Academic Engagement and Commercialisation

8/12/2019 Academic Engagement and Commercialisation

1/20

Research Policy 42 (2013) 423442

Contents lists available at SciVerse ScienceDirect

Research Policy

j ou rna l ho me pag e : www.e l sev i e r. com/ loca t e / r e spo l

Academic engagement and commercialisation: A review of the literature onuniversityindustry relations

Markus Perkmann a , , Valentina Tartari k , Maureen McKelvey b , Erkko Autio a , Anders Brostrm c ,Pablo DEste d , Riccardo Fini f , Aldo Geuna e , l , Rosa Grimaldi f , Alan Hughes m , Stefan Krabel h ,Michael Kitson g , Patrick Llerena i , Franceso Lissoni j, Ammon Salter a , Maurizio Sobrero f a Imperial College London, Business School, South Kensington Campus, London SW7 2AZ, UK b University of Gothenburg, School of Business, Law and Economics, Institute of Innovation and Entrepreneurship, P.O Box625, SE-405 30 Gothenburg Swedenc CESIS, Royal Institute of Technology, Drottning Kristinas vg 30, SE-100 44 Stockholm, Swedend INGENIO (CSIC-UPV), Polytechnic University of Valencia, Ciudad Politcnica de la Innovacin, Edif. 8E 4, 46022 Valencia, Spaine University of Torino, Department of Economics S. Cognetti de Martiis, Via Po 53, 10124 Torino, Italyf University of Bologna, Department of Management, Via Capo di Lucca, 34, 40126 Bologna, Italyg Judge Business School and UK Innovation Research Centre (UK IRC), University of Cambridge, Trumpington Street, Cambridge CB2 1AG, UK h University of Kassel, Institute of Economics, Economic Policy Research, Nora-Platiel-Str. 5, 34109 Kassel, Germanyi BETA University of Strasbourg, 61 avenue de la Fort Noire, 67085 Strasbourg, France j Gretha-Universit Bordeaux IV, av. Lon Duguit, 33608 Pessac, Francek Copenhagen Business School, Department of Innovation and Organizational Economics, Kilevej 14A, 2000 Frederiksberg, Denmarkl BRICK, Collegio Carlo Alberto, Via Real Collegio, 30, 10024 Moncalieri(Torino), Italym Centre for Business Research (CBR) and UK Innovation Research Centre (UK IRC), Judge Business School, University of Cambridge, Cambridge CB2 1AG, UK

a

r

t

i

c

l

e

i

n

f

o

Article history:Received 2 March 2010Received in revised form 22 June 2012Accepted 18 September 2012

Available online 23 October 2012

Keywords:Universityindustry relationsTechnology transferAcademic entrepreneurshipCommercialisationCollaborative researchAcademic consulting

a

b

s

t

r

a

c

t

A considerable body of work highlights the relevance of collaborative research, contract research, con-sulting and informal relationships for universityindustry knowledge transfer. We present a systematicreview of research on academic scientists involvement in these activities to which we refer as aca-demic engagement. Apart from extracting ndings that are generalisable across studies, we ask howacademic engagement differs from commercialisation, dened as intellectual property creation and aca-demic entrepreneurship. We identify the individual, organisational and institutional antecedents andconsequences of academic engagement, and then compare these ndings with the antecedents andconse-quences of commercialisation. Apart from being more widely practiced, academic engagement is distinctfrom commercialisation in that it is closely aligned with traditional academic research activities, and pur-sued by academics to access resources supporting their research agendas. We conclude by identifyingfuture research needs, opportunities for methodological improvement and policy interventions.

2012 Elsevier B.V. All r ights reserved.

1. Introduction

Universities are organisations that perform a key role withincontemporary societies by educating large proportions of the pop-ulation and generating knowledge. Recently, oftenon the initiativeof policy-makers, many universities have taken action to developa third mission by fostering links with knowledge users and

Corresponding author.E-mail addresses: [email protected] (M. Perkmann), [email protected]

(V. Tartari), [email protected] (M. McKelvey),[email protected] (E. Autio), [email protected](A. Brostrm), [email protected] (P. DEste), [email protected](R. Fini), [email protected] (A. Geuna), [email protected](R. Grimaldi), [email protected] (A. Hughes), [email protected](S. Krabel), [email protected] (M. Kitson), [email protected](P. Llerena), [email protected] (F. Lissoni), [email protected] (A. Salter),[email protected] (M. Sobrero).

facilitating technology transfer ( Etzkowitz et al.,2000b;Florida andCohen, 1999; Gulbrandsen and Sliperster, 2007 ).

Amongst the various channels available for establishing theselinks, the commercialisation of academic knowledge, involvingthe patenting and licensing of inventions as well as academicentrepreneurship, has attracted major attention both within theacademic literature and the policy community ( OShea et al., 2008;Phan and Siegel, 2006; Rothaermel et al., 2007 ). Commercialisa-tion is considered a prime example forgenerating academic impactbecause it constitutes immediate, measurable market acceptanceforoutputsof academicresearch( Markman et al., 2008 ). To supportcommercialisation, many universities have established specialisedstructures, suchas technology transfer ofces (TTOs), science parksandincubators ( Clarysse etal., 2005;Siegel et al., 2003 ), andcreatedsupportive internal rules and procedures ( Thursby et al., 2001 ).

Whilst commercialisation clearly represents an important wayfor academic research to contribute to economy and society,

0048-7333/$ seefrontmatter 2012 Elsevier B.V. All rights reserved.

http://dx.doi.org/10.1016/j.respol.2012.09.007
http://localhost/var/www/apps/conversion/tmp/scratch_7/dx.doi.org/10.1016/j.respol.2012.09.007http://localhost/var/www/apps/conversion/tmp/scratch_7/dx.doi.org/10.1016/j.respol.2012.09.007http://www.sciencedirect.com/science/journal/00487333http://www.elsevier.com/locate/respolmailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]://localhost/var/www/apps/conversion/tmp/scratch_7/dx.doi.org/10.1016/j.respol.2012.09.007http://localhost/var/www/apps/conversion/tmp/scratch_7/dx.doi.org/10.1016/j.respol.2012.09.007mailto:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]:[email protected]://www.elsevier.com/locate/respolhttp://www.sciencedirect.com/science/journal/00487333http://localhost/var/www/apps/conversion/tmp/scratch_7/dx.doi.org/10.1016/j.respol.2012.09.007


2/20

424 M. Perkmann et al. / Research Policy 42 (2013) 423442

there are multiple other ways in which university research istransferred ( Salter and Martin, 2001 ). In this paper, we focus onacademic engagement which we dene as knowledge-relatedcollaboration by academic researchers with non-academic orga-nisations. These interactions include formal activities such ascollaborative research, contract research, and consulting, as wellas informal activities like providing ad hoc advice and networkingwith practitioners ( Abreu et al., 2009; Bonaccorsi and Piccaluga,1994; DEste and Patel, 2007; Meyer-Krahmer and Schmoch, 1998;Perkmann and Walsh, 2008 ). Academic engagement is also some-times referred to as informal technology transfer ( Link et al., 2007 ),even thoughmost of these interactions tend to be formalised usingcontracts.

Academic engagement represents an important way in whichacademic knowledge is transferred into the industrial domain;many companies consider it signicantly more valuable thanlicensing university patents ( Cohen et al., 2002 ). Universitiesincome from academic engagement is usually a high multiple of the income derived from intellectual property ( Perkmann et al.,2011 ). It should be added that academic engagement is not a newphenomenon but has a long tradition, particularly at universitiesthat emphasise practical and technical relevance as part of theirmission, such as the US land grant universities who seek to providepractical educationwhilstassisting local rmsand agriculturalcon-texts ( Mowery and Nelson, 2004 ). Perhaps mirroring the recentpolicy and research interest in commercialisation, however, therehas been a surge in research published on this topic, yet the stateof knowledge remains relatively fragmented and tentative. In addi-tion, there havebeen fewefforts to underpinacademicengagementconceptually,which stands in contrast to commercialisationwhereentrepreneurship theory has been applied.

