CAPITALISMO CIENTIFICO ACADEMICO

This study addresses university-industry interactions for both educational and industrial outcomes. The results suggest that while academic faculty who are affiliated with cen-ters are more involved with industry than non-affiliated faculty, affiliates are also more involved with and supportive of students at the undergraduate, graduate, and doctoral levels.

Therehasbeenquitealotofstudyanddiscourseregardingtheconnec-tionbetweenuniversitiesandindustry.Thesubtitleofourpaperreflectsourviewofmuchof this literaturethat it is rifewith immoderation.Study of university-industry interactions is bifurcated,with some au-thorssuggesting that these interactionsarebeneficial toeconomicde-velopment and technology transfer (e.g., Etzkowitz, 1998) and otherauthors characterizing the interactions as disruptive and potentiallyharmfultotraditionaluniversitymissions,especiallytoeducation(e.g.,Slaughter & Rhoades, 2004). Really, the two perspectives constituteseparateliteratures.Seldomdoesempiricalstudyofuniversityinvolve-ment with industry address outcomes related to education, such asteachingandstudentsupport,andalmostasaruledoesthediscourseonthe potentially disruptive and harmful nature of university involvement withindustrygounconcernedwiththepotentialbenefitsofthisinvolve-mentforeducationaswellasforeconomicoutcomes.

Academic Faculty in University Research Centers: Neither Capitalisms slaves nor Teaching Fugitives

Barry Bozeman is Regents Professor and Ander Crenshaw Chair of Public Policy, De-partment of Public Administration and Policy, University of Georgia; [email protected]. Craig Boardman is Assistant Professor of Public Affairs, John Glenn School of Public Affairs, The Ohio State University.

Barry BozemanCraig Boardman

The Journal of Higher Education,vol.84,No.1(January/february)Copyright2013byTheOhioStateUniversity

Academic Faculty in University Research Centers 89

Bothviewsofuniversity-industry interactionscanbecorrect in iso-lation.Universitiesarecomplexandheterogeneousand thuslyexperi-encedifferentoutcomesfromorganizationalinnovationandinstitutionalchange.Doubtless,onecanidentifyinstanceswhereuniversity-industryinteractionsowing either to their commercial nature or to any of ahost of other differenceswith traditional academic activities andout-comeshave fractured the foundations uponwhich universities havebeen built. One can also find instanceswhere university-industry in-teractionsfitquiteeasily into thetraditionaluniversityanditshistoricmissions,enhancingratherthandetractingfromuniversitiesabilitiestoachievepublicdomainresearch,teaching,andservicegoals.Connectionsbetweenuniversities and industry in theUnitedStates

canbetracedbackwellover100years(Ruscio,1984;Mowery&Zie-donis,2002).Buttheyhavereceivedmuchattentionoflateduetotheestablishment of new institutional forms onAmerican campuses forgoverning university-industry interactions, specifically university re-searchcenters(URCs).TheprevalenceofURCsonAmericancampusesiswelldocumented (Dresser,1989;Coburn,1995).Aspolicymecha-nisms,URCshaveattractedattentiondueto theirpivotalroles ineco-nomic development and technology transfer. As organizations, theyhaveattractedattentionduetothedivergenceoftheirresearchmissionsfromthediscipline-basedmissionsoftraditionalacademicdepartmentstowardsmissionsfocusedonaddressingsocialandeconomicproblemswithscienceandtechnology(Boardman&Gray,2010),conjointlybe-tweenuniversitiesandindustryincludingtheinvolvementofindustrynotonlyinuniversity-basedresearchanddevelopmentbutalsoingrad-uateeducationandtraining(Bozeman&Boardman,2004).Whileweexpectawiderangeofoutcomesfromthecreationandin-

stitutionalizationofURCs,ourintuitionandourpreviousresearch(e.g.,Bozeman&Boardman,2003;Boardman&Bozeman,2007;Boardman,2009)haveledustobelievethatthetwodominantimagesofuniversity-industryinteractionsaretoostarkandtoofocusedonextremecasestoaptlycharacterize the impactsofURCsonacademic faculty. It isun-derstandablethatthoseviewingURCsandtheirinhabitantsasagentsoftechnologycommercialization focusoncenters andprograms that arecompatiblewiththeseaimsandthatsucceedinthismission.Similarly,thosewhoseeURCsandalternatearrangementsforuniversity-industryinteractionasdetrimentaltotraditionaluniversitymissionswillbeabletofindcaseswheretheyhavehadnegativeanddisruptiveoutcomes.SincewetakeitasaxiomaticthatnewinstitutionssuchasURCswill

havediverseoutcomes,ourconcerninthisstudywaswithunderstand-ing the distributionofoutcomesandthemechanismsaffectingthatdis-

90 The Journal of Higher Education

tribution. heretofore, representative data about URCs have not beenavailableand,thus,theliteratureonURCsisdominatedbycasestudiesof singleuniversitieswithmultiple centersorofoneor a fewURCs.OwingtoalackofgeneraldataaboutthenaturalpopulationofURCs,pastresearchhastendedtotreatURCsandacademicfacultyworkingintheminanundifferentiatedmanner.ScholarshaveunderstoodthatthereareconsiderablevariationsinURCs,butuntilnowextantdatadidnotpermitchartingthatvariation.ThusweexpectedsomeURCstofulfillexpectationsof commercially-relevantoutputsandeconomicbenefits,otherURCs tofulfillexpectationsfromcritics,buton theaverageforURCstocontributetoboththeentrepreneurialandeducationalmissionsof U.S. universities. OurstudyisoneofthefirsttousedataforabroadsampleofURCs.

WestartedwithquestionnairedatafromarepresentativesampleofU.S.academicresearchers,focusingspecificallyonscience,technology,en-gineering,andmathematics(STEM)facultyholdingpositionsinCarn-egiefoundationresearchextensiveuniversities.Since inouradminis-trationofthesurveywedidnot focusspecificallyonfacultywhowerecenteraffiliates,wewereabletoaskwhetherornotrespondentsareaf-filiatedwithaURCtodeterminethepercentageofacademicresearch-ersintheUnitedStateswhoareURCaffiliates.Additionally,wewereable to ask affiliated respondents to list theURC(s)withwhich theywereaffiliatedtoprovidearepresentativesampleofURCtypes,atleastforthesamplingpopulationexamined.fromthissampleofURCs,wecreatedanancillarydatasettrackingURCcharacteristics,includingin-formationaboutthesizeofthecenters,thenumberofuniversitiesrep-resented,andwhether thecenterswerevirtualorbricksandmortar,amongotherimportantfactors.TheURC-leveldatasupplementthequestionnairesprovidedbyuni-

versity facultyandprovide insight into thedistributionofURCaffili-atesandURCtypes.Thesedatapermitustoaddresssignificantgapsinthe literature.first,simplyknowingmoreabout the incidenceofcen-termembershipandthedistributionofcentertypesprovesusefulfromthestandpointofpopulationecology.Previousstudieshaveexaminedwell-known,high-profileURCs,especiallythosecreatedinpartnershipwithfederalagencies.however,asweshowhere,thesefamiliar,high-profileURCs are a relatively smallminority of allURCs.having anempirically-based typology ofURCs permits us to examinemember-shipinspecifictypesofURCsandtheeffectsofthesespecifictypesonacademicresearcherscareersandworkactivities.ThiseffortisabettedbythefactthatoursampleincludesamajorityoffacultywhoarequitesimilartothoseaffiliatedwithURCsbutwhoarenotthemselvesURC


affiliates.Thus,weare in theuniquepositionof testing long-standingassumptionsabouttheindividual-leveleffectsofURCsonuniversities.Weaimedour analysis at the twopredominant characterizationsof

centers.WeaddressedtheextenttowhichURCaffiliationaffects(1)theinvolvementofacademicfacultywithindustry,suchasthedevelopmentanduseofpotentiallycommercialknowledgeand technologyand (2)faculty involvementwith teachingandotherstudent-relatedactivities,includingstudentsupportandmentoring.

Divergent Characterizations of Boundary-Spanning Universities and University Faculty

The study of universities and university faculty as boundary-span-ningis itselfboundary-spanning, includingstudiesfromhighereduca-tion,sociologyofscience,researchmanagement,economicsofinnova-tion,andscienceandtechnologypolicy(hagedoorn,Link,&vonortas,2000).Acrosstheseliteratures,therehasbeensomeemphasisonURCs,butowingtothelackofgeneraldataforcenters,alargeproportionofresearchhasbeenfocusedmorebroadlyontheoutcomesofuniversity-industryinteractionsinanycontext.There seems tobe twodivergentviewsofuniversity facultywork-

ingwith industry. The first perspective comes from themanagementand policy literatures,which has taken a relatively descriptive stanceandgenerallyhas founduniversity faculty tobe involvedwith indus-try(Welsh,Glenn,Lacy,&Biscotti,2008).Thesestudieshavefocusedonuniversitypatenting(henderson,Jaffe,&Trajtenberg,1998)andli-censing(Thursby&Thursby,2002),spin-offcompanies(Link&Scott,2005),andfacultyinvolvementwithindustryrangingfrominformalinformation exchanges to consulting and employment arrangementsbetween companies and faculty (Bozeman & Gaughan, 2007). Themechanismsemphasized in thisbodyofworkhave includednotonlyinstitutional design such asURCs but also national tax, subsidy, andintellectual property policies (Mowery, Nelson, Sampat, & Ziedonis,2001), state-levelpoliciesand industry initiatives focusedon regionaleconomic development (Lee, 1996; Rogers, Takegami, &yin, 2001;Lin&Bozeman,2006),universitycharacteristics(Ponomariov,2008),andindividualinvestigatorgrantsandcontracts(Bozeman&Gaughan,2007).StudiesfocusedspecificallyonURCshaveshownaffiliatedfac-ulty to bemore likely than other faculty to collaboratewith industry(Boardman&Corley,2008).The secondperspectivecomes from thehigher educationand soci-

ologyof science literatures and essentiallyuses the industry involve-


ment of universities and university faculty, as evidenced in theman-agementandpolicyliteratures,asaplatformforlaunchingbroad-basedcritiques of university-industry relations, including but not limited toURCs.Mostnotableamongthesecritiques is theacademiccapitalismargument developed by Slaughter andRhoades (1996) and Slaughterand Leslie (1997), which suggests that boundary-spanning organiza-tions suchasURCsdetract from traditional educationalmissionsandgoalsbyinducingfacultytoengageinmarket-likebehaviorsandthere-fore to shirk traditional research and educational activities for moreprofitableonesor,at thevery least, toharness traditionalactivitiestowards profitable ends (Slaughter, Campbell, folleman, &Morgan,2002). There are related studies that similarly have questioned howtraditionaluniversityrolesarecompromisedbyURCsanduniversity-industryinteractionsmoregenerally(forexamples,seehackett,2001,andkrimsky,2003).Inourview,theextenttowhichtheindustryinvolvementofuniversi-

tiesanduniversity facultyaffects theuniversities traditional researchandeducationalmissionsremainsalargelyunansweredempiricalques-tion.Existingcase studiesare suggestivebutnotconclusive (Baldini,2008).Whileuniversityfacultymaybemoreactiveparticipantsinprob-lemsolvingforandjointresearchwithindustrythansuggestedbyBush(1945)orevenbyStokes(1997)characterizationofuse-inspiredbasicresearch, suchactivitydoesnotnecessitate a substitutionof commer-cialresearchandrelatedactivitiesforthegenerationofnewknowledge,theopenpublicationofresearchinreferredoutlets,orforteachingandotheractivitiesrelatedtouniversitieseducationalmissionsandgoals.While the extent to which URCs affect faculty involvement with

industry and students remains anopenquestion, a handful of empiri-calstudiesinformourendeavor.Abriefreviewoftheseworkshelpstobringfocusboththeoreticallyandempiricallytoouranalysis.

