulation density throughout the region to understandcurrent trends and use UrbanSim to understand the pos-

What Is Value Sensitive Design?

Value Sensitive Design is a theoretically grounded approach to technology design that accounts for human values, such as privacy,fairness, and democracy, throughout the design process. The methodhas three key features: an interactional perspective, attention to indirectas well as direct stakeholders, and a tripartite methodology.

An interactional perspective views values as stemming from a symbiosisof technology and social forces: People and social systems influencetechnological development, while technologies shape individual behaviorand social systems.

User-centered design methodologies focus their attention on directstakeholders—those who actually use the system.Value Sensitive Designemphasizes consideration of indirect stakeholders as well—those who donot use the system but are affected by its use. For UrbanSim, directstakeholders are urban planners and modelers; indirect stakeholdersinclude the residents of the region being modeled. Part of the UrbanSimvision is to empower indirect stakeholders to become direct stakeholders—to let citizens interact directly with UrbanSim’s output,and ultimately to run different simulations themselves.

Finally, the tripartite methodology consists of conceptual, empirical, andtechnical investigations.The application of these investigations is bothiterative and integrative; results from new investigations build on andintegrate earlier ones.

Conceptual investigations comprise philosophically informed analysesof the values at stake. Empirical investigations focus on the humanresponse to the technical artifact and on the larger social context inwhich the technology is situated. Technical investigations focus on thedesign and performance of the technology itself.

For example, in our UrbanSim conceptual investigations, we sharplydistinguished stakeholder values—those that some stakeholders, but notnecessarily all, would view as important—and explicitly supportedvalues—those for which we could make principled arguments and thatwe explicitly support in UrbanSim’s design. Explicitly supported valuesinclude freedom from bias, representativeness, and support for a democratic society.

As part of supporting a democratic society, we decided that the system should not a priori favor or rule out any given set of stakeholdervalues, but should allow various stakeholders to articulate the valuesmost important to them and to evaluate alternatives in light of thesevalues. Furthermore, our translation of Jürgen Habermas’s four elementsof legitimation potential into actionable design goals illustrates howdifferent kinds of investigations build on one another—in this case, atechnical investigation builds on a conceptual investigation.

68 Computer

sible impact of various policies on population density30 years later.

UrbanSim’s interacting component models representmajor actors and processes in the urban system.1 The sys-tem takes a highly disaggregated approach, modelingindividual households, jobs, and real-estate developmentand location choices, using relatively small grid cells (typ-ically 150 × 150 meters), such as those in Figure 1.

Most UrbanSim models are discrete-choice models, inwhich the probability that a given agent will make a par-ticular choice is a function of a set of variables that mea-sure the relative attractiveness of that choice. Forexample, in the Residential Location Choice model, theprobability that a particular household will choose tolocate to a residential unit within a particular areadepends on household attributes, such as income andnumber of children, as well as attributes of the potentialdwelling, such as cost and location.

An external travel model simulates trips between thelocations of various households and jobs. The resultingpatterns of transportation use and congestion then giverise to accessibility measures for different locations,which in turn influence the desirability of these locationsfor housing or jobs.

The most recent version, UrbanSim 4, is built on theOpen Platform for Urban Simulation, an object-orientedarchitecture and platform recently developed at theUniversity of Washington.4 OPUS and UrbanSim 4 areimplemented in Python, using highly optimized arrayand matrix manipulation packages written in C++ to

handle inner loop computations.The system is open source, underthe GNU public license.

Currently, regional planningagencies are transitioning Urban-Sim into operational use in thePuget Sound region (Seattle andsurrounding areas), Honolulu, andSalt Lake City. Urban plannersalready use UrbanSim to informdecisions in Houston, and therehave been research and pilot appli-cations in Detroit, Eugene, andPhoenix, as well as in Amsterdam,Paris, Tel Aviv, and Zurich. Urban-Sim also played a significant rolein the out-of-court settlement of alawsuit in Utah regarding a majorhighway construction project.5

UrbanSim refinement is ongoing,and it is evolving as we add newmodels and capabilities.

VALUES IN DEMOCRATICPLANNING

Commitment to three core val-ues helps UrbanSim achieve the goal of supportingdemocratic urban planning: democratic engagement,freedom from bias, and political legitimacy.