We address these gaps by presenting a systematic review of theliterature on academic engagement. The research question guidingour review is: What are the antecedents and consequences of aca-demic engagement? We will consolidate results from all existingstudies andextractgenerally applicable results. Ina further step,wecompare our ndings with what is known about the antecedentsand consequences of commercialisation, i.e. intellectual propertycreation and academic entrepreneurship ( Rothaermel et al., 2007 ).Thisanalysis allows us to address whether academicengagement isdriven by the same mechanisms as commercialisation, or whetherit represents a conceptually different type of phenomenon thatneeds to be treated separately by researchers and policy-makers.

Our work adds to existing research in four important ways. Weprovide the rst systematic review of academic engagement andcompare the latter with commercialisation. We paint a compre-hensive picture of the antecedents and consequences of academicengagement across various contexts. Our approach allows us toseparate factors and boundaryconditions thatmay be idiosyncraticand the patterns that apply to the phenomenon more generally.We also identify aspects that are either less well researched or

contested, providing direction for future research.Second, we synthesise our empirical results into a novel the-

oretical framework on academic engagement. We outline boththe differences and overlaps between academic engagement andacademics involvement in commercialisation and thereby hopeto facilitate a convergence between these two streams of theliterature.

Third, we make a methodological contribution, by discussingwhy studying academic engagement requires methodologicalapproaches that differ from those for studying commercialisation.We also identify the challenges posed by these approaches andsuggest how they may be overcome.

Fourth, our results are policy-relevant. In the last 30 years, uni-versities have experienced major changes that have affected their

objectives, sources of funding and modes of operation ( Geuna,

2001; McKelvey and Holmn, 2009 ). There have been importantmodicationsin universitiespolicyenvironmentsdue to initiativessuchas the BayhDole Act of 1980 in the US, and the abandonmentof the professors privilege in most European countries ( Baldiniet al., 2006; Grimaldi et al., 2011; Lissoni et al., 2008; Moweryet al., 2001 ). For policy-makers, it is important to know whetheracademic engagement is driven by similar mechanisms to com-mercialisation, or affected by factors that may not be activated byentrepreneurial incentives.

2. Conceptual background

Here we clarify the concept of academic engagement, andits relationship to the concept of commercialisation. Academicengagement is characterised by the following aspects whichrefer to organisation and objectives, respectively. First, aca-demic engagement represents inter-organisational collaborationinstances, usually involving person-to-person interactions( Cohenet al., 2002 ), that link universities and other organisations,notably rms ( Bonaccorsi and Piccaluga, 1994; Meyer-Krahmerand Schmoch, 1998; Schartinger et al., 2002 ). The quid-pro-quoagreed amongst the partners may be purely nancial, i.e. the aca-demic may work for a fee, ormay consist of non-nancial benetssuch as access to materials or data for academic research projectsor ideational input ( Manseld, 1995; Perkmann and Walsh, 2009;Senker, 1995 ). Second, generally the partners pursue goals that arebroader than the narrow connes of conducting research for thesake of academicpublishing, andseekto generate some kind of util-ity for the non-academic partners. For instance, the academic mayofferhis/herexpertise toprovidenew ideason application-orientedissues, solve problems and suggest solutions to collaborating orga-nisations.

How does academic engagement relate to commercialisation?First, in terms of organisation, while academic engagement rep-resents collaboration, commercialisation or technology transfer may occur via academic entrepreneurship, that is the foundingof a rm with the objective to commercially exploit a patentedinvention, or in some cases, a body of unpatented expertise ( Shane,2004 ). Alternatively, a patented or otherwise protected inventionmay be licensed out against the contracted receipt of royalties( Jensen and Thursby, 2001 ). For both processes, patenting repre-sents a preliminary step, indicating a disposition on the part of theacademic towards some kindof exploitation. Second, commerciali-sation means an academic invention is exploitedwiththe objectiveto reap nancial rewards; by contrast, academic engagement isbroader and is pursued for varying objectives.

Despite these differences, there are important links andoverlapsbetweenbothtypesof activity. Infact, commercialisationis oftenanoutcome or follow-onactivity, whether intended or unintended, of academic engagement. Working on common projects with indus-try may provide academics with insights into what ideas may be

commercially valuable, and hence the opportunity to develop orco-develop inventions that can be patented, licensed or enable anacademicspin-off. Inother words, academicengagementoftenpre-cedes commercialisation in time and can hence be regarded as aninput factor to the latter. It some cases, it may also accompanycommercialisation, for instancewhenspin-offcompanieswork col-laboratively with the university labs they originated from ( Meyer,2003 ).

Both academic engagement and commercialisation tend to beindividually driven and pursued on a discretionary basis. Uni-versities are professional bureaucracies ( Mintzberg, 1979 ) thatrely on the independent initiative of autonomous, highly skilledprofessionals to reach their organisational goals. While academicentrepreneurship as well as patenting as an often used proxy

for entrepreneurial behaviour are also primarily individual


3/20

M. Perkmann et al. / Research Policy 42 (2013) 423442 425

behaviours, licensing can be carried out by the university with-out the active participation of the academic inventor even thoughsuch participation enhances the probability of commercial success(Agrawal, 2006 ).

3. Methodology

We performed a systematic review of the available evidenceon academic engagement. Such a literature review establishes thestate of current knowledge in a eld ( Traneldet al., 2003 ). In med-ical elds, such reviews synthesise empirical evidence from largenumbers of studies and identify areas of consensus and disagree-ment between researchers within certain areas of research. Thisprinciple can be fruitfully adopted in the social sciences althoughthe overall number of existing studies tends to be smaller, meas-ures and variables are less tightly dened and conclusions may bemore contestable. For the current article, we followed a simpliedversion of the process outlined by Traneld et al. (2003) which wedetail below.

Our objective was to establish what is known about (a) theextent and types of academic engagement, (b) its determinants,and (c) its impact on academics, universities and other stakehol-ders. We focused our analysis on individual researchers becausethe decision to engage is a decision that, in the university context,is primarily taken on an individual level.

We applied the following procedure. We rst identied all therelevant research published on this topic from 1980 to 2011. Weconducted an extensive search in the titles and abstracts of pub-lished, peer-reviewed articles held by the bibliographical databaseservice EBSCO (EconLit included), using a series of keywords thatthe research team had identied (Appendix A). Subsequently, weperformed a manual search of the journals with the highest articlecounts over the past 22 years (19892011): Research Policy , Journalof Technology Transfer, and Technovation . This procedure allowedus to exclude possible bias towards newer studies and also to vali-date the search terms, given that there is little consensus on thekeywords used for classifying articles on academic engagement.

The above procedure yielded 413 results. We rst ltered thislist according to t. As we were particularly interested in studiesusing data on individual researchers, we removed all articles thatdidnot full this criterion. Forexample,we discarded studies at theaggregate levelof department, university or country,case studies of specicuniversities thatreferredonlyto theorganisational contextand historical analyses. We also excluded surveys of researcherswho had left academia, and articles dealing solely with commer-cialisation rather thanengagement. This procedure eliminated 280articles.We subsequently screenedtheremaining articles,applyingbasic qualitycriteriato ascertainwhetherdatahad beencollectedina systematic way and whether papers proposed intelligible results.At this stage we also eliminated those articles that had a strongpractitioner focus but containedlittle tangible data. This procedure

left us with a total of 36 articles.At this stage, we read and synthesised each remaining arti-

cle, compiling the following information in tabular form ( Pawson,2006 ): research questions, data used, methodology, variables, andndings (Appendix B). We synthesised this information into a pre-liminary report and held a full-day workshop, with all the authorspresent, to debate the ndings and implications of our discoveriesand used the results from the workshop to revise the report.

