The Science of Science PolicyA Handbook

Edited by Kaye Husbands Fealing, Julia l. Lane,John H. Marburger lll, and Stephanie S. Shipp

STANFORD BUSINESS BOOKSAn lmprint of Stanford University Press

Stanford, California2 tl

ir'r:i'': I' ,j.:.,

.

.

Editsrs' lntroductionKaye Husbands Fealing, Julia l' Lane, John H' Marburger lll,

and StePhanie S. SnP

1. lntroductionFederally funded basic and applied scientific research has had an enormousimpact on innovation, economic growth, and social well-being-but somehas not. Determining which federally funded research projects yield resultsand which do not would seem to be a subject ofhigh national interest, par-ticularly since the government invests more than $140 billion annually inbasic and applied research. Yet science policy debates are typically domi-nated not by a thoughtful, evidence-based analysis of the likeiy merits ofdifferent investments but by advocates for particular scientific fields or mis-sions, Policy decisions are strongly influenced by past practice or data trendsthat may be out of date or have limited relevance to the current situation, Inthe absence of a deeper understanding of the changing framework in whichinnovation occurs, policymakers do not have the capacity to predict howbest to make and manage investments to exploit the most promising andimportant opportunities.

This lack ofanalytical capacity in science policy sits in sharp contrast toother policy fields, such as workforce, healih, and education. Debate ,inthese fields is informed by the rich availability of data, high-quality analysisof the relative impact of different interyentions, and often computationalmodels that allow for prospective analyses, The results have been impres-sive. For example, in workforce polic the evaluation of the impact of edu-cation and training programs has been transformed by careful attention toissues such as selection bias and the development ofappropriate counterfac-tuals. The analysis ofdata about geographic dfferences in health care costsand health care outcomes has featured prominently in guidinghealth policydebates. And education policy has moved from a "spend more money" and"launch a thousand pilot projects" imperative to a more systematic analysisof programs that work and that could promote local and national reformefforts.

Husbands Fealing, Lane, Marburger, ana ilrpp

Each ofthose efforts, however, has benefited from an understanding ofthe systems that are being analyzed. In the case ofscience polic no such agree-ment currently exists. Past efforts to analyze the innovation system and theeffect that federal research has on it have typically focused on institutions(federal agencies, universities, companies, etc,) and/or outputs (bibliometrics'patents, funding levels, production of PhDs, etc.). Absent is a systemsJevelconstruct that those institutions and outputs function within and a failure tounderstand that science and technology innovations are created not by insti-

tutions but by people, often working in complex social networks. This socialdynamic, as well as the complex systemlevel interactions that result, is the

subject of increasing academic scrutiny. Science magazine recently devoted aspecial section to "complex systems and networks" and referenced studies

that examined omplex socioeconomic systems, meta-network analysis, scale-

free networks, and other analytical techniques that could be used to under-stand the innovation system.l

There is no fundamental reason why it is impossible to develop a sciencepolicy infrastructure that is similarly grounded in evidence and analysis asthe workforce, health, and education domains, It is true that it is difficultr theinstitutional and political environment is complex, and the scientific discoveryprocess is noisy and uncertain, Yet scientists should be excitecl, not deterred,

by interesting but hard problems. And the history ofthe scientific advance-ment of other policy fields, with their studies of equally complex, nois anduncertain prgcesses, is evidence that such efforts can suCceed. Indeed, an inter-

disciplinary and international community of practice is emerging to advancethe scientific basis of science policy through the development of data collec-tion, theoretical frameworks, models, and tools. Its advocates envision thatthey can make future policy decisions based on empiricaly validated hypothe-

ses and informed judgment.There are fundamental reasons why it is becoming criticai to develop such

an evidence basis, One is that the White House is requiring agencies to doso: the joint office of Management ancl Budget (oMB)/office of science andTechnology Policy (OSTP) R&D Priorities lnemo issued in preparation for theFY2011 budget asks agencies to "develop outcome-oriented goals for theirscience 4nd technology activities, establish procedures and timelines for eval-uating the performance of these activities, and target investnents towardhigh-performing programs. .Agencies should develop 'science of science policy'tools that can improve mangement of their research and development Port-folios and better assess the impact of their science anci technology investments.Sound science shouid inform policy decisions, and agencies should invest inrelevant science and technology as appropriate."2