The Relationship Between Industry Involvement and Publishing

Thenotionofuniversity-industryinteractionsaffectingacademicfac-ultyproductivitydoesnotoriginatewithinterest(positiveornegative)inURCs.Thoughtherehasbeenbutonedirectanalysisofthissortthatweareawareof,albeitasacasestudyofasingularURCusingpaneldata todemonstrateapositive impactofURCaffiliationon thedisci-plinaryandinstitutionaldiversityoffacultypublication(Ponomariov&Boardman,2010), therehas longbeenabroader literaturenotfocusedon URCs per se one anticipating and demonstratingmore broadlyapositive relationshipbetweenentrepreneurialoutcomes suchasuni-


versitypatentingandmoretraditionalmodesofdisseminationsuchaspublishing. Studiesofacademicpatentingandpublishinghaveexaminedtheex-

tent to which university faculty are supplanting traditional academicresearchactivitieswithmarket-likeones.Empirical assessments typi-callyhave showneither apositive (Meyer,2006;Breschi,Lissoni,&Montobbio,2005)ornull(Agrawal&henderson,2002)correlationbe-tweenthe twodisseminationmodes.Relatedly,universityfacultywhoperformcontractualworkforindustryaremoreprolificatpublishinginpeer-reviewedjournals(vanLooy,Ranga,Callaert,Debackere,&Zim-mermann,2004;vanLooy,Callaert,&Debackere,2006). In termsofimpact, the citation rates forpublications authoredconjointlybyaca-demicand industry researchersaregenerallyhigher thanare the ratesfor papers authored by university faculty alone (hicks & hamilton,1999).hence, there seems to be a complementary rather than substi-tutiverelationshipbetweenindustry involvementbyfacultyandtradi-tional modes of academic research. Moreover, industry involvementseemstoimprovenotjusttheresearchproductivityofuniversityfacultybutalsothequalityorimpactoftheirresearch(Baldini,2008).Examinationofdifferentdisseminationmodesandqualitydoesnot

addresstherelatedconcernoverthesubstantivenatureofacademicre-search.Theacademiccapitalistargumentsuggeststhatuniversitiesanduniversity faculty are increasingly taking cues from industry in theirvaluationandselectionofdifferentavenuesforresearch.Azoulay,Ding,andStuart(2006),forexample,providedsomesupportforthisconcernby showing academic researchers in biotechnology to favor researchquestions that are commercially relevant. The agencies, such as theNationalSciencefoundation (NSf) andNational Institutes ofhealth(NIh),whichhavehistoricallyprovidedmajorfundingforinvestigator-driven basic research, are increasingly emphasizing industry involve-mentinuniversityscience(Bozeman&Boardmanm,2004).yet,thoseconcernedwithsuchmovementinscienceandengineeringfieldshavefailedtodemonstrateorexplainhowsuchashiftmaybedetrimentaltothetraditionalresearchmissionofuniversities.Industryhashistoricallyprovidedcuestouniversityfacultyforresearchandeducation(Mowery&Ziedonis,2002),andsomefields,suchastheengineeringdisciplinesandcross-disciplinaryfieldsincludingbiotechnology,arerelativelyap-pliedandinherentlyclosertomarketthanothers.Inourview,therealconcernliesnotwithnewmodesofdissemina-

tion,suchaspatenting,sincesuchmodesshowgenerallytobecomple-mentarywithpublishingquantityandquality(Lin&Bozeman,2006).


Nordowefeelconcernshouldliewithnewandcommercially-relevantareasofresearchanddevelopmentinuniversities,asuniversity-industryrelationsarenothingnew(Ruscio,1984)andtheaccumulationofnewknowledgenecessarily forgesuniversity faculty intonew institutionaland epistemological realms (Corley, Boardman, & Bozeman 2006;youtie,Libaers,&Bozeman,2006).however,weagree that thecon-cernshouldbetheextent towhich(inanyway)thesenewmodesandnewareasofresearchanddevelopmentunderminetheeducationalmis-sions of universities.

The Relationship Between Research and Teaching

Thenotionofuniversityfacultyshirkingtheireducationalresponsi-bilitiesforresearchdoesnotoriginatewithURCsortheacademiccapi-talist argument. Numerous commentaries on the traditional academicrewardtemplate,specificallyitsheavyfocusonpublications,haveex-pressedconcernoveratradeoffbetweenresearchandteaching(Boyer,1990; Braxton, 1996; Diamond, 1993, 1999).Accordingly, there hasbeenconsiderableexaminationoftherelationshipbetweenresearchandteaching,thoughnonehasaddressedhowdifferenttypesofresearchanddissemination,suchasthoserelatedtoindustryinvolvementandURCaffiliation,mayaffectteachingdifferently.Thefindings for this body ofwork aremixed.While some studies

have observed a negative relationship between research and teaching(Cage, 1991;Massy& Zemsky, 1994), others have observed a posi-tiverelationship(Bowen&Schuster,1986).Thereisperhapsthemostempiricalsupportforanullrelationship(Braxton,1996),withtimeal-locatedtoresearchhavingnoeffectontimeallocatedto teachingandviceversa (feldman,1987).Ameta-examinationof these resultspro-vides some consistency, demonstrating particular university types tocorrespondwithparticularrelationships(Milem,Berger,&Dey,2000).Most relevant here are studies that have assessed samples of university facultyworkinginresearchextensiveuniversities,whichdemonstrateanullrelationshipbetweenteachingandresearch(Braxton,1996).Inourview,thesespecificfindingswarrantfurtherinvestigationinto

theextenttowhichtheinvolvementofuniversityfacultywithindustryaffectstheirinvolvementwithstudents,particularlyforfacultyaffiliatedwithdifferenttypesofURCs.Accordingly,inouranalysisbelowwead-dresstherelationshipbetweenfacultyinvolvementwithURCsandtheirinvolvementwith industryandstudents.Becauseourdataareunique,we are uniquely in a position to distinguish amongdifferent typesofURCs.Thus,wedonotexpectuniformlevelsofinvolvementwithin-dustryandstudentsacrossallfacultycenteraffiliates.


Data

Ourdata are from theNSf-fundedSurveyofAcademicResearch-ers (SAR)conducted in20042005andextended in20082009.Thepurposeofthesurveywastostudyavarietyofaspectsoffacultyworkandattitudesinuniversities,focusinginparticularonindustrialactivi-ties and research-center affiliations.Our initial targetpopulationwastenured and tenure-track faculty members in Carnegie foundation(2000)researchextensiveuniversities,historicallyBlackCollegesandUniversities(hBCU),anduniversitiesdesignatedasExperimentalPro-gramtoStimulateCompetitiveResearch(EPSCoR).Samplingframeswere constructed fromuniversity catalogs1 in the followingNationalScience foundation STEM disciplines: biology, computer science,mathematics,chemistry,physics,earthandatmosphericsciences,agri-culture,andsociology.Inaddition,samplesweredrawnfromfivesub-disciplines of engineering: chemical, civil, electrical, materials, andmechanical.Womenwereover-sampledfrommostdisciplines.Thiswasdoneto

make sure that sufficientwomenappeared in the sample and, specifi-cally,toensurethatwomenfromeverydisciplinewouldberepresented,eventhosewherewomenarefoundinquitesmallnumbers(e.g.,com-puterscience,electricalengineering).furthermore,stratificationbysexreplicatesthetenureandrankdistributionoffields,aswomenarelikelytobeatearliercareerages.Theresultofthisstratification,ofcourse,isthatvariablesstronglycorrelatedwithsexmaylead tospurious infer-enceaboutco-varyingvariables.Inthecurrentstudy,allourstatisticalmodelscontrolforgender,thusremediatingthispossiblebias(Winship&Radbill,1994). Ina recent study (Bozeman&Gaughan,2011),weused the alternative strategy of providingweights for allmodels butfoundnegligibledifferencesbetweenaweightingstrategyandcontrol-ling for gender. The survey was sent to 5,916 targets, yielding 2,086 respondents

afterthreewavesofadministration.Adjustingfordeceasedandretiredtargets, theeffectiveresponseratewas36%.Thecurrentstudyusedasub-sample of the initial survey, eliminating thehBCUandEPSCoRrespondentsaswellasthesociologydiscipline.Eachofthesesub-stratawasconsiderablylesslikelytobeaffiliatedwithURCsbut,moreimpor-tant,wenotedrespondentselectionbiasforhBCUfacultyaswellasaconsiderablylowerresponserate(22%).Theresponserateforthecur-rentsub-sampleofCarnegiefoundationresearchextensivefacultywas39.5%.forthisgroup,wefoundnoresponsebiasbasedondisciplineorrank.Thisparticularanalysisincluded1714respondents.Missingdata


wereaverysmallpercentage(lessthan.1%ofallobservations)andap-peared to be at random in sensitivity analyses. TheoriginalSARdataincludedvariablesindicatingwhetherrespon-

dentswereaffiliatedwithcentersandthenamesofthosecenters.Inthequestionnaire,weprovidedaspecificdefinitionofauniversityresearchcentersasaresearchinstitutionthathasfiveormorefacultyandpost-doctoralresearchersandincludesparticipantsfrommorethanonedis-ciplineandmorethanoneacademicdepartment.followingthatdefini-tion,weaskedrespondentstocheckaboxindicatingthattheywereaf-filiatedwithaURCoranotherboxindicatingthattheywerenot.Iftheywereaffiliatedwithmorethanonecenter,weaskedthemtoindicatetheaffiliationtheyviewedasmostimportant.Ina recentextensionof thedata,wesetout toverifycentermem-

bership.We did this in threeways. first, we examinedwebsites foraffiliate rosters. Second,we examined curricula vita for respondents(another part of the overall project involved collection of Cvs fromrespondents;wedidnotusetheCvdatainthisstudy).finally,inafewcases,weeithermade telephonecallsor sent emails toURCperson-neltodevelopinformationaboutavarietyoffactors,includingcentermembership. BeforeprovidinghypothesesabouttheeffectsofURCsonindustrial

involvement and teaching and student-related activities, we begin byshowingdescriptiveresultsforURCaffiliationandcentertypes.