The majority of articles we considered were published in theperiod from 2006 onwards ( Fig. 1). In their review, which covereda 25-year periodup to2005, Rothaermel et al. (2007) , notedthat theliterature on university entrepreneurship neglected the analysisof individual researchers involvement in the process. Indeed, thenumber of articles addressing particularly academic engagement

has increased signicantly since 2005, strengthening the case for

Fig. 1. Articles published peryear. For2011, thegraph shows thenumber of publi-cations forJanuary to March only.

carrying out a fresh literature review ( Perkmann and Walsh, 2007;Phan and Siegel, 2006; Rothaermel et al., 2007 ). Regardingthe jour-nals in which these articles are published we observe a skeweddistribution ( Table 1 ): two journals ( Research Policy and Journal of Technology Transfer ) account for 63% of all output. Other outletsinclude medical and education journals while articles in generalmanagement, economics and sociology journals are relativelyrare.Whilst these latter journals emphasise theory building and theorytesting, research on academic engagement has produced predom-inantly phenomenon-focused studies.

As a nal step in our analysis, we systematically comparedthe results obtained on academic engagement with what isknown about academics involvement in commercialisation, i.e.academic entrepreneurship and IP-related activities. Unlike aca-demic engagement, research on commercialisation has previouslybeen systematically documented and synthesised by publishedreviews ( Geuna and Muscio, 2009; Larsen, 2011; Phan et al., 2005;Rothaermel et al.,2007 ) andwe therefore use that literature forourcomparison.

4. FindingsIn this section, we present the ndings from the system-

atic review. Each subsection details a specic aspect of academicengagement and we also report how this compares with commer-cialisation.

Synthetic results are presented in Table 2 .

4.1. Extent of academic engagement

Academic engagement is varied and includes collaborativeresearch, contract research, consulting and other forms of

Table 1Synthesis of articles.

Number of articles

Research Policy 13The Journal of Technology Transfer 10Technovation 3The Journal of Higher Education 2Others 8

Quantitative data 33Qualitative data 3

US 18UK 5Other Europe 11Asia 1Other countries 1

Sum 36

Breakdown of articles according to journal, type of data, and empirical focus.


4/20


Table 2Comparison between academic engagement and commercialisation.

Variable Engagement Commercialisation

Individual determinantsMale + +Age o oSeniority + oPrevious commercialisation experience o +Grants awarded (government) + o

Contracts awarded (industry) + oScientic productivity + +

Organisational determinantsQuality university/department +Organisational support o +Incentive system o oOrganisational c ommercialisation e xperience o +Peer effects o +

Institutional determinantsApplied discipline + +Life-science/biotech o +Country-specic r egulations/policy o +

ImpactScientic productivity o +Commercial productivity o n/aShift towards applied research o oIncreased secrecy o +Collaborative behaviour + +Teaching o o

Notes:The table reports the effects of independent variables (vertical) on outcomevariables (individual-levelacademic engagementand commercialisation).Commer-cialisation includes academic entrepreneurship and IP-based technology transfer.Key: (+) Positive effect in at least some studies. ( ) Negative effect in at least somestudies. (o) ambiguous effect/not enough empirical evidence. (n/a) not applicable.

knowledge exchange ( DEste and Patel, 2007; Louis et al., 1989;Nilsson et al., 2010; Schartinger et al., 2002 ). In Table 3 , we reportthe proportion of academics engaged in diverse types of activitiesin some of the studies reviewed.

For instance, almost half of UK investigators in the physical andengineering sciences engaged in collaborative research, contractresearch or consulting at least once over a two-year period while12% and 22% engaged in academic entrepreneurship and patenting

respectively ( DEste andPerkmann,2011 ). Accordingto an older USstudy, the supplemental income of academic life scientists gainedfrom consulting amounted to around 10% of their salaries, withfewer than 5% of respondents reporting consulting income of over40% of their salaries ( Louis et al., 1989 ). Median research incomefrom industry grants as a proportion of total grants of respondentswasapproximately8% ( Louis etal.,1989 ). According to theResearchValue Mapping Survey, 17% of academics at US research-intensiveuniversities obtained an industry grant in the 12 months preced-ing the survey while 18% consulted with a rm ( Bozeman andGaughan, 2007 ). Similarly, in Germany 20% of academics published jointly with industrial partners and 17% served as formally paidconsultants in the 12 months preceding the survey ( Grimpe andFier, 2010 ). A similar gure for consulting is obtained by Haeusslerand Colyvas (2011) f or German and UK life scientists. Accordingto a comparative study of science, engineering and medicine fac-ulty in Sweden and Ireland, 5168% of respondents were involvedin soft collaboration activities, such as consulting, while 1219%engaged in spin-off creation at least once during their academiccareer ( Klofsten and Jones-Evans, 2000 ).

Given these differences in participation rates, it appearsworth investigating the determinants of academic engagementbehaviours and what consequences these behaviours have for thetraditional activities of research and teaching.

Unlike academic engagement, fewer academics are involved incommercialisation. Lissoni et al.s (2009) report for three Europeancountriesshowed thatin the totalpopulation of academics, approx.45% of individuals had led a patent. A roughly equivalent gurefor the US is provided by the Research Value Program, with about5% (Bozeman and Gaughan, 2007 ). In the studies we reviewed theproportion of academics involved in patenting ranges from 5% to40%,due todifferentsampling strategies, illustratinghow patentingrates vary strongly with scientic discipline, as well as universityculture ( Agrawal and Henderson, 2002 ). The proportion of aca-demics taking part in a commercial enterprise in the studies wereviewed was generally below 10%; here we see, however, equallylarge differences for specic samples, such as certain departmentsat Stanford ( Kenneyand Goe, 2004 ). Despite these outliers,it canbeestablished that the proportion of academics involved in academic

Table 3External engagement: comparison across different studies.

Population Time frame analysed Collaborativeresearch

Consu lting Sponsoredresearch

Contractresearch

Patenting Academicentrepreneur-ship

Klofsten and Jones-Evans(2000)

Academics inSweden

Entire career 51% 44% 45% 12% 12%

Klofsten and Jones-Evans(2000)

Academics inIreland

Entire career 68% 68% 69% 26% 19%

Gulbrandsenand Smeby(2005)

Tenureduniversityprofessors inNorway

5 years 21% 31% 21% 7% 7%

Bozeman andGaughan(2007)

Academic at USresearcheruniversities

12 months 17% 18% 5% 3%

DEste andPerkmann(2011)

UK Physical &EngineeringSciencesInvestigators

2 years 44% 38% 47% 22% 12%

Grimpe andFier (2010)

Academics inGermany

12 months 20% (jointpublications)

17%

Haeussler andColyvas,2011

Life scientistsin Germanyand UK

12 months 20% 40% 9%

The gures indicate the percentage of academics involved in the specied activities unless otherwise indicated, according to different studies. Figures on patenting and

academic entrepreneurship are included for comparison.


5/20


engagement is generally a multiple of individuals involved in com-mercialisation. This means a signicant proportion of academicspursue academic engagement, without conducting commerciali-sation.

4.2. Antecedents of external engagement

4.2.1. Individual characteristicsIndividual characteristics play an important role in predict-

ing academic engagement. Male academics are signicantly morelikely to engage with industry ( Azagra-Caro, 2007; Boardman,2008; Giuliani et al., 2010; Goktepe-Hulten, 2010; Link et al.,2007 ). Age has an ambiguous effect, even when controlling forseniority. Some studies nd a positive relationship ( Boardmanand Ponomariov, 2009; Haeussler and Colyvas, 2011; Link et al.,2007 ), whileothers nd a negative relationship ( Bekkers and BodasFreitas, 2008; DEste and Patel, 2007; DEste and Perkmann, 2011;Giuliani et al., 2010 ) or no relationship at all ( Boardman andPonomariov, 2009; Gulbrandsen and Smeby, 2005; Renault, 2006 ).The negative impact of age found by some studies may reect atraining effect: individuals who were trained in earlier periodswhen universities engagement with industry was less relevantor even discouraged, may be attached to norms not compatiblewithcollaborationwithprivatecompanies ( Bercovitzand Feldman,2008 ).