Editors: lntroducton

Another is the looming imperative to document the impact of theneaily $20 billion in R&D investments embodied in the 2009 AmericanRecovery and Reinvestment Act (ARRA)' As Kei Koizumi points out in hischapter:

Policymakers and evaluators can demonstrate easily the short-term economic

efiects of highway proiects, of which there are billions of dollars worth inthe Recovery Act; miles ofasphalt poured, construction jobs created, and dol-lars introduced into local economies are well developed and easily produced

measures for these investments. But what are the sirnilar indicators for R&Dinvestments?

Finall the federal budget environment is likely to be extremely competi-tive for the foreseeable future. For a case to be made that investments in sci-ence have value relative to investments in education, health, or the workforce,an analytical and empirical link has to be made between those investmentsand policy-relevant outcomes, It is likely that that link will need to be made atmultiple levels, since the macro link between R&D investments and economicgrowth is less convincing given the international evidence provided by theJapanese and Swedish experience,3

The federal agencies have begun to respond in two ways, One is to advancethe theoretical and empirical research frontier through investigator-initiatedresearch and new data collection. The second is to develop a federal commu-nity ofpractice among the seventeen science agencies involved in funding andadministering science research.

In the former case, by mid-2010, the National Science Foundation's (NSF)Science of Science & Innovation Policy (SciSIP) prograrn has made overninety awards to social scientists and domain scientists. Ten of these are ex-plicitly to use the ARRA. stimulus as a way to examine the impact of scienceinvestments. The SciSIP program, through the Division of Science ResourcesStatistics, is also investing in the development and collection ofnew surveysto better inform the biennial Science and Engineering Indicators that are thebasis for many policy decisions, This includes the new Business R&D Innova-tion Surve which involves a complete redesign of the collection of R&D data,as well as the coilection of innovation data,

In the second case, the National Science and Technology Council (NSTC)established, under the Social, Behavioral and Economic Sciences Subcommitteeof the Committee on Science, a federal interagency task group on the Science ofScience Policy interagency task group (SOSP ITG). lhis task group produced aroad map for federal investmentsa and held a major international conference tohighlight the findings in that road map.

Husbands Fealng, Lane, Marburger' and Shipp

Both the scislP program and the sosP subcommittee have worked to

foster a community of practice in a number of ways. The interagency group has

organized major annual workshops on the implementation of science poiicy'A flourishing Listservfor the exchange ofideas and information has been estab-

lished. And a new sosP ITG/ScisIP website has been developed,s which has

begun to provide an institutional basis for the development of a community

ofpractice.Of course, SOSP will not solve all science policy problems' It is intended to

provide an intellectual framework upon which to make decisions, Indeed, as

Goldston notes in his chaPter:

Science ofscience Policy research will never be definitive, and Congress certainly

always would and should draw on more than social science results in makilrg its

decisions. But there is plenty of room to improve the current state of affairs' In

other areas of policy-macroeconomics, health care, environmental Protection,to nme a few-there is at least a semblance of an ability to project the outputsthat will result from a given set gf inputs, aud a range of studies to draw on in

discussing what has worked and what has failed, Reaching a similar level of un-

derstancling for science policy would be a welcome change, if hardly a panacea.

2, What the Science of Science Policy EntailsOne of the aims of recent science of science policy activities is to deveiop the

evidentiary basis for decisin making by poiicy practitioners. There is also an

organic development or reshaping of frameworks that pushes the boundaries

of discovery in several fields and disciplines. While some debate whether the

science of science policy is itself a discipline, there is wide agreement that

there is a coalescing community of practice, which Feller, in his chapter, de-scribes as a distributed association ofpolicymakers (public and private) andresearchers in a variety of fields and disciplines. This community s interdisci-

plinary and includes economics, engineering, the history of science, operations

research, physics, political science, psycholog and sociology-and this list is

not exhaustive.6Fecleral science investments are dfiven by a political context, so the insights

provided by political scieirtists are critical, Sapolsky and Taylor argue in their

chapter that

governments supPort the advancement of science and technology (S&T) mostlythrough their support of specific missions such as defense or health, and it is the

politics of these missions, and the many contextual goals of governmetrt, that de'

termines the rate and direction of its research and development investrnents.