Towards a Taxonomy of University Research Centers

Because therehavebeen so fewgeneral assessmentsofURCs, ex-aminationofbasicdifferencesbetweendifferentURCsisnecessarybe-foreanalyzingtheir impactsonfaculty involvementwith industryandstudents.Webeginwithabriefdiscussionof theURCcodingschemeandthenassessdifferencesacrossthecentertypes.

Coding the URCs

Thefirst step in thecodingprocesswas tovalidateURCaffiliationforSARrespondents.Whereas632 respondents to theSAR indicatedtheywerecenteraffiliates,wewereable tovalidate500.Ouranalysisreflectsonlyvalidatedcenteraffiliates.The132notcodedasvalidcenteraffiliatesweredroppedforseveral

reasons.ThemostcommondisqualificationwasthatrespondentslistedaffiliationswithresearchunitsnotmeetingourstipulatedcriteriaforaURC(fiveormorefacultyandpostdoctoralstudents,participantsfrommorethanonedisciplineandmorethanoneacademicdepartment).2 Our


objectivewastogetthebestpossibleestimate,giventhemanyvagariesassociatedwithcenteraffiliation.InourURCcodingprotocol,seniorinvestigatorsworkedcloselywith

graduatestudents tohelpclassifyURCswhenthecenter typewasnotobvious. In some cases, therewas little possibility for error. for ex-ample, theNSfandmanyotherfederalagenciesprovide listsof theircenters.Acenterwasdesignatedasbeingassociatedwithaparticularfederal agency not on the basis of the amount of funds provided viaconventional grants but on the basis of formal affiliations, programcompetitionawards,cooperativeresearchagreements,andsuch.Somefederalagencyprograms,includingtheNSf,havesunsetperiodsafterwhichdirectfundsarewithdrawnandthecenterisexpectedtobeself-sustaining.Inthecaseofcentersthathadreachedtheirsunsetperiodwithafederalagency,we listed themasUniversity if theURChadcontinuedwithuniversity-basedfunding.Thegreaterdifficultywasdis-tinguishingbetweenStateandUniversityURCs.Thegeneralrulewasthat ifcenterswerecreatedbystategovernmentwithspecialpro-gram initiatives, suchas centersof excellenceprograms, theywereclassifiedasState.Naturally,mostofthepublicuniversityURCshadatleastsomebasestatefunding;absentsomespecialdesignationorspe-cialprogram,thesewerecodedasUniversity.3

Different Types of URCs

AsTable 1 shows, three-fifths of theURCs in our studywere cre-atedbyuniversitiesandsustainedbyacombinationofuniversitybaseresourcesand individual investigatorgrantand foundation funds (andmore rarely industry funds) that center affiliates typically bring in tosupporttheirresearch.MostoftheliteratureonURCshasbeenlessat-tentivetothesetypesofcenters,despitetheirnumericaldominance.Tothebestofourknowledge, all case studiesofURCshave focusedoneitherfederal-orstate-establishedcenters,particularlythoseestablishedbytheNSf.Otherdata-basedstudiesofURCshaveemphasizeddiffer-encesbetweengovernmentandindustrycenters(e.g.,Boardman&Cor-ley,2008;Boardman,2009)buthavenotaddressedthisdominantcentertype(atleastintermsofrawnumbersofURCs).ThesecondmostcommonURCsinourstudywereonescreatedby

theNSfunder theirvariouscentersprograms.Therewasnodifficultyclassifyingthe81NSfURCs,astheywereonlistsofcentersandtheNSfwebsite, inmostcases,providedclear informationabout theori-gins,sponsorship,andprogrammaticstructureof theNSfcenters.4 In addition,manyof theprogramsofNSfcentershaveestablished theirownprogram-widewebsites,independentoftheNSf,detailingpastand


presentcenters,informationoncentercharacteristics,andbestpractices,amongotherresources.5

TheURCsweclassifiedasStatearegenerallythosecreatedbyspe-cialprogramsfromstategovernment,oftenprogramsaimedspecificallyat engendering technology-based regional economic development. Incomparison to theUniversityURCs, these centers have been amuchmorecommonfocusofresearch(Lee,1996;Rogers,Takegami,&yin,2001;Lin&Bozeman,2006),chieflybecauseoftheirexpliciteconomicdevelopmentandtechnologytransfermissions.The University, State, and NSf centers accounted for about 89%

of theURCsinourdatabase,dominatingthepopulationofcentersatCarnegie research extensive universities.6 Other URCs included inourstudyweresponsoredby,amongothers, theNational Institutesofhealth(NIh),theDepartmentofEnergy(DOE),theU.S.Environmen-talProtectionAgency(EPA), theNationalAeronauticsandSpaceAd-ministration(NASA),and theU.S.DepartmentofDefense(USDOT).ManyotheragenciesalsosponsorURCsbutnotinsufficientnumberstoensuretheirpresenceinalargesampleofuniversityresearchers.

Empirical Differences Among the URC Types

Inlightoftheempiricalevidence,developingacrudetaxonomypre-sentednogreatchallenge.Ourbasictaxonomyisfour-fold:University,State,NSf,andOther.This isnot entirely satisfying inasmuchas the

TABLE1

NumberofCenterAffiliates,byCenterType

Center Type NumberofvalidatedCenterAffiliates

University 307

State 55

NSf 81

NIh 10

DOE 7

EPA 5

NASA 7

DOD 6

USDOT 4

Other 18

Total 500


Other category lumps together quite disparate centers.We also un-derstand that there may be considerable within-type variation. Buteven if the approach remains crude, it is important to remember thattheOthercategoryconstitutesonlyabout10%ofURCs inour study.WhenwespeakofuniversityfacultyparticipatinginURCs,wealmostalways refer to just three types:University,State, andNSf.This is adistortionofneitherempirical realitynorhistory.Theimportantques-tioniswhetheracrudefour-foldtaxonomycandiscriminateandpredictbehaviors. Webegintheanalysisofthetaxonomybyconsideringpossiblediffer-

encesamongthefourtypes.Table2providesthemeansforthe500cen-ters includedinouranalysisaccordingtonumberofaffiliatedfaculty,founding year, and number of universities formally affiliatedwith thecenter.AsweseefromTable2,thecenterstendedtobequitelargewiththeaveragehavingabout68 faculty. Interestingly, theOther categorytendedtobelargest in termsoffacultypercenter.This isperhapsnottoosurprisingsincealmostallof thesearesponsoredbyfederalagen-cies, including the largecenterscreatedby theDepartmentofEnergyand Defense. The university-sponsored URCs, while most prevalent,wereonaveragethesmallestintermsoffacultypercenter.Onaverage,theNSfcenterswereofmostrecentorigin.Thisislikely

due, inpart, to the fact thatmostNSfcentershavesunset featuresandthat theyarealsodefundedonoccasionprior tosunset.TheStatecentersweretheoldestand,onepresumes,themoststableintheirfund-ing.ManyState centers have a line itembudget allocation.TheNSfandStateURCshavea largernumberof cooperatinguniversities.Asonemight expect, theStateURCs often includemultiple universitiesfromwithin the sponsoringstate.NSfURCs tend to linkuniversities

TABLE2

CharacteristicsofCenters,MeansbyType

Center Type Centerfaculty yearCenterfoundedNumberofUniversitieswithformalAgreements

University 54.9 1986 1.1

State 91.9 1976 3.1

NSf 67.3 1990 3.8

Other 112.1 1982 1.5

All 68.9 1985 1.8


thataregeographicallydispersed,partlybecausemanyofthesecentersareintendedasnationalresources.Wewereparticularlyinterestedindistinguishingbricksandmortar

URCsfromvirtualones.Literatureonscientificcommunication leadsonetoexpectquitedifferentcollaborativebehaviorsbasedonproximityandvirtuality(hinds&kiesler,1995;Iansiti,1995;Bouty,2000).ThedatapresentedinTable3showthatUniversityURCsaretheonesmostlikelytobevirtual,perhapsindicatingarelativelymodestinvestmentinresourcesbasedonexistingcapabilitiesratherthannewinitiatives.TheStatecentersareleastlikelytobevirtual.ThisisperhapsowingtothefactthattheStatecenterstendtobeespeciallyfocusedonaparticulartypeoftechnologydevelopmentandthushaveappliedmissionsrequir-ingface-to-faceinteractionfortheiraccomplishment.WithabasicunderstandingofthedifferenttypesofURCs,webegan

our examinationbyassessing self-reports fromURCfaculty affiliatesregardingtheimpactsofURCs,byURCtype,onindividualcareersandproductivity, including individualoutcomesrelated toresearch, teach-ing,andcommercialinvolvement.