By contrast, seniority is often positively related to collaboration(Boardman,2008, 2009; Boardman and Corley, 2008; Bozeman andGaughan,2007;DEsteand Perkmann,2011;Haeusslerand Colyvas,2011; Link et al., 2007; Ponomariov, 2008 ). Given that engagementis oftenseeded by personal contacts, moreexperienced researchersare likely to have larger networks, and hence more social capi-tal, enabling them to nd potential partners in the private sector(Giuliani et al., 2010; Haeussler and Colyvas, 2011; Landry et al.,2006 ). Such network effects are reinforced by routine interactionwith industry partners. Previous experience with industrial col-laborators positively affects the attitudes of academics towardsindustry ( Van Dierdonck et al., 1990 ) and also their collaborativebehaviour ( DEste and Patel, 2007 ). Thesendingsare supportedbythe observation that previous experience with commercialisation,patenting or venture creationincreases the likelihood of academicsparticipation in collaborative activities ( Bekkers and Bodas Freitas,2008 ).

A further individual characteristic predicting academic engage-ment is related to scientists quality and success. Academicsscienticproductivity is generallypositively related to engagement(Bekkers and Bodas Freitas, 2008; Gulbrandsen and Smeby, 2005;Haeussler and Colyvas, 2011; Louis et al., 1989 ). In other words,the best and most successful scientists are also those who engagemost with industrial partners. In addition, individuals ability tomobilise resources for their research is also positively linked tocollaboration with industry. Various studies nd complementar-

ities between academics volume of government grants and fundsraised fromindustry( Boardman, 2009; BoardmanandPonomariov,2009; Bozeman and Gaughan, 2007; Lee and Bozeman, 2005; Linket al., 2007 ). Grant funding is predominantly based on a peer-review process and is therefore indicative of the scientists successin the eld. It appears that scientists productivity and success infund raising acts as a signal for private companies when identi-fying potential collaborators, leading to more opportunities andconsequently more engagement activities. Moreover, the ability toacquire public resources may indicate a general ability to attractfunds, which will also increase the likelihood of moving into col-laborative projects with industry. This dynamic may however alsobe driven by the increasing tendency of government funding agen-cies to look positively upon grant proposals that involve industry

interaction.

When comparing academic engagement with commercial-isation, individual determinants play a similar role for bothsets of activities. Notably, male researchers are more likely toengage in both, while the role of age appears to be ambigu-ous for both engagement and commercialisation. Importantly,both engagement and commercialisation are more likely to bepursued by individuals that are more scientically productivethan their colleagues. However, academic engagement is clearlyassociated with higher seniority and success in obtaining govern-ment grants while for commercialisation the role of these factorsis ambiguous. Some studies suggest that engagement in com-mercialisation behaviour may be associated with being youngerbecause the lower age range were socialised in contexts wherecommercialisation had become more legitimate ( Bercovitz andFeldman, 2008 ).

4.2.2. Organisational context The most salient organisational-level determinant for academic

engagement is represented by the quality of academics universityor department. The overall effect of organisation-level academicquality on participation in collaboration activities appears to benegative ( DEste and Patel, 2007; Ponomariov, 2008; Ponomariovand Boardman, 2008 ). To some extent, the nding that faculty sup-port for universityindustry interaction is negatively inuenced bytheage of theiruniversity pointsin a similar direction( Azagra-Caroet al., 2006 ). This result seems counterintuitive when comparedto the ndings on the impact of individual scientic quality, butmay arguably be related to a lower degree of resource muni-cence at lower quality research institutions, motivating academicsto embrace industrycollaboration as a means of acquiring researchfunds.

Organisational factors are likely to moderate the impactof indi-vidual characteristics on external engagement. Louis et al. (1989)found that individual characteristics are strongly moderated bythe effect of group-level norms. These ndings are backed bymore recent research on UK and German scientists ( Haeussler andColyvas, 2011 ). If their colleagues value patents and awards, aca-demics are more likely to consult for private companies, while theopposite is true if their peers value traditional academic values.

Finally, academics afliation with special entities within theiruniversities, such as research centres, has been found to posi-tively inuence engagement ( Bozeman and Gaughan, 2007 ). Neworganisational structuresdrawingon expertise frommultipleeldsappear to be instrumental in facilitating interactions between thepublic and private sector.

With respect to organisational determinants, there are pro-nounced differences between academic engagement and com-mercialisation. An extensive literature has analysed the role of university and departmental features ( Owen-Smith and Powell,2001 ) and technology transfer infrastructures ( Lockett and Wright,2005; Siegel et al., 2003 ) for commercialisation. This has shown

that the research quality of the afliate university increases thelikelihood of researchers participating in commercialisation ( DiGregorio and Shane, 2003; Manseld, 1995; OShea et al., 2005;Owen-Smith and Powell, 2001; Sine et al., 2003 ). Moreover, thepresence of formal technology transfer mechanisms is generallypositively related to commercialisation ( Markman et al., 2005a,b; Phan and Siegel, 2006 ). Research has also found local peereffects imply that academics are more likely to be entrepreneurialif departmental colleagues of the same rank are entrepreneurial(Bercovitz and Feldman, 2008; Stuart and Ding, 2006 ). For aca-demic engagement, these relationships do not appear to hold.Most notably, in contrast to commercialisation, individual aca-demic engagement tends to be negatively correlated with theresearch quality of departments or universities. Simultaneously,

there is no conclusive evidence on the role of formalorganisational


6/20


support structures or peer effects for stimulating academicengage-ment.

4.2.3. Institutional context We consider two aspects of the institutional context in which

academics operate: the afliation to a scientic discipline and theeffect of specic national regulations and public policies. Both fac-tors inform academic engagement as they shape the norms andrules relevant for researchers, either because they are ofcial gov-ernment regulations or because they are the rules of conductprevailing within the invisible colleges in which academics oper-ate( Crane, 1972 ).

Disciplinary afliation is an important variable informingengagement with industry ( Bekkers and Bodas Freitas, 2008;Martinelli et al., 2008 ). Applied elds of research, such as engi-neering, make collaboration or engagement in entrepreneurialactivities more likely ( Bekkers and Bodas Freitas, 2008; Boardman,2008, 2009; Bozeman and Gaughan, 2007; Lee and Bozeman,2005; Lee, 1996; Ponomariov, 2008 ). Scientic disciplines affectthe selection of knowledge transfer channels from university torms: Bekkers and Bodas Freitas (2008) nd that in biomedi-cal and chemical engineering the most important channels arepatentsand licensing, scienticoutput,studentsplacements, infor-mal contacts and contract research. For researchers in computerscience, patents and licenses do not seem a relevant transferchannel, while they are very important for material scientists.Finally, in social sciences, knowledge seems to be transferredthrough personal contacts and labour mobility. In the medicaleld, clinical researchers are more likely to partner with industrybut the non-clinical researchers are more likely to commercialise(Louis et al., 2001 ).

In terms of the role of national policies, comparative empiri-cal evidence is limited. Most studies focus on North American andEuropean countries including the UK, Spain, Germany and Swedenwhile little is known about other geographical contexts. At leastamong these countries, there appear to be no major differences inthedeterminantsof academicengagement found in comparisons of the UK and Germany ( Haeussler and Colyvas, 2011 ), Germany andUS (Grimpe and Fier, 2010 ), or Sweden and Ireland ( Klofsten and Jones-Evans, 2000 ). Lee (1998) speculates that increased academicengagement, rooted in the US academic career system, providesincentives to academics to acquire resource from industry in orderto further their career. Suchinstitutional pressures may be lower insystems where funding is allocated on a less competitive basis andwhere university positions are endowed with discretionary funds(Haeussler and Colyvas, 2011 ).