Editors' lntroducton

Governments can also affect fhe supply and demand conditions for science and

tech'ology outside the buclgetary process via regulatory regimes, anti-trust, taxes,

standards, etc.

Understanding the institutional and sociological environment is alsocritical, which is why sociologists make an important contribution, Powell,Owen-Smith, and Smith-Doerr indcate in their chapter that the "sociological

science of science policy will theorize the link betweeu the origins and latertrajectories of social systems that will provide guidance for policymakers eagerto intervene."

The economics of science policy is evolving beyond the initial constructsof maooecoromic linkages of inputs and productivity outcomes. Recent modelsutilize network anais, bibliometric tools, and behavioral models to unoverlatent relationships between the levels and rates ofnew scientific discoveriesand the nancial, human capital, organizational, and infrastructural inputs.While these models have historically made important contributions to policydecisions, Feller, |affe, and Freeman each caution in this volume that there is aneed to understancl the limitations of incentive structures and the require-ment for careful empirical analysis to understand the system of scientificknowledge creation. Morgan, in his chapter, describes several systems model-ing approaches, some of whith originate outside of the social sciences. Thismigration and synthesis of ideas is precisely what creates a dynamic communityofpractice.

One area of the science of science policy that is often overlooked is thatconceptualization of scientific development at the cognitive level. This verymicro-examination of science policy is an emerging field, with coliaborationbetween psychologists and engineers. Both disciplines are eager to under-stand the elements of the creative process, Gero describes frameworks that areused to understand creative cognitive processes, which may lead to new ideasthat are marketable-innovation.

And, of course, science investments are ultimately predicated on contrib-uting to innovation, Gault's chapter connects the work on the understandingof the science system to the need for work on delivering value to the market inthe form of new goods and services and contributing to economic growth andsocial welfare,

3, The Need for the HandbookOur review of the science policy curricula and syllabi in major research pro-grams suggests that the emerging field lacks a cornerstone document that

Husbands Feating, Lane, Marburger' and Shipp

describes the current state ofthe art from both a practitioner and an academic

point of view.Thishandbookisintendedtofillthisgapbyprovidingin-depth,scholarly

essays authored by leading scientists and policy practitioners' We recognize

thatthefieldhasmultipledimensions,andassuch,thisbookisdividedintothree sections: theoretical issues, data and measurement, and policy in prac-

tice. Each author has been asked to provide a survey ofa different aspect ofthe

field, based on his or her domain expertise, which explores the plausible foun-

dations of an evidence-based platform for science policy. The interdisciplin-

ary nature ofsuch a platform is evident from the nature ofthe questions asked

by the authors: What are the essential elements of creativity and innovation'

and how can they be defined to serve a truly scientific approach to policy?

Howcanthetechnicalworkforcebequantifiedandmodeied_whatisitslikely future, and how does it respond to the multiple forces that could be tar-

getsofpolicy?Whatistheimpactofglobalizationoncreativityandproduc-tivity in the science and engineering fields? What are the optimal roles of

government and private investments in R&D, and how do their different out-

comes influence R&D and innovative activities? As such, the contributors

span a variety of disciplines, including economics, sociolog psycholog and

political science.It is wor.th noting that this handbook focuses on the science of science

polic which we feel is aD understudied and underresearched area, There has

been a great deal more research on the science of innovation polic although'

inevitabl some of that research is alluded to in different chapters. In adcli-

tion,thefocusisonU.S.federalsciencepolicy,Werecognizethattherearevibrant and important research areas that study both business R&D invest-

ments and regional science and innovation policies. And while managers of

large research enterprises, such as Microsoft, and state agencies face substan-

tial resource allocation decisions, our sense is that these decisions are funda-

mentally clifferent from those in the federal science arena, And, aithough the

science of science policy has garnered important attention on the interna-

tional stage, it is impossible to do full justice to the complexity of the interna-tional issues-that deserves another volume in its own right