Descriptive Results for the Impacts of URC Affiliation on Faculty Involvement with Industry and Students

OurchiefconcernwasdeterminingtheimpactofURCaffiliationonfaculty teaching-related activities and industrial interactions.We fo-cusedonthese twooutcomesbecausewefelt thatcharacterizations inpreviousresearchandcommentaryhavebeentoostarkandtoofocusedonextremecases.SincewehadrepresentativedataforuniversitySTEM

TABLE3

virtualandNotvirtualURCs,byType

Center Type Not Virtual Virtual Total

University 217 85(28.1%) 302

State 49 5(9.2%) 54

NSf 68 12(15%) 80

Other 43 13(23.2%) 56

Note. Chisquare=12.97. * p


faculty,withacloseapproximationinthepercentagesandcompositionofURCaffiliates, ourdatapermittedus toprovide analysisof centerimpactswithadegreeofconfidenceinthevalidityandgeneralizabilityof thefindings.Overall,weexpectedURCs toaffect the industryandstudent involvementofaffiliatedfaculty, thoughnot in thewayschar-acterizedbyprior research focusedonspecificcases.Specifically,weexpectedcertain typesofURCs tohavemore impact thanothers.Webeginwithdescriptivedata for the industrial and student activitiesofURC-affiliated faculty.Wereserveexaminationof the impactsof fac-ultyaffiliationwithdifferenttypesofURCsfortheregressionanalysesinthefollowingsection.

URC Affiliation and Faculty Industrial Activities

TheSARquestionnaireaskedabranchingquestion,first asking re-spondentsiftheyhadhadanyworkingrelationswithprivatecompaniesduring the past 12months.While only 12%of respondents indicatedtheyhadindustryfundingthroughgrantsandcontracts,morethanhalfofrespondentsreportedthattheyhadhadsometypeofindustryinterac-tion.Theserespondentswerereferredtoachecklisttoindicatethespe-cifictypesofactivities.Table4providescrosstabulations,showingthenumber and percentages of respondents engaged in various industrialactivities,forbothvalidatedURCaffiliatesandnon-affiliates.Themostapparentgeneralizationwecouldmakefromtheseresults

isthatURCaffiliatesseemtodifferfromothersinthattheaffiliatesaremorelikelytobeinvolvedinawiderangeofindustrially-relevantactiv-ity.foronlyoneoftheitems,whethertherespondentactuallyownedorworkedforacompanyduringtheperiodofthesurvey,istherenostatis-ticallysignificantdifferencebetweenURCaffiliatesandnon-affiliates.Thisislikelyduetothefactthatsofewinthesampleansweredyeson this item. There issomedifferenceamongthe itemsin termsof thedegreeof

differencecenteraffiliationseemstomake.URCaffiliatesaretwiceaslikely as non-affiliates to have coauthored a publishedpaperwith in-dustrypersonnelandtohaveplacedgraduatestudentsandpostdoctoralstudentsinindustryjobs.URCaffiliatesaremorethantwiceaslikelytohaveservedasapaidconsultanttoaprivatecompany.Despitenoteworthyandstatistically significantdifferencesbetween

the twogroups, it is important not to overemphasize differences. forevery activity, both URC affiliates and non-affiliates were involved.Perhapsmoreimportant,fornoneoftheactivitieswasaclearmajorityengaged.Withrespecttothemostcommonlystudiedcommercialactivi-


ties,patenting,morethan90%ineachgroupdidnot engage in patent-ing(atleastduringtheyearoftheSARsurvey).Even the relativelymodest level in industrial involvementdepicted

inTable4cannotbecausallyattributedtocenteraffiliation.Sincecen-teraffiliationisnotarandomlydistributedvariable,unrelatedtoothercharacteristicsof researchers, it ispossible that someof theobserveddifferencesareowingtovariablescorrelatedwithcenteraffiliation.forexample,engineerstendtobemoreinvolvedwithindustryandarealsooverrepresentedinURCs(at least inoursample).Moreover, insubse-quentmultivariatemodels,wecontrol forfieldandseveralotherpos-siblyconfoundingvariablesinordertogetamorevalidassessmentofURCeffects.

TABLE4

Cross-TabulationsforIndustryInvolvementandURCAffiliation

IndustrialInvolvementActivityaNumber and %Non-URC

Number and %URC p valueb

Personsfromaprivatecompanyaskedforinforma-tionaboutmyresearch,andIprovidedit.

372 (31.0%)

259 (50.3%)

0.000

Icontactedpersonsinindustryaskingabouttheirresearchorresearchinterests.

187 (15.6%)

135 (26.2%)

0.000

Iservedasaformalpaidconsultanttoanindustrialfirm.

192 (16.0%)

117 (37.9%)

0.002

Ihelpedplacegraduatestudentsorpost-docsinindustryjobs.

242 (20.2%)

186 (43.5%)

0.000

IworkedatacompanywithwhichIamowner,partner,oremployee.

36 (3.0%)

24 (4.7%)

0.115 (n.s.)

Iworkeddirectlywithindustrypersonnelinworkthatresultedinapatentorcopyright.

51 (4.3%)

42 (8.2%)

0.004

Iworkeddirectlywithindustrypersonnelinanefforttotransferorcommercializetechnologyorappliedresearch.

159 (13.3%)

114 (22.1%)

0.000

Icoauthoredapaperwithindustrypersonnelthat hasbeenpublishedinajournalorrefereed proceedings.

139 (11.6%)

115 (22.3%)

0.000

a n=1714(1199NonURC-Affiliated,515URC-Affiliated).b BasedonGammacoefficient.


URC Affiliation and Faculty Teaching and Student Activities

TheSARquestionnaireincludesanumberofitemsforattitudesandbehaviorsrelated tofaculty involvementwithstudents.Theattitudinalquestionspertaintomotivesforcollaboratingandpursuinggrants.Theitems for behaviors pertain to teaching graduates and undergraduatesand student support.Aswith the descriptive results for industrial ac-tivities,wemusttreattheinferencesfromthisinitialexaminationwithcaution,astheresultsmaybeafunctionofinter-correlationwithothercausalfactorsthatdiminishtheimportanceofURCaffiliation.

Student-Related Motives for Pursuing Grants and Research Collabora-tions.OnequestionintheSARasksthefollowing:Ifyouspendtimewritingorparticipatinginthepreparationofproposalsforcontractsorgrants,please indicateyouragreementordisagreementwith thestate-mentsbelow, 7allofwhichprovidereasonsforpursuinggrants,includ-ing(amongothers)formal jobrequirements,administrativedirectives,intellectualpursuits,summersalary,and,relevanttoourstudy,amajormotivationformypreparingproposalsistosupportgraduatestudents.TheresponsesfortheitemsuseaLikert-typescale,includingStronglyAgree,AgreeSomewhat,DisagreeSomewhat,andStronglyDis-agree.ThedifferencesbetweenURCaffiliatesandnon-affiliatesonthesupport students factorwere statistically significant (Gamma0.268,p


ever,isthatURCaffiliationseemedtohaveapositiveratherthannega-tiveimpactonstudent-relatedattitudesandmotives.

Teaching and Student Support. In addition to the motivational vari-ables,weconsideredself-reportsof teachingandotherstudent-relatedactivities.Table5providestestsfordifferencesofmeansbetweenURCaffiliates and non-affiliates for time allocations to teaching and stu-dent advisingand for thenumberofundergraduate students,mastersstudents, and doctoral students supported by respondents grants orcontracts.The general pattern was that university faculty not affiliated with

URCs were more engaged teaching undergraduate students, spend-ingonaveragemoreworkhoursinundergraduateteaching(10.5hoursperweekonaverage)whencomparedtoURCaffiliates(8.5hoursperweek).however,URC-affiliatedfacultyspentslightlymoretimeteach-inggraduatestudents,but thedifferenceisnotstatisticallysignificant.TheothermajorfindingisthatURCaffiliatessupportedalargernumberof students, not onlydoctoral students but alsomasters students andundergraduates.ThelessonseemstobethatURCaffiliationlessensun-dergraduate teaching but has the significant educational advantage ofprovidingmorefinancialsupporttostudentsateverylevel.

TABLE5

MeansforTeaching-RelatedActivities,byURCAffiliation

n Mean Std. Deviationp value (f-Test)

Workhours: 0(notURC) 1197 10.51 8.81 0.012TeachingUndergraduates 1(URC) 512 8.50 7.74

Workhours: 0(notURC) 1196 6.26 6.33 n.s.TeachingGraduates 1(URC) 512 6.52 6.10

Workhours: 0(notURC) 1196 2.43 3.01 0.074Advising 1(URC) 512 2.57 2.91

# of Undergraduates 0(notURC) 1198 1.02 2.04 0.089Supported 1(URC) 514 1.38 1.95

#ofMastersStudents 0(notURC) 1198 0.76 1.25 0.000Supported 1(URC) 515 1.14 1.90

#ofDoctoralStudents 0(notURC) 1198 1.53 2.19 0.000Supported 1(URC) 515 3.00 2.89


Dependent Constructs and Hypotheses

BecausetherearesofewgeneralassessmentsofURCs,webasedourhypothesesforURCaffiliatesindustryandstudentinvolvementonourdescriptiveanalysisaboveandontherelevanttheoriesofresearchcol-laboration and faculty time allocations rather than on past results forcenters based on single case studies or those studies using anecdotalevidence.WehadtwosetsofhypothesesabouttheimpactsofURCaffiliation,

onepertainingtoteachingimpactsandtheothertocommercialimpacts.Beforepresentingeachsetofhypotheses,wedescribetherelevantde-pendent constructsfor industry and student involvement, which in-cludeavarietyofteaching-(e.g.,timeallocatedbyacademicfacultytoteaching, facultysupportofgraduateandundergraduatestudents)andcommerce-related (e.g., different types and intensities of industry in-volvement)outcomesandactivities.

URCs and Faculty Involvement with Industry

The Industrial Involvement Index.forcommercialactivities,weem-ployedascaledevelopedpreviously(Bozeman&Gaughan,2007),onethathasbynowbeenvalidatedinanumberofstudies(e.g.,Bozeman&Gaughan,2011;Gaughan&Corley,2010;Boardman,2009;Pono-mariov,2008).TheIndustrialInvolvementIndexwascreatedbyex-aminingthepercentagesforeachoftheindustrialinteractionitemsandthen using the inverseasaweight.Togiveanexample,14.8%of thissamplereportedhavingcoauthoredapaperwithsomeonefromindustryand,thus,engaginginthatactivityreceivesaweightof85.2.Similarly,6.2%of thesamplereportedhavingdevelopedapatentwithsomeonefromindustryand, thus,engagingin thatmoreuncommonactivityre-ceivesaweightof93.8.Theweightsforalltypesofactivitiesinwhichtherespondentswereengagedwerethensummed,creatingaweightedIndustrialInvolvementIndex.forthepresentsample,theindexrangesbetween0and6.68withameanof1.27.Thescalehas the followingproperties:analphareliabilityof0.78andnormaldistribution(skew=1.3; kurtosis = .99). It thereforemeets the requirements forOrdinaryLeast Squares (OLS) Regression. Additional information about thescaleisprovidedinBozemanandGaughan(2007),includingresultsofanoblique(Oblimin) factoranalysisusedasanalternative,validatingscale.