Again comparing academic engagement with commerciali-sation, a large body of evidence exists on institutional-leveldeterminants of the latter, due to the presence of clear policychanges such as the BayhDole Act in the US and the abolition of

the so-calledprofessors privilege in European countries ( Moweryand Sampat, 2005; Powers and McDougall, 2005; Sampat et al.,2003 ). Mowery and co-workers argue that the rapid growth inuniversity patenting in the period following the BayhDole Actis partly due to the supply of technological opportunities in sci-entic elds such as in biomedicine. Authors comparing differentinstitutional contexts found that commercialisation of univer-sity inventions is more likely in environments characterised byintense competition (such as in the US) while more rigid envi-ronments discourage these initiatives ( Goldfarb and Henrekson,2003; Henrekson and Rosenberg, 2001 ). By contrast, research onacademic engagement has rarely addressed the role of institu-tional environments and national policies, partly because it hasenjoyed less attention from policy-makers than commercialisa-

tion.

4.3. Consequences of academic engagement

Academicengagement canhave an impact on various universityactivities. A rst issue concerns the effect of engagement on aca-demics research productivity. Most authors nd that faculty withindustrial support publish at leastas manyscienticarticlesas theircolleagues, if not more ( Blumenthal et al., 1996; Gulbrandsen andSmeby, 2005 ). Collaborative projects oftenyield new,academicallyvaluable insights and ideas even if they are relatively applied anddo not directly result in publishable results ( Lee, 2000; Manseld,1995; Perkmannand Walsh, 2009 ). However, there is also evidencethat researchers with industrial exposure may publish less overtheir career as a whole, and that publishing may have an inverseU-shape-relationship with engagement ( Lin and Bozeman, 2006 ).

A second issue refers to the impact of external engagement onacademic scientists research agendas. Some observers fear thatengagement with industry shifts researchers agendas towardsmore applied topics at the expense of the long-term benets of basic science. For example, Blumenthal et al. (1996) in their studyof US life science faculty showedthat academics with industry sup-port are more likely to report that their choice of research topic isinuenced by the projects commercial potential. Evidence indi-cates that industry-funded research is more applied, but also morecollaborative, both with private and public partners ( Boardmanand Corley, 2008; Gulbrandsen and Smeby, 2005 ). A bibliometricstudy ( Godin and Gingras, 2000 ) reported no evidence of industrialinuence on the direction of research. Third, academics involvedin engagement activities may restrict communication with theircolleagues. Empirical evidence to support this statement is ratherlimited, with only one quantitative and one qualitative studyreporting a positive relationship between engagement and secrecy(Blumenthal et al., 1996; Welsh et al., 2008 ).

Theimpact of academics engagement withindustry on teachingis not clear and the question has not been addressed in the litera-ture. The only exception is Lin and Bozeman (2006) , who observethat academics with industry exposure support more students.

Turning again to the impact of commercialisation activities,there is relative agreement that academic inventors publish moreand better papers than their non-patenting colleagues ( Agrawaland Henderson, 2002; Azoulay et al., 2007; Breschi et al., 2007;Fabrizio and Di Minin, 2008 ). Although involvement in commer-cialisation may not directly impact on scientists academic careers,theybelieve it canincreasetheir prestigeandreputation ( Moutinhoet al., 2007; Owen-Smith and Powell, 2001; van Rijnsoever et al.,2008 ). Concerning the possible shift of research towards moreapplied topics, Hicks and Hamilton (1999) f ound that the shareof basic research at US universities remained unchanged between1981, the year in which the BayhDole Act was passed, and 1995while university patenting increased signicantly. Thursby andThursby (2002) f ound that increases in university licensing werelargely due to universities greater commercialisation efforts rather

than changes in research direction.Finally, there is some evidence that academic researchers

involved in commercialisation activities practice higher degreesof secrecy than their non-commercialising colleagues ( Campbellet al., 2000, 2002 ) and that academic entrepreneurship may ham-per the accumulation of knowledge in the public domain ( Tooleand Czarnitzki, 2010 ). Related research suggests that increasedacademic patenting may slow the unencumbered diffusion of aca-demic knowledge ( Huang and Murray, 2009; Murray and Stern,2007 ). Most research on this issue focuses on commercialisation inthe lifesciencesbecausehere, intellectualpropertyis of highpoten-tial value. This limited evidence suggests, however, that academicresearchers with an interest in commercialisation may employgreater levels of secrecy about theirresearchresultsthan their open

science oriented colleagues.


7/20


Comparingthe consequences of individual involvement in com-mercialisation withthe consequences of academicengagement,wecan conclude the following. First, whilst participation in commer-cialisationappears to have positiveeffects onresearchproductivity,the evidence for academic engagement is ambiguous. Second, nei-thercommercialisation nor engagement seems to skew academicsresearchtowardsmore applied topics.Thirdly, there is tentative yetno conclusive evidence that both types of industry exposure lead toincreased secrecy and restricted communication of open researchndings.

5. Discussion

5.1. Implications for the concept of academic engagement

The systematic literature review and comparison with extantresearch on commercialisation enable us to further develop theconcept of academic engagement. This type of universityindustrylink may take different forms, including collaborative research,contract research and consulting, and is practiced by a far largerproportion of academics than commercialisation. The review sug-gests that academic engagement is a multi-level phenomenon, inthe sense that it is determined by both the characteristics of indi-viduals as well as the organisational and institutional context inwhichthey work. Below, we synthesise our ndings with a view todeveloping a fuller understanding of this phenomenon.

On the individual level, academic engagement is pursued byscientists who are well established and well connected in the aca-demic community. Individuals who are more senior, have moresocial capital, more publications and more government grants, alsoworkmostprolicallywithindustrialcollaborators.Particularlythepositive correlation between engagement and government grants,and engagement and scientic productivity found by most stud-ies, suggests that academic engagement goes hand-in-hand withacademic success. Engagement appears to be part of the activi-ties responsible for the Matthew effect in academia, according towhich individual success is reinforced through a virtuous cycle of achievements and returns on those achievements ( Merton, 1968 ).This pattern also chimes with the result that male academics aremore likelyto engage withindustry, mirroringthe insight obtainedfrom other studies suggesting that male scientists occupy moreprominent positions than women and are hence in a better posi-tionto mobilise resourcesandestablishwider networks( Etzkowitzet al., 2000a; Gupta et al., 2005; Murray and Graham, 2007 ). Thepositive correlation between academic success and engagementdemonstrates that engagement is highly complementary withstrictu sensu academic activities, i.e. publishing and writing grants.For commercialisation, the degree of complementarity with aca-demic activities is less clear cut; while commercialisation tends tobe positivelyrelated to scientic productivity, there is a contingent

relationship with grant and contracts ( Table 2 ).Evidence presented by research on the attitudinal and

motivational aspects of both academic engagement and com-mercialisation supports a view of academic engagement ascomplementary to academic research. Lee (1996) investigatedthe attitudes of US academic researchers towards universityentrepreneurship. Most members of faculty were in favour of transfer activities but objected to the most radical forms of com-mercialisation suchas start-up assistance to new technology rms,and equity investment by universities. In turn, other studies showthat academic engagement tends to be viewed by academics as anatural extension of publication-driven open science, while com-mercialisation is seen as a distinct type of activity. Subscription totraditionalscienticnormsis notnecessarilyat oddswith academic

engagement ( Boardman and Ponomariov, 2009 ) while a study of

scientists at the Max Planck Society indicates that those believingthat science is a public good are signicantly less likely to pursuecommercialisation ( Krabel and Mueller, 2009 ).

The notion of engagement as complementary to traditionalacademic science is also supported by research on academicsunderlying motivations. Lee (2000) f ound that science and engi-neering faculty at US research universities collaborate with privatepartners for two main reasons: (a) to access resources relevant totheir research activities via additional funds, equipment and sup-port for students, and (b) to access learning opportunities, suchas eld-testing opportunities for their own research and obtainingnew insights. Similarly, a study of UK physical sciences and engi-neering faculty found that academic engagement was driven byresearch considerations (i.e. learning and resource access), whilecommercialisation was motivated by monetary incentives ( DEsteand Perkmann, 2011 ).