4. Goncluding Goalswe hope that this handbook will contribute to the overarching goal for science

polic namel the development of "comtnon, high-quaiity data resources and

interpretive framewol'ks, a corps of professionals trained in science policy meth-

ods and issues, and a network of high-quality communication and discussion

, irat .un encomPass all science policy stakeholders"'7 s such' the purpose of

-,il,uoot is to Providel. an overview of the current state of the science of science policy in four

kev soc" scienee

psychologY;

areas: economics, sociology, political science, and

2. aperspective from the broader social and behavioral science colrlnu-

emging field;3, a review of the empirical-measurement and data-challeqrges inher-

ent in describing and assessing the scientific enterprise; and

4, a perspective from the federal science and policy cornmunity on thecritical science policy questions that create the demand for a science of

science PolicY.

Notes

L Science,l,'tb 24,2tA9, pp' 405-432'2. M-0g-27, Memorandum for the Heads of Bxecutive Department$ and Agen-

cies, ugust 4 2009.3. Iulia Laqe, "Assessing the Impact of science Fun{lng,' Scienee 5 (Tune 2009),

vol. 324, no, 5932, pp. 12 a275, DAk 10.1126/science'1175335.4. ,.The Science of science Poliry: A Fedeal Research Roadmap," November 2008.

5. See http://scienceofsciencepolicnet'6. For example, all of these aieii are represented arnong the SciSlP awardeesi Soe

wwwscienceofsciencepolicy,net/scisipmernbers.aspx,7. See ldarburger, Chap. 2, in this book.

Topical Guide to This Handbook

Topic

Based Models and Decision AnaI

Culture of Science and Institutions

and Knowledge Creation

conomic Models

and the Environment

valuation Studies

Fundi and Incentive Structures

Histo ofScience Polic

Innovation

fnstitutional Networks, Including lnter- andIntramural

Intenational Part

Orsanizations

Pospective and Retrospective

Quantitative Methods

Science dministration and

lovers and Clusters

Sustainable Growth, Including Wealth Creation

Chapter 2

Xs

fi

iI

Technology Cornmercialization/Diffrrsion,Includine Bridging the "Valley of Death"

Triple Helix-Partnerships Between Academic,Industry and Covernlnent Entities

Visual Analytics

pAffiT #rwffiThe Theory of Science PolicYEditors' Overview

The foundations ofthe evidence-based platform ofscience policy span several

disciplines. For decades, the core social scienie disciplines ofeconomics, soci-

olog and political science have given us frameworks that attempt to explain

the dynamics !f science and innovation activities. Together their methodolo'gies provide a(understanding ofthe stocks and flows ofinputs and outputs inthe systerg,rhl institutional structures that promote or impede scientifrcprogress, lnd the power relationships that determiDe distributioral outcomes,

Recent calls for a social science of science policy have provided the impetus

for a resurgence of researchers collaborating across disciplinary boundaries

in search of a systems approach to answering age-old science policy questions.

Engaged in this process as well are psychologists, whose frameworks add nec-

essary dimensions to developing an understanding of the creativity Processleading to scientifrc discoveries and downstream innovations. complexity

theorists and modelers have also expanded our view of the science and inno-

vation enterprise, with architectural frameworks that provide scaffolding for

policy sirnuiations, Such empirical exercises complement those that are devel-

oped in the core social science disciplines. The following chapters synthesize

the theoretical knowledge bases from which the science of science policyemerges and for which there is call for an engaged community of practice'

The politics of distribution related to scientific and technological endeav-

ors is the focus of the Harvey M, Sapolsky and Mark Zachary Taylor chapter.