Hypotheses for URCs and Industry Involvement. Wehadtwooverarch-inghypothesesfortheeffectsofaffiliationwithURCsontheindustryinvolvementoffaculty:


hypothesis1.1:ComparedtofacultywhoarenotaffiliatedwithURCs,affili-atedfacultywillbe no more (or less) involvedwithindustryaftercontrollingforvariablesmitigatingtherelationshipbetweenURCaffiliationandindus-trial involvement.

hypothesis 1.2:While overall there is no significant relationship betweenURCaffiliationandindustrialinvolvement,therewillbemodestdifferencesintheindustrialinvolvementwhenonetakesintoaccountURCtype.Specifi-cally,affiliationwithNSfcentersandStatecenterswillincreaseindustrialinvolvement.

It isnotuncommon,asamatterofcourse, toposit thenullhypoth-esis.however,ourhypothesisofnoeffectwasnotsimplyamatterofform.Rather,webelievedthatmostofthevariancebetweenURCaffili-ationandindustrialinvolvementcouldbeaccountedforbytwofactors.first,onlyasmallpercentageofURCshaveexplicitindustrialoutreachandcommercializationmissions,specificallytheNSfandStateURCs.Examiningtherepresentative listofURCsinourdataset, therestarerelativelysmallinstitutionsfocusedoncreatingfundamentalknowledgeinthesciences.Thus, inourexpectationsforfacultyindustryinvolve-ment,we followed both the institutional (Clark, 1998) and resource-basedviews(e.g.,Barney,1991;Grant,1996)ofresearchcollaborationbypositingapositiveeffectforonlycertaintypesofURCs.WhileNSfandStateURCsarelikelytohavetheknowledge-integrationandcapac-ity-developmentmotives conducive to institutional norms and expec-tations for industry involvementby affiliated faculty, theother centertypesare likely toharbormotives andnorms towards traditional aca-demicresearchandrelatedactivities.Second,weknowfromcasestudies thatengineeringfacultytendto

beoverrepresentedinURCsandthatengineeringfacultygenerallyaremuchmorelikely,whether or not they are associated with URCs, to be involvedwithindustry(see,forexample,Bozeman&Boardman,2003;Boardman&Bozeman, 2007). Engineering and other disciplines areimportant controls, as different disciplines have properties that resultin variable propensities for industry involvement. for instance, somedisciplines require broader skill sets and access to diverse resources(vanRijnsoever,hessels,&vandeberg,2008),whichmaybegainedbyexpandedprofessional ties andnetwork linkages (Bozeman,Dietz,&Gaughan2001),includingbutnotlimitedtoindustry.Wealsocontrolledforothervariablespossiblyconfoundingtherela-

tionshipbetweenURCaffiliationandindustrialinvolvement,includingtenure,age,U.S.citizenship,andgender.Weexpected that thosewho


areolderandtenuredwouldbemorelikelytobeinvolvedwithindus-try;ourreasoningisthatindustrialactivitiesgenerallydonotcontributemuchin tenuredecisionsand, thus, those in tenure jeopardywill tendtofocusmoretimeontraditional(non-commercial)academicresearch(vanRijnsoever,hessels,&vandeberg,2008;Boardman&Ponomar-iov,2007).WithregardtoU.S.citizenship,weexpectedthat this is tosomeextentanindicatorofEnglishlanguagefluencyandthatindustry,all else equal, prefers to avoidcommunicationbarriers.Naturally,weunderstandthatmanynon-citizensarefluent,butcomparedtocitizens,manyfewerarefluent.U.S.citizenshipmayalsobeaproxyforsocialcapitalandforinsiderknowledgeabouthowtodealwiththeuniver-sityandbusinessbureaucraciesthatbufferrelationshipsbetweenfacultyandindustry(Lee&Bozeman,2005).finally,wecontrolledforgenderbecauseweknowfromotherresearchusingthisdatathatwomen,boththosewhoareURCaffiliatesandothers,tendtobelesslikelyinvolvedwithindustryandalsobecauseourdataover-sampledwomenandsuchacontrol, tosomeextent,remediates theover-sampling(seeCorley&Gaughan,2005,fordiscussionofbothpoints).

URCs and Faculty Involvement with Students

The Student Involvement Indices. Aswasthecasewiththeindustrialinvolvementvariables,itwasusefultodevelopmultivariateindicessoas to vitiate the need for a large number of separate regression models. In this case,wedevelopedStudent Involvement Indicesbasedonafactoranalysisof the twomotivationalandsix itemsdiscussed in thedescriptiveanalysisabove.Specifically,weemployedaprincipalcom-ponentsanalysis,usingvarimax rotation toensureoptimal loadingofvariables on the dimensions. Table 6 provides the results of the analysis. As is customary,wenamed the resultant factordimensionsaccordingtothevariableswiththehighestmagnitudeloadingsfortherespectivefactors.ThefirstfactordimensionwetermedStudentsSupportedbe-cause all three loadings in excess of +/- .50 pertained to the numberof students supported.ThesecondStudent Involvement Index,Men-torMotivation,includeshighloadingsonthetwomotivationvariables,themotivetoseekgrantstosupportstudentsandthemotivetocollabo-rateonresearchtohelpgraduatestudents.ThethirdIndexwastermedTeachingUndergraduatesbecausethestrongestpositive loadingwaspercentageoftimeteachingundergraduates;teachinggraduatestudentsloads strongly but negative on the dimension. It is useful to note that the three factors range in their eigenvalues

from1.09to1.83,indicatingthatnofactorisdominantandthatnoneisaspuriousorshadowfactor(Green,1983).furthermore,letusnotethat


eachofthefactorsisanintervalscalevariableand,thus,canbeusedasadependentvariableinanOLSregressionmodel.

Hypotheses for URCs and Student Involvement.WhereasweexpectedURC affiliation to have only modest overall effects on industry in-volvement, we expected greater effects on certain aspects of studentinvolvement:

hypothesis2.1:ComparedtofacultywhoarenotaffiliatedwithURCs,affili-atedfacultywill,ingeneral,bemore involved with students per the Students Supported index,aftercontrollingforvariablesmitigatingtherelationshipbetweenURCaffiliationandstudentinvolvement.

hypothesis2.2:ComparedtofacultywhoarenotaffiliatedwithURCs,affili-atedfacultywill,ingeneral,bemore involved with students per the Mentor Motivation index,aftercontrollingforvariablesmitigatingtherelationshipbetweenURCaffiliationandstudentinvolvement.

hypothesis2.3:ComparedtofacultywhoarenotaffiliatedwithURCs,af-filiatedfacultywill,ingeneral,beless involved with undergraduate students and more involved with graduate students per the Teaching Undergradu-ates index, after controlling for variablesmitigating the relationship be-tweenURCaffiliationandstudentinvolvement.

TABLE6

factorAnalysisofStudentImpactvariables

factorLoadingsStudents Supported

Mentor Motivation

Teaching Undergraduates

Grants:SupportGraduateStudents 0.176 0.687 0.256

# of Undergraduates Supported 0.799 -0.070 -0.149

#ofMastersStudentsSupported 0.659 0.101 0.044

#ofDoctoralStudentsSupported 0.624 0.142 0.186

Collaboration:helpGraduateStudents 0.058 0.783 0.075

%timeonalladvising -0.020 0.468 -0.247

%timeonundergraduateteaching -0.177 -0.096 -0.734

%timeongraduateteaching -0.101 -0.031 0.773

Note. Specifications:(a)ExtractionMethod:PrincipalComponentAnalysis;(b)RotationMethod:varimaxwithkaiserNormalization;(c)Eigenvalues:StudentsSupported(1.83),MentorMotivation(1.27),TeachingUn-dergraduates(1.09).


Exceptfortheexpectationforteachingundergraduates,thehypothe-sesfortheStudentInvolvementIndicesruncountertocharacterizationsof URCs having a negative impact on faculty involvement with stu-dents.SinceURCstypicallyareholdinginstitutionsfornewordiscre-tionaryresourcesandsincestudentsareoftenanimportantpartof theresearchproductionmissioncommon toallURCs, it stands to reasonthatURCaffiliateswouldhaveatleastsomestudentsupportresourcesbeyond those available to non-affiliated faculty. This general expec-tation is supported by previous case study and empirical research onURCfacultyteachingandstudentinteractions(e.g.,Corley&Gaughan,2005;Bozeman&Gaughan,2011).Moreover,someoftheURCtypesinthetaxonomypresentedabovehaveexpliciteducationalmissions,in-cludingthecenterprogramsattheNSf(Bozeman&Boardman,2004).Ponomariov (2009)hasprovidednotonly themostgeneral empiri-

calsupport(andthereforethestrongestevidence)ofapositiverelation-shipbetweenuniversity-industry interactionsandstudentoutcomesatthefacultylevelofanalysis,andcontrollingforURCaffiliationatthat,butalsoacompellingexplanationof thepositiverelationshipin termsofcumulativeadvantage(albeitattheindividualfacultyratherthantheuniversity level of analysis). Specifically, Ponomariov explained stu-dentsasan importantdimensionofacademic researchershumanandsocialcapitalthatisparticularlysuitablefortheworkthatcharacterizesmanyuniversity-industryinteractions,inURCsaswellasinothercon-textsandarrangements.It is important tonote thatourgeneralexpectationofapositiveef-

fectwasforgraduate students.Incontrast,weexpectedthatURCsmayhaveanegative impactonfaculty involvementwithstudents insitua-tionswhen faculty researchmay be less integratedwith students, forinstance when teaching undergraduates. however, if prior empiricalfindingsontherelationshipbetweenfaculty timeallocatedtoresearchversus that allocated to teaching is to serveasaguide,weknow thatforfacultyinCarnegiefoundationresearchextensiveuniversities(fromwhichtheSARsamplewasdrawn)generallyanullrelationshipisfound(feldman, 1987; Braxton, 1996;Milem, Berger, &Dey, 2000; Link,Swann,&Bozeman,2008).As with our expectations for the Industry Involvement Index, we

expected that a number of individual factors besidesURC affiliationwouldexplainvariabilityinfacultystudentinvolvement.forinstance,tenured faculty may be more likely involved with teaching graduatethanundergraduatecourses,whichis trueformanydepartments inre-searchextensiveuniversitiesthelatterbeingtaughtbyjuniorfacultyandadjuncts.Thus,whilewehadsomeexpectationofanegativerela-


tionshipbetweenURCaffiliationandstudent involvement(i.e., teach-ingundergraduatestudents),thisexpectationwastentative.Ourgeneral expectationofa complementary rather than substitutiv

relationshipbetweenURCaffiliationandstudent involvementwasnottheresultofoptimismorotherpositiveattitudestowardsURCsdevel-opedduringourpastresearchonthe topic. Indeed,muchofourworkoncentershasexaminedtheproblemsthataffiliationposesforfaculty(e.g.,Bozeman&Boardman,2003;Boardman&Bozeman,2007).ConsideringthatmanyURCshaveexpliciteducationalmissions,we

expectedapositiveconnectionnotjustbetweenURCaffiliationandstu-dent support but also between affiliation and teaching.This latter re-lationship (betweenURCs and teaching) has received the least atten-tion in past empiricalstudyandconstitutesperhapsthemostimportantcomponentofouranalysis.WhileapositivecorrelationbetweenURCaffiliationand student supportmay still be interpretednegativelyasacademicfacultycapitalizingonstudents toensure theirownprofitintermsofgrantsandcontractsprocuredfromgovernmentandindus-try8findingatandempositivecorrelationbetweenURCaffiliationandteachingwillmitigatesuchaninterpretation.