On the organisational level, a noteworthy nding of ourreview is that academic engagement is negatively associatedwith organisation-level research quality in some studies anduncorrelated in other studies, in contrast to its positive associ-ation with high individual research quality and the commonlydiagnosed positive relationship between engagement in commer-cialisation and university/department-level research quality. Thismeans academic engagement is pursued by highly motivated andsuccessful individuals who are, however, not necessarily afliatedto higher quality research institutions. One may hypothesise thatacademic engagement acts as a resource mobilisation device forhigh-performing individuals at lower ranked institutions, wherefewer resources are available. Our results further suggest, thatorganisation-level support appears far more relevant for commer-cialisation activities thanacademic engagement, which tends to bedriven by individuals and teams with little central support. Finally,on the institutional level, engagement is strongly associated withafliation to engineering and applied sciences.

Synthesising our results, one may postulate three key insightsfor academic engagement. First, academic engagement is posi-tively correlated with individual characteristics that dene senior,scientically productive individuals, indicating that it is in linewithfurthering their academic research activities. Second, engage-ment is less organisationally embedded than commercialisationactivities, and is more autonomously driven by individuals. Third,engagementappears tobe aneffectivetool formobilising resources,and may function as a substitute for generous resource endow-ments at highly ranked institutions. One needs to exercise cautionas the positive effects of engagement on scientic productivity areat this stage hypothetical, because the existing research is almostentirely based on cross-sectional analysis. The direction of causal-ity is uncertain, therefore, leaving open the possibility that goodresearch performance by individuals leads them to engage more.We will return to this issue in the discussion of methodologicalchallenges.

The insights gained from the literature review are representedin a stylised model, outlining the various antecedents of academicengagement at the individual, organisational and institutional lev-els, as well as itsconsequences ( Fig. 2). We alsohighlight the factors indicated by dashed lines that require more research eitherbecause they were relatively neglected by previous research orbecause studies yielded conicting or ambiguous results.

5.2. Implications for conducting research

Our conceptual considerations and review of the literaturesuggest that the study of academic engagement requires distinctanalytical and methodological approaches. Analytically, given thatacademicengagementmeans collaboration, approaches focused on

studies of how individuals within organisations initiate, build and


8/20


DemographicsCareer trajectory

Productivity

SCIENTIFIC OUTPUT

ACADEMIC ENGAGEMENT

INDIVIDUAL FACTORS

AttitudesMotivations

Identity

Scientific discipline RegulationPublic policy

INSTITUTIONAL FACTORS

Productivity

Agenda

EDUCATIONALOUTPUT

TT supportFormal

incentives

University /department

quality

ORGANIZATIONAL FACTORS

LeadershipDepartment

climate

COMMERCIALOUTPUT

Fig. 2. An analytical framework of external engagement by academic researchers.

maintain collaborations with other organisations (e.g. rms) arerelevant for studying it ( Bouty, 2000; Kreiner and Schultz, 1993 ).Hence aspects such as the motivation for individuals to engage incollaboration, the formal characteristics of the collaboration, andoutcomes both at individual and organisational levels are salient.Specically, in terms of motivation, collaboration may be pursuedfor a variety of reasons, such as attracting resources, obtainingknowledge, or building social capital. Commercialisation, by con-trast, implies a more narrowly focused interest in the exploitationof a specic technology. Hence, analytically, commercialisation ismore akin to an entrepreneurialact or at least,in thecase of patent-ing, a rst step towards it. Researchers have therefore used thelanguage and analytical framework of entrepreneurship to explorecommercialisation, involving an emphasis on the aspects of oppor-

tunity recognition and individual economic incentives ( Lach andSchankerman, 2008; Shane, 2004; Wright et al., 2007 ).

Empirically, academic engagement leaves distinct traces. Forinstance, academicentrepreneurship canbe measured by counts of university spin-offs or company directorships maintained by aca-demic researchers. Information on patents is accessible via publicpatent directories; various studies have successfully attempted toidentify university-invented patents that are not assigned to uni-versities, in addition to patents assigned to universities ( Lissoniet al., 2008; Thursby et al., 2007 ). Even though more widelypracticed, academic engagement is empirically more difcult todetect because it includes collaboration instances that may notbe documented by generally accessible records. Researchers haveapproximated engagement via instancesof co-authorship between

university researchers and industry scientists ( Liebeskind et al.,1996; Murray andStern, 2007 ). This procedure however is likely tounder-represent collaborations thatare moreapplied in nature anddo notresult in publications, suchas contractresearchor consultingassignments. Records held by universities on industry contractswould represent an ideal source of information but are not readilyavailable because they are often considered commercially sensi-tiveby university administrators, in addition to which suchrecordsare likely to underestimate consulting activities. Moreover, theyare difcult to standardise across large numbers of universities.Nevertheless, studies using record-based information for singleuniversities or a small number of institutions can offer powerfulinsights with a high level of granularity ( Rawlings and McFarland,2011 ).

In facing the above challenges, the overarching majority of studies have traced such engagement by asking academics for self-reported information, usually via questionnaires. Questions tendto be structured around the essential characteristics of collabora-tion instances, such as underlying motives, resources exchangedand the type of collaborative arrangements. It is clear that rely-ing on self-reported information is fraught withspecic challengeswhich future work should address in order to improve the quality,reliability and validity of research results.

A rst issueregards the lack oflongitudinaldata.In fact,all large-scale survey-based studies are based on cross-sectional data andtherefore pose limitations in termsof inferringcausal relationshipsbetween variables. For instance, it is unclear whether individualresearch performance is enhanced by academic engagement, or

engagement is a mere consequence of high research performance.On the basis of existing evidence it is difcult to say whetherengagement in one period is linked to activities in a prior period.While a new generation of studies using panel data on academicpatents and publications has taken into account the time dimen-sion ( Stuart and Ding, 2006 ), this has yet to be accomplished byresearch using survey data, the main source of information onacademic engagement. The only studies acknowledging the longi-tudinal dimension are qualitative contributions ( Etzkowitz, 1998; Jain et al., 2009; Kenney and Goe, 2004; Shinn and Lamy, 2006 ).Future research should conduct surveys repeatedly, or at leastadminister subsequent surveys containing some identical ques-tions, across a comparable population of academic researchers.

Second, the validity and comparability of results are hampered

by the way

measures are constructed. Across the studies we exam-ined, the measures for the key dependent variable (i.e. academicengagement) varied considerably. While most authors capturethe frequency of engagement, others measure the importance of specic channels perceivedby individuals. The latter is more prob-lematic because individuals vary in how they dene importance.Approaches that measure the frequency of engagement, possiblygoing beyond yes/no statements and accounting for the numberof engagements over a period of time, are preferable because theyare more likely to reect facts. The use of different measures alsoimpedes comparisons across studies. BozemanandGaughan(2007)proposed an index that takes into account not only the variety of interaction mechanisms, but also their difculty or rareness.This index could provide the starting point for a harmonisation of efforts across different studies.


9/20


Similar considerations apply to measuring other variables,such as motivation, organisational conditions or outputs resultingfrom academics external engagement. Items not well groundedin theory may produce meaningless or even misleading results,and hence hamper comparison between different datasets. Hereresearchers should use well-established scales, like those foundin motivation research, or combine surveys with eld studies tothoroughly understand the phenomenon ( Bearden and Netemeyer,1999 ). Some studies also fail to differentiate between the mea-surement of perceptions and structural factors. For instance, whilesome surveys simply measure academics perception of organisa-tional support for engagement activities, others attempt to capturestructural features by requestingfact-derived information. The lat-ter approachis preferable because it reduces social desirability bias(Podsakoffet al.,2003 ) as respondents may believe that it is sociallydesirable to be entrepreneurial and engage with industry.