While economists evoke the public good rational for government funding of

R&D (see Richard B. Freeman's chapter), sapolsky and Taylor describe a dif-ferent concept of sciencer the public good. They argue that "science andtechnology create winners and losers, especially in the long run." In theirparadigm, the "losers" are endowed with assets such as skills, capital, land,

and other resources. Distributive innovation hurts them because it changesthe status quo. Ifthe potential "lpsers" are also power holders, then they have

an incentive to create institutions and to exercise policy mechanisms that

24 The Theory of Science Policy: Editors' Overview

enable them to retain power. Sapolsky and Taylor conclude that this potential

for redistribution of wealth and power could slow technological change-butnot in every case. Interestingly, political leaders who couple the technological

enterprise with national security concerns or to nationalistic economic com-petitiveness races are able to forestall the drag on technological progress by

stakeholders who stand to lose ground in a new tecirnological equilibrium.The evolutionary nature ofscience and technological innovation necessarily

means realignment of power, Understanding these dynamics is critical to our

understanding of the ecological system of innovationSapolsky and Taylor avoid reference to the traditional parlance of"national

innovation system." Instead, they discuss the role of government in a globaliz-

ing world. They posit two different scenarios for deveioping and developedcountries, Governments in developing countries, they argue, must be strategic

in their capital investments (including human capital). with far-flung func-tions of corporations, policymakers in developing countries should move away

from targeting entire industries to targeting specific functions within a givenindustry. In the developed country context-presumed to be lead innovatingnations-the authors argue for the importance of technological rnodularityand interoperability with global technical infrastructures. Globalization, there-

fore, requires technological and managerial flexibility within industry and gov-

ernment, Science and technologypolicy decisions will foliow paths quite differ-ent from those in the mid-twentieth century.

Institutions are the focus of the chapter by Walter W. Powell' Jason Owen-

Smith, and Laurel Smith-Doerr, 'Ihis chapter focuses on the importance ofsocial systems to the study ofscience and innovation policy. Identiable link-ages between inputs and outcomes in the science and innovation systemchannel through ethical, politicai, environmental, and other social constructs,Even if a causal link is empirically established between, sa human capitalinputs and productivity in an industry, in a specific region, at a certain pointin time, it would not be necessarily true that the same results would be ob-tained if any of the structural factors were changed.

Ethical, power, and network relationships critically determine outcomes'Powell, Owen-Smith, and Smith-Doerr give specific examples to make theirpoint. In the case ofstem cell research, the authors show how changes in presi-

dential administrations affected the methods that researchers could use to

extract, store, and use embryonic stem celis in laboratory experiments, Fed-eral funding of research in this area aliows politics (a derivative of social eth-ics and power) to affect lab science practices, The authors take this argumenteven farther. Since institutions vary by nation, international competition in

fhe Theory of Science Policy: Editors' Overview

science and innovation is appreciably affected by established institutions and

institutional change. This is an important dimension to the science of science

policy. Many countries use a benchmark rate of 3 percent of GDP for R&D

iunding. Since the institutional structures of these countries var it is not

likely that a one-size-fits-al1 rule should apply. The authors also use this rea-

soning to explain why technology clusters form in some cities and not in

others'power relationships, institutional contexts, and network structures, there-

fore, impact behavior in fields ofresearch. The authors discuss the importance

of this triad on the balance of basic and translational research at funding

agencies, such as the National Institutes of Health. Additionall they high-

light the debate about the need for interdisciplinary research to achieve trans-

formative outcomes. l,astl they link the work on interdisciplinarity to gen-

der and racial diversity in the sciences, The institutional frameworks determine

differences in access and, therefore, contribution ofindividuals from a variety

of backgrounds. If increased variety is super-additive (as Martin Weitzmanshows in hi's journal article "on Diversity"r), then institutional barriers to di-versity in science and innovation networks could reduce the efficacy of R&D

funding.The economics ofscience policy has several contributing fields and areas.