Regression Results

URCs and the Industry Involvement Index

Table7providestheresultsforanOLSregressionmodelfortheIn-dustrial Involvement Index on the center taxonomy dummies aswellasoncontrolsforscientificfield(engineeringis thereference), tenure(untenuredisthereference),gender(femaleisthereference),age,andU.S.born(non-nativeisthereference).Theresultsshowthatthemodelexplainsabout16%ofthevarianceintheIndustrialInvolvementIndex(adjustedR2=0.165)andthateachoftheURCtaxonomyvariablesisstatisticallysignificant.Asweanticipated,thetypeofURCwithwhichone isaffiliateddoesmakeadifference.Whileeachof thebetacoef-ficients is significant and positive, theUniversity-sponsored center ismoststronglypredictiveof industryinvolvementandtheStatecenters(0.052) the least. The more important story is perhaps the one told by examining theresults forfield.While theengineeringfieldvariable isexcludedfromthemodelowingtocollinearity tolerances,eachof theincludedfieldsissignificantlyandnegativelyassociatedwithindustrialinvolvement, indicating thatrelationsbetweenURCaffiliationandin-dustrialinvolvement,whilesignificant,areeclipsedbyfieldeffects.Theimplication is thatwhileURCaffiliation appears to increase industry

TABLE7

RegressionResultsforth

eIndustryInvolvementandStudentInvolvementIndices

(1) I

ndus

try In

volv

emen

t(2

) Stu

dent

s Su

ppor

ted

(3)MentorMotivation

(4)TeachingUndergraduates

Unstandardized

Coefficients

Standardized

Coefficients

Unstandardized

Coefficients

Standardized

Coefficients

Unstandardized

Coefficients

Standardized

Coefficients

Unstandardized

Coefficients

Standardized

Coefficients

1 if

ever

tenu

red

0.434**

0.137

0.450**

0.203

0.189**

0.086

0.099**

0.04

5(0.086)

(0.0

62)

(0.0

64)

(0.0

65)

Gender

0.302**

0.10

50.

010

0.00

5-0.120*

-0.0

600.075**

0.037

1=M

ale,0=fem

ale

(0.0

70)

(0.0

51)

(0.053)

(0.053)

Native-BornU.S.C

itizen

0.164*

0.05

20.096

0.043

-0.029

-0.013

-0.254**

-0.1

15(0

.075

)(0

.055

)(0

.056

)(0

.057

)

Agein

years

-0.008*

-0.0

57-0.015**

-0.1

54-0.003

-0.036

-0.007**

-0.0

75(0

.004

)(0.003)

(0.003)

(0.003)

Biologyoragriculture

-0.653**

-0.1

67-0

.075

-0

.027

0.20

70.

075

0.378**

0.138

(0.1

00)

(0.137)

(0.1

41)

(0.143)

Mathorcom

puterscientist

-0.790**

-0.197

-0.2

02-0

.070

0.081

0.028

0.597**

0.209

(0.1

01)

(0.138)

(0.1

42)

(0.1

44)

Physicistorchem

ist

-1.104**

-0.346

-0.069

-0.031

0.446**

0.198

0.520**

0.230

(0.083)

(0.131)

(0.135)

(0.137)

Engineera

0.320*

0.15

60.631**

0.308

0.476**

0.232

(0.130)

(0.134)

(0.135)

Uni

vers

ity-S

pons

ored

Cen

ter

0.372**

0.10

20.325**

0.12

60.

106

0.04

10.137*

0.053

(0.086)

(0.063)

(0.0

65)

(0.0

66)

(con

tinue

d)

TABLE7(c

ontin

ued)

RegressionResultsforth

eIndustryInvolvementandStudentInvolvementIndices

(1) I

ndus

try In

volv

emen

t(2

) Stu

dent

s Su

ppor

ted

(3)MentorMotivation

(4)TeachingUndergraduates

Unstandardized

Coefficients

Standardized

Coefficients

Unstandardized

Coefficients

Standardized

Coefficients

Unstandardized

Coefficients

Standardized

Coefficients

Unstandardized

Coefficients

Standardized

Coefficients

Stat

e-Sp

onso

red

Cen

ter

0.430*

0.05

20.

140

0.02

40.218

0.038

0.433**

0.07

6(0.192)

(0.139)

(0.143)

(0.1

45)

NSf-SponsoredCenter

0.528**

0.079

0.435*

0.091

0.06

10.013

0.096

0.02

0(0

.157

)(0

.117

)(0

.120

)(0

.122

)

Otherfederal-SponsoredCenters

0.558**

0.073

0.518**

0.094

-0.0

12-0

.002

0.310*

0.05

6(0.178)

(0.134)

(0.138)

(0.139)

(Con

stan

t)1.284

0.153

-0.326

-0.108

(0.169)

(0.168)

(0.173)

(0.1

75)

Obs

erva

tions

1564

1532

1532

1532

AdjustedRsquared

0.16

50.119

0.06

50.037

Not

e.Robuststandarderrorsin

parentheses.

a Excludedvariable:E

ngineer(Tolerance.000)

*Significantat5

%.**Significantat1

%orbetter.


involvement, theoverrepresentationofengineers inURCsmaybe themoreimportantcharacteristic.

URCs and the Students Supported Index

Table7alsogivestheresultsfortheStudentInvolvementIndices,thefirstofwhichistheStudentsSupportedIndex.ThestandardizedbetacoefficientsforaffiliationwithURCsintheUniversity,NSf,andOthercategoriesarepositiveandsignificant.ThefindingsforState-sponsoredURCs,whileinamodestlypositivedirection,arenotstatisticallysignif-icant.WespeculatedthattheseStateURCstendtohavelessrelianceonfederalgrantsand,asaresult,lessexternalfundingforsupportofstu-dents.Alsotheyaresomewhatmorelikelytobecommerciallyfocusedand, thus,studentsmayplayalesserproductionrolethanincentersmorefocusedonfundamentalorpre-commercialresearch.WhiletheURCvariableswereourchiefinterest,themodelsresults

offerotherpotentiallyinterestingfindings.Aswiththefindingsforin-dustryinvolvement,engineersaremorelikelytosupportstudentsthanarefacultyinotherfields.Thisfinding,whichis inaccordwithprevi-ousstudieswithadifferentbutcomparabledatabase(Lin&Bozeman,2006), isnot surprising forat least two reasons.first, the largestandbest-fundedURCs (e.g., theNSfcenters),which tend tohaveeduca-tionalmissions,alsotendtohaveadisproportionatepercentageofengi-neers.Second,engineerstendtobeconsiderablymoreactiveinsupport-ingmastersstudentsandnolessactiveinsupportingdoctoralstudents(Lin&Bozeman,2006;Bozeman&Gaughan,2007).Unexpectedly,ageisnegativelyrelatedwithstudentsupport,butten-

ureis(asexpected)positivelyrelated.Aftercheckingfrequencies,dif-ferencesofmeans,andzeroordercorrelations,weobservedthatthesefindingsarevalidones.Ascanoftheoriginaldatasuggeststhatthere-lationship is curvilinearwith theyoungest respondents and theoldestrespondents supporting fewerstudents.But since thecurveof studentsupport and age is steeper at the front end of the age distribution than in thebackend,thebetaforlinearregressionisaswehaveobserved.

URCs and the Mentor Motivation Index

ThemodelforasecondStudentInvolvementIndex,MentorMotiva-tion, isnotwellaccounted forbyURCaffiliationofany type.Table7showsthatonesaffiliationwithaURCdoesnotseemtoaffectstu-dent-orientedmotivationforgrantsupportorresearchcollaboration.In-stead, themodest amount of variance explained in thismodel is bestaccountedforbytenurestatus,with tenuredpersonshavingapositiveassociationwith thementordimensionand,byfield,withengineers


andphysicists/chemistsmorelikelytohaveapositiveassociationwiththe dimension.

URCs and the Teaching Undergraduates Index

finally,Table 7 shows the results for the last Student InvolvementIndex,TeachingUndergraduates,tobemorerobust.AmongtheURCtypes,onlyaffiliationwithNSfcentersdoesnothaveapositiveasso-ciationwith the teachingdimension, implyingnotonly that faculty inNSfcentersspendlesstimeteachingundergraduatesbutalsothattheseURCaffiliatesspendmoretimeteachinggraduatestudents.TheState-sponsored centers have the strongest associationwith the dimension.AswiththefindingsfortheStudentsSupportedIndex,thismaybeduetoasomewhatlesserdegreeofdependencebythesecentersonfederalgrantsand,asaresult,lessfundsavailableforteachingbuyout.Thefinding thatnative-bornU.S. citizenshaveanegative relation-

shiptoteachingundergraduatesispossiblyexplainedbythemeasureofthekeyvariablesinthedimension.Theteachingitemsdidnotaskhowmanycoursesare taughtorhowmanystudentsbuthowmuchtimeisspenton teaching.Sinceamajorityof thenon-nativeU.S.citizensdonot speakEnglish as afirst language, it is at leastpossible thatmorepreparationtimeisrequiredof them.Anotherpossibility is thatnativeEnglishspeakersaremoreadeptatobtaininggrantsandteachingbuyoutandthattheyprefertobuyoutofundergraduateclasses.