Reaching a consensus on the central measures relating to aca-demic engagement, such as activities, motivations, barriers andoutcomes, would improve the quality and comparability of stud-ies. Researchers should therefore disclose the scale items they useas well as scale means, individual item means and standard devi-ations across different groups, providing a frame of reference forpotential scale users ( Bearden and Netemeyer, 1999 ).

Finally, existingstudies areafictedby incoherence in thepopu-lations studied, raising questions over the external and internalvalidityof results. Ideally, sampling procedures should ensure pop-ulation representativeness and avoid sampling bias. Many studiesfocus on specic disciplines whilst others address populations of academics at particular universities, possibly representing the out-come of convenience sampling that risks skewing results. Futureresearch should ensure population representativeness and alsolimit selection bias. For instance, the disciplinary scope of the sur-veyed populationcouldbe enlarged,and moreuniversitiescould betargeted. The central practical challenge for researchers success-fully addressed by some studies ( Abreu et al., 2009 ) is togeneratelarge lists of researchers that are eitherreective of the whole pop-ulation, or a random non-biased sample. Some of our suggestionswill be more costly to implementthanothers.Yet, designingappro-priate measures and factual questions do not come with increasedcost, so researchers should start here. Similarly, response bias maybe detectedandremedied withthe appropriate statistical methods.

We provide an overviewof the samplingtechniqueswe encoun-tered in theliterature in Table 4 , summarising their advantages anddisadvantages.

The existence of a reliable and both nationally and internation-ally comparable evidence-base on academic engagement wouldserve as a valuable resource for policy-makers. In the UK, Sweden,Spain and Italy, for instance, data on academic engagement is, atthe time of writing, collected nationally but only for universities inaggregate, 1 and not for individuals. Efforts to standardise measuresand ensure individual-level data collection exercises encompass

larger populations of universities would allow policy-makers toreach better judgements of the strengths and weaknesses of typesof collaboration, proles of individuals, organisational specialismand universities. It appears that research funding bodies are bestpositioned to sponsor or even carry out data collection, as theresource dependence of academic researchers makes them morelikely to respond to requests extended by these kinds of organisa-tions.

1 See: the HEBCI (Higher Education Business and Community Interac-tion) Survey in the UK ( http://www.hefce.ac.uk/econsoc/buscom/hebci ), Statis-tics Sweden ( http://www.scb.se/Pages/Product 8793.aspx ), RedOTRI in Spain

(http://www.redotriuniversidades.net )andANVURinItaly( http://www.anvur.org ).

5.3. Agenda for future research

Our analysis of the literature has not only generated novelinsights into the nature of academicengagement but also indicatesareas that require further research. First, while the individual-level determinants are relatively well explored, more research isneeded on the organisational context in which industry engage-ment occurs and how it moderates individual characteristics.Our nding that traditional technology transfer infrastructuresplay a lesser role in facilitating academic engagement does notnecessarily imply that other organisational characteristics maynot be instrumental. Rather than centralised support mecha-nisms, characteristics of the department or the research teamsin which academics operate, may be more salient in determin-ing engagement. Research should explore aspects of organisationalambidexterity, i.e. whether engagement is associated with a diver-sication of skills, roles, and organisational units adept at dealingwith academic and industrial requirements. In other words, stud-ies shouldinvestigatewhetherthe structure and proleof researchteamsand departments explains engagement withindustry, ratherthanthe individual-level variables thatare commonly perceivedascritical. This kind of analysis would also enable us to substantiatethe above claimed link between engagement and resource mobili-sation, given organisational contexts vary signicantly in terms of resource municence. Furthermore, for organisation-level factors,extant research has focused on technology transfer or licensingofces while the role of organisational support in the guise of industrial liaison ofces or business relationship ofces has beendisregarded. Future research should take these structures intoaccount when analysing engagement, with the possible outcomethat organisation-level (i.e. university-wide) variables may play amore important role than hitherto thought.

Second, theconsequencesandimpactsof academicengagementneed to be further explored. Extant analyses have neglected toconsider its impact on educational outputs, such as time devotedto teaching, curriculum and courses development, and teachingquality. Insights into this aspect of engagement would be highlyvaluable in extending our knowledge of the benets or costs of third stream activities within the context of universities othermissions.

Third, future research should explore the relationship betweenacademic engagement and commercialisation. Whilst our compar-isonsuggeststhat bothtypes of activities may be driven by differentunderlying mechanisms, we cannot infer possible complementari-ties or contradictions between them. Research here should addresstwo issues. On the one hand, there may be a temporal relation-shipbetween engagement andcommercialisation, in the sensethatprior involvement in collaboration with industry may lead to com-mercial output later in time, either individually or within researchgroups. On the other hand, researchers should investigate the pos-sibility that some types of collaboration are complementary with

commercialisation outputs while others may be neutral or evencompetewiththem. Knowingmore about therelationshipbetweenacademic engagement and commercialisation would also benetpolicy debates by clarifying whether the policies designed to stim-ulate entrepreneurship also stimulate academic engagement, orwhether more focused approaches are needed.

Fourth, institutional aspects should be further investigated.Most studies focus on the United States and selected Europeancountries while contributions covering other geographic contextsare rare ( Giuliani et al., 2010; Walsh et al., 2008 ). Furthermore,few studies offer cross-national comparative analyses ( Dutrenitand Arza, 2010; Grimpe and Fier, 2010; Haeussler and Colyvas,2011; Klofsten and Jones-Evans, 2000 ). The context-specic natureof most published research makes it difcult to form generalised

conclusions. It seems plausible thatcountries with different higher
http://www.hefce.ac.uk/econsoc/buscom/hebcihttp://www.scb.se/Pages/Product_8793.aspxhttp://www.redotriuniversidades.net/http://www.anvur.org/http://www.anvur.org/http://www.redotriuniversidades.net/http://www.scb.se/Pages/Product_8793.aspxhttp://www.hefce.ac.uk/econsoc/buscom/hebci


10/20


Table 4Sampling techniques.

Type of sampling Implementation Benets Drawbacks

Census (whole academic population) Ofcial list of academics (i.e.fromthe Ministry of Education)

No biastowards a singleuniversity or discipline

Resource intensive

Manual search on universitieswebsite

Completeness Ofcial lists only available insome countries (i.e. whereacademics are civil servants)

Selected universities List of academics from theuniversity centraladministration ofce

No biastowards a singlediscipline

Organisation bias


Large institutionstend to bepreferred

Selected disciplines List of academics located inspecic departments from theuniversity centraladministration ofce

No biastowards a singleuniversity

Discipline bias


Denition of scienticdisciplines can be ambiguous

Selectedgroups (i .e. grant holders) List ofacademics from theorganisation which denes thegroup (i.e. the organisationawarding the grants)

Relatively cheaper Group biased

No biastowards a singleuniversity

Possibly discipline biased

education and public science systems, different stages of econom-ical development and different innovation systems will exhibitdifferent patterns of universityindustry relations, and differentantecedents and consequences. For instance, the regulations con-cerning intellectual property ownership for university-generatedknowledge vary between countries. These and other differencesare likely to systematically shape academics response to externalfactors, as well as the outcomes of academic engagement. Futureresearch should therefore seek to cover multinational settings, ide-ally by deploying harmonised survey tools. The tools currentlyavailable, such as university-level indicators on TTO activity, aretoo coarse for meaningful international comparison, so micro-measures of engagement will provide superior information.