Feeman's chapter focuses on the market for labor-the supply and dernandfor scientists and engineers, particularly under uncertainty. In an effort toatrswer the cluestion of how to facilitate a market for scientific output, Free-

man turns to a fundamental neoclassical tool-incentive structure' Pecuniary

and nonpecuniary incentives can be used to encourage welfare-improvingredistributions of human capital inputs into various scientic elds' Incen.tives can also be used to encourage various types of research initiatives' in-

cluding the balancing of incremental and potentially transformative projects.Here Freeman embraces the sociology literature, acknowledging that theincentive structure must be cognizant of social networks. Networks can be

intradisciplinary and interdisciplinar as weli as intramural and extramural(e.g., linkages between the academy and industry)' One obvious question iswhether the tournament nature of competition for scientists and engineersaffects ail demographic groups the same way. Freeman addresses this question

with respect to women and immigrant scientists'Government expenditures on scientic and technological development

affect outcomes directly and indirectly. Awards and grants are direct dispen-

sations to scientists and engineers at universities and other research institutions,

while inyestment tax credits and other R&D subsidies may increase yields

The Theory of Science Policy: Editors' Overview

through businesses (crowding out notwithstanding)' Freeman suggests thatgovefnment,sdemandforscienceshouldcompriseadiversifiedportfolio.Moreimportantl government funding institutions must recognize the impact that

uncertainty has on their decision making, but also that fluctuations in their

decisions have measurable efiects on the development of R&D infrastructure,

on the supply of human capital, and on the achievement of social goals'

Freeman highlights a critical aspect of science funding decisions-the un-

certainty of outcomes. He draws on frnance-based analytical tools for policy

guidance. options models and other portfolio allocation tools traditionally

used to develop private-sector frnancial strategies have powerful applications

in the context of the science of science policy. Research portfolios at large fed-

eral R&D labs or at academic institutions are likely test beds for such mod-

els. Freeman asserts that modeling the supply and demand of scientists and

engineers-and the incentive structures that close the loop-requires a sys-

tems approach with options modeling to capture decision making under un-

certainty. Measuring causai outcomes is not an exact science. The knotty prob-

lemishowtomeasureR&Dspillovers_theexternalitiesnotonlytosocietyfrom new discoveries and innovations, but also to other scientific endeavors'.

Establishing metrics on these linkages is critical to weighing the economic

returns to R&D, but it is precisely the uncertainty in the system that makes

measuremet difficult yet highly sought after'The economics, sociolog and political science chapters have focused on

understanding the science and technology enterprise at. a mezzo or maclo

level. john s. Gero's chapter focuses instead on the extreme micro level-human cognition and innovation, He distinguishes creativity from innova-

tion, in much the same way other social scientists writing in this section

have done. creativity is the process that yields new and useful ideas, while

innovation is the process that turns creative ideas into products or pro-

cesses,z The exploration of innovation and innovation policy is a relatively

recent area of exploration in neuroscience, As such, this chapter gives an

overview of the emerging eld, including a thorough mapping from tradi-

tional literature and nomenclature on cognition to the application of tech-

nological innovation.There is a clear theoretical iink to the other chapters in this section of

the science of science policy handbook, Gero directly links the cognitive

definition of innovation to Schumpeter's concept of "creative destruction,".ivhere innovation displaces existing products or processes. In his rubric,there are three types ofinnovation: "augmentation, partial substitution, and

displacement through total substitution," The latter is schumpeterian' This

:l:,.;jt.'.'.'.i lt lt. il r'nu nneory of scienco

Policy: Editors' overview

..''..-...'.

'..'.:iru,po*erful connection' Using a computational sociology technique' cog-

.,,;;;. scientists can test the behavioral interaction between innovators and

adoPters of innovation' This is not only the development of neural networks

.,,irn. evolution of social net\dorks in the creativity-cum-innovation pro-

' ess, In these experiments' the impetus for change is the science of innova-

i"'n"t" y' nalyzilgbehavioral resPonses io poiicy decisions is criticai to

'i"jt"tng linkages and the dynamics within science and innovationsYstens'

M. Granger Morgan's chapter outlines a wide range of analytical models

or.U ,n the -science

policy arena' The main emphasis of his chapter is that

policy tools are not one-size-fits-all' Policy problems-particularly science

andtechnologypolicyconundrums_requireexpertiseintechniquesthataremore commony utilized in operations research, decision analysis,

technology

assessment, options modeling, benefit-cost analysis, and life-cycle analysis.