Discussion and Conclusions

Givenourdual interest in theeffectsofURCsonboth the industry andstudent involvementoffacultyaffiliates, theregressionresultsaremost appropriately discussed collectively. To discuss each individu-ally does not exclude necessarily the assumptions and interpretationsfound in prior research and commentaryspecifically those from theacademiccapitalist literature that industry involvementby facultyne-cessitateseitheratradeoffwithstudentinvolvementorsomehowaltersfaculty involvementwith students towardsprofitableendsthatweavoidbyempiricallyaddressingbothindustry-andstudent-relatedout-comesforURCaffiliates.By themselves the regression results for industry involvementmay

givetheimpressionthatURCaffiliatesareindeedcapitalismsslaves,since respondents to our survey affiliated with each URC type havehigher scores on the Industry Involvement Index when compared totheirnon-affiliatedcounterparts.however,evenlimiting interpretationtothissinglemodel,thiscapitalismsslavesconclusionisquitealeap


from the empirical results.The relative effectsof theURCaffiliationonindustryinvolvementwerequitesmallwhencomparedtothescien-tificandengineeringfieldeffectsandalsowhencomparedtotheeffectoftenurestatus.Thus,totheextentthatonemayusethesefindingstocharacterizeURCaffiliatesasbeholdentomarketdemands,somustse-niorfacultyandengineersbecharacterized.TheappropriateconclusiontodrawfromthissinglemodelisthatonaveragealluniversityfacultyareinvolvedwithindustryatsomelevelandthatURCaffiliatesareonlyslightlymoreinvolved,atleastforarepresentativesampleoffacultyinCarnegieresearchextensiveuniversities.Theresultsdonotshowacademicresearchersasnecessarilybeholden

tomarketdemandsat the expenseofuniversities traditional researchandeducationalmissions,whichisthepredominantconcernbehindtheacademiccapitalistcritique.Interpretingtheresultsforstudentinvolve-mentinlightofthoseforindustryinvolvement,weknowthatcenteraf-filiationdoesnotresultinfacultyflightfromstudentsand,importantly,fromteaching.Rather,theregressionfindingsshowthatthesameURCaffiliateswhoaresomewhatmore involvedwith industry thannon-af-filiatedfacultyarealsomore likely tosupportbothundergraduateandgraduate students. While this may be interpreted as students being dealt as tokens of exchange (Slaughter,Campbell, folleman,&Morgan,2002)fromfacultytoindustry,suchaninterpretationoffacultymotivesiscalledintoseriousquestionwhenoneadditionallyconsidersthatURCaffiliatesarealsomoreinvolvedwithteachingundergraduatesperhapstheeducationalactivitywiththeleastopportunityforfacultyprofit.Wedonotwish tooverstateourfindings.While thenightmaresce-

narios,thoseseeingURCsasfundamentallyunderminingthetraditionalmissions of universities, obtain no support from the data presented here, thisisnottosaythatsuchscenariosdonotexist.Oursisanaggregateanalysis,andcasestudiessuggestarangeofpossibleoutcomes.Wedonotwish to imply thatURCs are an unalloyed benefit for education,andwehavenodoubt that therearecircumstanceswhereURCaffili-atesaretorninmultipledirections,resultinginlessattentiontoteach-ing,regardlessoftheteachingload(seeBoardman&Bozeman,2007).ThisvariabilityintheunintendedeffectsofURCs,webelieve,hastodovariabilityatthecenterlevel.OneofthechiefthemesofthispaperhasbeenthatnotallURCtypesarethesameandthatbeingacenteraffiliatemayhavedifferentmeaningdependingonthetypeofURC.WhetherornotourassessmentofthedifferencesbetweenURCsisthecorrectone,wehavemadeclearthatURCsarenothomogenousandmuchremainsto be learnedbydeveloping a deeper knowledgeof their differences.Inparticular, thepervasivenotionthatURCsarecentersforacademic


commerce seems clearly overstated.MostURCs and the faculty per-formingresearchinURCsareorientedtotraditional,publicdomain,re-searchpublications.Academiccapitalismmeritsstudynotbecauseitisthenormbutbecauseofitsexceptionalism.

Notes

ThismaterialisbasedonworksupportedbytheNationalSciencefoundation:Assess-ingRandDProjectsImpactsonScientificandTechnicalhumanCapitalDevelopment(SBR9818229,BarryBozeman,PI)andNSfCAREER:UniversityDeterminantsofWomensAcademicCareerSuccess(REC-0710836,MonicaGaughan,PI).TheauthorsgratefullyacknowledgethesupportoftheNationalSciencefoundation.Anyopinions,findings,conclusions,or recommendationsexpressed in thismaterialare thoseof theauthorsanddonotnecessarilyreflecttheviewsoftheNationalSciencefoundation.WethankRebeccaArce,Min-weiLin,TolgaArslan,andBrancoPonomariovfortheirworkduringdatacollection.TheauthorsaregratefultoMonicaGaughanforhercontributionstothedataandconceptualizationoftheproject.

1Orwereferredtodepartmentalwebpagesiftheuniversitycatalogwasnotavail-able online.

2Generally,theserespondentsreportedaffiliationswithlaboratoriessupportedbyasingleacademicdepartment.Inafewcases,respondentsreportedaffiliationswithquali-fyingURCsbutforperiodsinwhichtherespondentsweregraduatestudentsorpostdoc-toralresearchers(wefocusedonlyonfacultyaffiliates).Inadditiontothesefactors,itis likely(thoughwehavenodirectevidence)thatsomediscrepancyisduetothefactthatcenteraffiliation,aseeminglydiscretebehavior, isnotaclear-cutphenomenon.Whereas reporting that one is, say, an assistant professor, has formal meaning and is basedultimatelyon formaldocuments,centeraffiliation is rarelyso formal.Atmanycenters,thereareseverallayersofinvolvementrangingfromname-onlytoday-to-dayactivityandinteraction.

3Onthoseoccasionswhereitdidnotseemclear,evenaftersomeconsiderableinves-tigationastowhetheraURCshouldbeclassedStateorUniversity,wetelephonedorwroteemailsaskingforclarification.Allthosewithwhomwecommunicated(usuallycenterdirectorsortheiradministrativestaff)responded.

4 See, for example, http://www.nsf.gov/pubs/2009/nsf09545/nsf09545.htm, down-loadedSeptember5,2009.

5See,forexample,theindependentsitesfortheNSfEngineeringResearchCentersprogram,http://erc-assoc.org/,andtheNSfMaterialsResearchScienceandEngineer-ingCenters,http://www.mrsec.org/.

6Wefeel thisset isagoodapproximationofourURCsamplingpopulation.how-ever,thereisoneartifactofthesamplingpopulationthatrendersitanimperfectindica-toroftheentirepopulationofURCs.Sincewedidnotincludemedicalfacultyandmed-icalschoolsinourSTEMsample,wehaveunderestimatedNIhcenters.however,theNIhbeganitscentersprogramsrelativelylatecomparedtomanyofthefederalagenciesrepresentedinTable1,andthusitislikelythattheunderrepresentationisamodestone.Today,NIhhas27intramural centersandinstitutes,andtheseremainastrongerfocusthanuniversity-basedcenters(NationalInstitutesofhealth,2009).ExaminingtheP50classification center awards fromNIh leads us to believe thatNIh represents about34%ofallURCs.


7Among the1714 respondents in the sample, 23 reported theywerenot involvedinpursuinggrantsandcontracts.Theseindividualsaretreatedasmissingdataforthisvariable.

8Thewording, though,of thegrantsandcollaborationmotivation itemsdiscussedabovefortheStudentInvolvementIndicesstronglyimplythatstudentsarethemotiva-tiontopursuegrantsandcollaborators,nottheotherwayaround.

References

AgrawalA.,&henderson,R. (2002).Puttingpatents incontext:Exploringknowledgetransfer from MIT. Management Science, 48,4460.

Azoulay,P.,Ding,W.,&Stuart,T.(2006).The impact of academic patenting on the rate, quality, and direction of (public) research output (WorkingPaper #11917).NationalBureauofEconomicResearch,Cambridge,MA.

Baldini,N. (2008).Negative effects of university patenting:Myths and grounded evi-dence.Scientometrics, 75(2), 289311.

Barney,J.(1991).firmresourcesandsustainedcompetitiveadvantage.Journal of Man-agement, 17(1),99120.

Boardman,C.(2009).Governmentcentralityinuniversity-industryinteractions:Univer-sityresearchcentersandtheindustryinvolvementofacademicresearchers.Research Policy, 38,15051516.

Boardman,C.,&Bozeman,B.(2007).Rolestraininuniversityresearchcenters.The Jour-nal of Higher Education, 78(4),430463.

Boardman,C.,&Corley,E.(2008).Universityresearchcentersandthecompositionofresearchcollaborations.Research Policy, 37,900913.

Boardman,C.,&Gray,D. (2010).Thenewscienceandengineeringmanagement:Co-operative research centers as government policies, industry strategies, and organiza-tions. The Journal of Technology Transfer, 35(5),445459.

Boardman,C.,&Ponomariov,B.(2007).RewardsystemsandNSfuniversityresearchcenters:Theimpactoftenureonuniversityscientistsvaluationofappliedandcommer-cially-relevantresearch.The Journal of Higher Education, 78(1),5170.

Bouty,I.(2000).InterpersonalandinteractioninfluencesoninformalresourceexchangesbetweenR&Dresearchersacrossorganizationalboundaries.Academy of Management Journal, 43(1),5065.

Bowen,h.R.,&Schuster,J.h.(1986).American professors: A national resource imper-iled. Newyork:OxfordUniversityPress.

Boyer,E.L.(1990).Scholarship reconsidered: Priorities of the professoriate.Princeton,NJ:CarnegiefoundationfortheAdvancementofTeaching.

Bozeman,B.,&Boardman,C.(2003).Managingthenewmultipurpose,multidisciplineuniversityresearchcenter:Institutionalinnovationintheacademiccommunity.Wash-ington,D.C.:IBMEndowmentfortheBusinessofGovernment.

Bozeman,B.,&Boardman,C. (2004).TheNSf engineering research centers and theuniversity-industryresearchrevolution.The Journal of Technology Transfer, 29(34),365375.