Fifth, as most research hitherto on academic engagement isphenomenon-driven, future research should deploy data on thisphenomenon to build and test theory. For instance, studies mayviewacademic engagement as pro-active behaviour in knowledge-intensive organisations ( Crant, 2000 ). Academia is an ideal contextto study this kind of individual behaviour likely to be benecialto overall organisational performance because academics enjoy alarge degree of professional autonomy, and hence their individualperformance as wellas their contribution to organisational goods isdrivenlargelybyself-motivation rather thancommandandcontrol.Moreover, relative to professional service organisations, academicsettings are richer in publicly available data on individual charac-teristics such as performance and career histories, enabling more

detailed studies. Another opportunity for contributing to theorymay arise from studying how individuals respond to local norms,such as those prevailing in their immediate, departmental workcontexts ( Bercovitz and Feldman, 2008 ), and how these relate toglobal norms for instance at the level of scientic disciplines ( Finiand Lacetera, 2010 ). This may also provide clues for determin-ing the locus of possible policy interventions aimed at supportingengagement or mitigating adverse effects. Finally, within institu-tional theory, academic engagement may offer insights into howindividuals within organisations manage exposure to different log-ics ( Friedland and Alford, 1991 ), that of academic science and of commercial R&D ( Murray, 2010 ). Working with industry is likelytogenerate conicting pressures including whether research resultsshould be public or private, and whether research should be ori-

ented towards publication or technical application ( David, 2004 ).

Although we know that ambidextrous management of both log-ics is commonplace, the study of academic engagement is likely toexpand our understanding of how individuals accomplish this andwhat factors enables these efforts.

6. Conclusions and policy implications

Exploring external engagement by academic researchers isof undoubted interest to practitioner audiences, notably policy-makers and university managers. Government agencies anduniversities themselves have made concerted efforts to increaseacademicengagement, forreasonsrangingfrom generatingsocietallegitimacy for publicly subsidised scientic research, stimulatingeconomic activity to raising revenue for universities. There arethree policy-related lessons that can be drawn from our review.

First, our analysis suggests a partial lack of understandingabout the consequences of academic engagement. Evidence on theimpact of these collaborations on other university activities, suchas research and teaching, is scarce so it cannot be assumed thatengagement activities are always benecial and should thereforebe promoted. Further research on the consequences of engage-ment will allow policy-makers to derive considered judgementsas to what behaviours and organisational forms to promote, andunder which conditions they are likely to further scientic and/oreconomic objectives. For instance, a better grasp of the causal rela-tionship between engagement and research performance is crucial

for designing policy interventions. If engagement spurs researchperformance, it is obvious that engagement should be promotedif policy-makers wished to promote better research. However, if the opposite is true research performance drives engagement interventions would need to promote research excellence lead-ing to further engagement. This issue is particularly salient inthe context of increasing pressures exerted by democratic con-stituencies to evaluate the impact of academicresearch. As fundingbodies in many countries require bidding academics to provideevidence of societal impact (and not just on the academic commu-nity), understanding how engagement results in such benets, andsimultaneously maintain scientic quality, appears now of greaterrelevance.

Second, research on universityindustry interaction has

strongly emphasised the role of TTOs or liaison ofces, so some


11/20


policy-makers have consequently resorted to subsidising tech-nology transfer operations at universities. This instituted anorganisational focus at universities on formal mechanisms of com-mercialisation, i.e. patenting, licensing and entrepreneurship. Asour review suggests, these technology transfer structures areless adept at fostering academic engagement. From a policy per-spective, it is important to recognise that different transfer orcollaboration mechanisms may require different support struc-tures and incentive mechanisms. As individual discretion seemsthe main determinant of academic engagement with industry,policy measures should address individuals, in addition to inu-encing university practices and structures. For instance, fosteringindividual-level engagement skills would appear to be a poten-tially powerful lever, not only for increasing the volume of universityindustry relations but also their quality. In this respect,policy should not implicitly assume that more is better but seekto differentiate the conditions under which engagement gener-ates both academic and industrial benets, so minimise the riskof failure.

Third, considering the higher volume of engagement activ-ities than patenting and entrepreneurship, it is essential thatrms be well-equipped to effectively participate in collaboration(Perkmann and Salter, 2012 ). If policy aims to successfully increasethe impact of academic research through fostering engagement,not only academics but rms too need to be skilled in initiatingandmaintaining suchcollaborations, cruciallyrecognisingthat col-laborating with academia presents distinct challenges, separate tothose of customers or suppliers. Particularly when collaboratingwith the best academic researchers, rms need to take into accountthat these academics will under most circumstances only workwith them if there is also some academic benet to be derived. Inorder to reduce transaction costs, policy-makers should addition-allyseek to establishsome guidelines or standardcontracts to guideboth partners during the negotiations, such as those of the Lam-bert agreements in the UK or the arrangement between Procter &Gamble and the University System of Ohio ( Lambert, 2003 ).

A nal issue relates to the theoretical underpinning implied inmuch of the current research. Studies commonly conceive aca-demic science as a discrete institutional order that differs fromindustry on the basis of academic values, normsand conventions aslaid out by Mertonian sociology of science ( Merton, 1973 ). Even if such a discrete, binary characterisation may be analytically appro-priate in some instances, it may underestimate the diversity of organisations and institutional orders found within the interna-tional higher education and public R&D sectors. The archetype of the research-intensive university represents, in fact, only one typeof organisation within a eld populated by entities as diverse as

professionallyoriented polytechnics,national R&Dlaboratoriesandliberal arts colleges. A considerably proportion of academics donot conduct research, or at least not research close to the frontiersof knowledge production. It is important, therefore, to differenti-ate the frontier researchers and their external engagement, fromthose who are removed from it when conducting their research.Furthermore, differences in how universities are governed andmanaged across different countries may be important in determin-ing individuals engagement with non-academic organisations.

An important objective for future research is, therefore, toquestion the pervasiveness and purity of Mertonian norms, andshed light on the differences characterising the diverse patterns of university-society interactions in various settings. If organisationalor even group-level variance is more important than previouslythought, a view supported by recent studies using data on patent-ing or disclosures data ( Bercovitz and Feldman, 2008; Stuart andDing, 2006 ), then relaxing the assumption of relative homogeneityacross the organisational population becomes a crucial aspirationfor research in this area.

Acknowledgements

We acknowledge support from the DIME Network, fundedby the 6th Framework Programme of the European Commis-sion (CIT3-CT-2005-513396). Additional support was providedby IMIT (Institute of Management of Innovation and Technol-ogy), the Advanced Institute of Management Research (AIM)(RES-331-27-0063), the UK Innovation Research Centre (UK-IRC, RES-598-28-0001), University of Gothenburg, Alma GraduateSchool/University of Bologna and ANR (ANR-06-APPR-003).

Appendix A.

Keyword: combinations used for literature search:

joint research and university*/academi*/facult* and indus-try*collaborative research and university*/academi*/facult* andindustry*contract research and university*/academi*/facult* andindustry*technology transfer and university*/academi*/facult* andindustry*commercialize and university*/academi*/facult* and indus-try*academic entrepreneurshipuniversityindustry


12/20


13/20


14/20


15/20


16/20


17/20


18/20
http://www.cbr.cam.ac.uk/pdf/AcademicSurveyReport.pdf


19/20


20/20


Toole, A.A.,Czarnitzki, D., 2010.Commercializingscience:is there a universitybraindrain from academic entrepreneurship? Management Science 56, 15991614.

Traneld, D., Denyer, D., Smart, P., 2003. Towards a methodology for developingevidence-informed management knowledge by means of systematic review.British Journal of Management 14, 207222.

Van Dierdonck, R., Debackere, K., Engelen, B., 1990. Universityindustry relation-ships: how doesthe Belgian academic communityfeel aboutit? Research Policy19, 551566.

van Rijnsoever, F.J., Hessels, L.K., Vandeberg, R.L.J., 2008. A resource-based view onthe interactions of university researchers. Research Policy 37, 12551266.

Walsh, J.P., Baba, Y., Goto, A., Yasaki, Y., 2008. Promoting universityindustry link-ages in Japan:faculty responsesto a changing policyenvironment. Prometheus26, 3954.

Welsh, R., Glenna, L., Lacy, W., Biscotti, D., 2008. Close enough but not too far:assessing the effects of universityindustry research relationships and the riseof academic capitalism. Research Policy 37, 18541864.

Wright, M., Clarysse, B., Mustar, P., Lockett, A., 2007. Academic entrepreneurship inEurope. Edgar Elgar, London.

Academic Engagement and Commercialisation

Documents

Transcript of Academic Engagement and Commercialisation