Democratic engagementDemocratic urban planning requires that citizens be

engaged in decision making. While acknowledging thatengaged citizenship is not simple to characterize, MichaelX. Delli Carpini6 offers this definition: “A democrati-cally engaged citizen is one who participates in civic andpolitical life, and who has the values, attitudes, opinions,skills, and resources to do so effectively.”

To foster engaged attitudes, consistent opinions, andenthusiastic participation, a planning system must pro-vide information about the issues that form the sub-stance of political life. UrbanSim helps fulfill thatrequirement by providing information about the poten-tial impacts of land-use and transportation alterna-tives—a major political issue.

But providing information is not enough. Citizens mustwant to use it. Through UrbanSim, we seek to foster suchdemocratic engagement, not only to help citizens makemore informed decisions, but also to encourage an atti-tude that can lead to participation in public decision mak-ing. Information systems, such as online discussion forumsor tools for citizens to propose new policy and investmentpackages, for example, could provide new opportunitiesfor citizen participation in urban planning. Of course, sys-tems such as UrbanSim supplement—not replace—infor-mal discussions, town meetings, and voting.

Figure 1. UrbanSim’s geographic representation.The small white polygons are individ-

ual lots, while the blue squares are the 150-meter grid cells that form the core of Urban-

Sim’s geographic representation.The data is for the Green Lake neighborhood in Seat-

tle, Washington.

September 2006 69

Freedom from biasBatya Friedman and Helen Nissenbaum7 refer to bias

in computer systems as

computer systems that systematically and unfairly dis-criminate against certain individuals in favor of others.A system discriminates unfairly if it denies an opportu-nity or a good or if it assigns an undesirable outcome toan individual or a group of individuals on grounds thatare unreasonable or inappropriate.

To warrant the term biased, then, dis-crimination must be both systematicand unfair.

We first identified freedom frombias as an explicitly supportedvalue—one that we wanted the sim-ulation to support—because it is amoral good in itself. However, thereare other reasons to support thisvalue: Freedom from bias is instru-mental in providing an equal opportunity to participatein a democratic society; stakeholders whose concernsare represented in the system could have a privilegedplace in deliberation relative to those whose concernsare not represented.

Political legitimacyUrbanSim’s legitimacy is crucial for its effective use in

urban planning. Unresolved disagreements about itslegitimacy might disenchant some stakeholders or causethe agency to stop using the system.

Our conceptual investigation of political legitimacydraws primarily on the work of Jürgen Habermas.8 Theuse of modeling software is just one part of the planningprocess, and even the best-designed system could be usedin a process lacking in legitimacy. Because most factorsare beyond our control, we focus on the modeling sys-tem’s “legitimation potential”8 rather than the legiti-macy of the entire decision-making process.

Communicative action plays a key role in legitimationpotential. Habermas defines communicative action asspeech in which all parties aim for mutual understandingwithout manipulative or strategic designs. In commu-nicative action, each utterance raises four validity claims,which we have mapped to testable design goals. Achievingthese goals helps establish UrbanSim’s legitimacy.

• Comprehensibility. Can a wide range of stakehold-ers understand the information provided?

• Accuracy and transparency. Are the models and dataa reasonable representation of reality? Are the innerworkings and design of UrbanSim transparent tostakeholders so they can assess its accuracy?

• Clarity of intent. Is the intent behind the informa-tion—to advocate for a particular position, or to pro-

vide relatively neutral, factual information—clear tothe users?

• Appropriateness with respect to values and norms.Is the information relevant to the stakeholders’ val-ues in the decision-making context?

Although comprehensibility, accuracy, transparency,freedom from bias, and relevance to decision makingare not new goals for operational models,9 tying thesegoals to the potential for achieving legitimacy helps us

understand their significance formodels that must support democra-tic decision making.

INTERACTION DESIGNThe part of our work described in

this article is intended primarily tocreate an interaction design aroundUrbanSim indicators through tech-nical documentation, indicator per-spectives, and household indicators.