Morganintroducesthereadertothecanonicalliteratureintheseafeasanddraws liberally on examples from energy and environmental policy'

This

chapter is particularly useful for pedagogical purPoses' It atso highlights the

agenciesandthinktanksthatareleadersinthedevelopmentanduseofsys.tems modeling aPProaches'

Irwin Feller introduces the theoretical paradigms of the new science of

sciencepolicdeflningthescience,(andart)ofsciencepoliccautioningthebuildersandusersoftheenterpriseaboutthelimitationsofmodels,tools,and

',

dur. that currently exist, and encouraging researchers and policymakers to

cultivateadynamiccommunityofpractice.Fellerdrawslinkagesbetweendisciplinesandacrossgenerationsthatcontributetotheunderstandingofhowprioritiesshouldbesetasorganizationsfundscientificdiscoveryandtechnologicalinnovation.Theprimaryutilityofthischapter,therefore,isthecollection of works and the relationships among the works by thought leaders

on evidence-based science policy'

Predictabitityisoftendemandedbydecisionmakers'Purveyorsofsciencepolicy models and tools are required to suPply the best estimates under a va-

retyofconditions,includinguncertainty'Felleraddressestheprecariousdance between knowns and unknowns' highlighting what has been shown to

bemosteffectivebutcautioningagainstignoringthepossibilityofT}peIorType II errors. This is an important point' particularly in an era ofincreased

dependency on empirically driven decision making' Noted in this chapter'

however, is the need for concurrent data on the science and engineering enter-

prise,particularlydatathatareconsistentacrosscountries.FellerrecognizestherecenteffortsattheNationalscienceFoundationtomeasureR&Dand

The Thoory of Scionc Policy: Editors' Overviow

innovation separatel but he notes that measurements of the related inputs

and outcomes require new and distributed vantage points'

Feller cautions researchers not to refine their models and tools af to de-

velop new data sets in a vacuum, A community of practice, where researchers

and practitioners engage in periodic discourse about new pardigms andoutcomes, should minimize, as Feller puts it, "buyer's remorse," This is not to

say that all social science of science policy belongs in Pasteur's Quadrant, But

researchers and practitioners alike may increase respective utilities from such

engagement. Referring to Bozeman and Sarewitz, Feller acknowiedges that

these activities ale not merely to achieve increased economic efficiencies but

to enhance social value,

Fred Gaultt chapter provides a history of how innovation is measured andthe components of innovation strategies (markets, people, activities, and pub-lic and international institutions), Gault identifies how innovation strategiesinform policymaking as well as actions that can be taken to advance innova-

tion. International organizations, such as the organization for Economic co-

operation and Development (OECD), United Nations Educational, scientific,and Cultural Organization (UNESCO), and New Partnership for Africa'sDevelopmenr (NEPAD) (all described in this chapter), faciiitate this coordi-nation and evaluation, which involves multiple dimensions such as the char-

acteristics ofthe population, the history ofthe country, and the type ofpolicyto be undertaken. Involving stakeholders at each stage ofpolicymaking en-sures the implementation of policies relevant to country or a group of coun-

tries. International organizations provide assistance with coordination andevaluation, including the creation and analysis ofsurvey data and case studies,

the dissemination of these frndings, and the formulation and review of policies'

The ultimate goal r:f this support is to build a country's capacity "to identiffind, acquire, adapt, and adopt" knowledge and to incorporate this knowledge

as an "indispensable component" to create and implement a science and tech-

nology innovation strategy'All of these chapters have some mention of a system of science and inno-

vation activities. Morgan's chapter describes the use of operations research

tools for policy analysis. This is au important addition to the benefrt-cost mocl-

els and options lnodels described in Freeman's chapte the social networkmodel alluded to in the chapter by Powell, owen-smith, and smith-Doerr, the

stakeholder analysis discussed in Sapolsky and Taylor's chapter, and the com-

putational sociology rrrodel mentioned by Gero. Felier and Gault dene the

spaces of the science of science policy and the science of innovation policrespectiveiy, Taken together, these chapters outline the critical questions and

' l, Martin Weitzman, "On Divelsiry" Quatteily loutnal of Econonnics' vol' l0Z

no.2 0992): 363-405'2. Business model innovation is typically included

in the definition a's well'