Bozeman,B.,Dietz,J.,&Gaughan,M.(2001).Scientificandtechnicalhumancapital:Analternativemodelforresearchevaluation.International Journal of Technology Manage-ment, 22(78),636655.

Bozeman,B.,&Gaughan,M. (2007). Impactsofgrantsandcontractsonacademic re-searchersinteractionswithindustry.Research Policy, 36,694707.

Bozeman,B.,&Gaughan,M. (2011). Job satisfaction among university faculty: Indi-vidual,work,andinstitutionaldeterminants.The Journal of Higher Education, 82(2), 154186.

Braxton,J.M.(1996).Publictrust,researchactivity,andtheidealofservicetostudentsasclientsofteaching.InJ.B.Braxton,(Ed.),Faculty teaching and research: Is there a conflict? (pp.122). NewDirectionsofInstitutionalResearch,No.90.Sanfrancisco:Jossey-Bass.

BreschiS.,Lissoni,f.,&Montobbio,f.(2005).frompublishingtopatenting:Doproduc-tive scientists turn into academic inventors?Revue dEconomie Industrielle, 110(2), 75102.

Bush,v.(1945).Science: The endless frontier. WashingtonDC:UnitedStatesGovernmentPrintingOffice.

Cage,M.C. (1991).Statesquestioninghowmuch timeprofessorsspendworkingwithundergraduate students. Chronicle of Higher Education, 37,120.

Carnegie foundation. (2000).Classification of higher education. http://www.carnegief-oundtion.org/Classification/index.htm.

Clark, B.R. (1998).Creating entrepreneurial universities: organisational pathways of transformation.Oxford:Pergamon/ElsevierScience.

Coburn,C.(1995).Partnerships.Columbus,Ohio:BattellePress.Corley,E.,Boardman,C.,&Bozeman,B.(2006).Designandthemanagementofmulti-institutionalresearchcollaborations.Research Policy, 35,975993.

Corley,E.,&Gaughan,M.(2005).Scientistsparticipationinuniversityresearchcenters:Whatarethegenderdifferences?Journal of Technology Transfer, 30(4),371381.

Diamond,R.M.(1993).Institutingchangeinthefacultyrewardsystem.New Directions for Higher Education, 81, 1322.

Diamond,R.M.(1999).Aligning faculty rewards with institutional mission: Statements, policies, and guidelines.Boston:Anker.

Dresser,P.(1989).Research centers directory (13th ed.). Detroit:GaleResearch.Etzkowitz,h.(1998).Thenormsofentrepreneurialscience:Cognitiveeffectsofthenewuniversity-industrylinkages.Research Policy, 27(2),109123.

feldman,k.A.(1987).Researchproductivityandscholarlyaccomplishmentofcollegeteachersas related to their instructionaleffectiveness:A reviewandexploration.Re-search in Higher Education, 26(3),227298.

Gaughan,M.,&Corley,E.A.(2010).SciencefacultyatU.S.researchuniversities:Theimpactsofuniversityresearchcenteraffiliationandgenderonindustrialactivities.Tech-novation, 30,215222.

Grant,R.M.(1996).Prosperingindynamically-competitiveenvironments:Organization-alcapabilityasknowledgeintegration.Organization Science, 7(4),375387.


Green,S.G. (1983). Identifiabilityof spurious factorsusing linear factoranalysiswithbinary items. Applied Psychological Measurement, 7(2),139147.

hackett,E. J. (2001).Scienceasavocation in the1990s:Thechangingorganizationalculturesofacademicscience.InJ.Croissant&S.Restivo(Eds.),Degrees of compro-mise: Industrial interests and academic values (pp.101138).Albany:StateUniversityofNewyorkPress.

hagedoorn, J.,Link,A.N.,&vonortas,N.S. (2000).Researchpartnerships.Research Policy, 29,567586.

henderson,R.,Jaffe,A.B.,&Trajtenberg,M.(1998).Universitiesasasourceofcom-mercial technology:A detailed analysis of university patenting, 196588.Review of Economics and Statistics, 80,119132.

hicks,D.,&hamilton,k.(1999).Doesuniversity-industrycollaborationadverselyaffectuniversityresearch?Issues in Science and Technology, 7475.

hinds,P.,&kiesler,S.(1995).Communicationacrossboundaries:Work,structure,anduseofcommunicationtechnologiesinalargeorganization.Organization Science, 6(4), 373391.

Iansiti,M.(1995).Technologyintegration:Managingtechnologicalevolutioninacom-plexenvironment.Research Policy, 24(4),521542.

krimsky,S.(2003). Science in the private interest: Has the lure of profits corrupted bio-medical research?Lanham,Maryland:RowmanandLittlefieldPublishers.

Lee,S.,&Bozeman,B.(2005).Theimpactofresearchcollaborationonscientificproduc-tivity. Social Studies of Science, 35(5),673702.

Lee,y.(1996).Technologytransferandtheresearchuniversity:Asearchforthebound-ariesofuniversity-industrycollaboration.Research Policy, 25(6),843863.

Lin,M.,&Bozeman, B. (2006). Researchers industry experience and productivity inuniversity-industryresearchcenters:Ascientificandtechnicalhumancapitalexpla-nation. The Journal of Technology Transfer, 31(2),269290.

Link,A.,&Scott,J.T.(2005).Openingtheivorytowersdoor:Ananalysisofthedeter-minantsoftheformationofU.S.universityspin-offcompanies.Research Policy, 34(7), 11061112.

Link,A.N.,Swann,C.,&Bozeman,B.(2008).Atimeallocationstudyofuniversityfac-ulty. Economics of Education Review, 27(4),363374.

Massy,W.,&Zemsky,R.(1994).facultydiscretionarytime:Departmentsandtheaca-demicratchet.The Journal of Higher Education, 65, 122.

Meyer,M.(2006).Arepatentingscientiststhebetterscholars?Anexploratorycomparisonofinventorauthorswiththeirnon-inventingpeersinnanoscienceandtechnology.Re-search Policy, 35,16461662.

Milem,J.f.,Berger,J.B.,&andDey,E.L.(2000).facultytimeallocation:Astudyofchangeovertwentyyears.The Journal of Higher Education, 71(4),454475.

Mowery,D.C.,Nelson,R.R.,Sampat,B.,&Ziedonis,A.(2001).ThegrowthofpatentingandlicensingbyU.S.universities:anassessmentoftheeffectsoftheBayhDoleactof1980.Research Policy, 30(1),99119.

Mowery,D.C.,&Ziedonis,A.(2002).AcademicpatentqualityandquantitybeforeandaftertheBayhDoleactintheUnitedStates.Research Policy, 31(3),399418.


NationalInstitutesofhealth.(2009).OfficeoftheDirector(Description).Retrievedfromhttp://www.nih.gov/icd/od/index.htm.

Ponomariov,B.(2008).Effectsofuniversitycharacteristicsonscientistsinteractionswiththe private sector:An exploratory assessment. The Journal of Technology Transfer, 33(5),485503.

Ponomariov,B.(2009).Studentcentralityinuniversity-industryinteractions.Industry and Higher Education, 23,5062.

Ponomariov,B.,&Boardman,C.(2010).Influencingscientistscollaborationandproduc-tivitypatternsthroughnewinstitutions:Universityresearchcentersandscientificandtechnicalhumancapital.Research Policy, 39(5),613624.

Rogers,E.M.,Takegami,S.,&yin,J.(2001).Lessonslearnedabouttechnologytransfer.Technovation, 21(4),253261.

Ruscio,k.P.(1984).Thechangingcontextofacademicscience:University-industryrela-tionsinbiotechnologyandthepublicpolicyimplications.Policy Studies Review, 4(2), 259275.

Slaughter,S.,Campbell,T.,folleman,M.h.,&Morgan,E.(2002)Thetrafficingradu-atestudents:Graduatestudentsastokensofexchangebetweenacademeandindustry.Science, Technology and Human Values, 27(2),282313.

Slaughter,S.&Leslie,L.L.(1997).Academic Capitalism: Politics, Policies, and the En-trepreneurial University.Baltimore:JohnshopkinsUniversityPress.

Slaughter,S.,&Rhoades,G.(1996).Theemergenceofacompetitivenessresearchanddevelopmentpolicycoalitionandthecommercializationofacademicscienceandtech-nology. Science, Technology and Human Values, 21,303339.

Slaughter,S.,&Rhoades,G.(2004).Academic capitalism and the new economy. Balti-more: JohnshopkinsUniversityPress.

Stokes,D.(1997).Pasteurs quadrant.Washington,D.C.:BrookingsInstitutionPress.Thursby, J.G.,&Thursby,M.C. (2002).Who is selling the ivory tower? Sources ofgrowthinuniversitylicensing.Management Science, 48,90104.

vanLooy,B.,Callaert,J.,&Debackere,k.(2006).Publicationandpatentbehaviourofacademicresearchers:Conflicting,reinforcingormerelyco-existing?Research Policy, 35,596608.

vanLooy,B.,Ranga,M.,Callaert,J.,Debackere,k.,&Zimmermann,E.(2004).Combin-ingentrepreneurialandscientificperformanceinacademia:TowardsacompoundedandreciprocalMatthew-effect?Research Policy, 33,425441.

vanRijnsoever,f.J.,hessels,L.k.,&vandeberg,R.L.J.(2008).Aresource-basedviewontheinteractionsofuniversityresearchers. Research Policy, 37,12551266.

Welsh,R.,Glenn,L.,Lacy,W.,&Biscotti,D.(2008).Closeenoughbutnottoofar:As-sessingtheeffectsofuniversity-industryresearchrelationshipsandtheriseofacademiccapitalism.Research Policy, 37,12551266.

Winship,C.,&Radbill,L.(1994).Samplingweightsandregressionanalysis.Sociological Methods & Research, 23,230257.

youtie,J.,Libaers,D.,&Bozeman,B.(2006).Institutionalizationofuniversityresearchcenters:ThecaseoftheNationalCooperativePrograminInfertilityResearch.Techno-vation, 26(9),10551063.

Copyright of Journal of Higher Education is the property of Ohio State University Press and its content may notbe copied or emailed to multiple sites or posted to a listserv without the copyright holder's express writtenpermission. However, users may print, download, or email articles for individual use.

CAPITALISMO CIENTIFICO ACADEMICO

Documents

Transcript of CAPITALISMO CIENTIFICO ACADEMICO