Just the factsWe designed the technical documentation to provide

comprehensible, useful, factual information about theindicators to urban planners and other stakeholders,with an eye toward minimizing both actual and per-ceived bias. By putting technical information about indi-cators in a single place on the Web, we intended to makethis information easy to access, or “ready to hand,”10

during the deliberation that precedes decision making.Feedback from urban planners, modelers, and policy

experts led us to standardize the technical documenta-tion for each indicator to consist of 11 sections, includ-ing (among others) its name, an informal definition, amore formal specification, known limitations, andadvice for interpreting results. The technical documen-tation also includes the Structured Query Language(SQL) code used to compute the indicator from data-bases of simulation results, as well as input and expectedoutput for a unit test to check the code’s correctness.The technical documentation is “live” in that the Webserver extracts SQL code and tests directly from the codebase each time it displays it, guaranteeing that what theuser reads in the technical documentation is current. Wealso provide similar functionality for the new Python-based indicator computations in UrbanSim 4.

We conducted a study11 of our technical documenta-tion design with eight urban planners interested inUrbanSim, who currently represent its primary usercommunity. Through interviews in which the plannersinteracted with the system, we learned that they requiredmuch less time to complete each of four tasks using thetechnical documentation as compared with their cur-rent work practices. This result is evidence that we haveat least partially solved the information fragmentation

Democratic

urban planning requires

that citizens

be engaged in

decision making.

70 Computer

problem in terms of both consolidating information andmaking it readily accessible. This in turn would improvetask performance by increasing comprehension andmaking indicator evaluation more meaningful. Theresults also supported our hypothesis that including liveSQL code, unit test information, and limitationsincreases indicator transparency and comprehensibility.

Room for advocacyIn formative evaluations of our design for technical

documentation, much of the strongest feedback wasabout neutrality. Earlier documentation versionsincluded a section describing thedesired direction of change for theindicator, which we had thought use-ful in decision making. However,information in this section reflectedwidespread disagreement about thedesired direction for many indica-tors. Some participants indicated thateven the name, “Desired Direction,”conveyed bias. On the basis of thisfeedback, we replaced the problem-atic section with “Interpreting Results,” which relateshow the indicator might be used in policy evaluation.

Given the goal of democratic urban planning, havingstakeholders advocate values and put forth opinions isan essential and integral part of the overall process, notan inconvenient blemish on an otherwise clean techni-cal exercise. How then could we enable stakeholders touse indicators to represent and express their views, yetmaintain the informative role of technical documenta-tion? Our solution was to construct indicator perspec-tives, which stakeholders can use to tell a story andadvocate particular values and criteria for evaluatingoutcomes.

We have partnered with three local organizations toconstruct prototype perspectives: a government agency,a business association, and an environmental group—King County Budget Office, Washington Association ofRealtors, and Northwest Environment Watch, respec-tively. Each organization provides content for each per-spective, and the UrbanSim team provides technicalsupport. In keeping with our emphasis on fair repre-sentation, we chose partners that cover a range of views.We plan to provide opportunities for broader involve-ment, actively soliciting partners as needed to ensurethat the perspectives cover a wide range of politicalviews and economic interests.

We are currently analyzing the results of a study toassess our success at simultaneously providing facts andrelatively neutral technical information through docu-mentation, while also supporting value advocacy andopinion through the indicator perspectives. We alsoinvestigated how these systems approach our ideal offreedom from bias.

In the study, we engaged 20 Seattle citizens in inter-acting with both the indicator perspectives and techni-cal documentation and in reflecting on their per-ceptions. Preliminary results confirm that the indicatorperspectives framework is indeed useful in advocatingfor specific views and values and will be a valuablesource of information about UrbanSim indicators.

Personal touchWhile indicators such as population density and

total vehicle miles traveled are familiar to urban plan-ners who monitor or model regional trends, such

aggregate measures are probablyless compelling to citizens not wellversed in urban planning. To reachthese citizens, we created householdindicators,12 tailored simulationresults that show how policy alter-natives could affect their ownhouseholds. Through such indica-tors, we hope to encourage citizensto become involved in evaluatingthe impact of transportation and

land-use choices. When users launch the Web application, they supply

personal information. On the basis of this information,the application answers questions the users might pose:Where could I afford to live in the region? How longwould it take to get to work? How long would I have totravel to get out of the city?

Initial user study results12 support the hypothesis thatcitizens can more readily understand household indica-tors because they can compare such indicators directlyto living, working, and getting around. Household indi-cators also aim to engage citizens by showing them howpolicy decisions could affect their lives in the long term.In that sense, they answer the broader question, Howwill this decision affect my future?

Earlier work on the technical documentation aimedto provide comprehensible, accurate, transparent, use-ful, and relatively neutral technical information to urbanplanners. Household indicators, in contrast, focus onproviding information that is comprehensible to citizensand clearly relevant to their own lives. However, ourcommitments to accuracy, transparency, and freedomfrom bias remain.

In meeting our accuracy commitment, we were care-ful not to oversimplify to enhance comprehensibility.The commitment to transparency presents the challengesof conveying uncertainty in simulation results and hav-ing results explanations at hand when questions arise. Afocus on personal impacts might bias deliberationtoward individual rather than societal or environmentalimpacts—a focus we hope to balance by linking house-hold indicators to related regional indicators and indi-cator perspectives.

Including live SQL code,

unit test information,

and limitations increases

indicator transparency

and comprehensibility.

Because one of ourmain aims is to engagecitizens, users begin bycreating a householdprofile. Seeing your ownname and personal in-formation throughoutthe system fosters asense of identity thatpersists while viewingdifferent indicators. Thehousehold profile in-cludes the household’sname, its approximatelocation, a set of placesimportant to the house-hold, and descriptionsof trips between thoseplaces.

As Figure 2 shows,the current trip inter-face lets the user spec-ify a start location, adestination, a time ofday, and a travel modefor a particular trip.On the basis of thisinformation, the usercan view indicatorssuch as travel times, employment and population den-sity in their neighborhoods or other important places,land-use mix in their neighborhood, and property val-ues. Each indicator display includes a question or agroup of questions the indicator is intended toaddress—a strategy that encourages users to ponderhow the information might apply to their own lives.We designed the indicator displays to let the user com-pare the same indicator for all the alternatives underconsideration.

Our central hypothesis in designing to inform citizensis that simulation results in familiar terms will be easierto understand. Early user studies and informal feedbacksupport this hypothesis; indeed, there has been a pushfor greater realism and detail in reporting householdindicators. In early paper prototypes, the Travel Timesindicator showed only the round-trip travel timebetween home and work at peak travel times. Some par-ticipants—those who travel at nonpeak times, from loca-tions other than home, to destinations other thanwork—found this difficult to relate to. In response, werevised this indicator to use trip profiles that include astarting location, destination, travel mode, and time ofday. Colleagues have also noted that a trip might includeseveral stops along the way, for example, to buy coffeeor leave a child at day care, a phenomenon that urbanplanners call trip chaining. The current interface, which

focuses on one-way trips, better supports queries involv-ing trip chaining.

We have recently completed a prototype implemen-tation of household indicators and conducted focusgroups to further inform the design; a summative eval-uation of the concept and design is planned for futurework.

W e have presented but a snapshot of our work toinform public deliberation and decision makingabout major land-use and transportation issues.

Our future work will include further research on apply-ing UrbanSim in real planning contexts and on develop-ing tools to support citizen discussion and comment. Allthese efforts are part of an overall agenda: to better sup-port informed public deliberation and democraticengagement in the urban planning process. ■

Acknowledgments

This work has been supported in part by the NationalScience Foundation under grant nos. EIA-0121326, IIS-0325035, and IIS-0534094. We thank current and for-mer members of the UrbanSim team and the ValueSensitive Design Research Lab for their contributions tothe work presented here.

September 2006 71

Figure 2. Household indicators prototype, showing the trip configuration page. Household

indicators aim to show citizens how transportation and land-use alternatives will affect their

everyday lives—for example, how their travel time between places could change under different

alternatives.

72 Computer

References1. P. Waddell, “UrbanSim: Modeling Urban Development for

Land Use, Transportation and Environmental Planning,” J.Am. Planning Assoc., vol. 68, no. 3, 2002, pp. 297-314.

2. G.C. Gallopín, “Indicators and Their Use: Information forDecision Making,” Sustainability Indicators: Report on theProject on Indicators of Sustainable Development, B. Moldanand S. Billharz, eds., John Wiley & Sons, 1997, pp. 13-27.

3. B. Friedman, P. Kahn, and A. Borning, “Value Sensitive Designand Information Systems,” Human-Computer Interaction inManagement Information Systems: Foundations, P. Zhangand D. Galletta, eds., M.E. Sharpe, 2006.

4. P. Waddell et al., “OPUS: An Open Platform for Urban Simulation,” Proc. 9th Conf. Computers in Urban Planning and Urban Management, 2005; www.urbansim.org/papers/opus-2005.pdf.

5. P. Waddell and A. Borning, “A Case Study in Digital Govern-ment: Developing and Applying UrbanSim, a System for Sim-ulating Urban Land Use, Transportation, and EnvironmentalImpacts,” Social Science Computer Rev., vol. 22, no. 1, 2004,pp. 37-51.

6. M.X. Delli Carpini, “Mediating Democratic Engagement: TheImpact of Communications on Citizens’ Involvement in Polit-ical and Civic Life,” Handbook of Political CommunicationResearch, L.L. Kaid, ed., Lawrence Erlbaum Assoc., 2004, pp.395-434.

7. B. Friedman and H. Nissenbaum, “Bias in Computer Sys-tems,” ACM Trans. Information Systems, vol. 14, no. 3, 1996,pp. 330-347.

8. J. Habermas, Communication and the Evolution of Society,T. McCarthy, translator, Beacon Press, 1979.

9. K.R. Fleischmann and W.A. Wallace, “A Covenant withTransparency: Opening the Black Box of Models,” Comm.ACM, vol. 48, no. 5, 2005, pp. 93-97.

10. T. Winograd and F. Flores, Understanding Computers andCognition: A New Foundation for Design, Ablex, 1986.

11. A. Borning et al., “Informing Public Deliberation: Value Sen-sitive Design of Indicators for a Large-Scale Urban Simula-tion,” Proc. European Conf. Computer-Supported Coopera-tive Work (ECCSCW 2005), Springer, 2005, pp. 449-468.

12. J. Davis, “Household Indicators: Design to Inform and EngageCitizens,” CHI 2006 Extended Abstracts on Human Factorsin Computing Systems, ACM Press, 2006, pp. 688-693.

Janet Davis is an assistant professor of computer science atGrinnell College. Her research interests include applyingValue Sensitive Design to technologies such as ubiquitouscomputing and tools for civic engagement as well as the roleof values in computer science curriculum. Davis received aPhD in computer science from the University of Washing-ton. She is a member of the ACM. Contact her at davis-jan@grinnell.edu.

Peyina Lin is a PhD student in the Information School atthe University of Washington. Her research interests focuson information technology, youth community-basedengagement, and loss of public space. She received an MSc in information management from Syracuse Univer-sity and an MA in telecommunication digital media arts from Michigan State University. Contact her at pl3@u.washington.edu.

Alan Borning is a professor in the Department of ComputerScience and Engineering, an adjunct professor in the Infor-mation School, and codirector of the Center for Urban Sim-ulation and Policy Analysis at the University of Washington.His research focuses on human-computer interaction anddesigning for human values, particularly as applied to landuse, transportation, and environmental modeling. Borningreceived a PhD in computer science from Stanford Univer-sity. He is a fellow of the ACM. Contact him at borning@cs.washington.edu.

Batya Friedman is a professor in the Information Schooland an adjunct professor in the Department of ComputerScience and Engineering at the University of Washington,where she is codirector of the Value Sensitive DesignResearch Laboratory. Her research interests include ValueSensitive Design, informed consent, privacy in public, trust,freedom from bias, moral agency, and human dignity. Shereceived a PhD in science and mathematics education fromthe University of California, Berkeley. Contact her atbatya@u.washington.edu.

Peter H. Kahn Jr. is an associate professor in the Depart-ment of Psychology and an adjunct associate professor inthe Information School at the University of Washington,where he is also codirector of the Value Sensitive DesignResearch Laboratory. His research interests include socialand moral relationships with robotic others, the psycho-logical effects of digitized natural information, and valuesensitive design. Kahn received a PhD in educational psy-chology from the University of California, Berkeley. Con-tact him at pkahn@u.washington.edu.

Paul A. Waddell is a professor in the Daniel J. Evans Schoolof Public Affairs and the Department of Urban Design andPlanning at the University of Washington and the directorof the Center for Urban Simulation and Policy Analysis.His research interests include developing simulation mod-els to support land use and transportation and environ-mental policy analysis. Waddell received a PhD in politicaleconomy from the University of Texas at Dallas. Contacthim at pwaddell@u.washington.edu.