University–industry collaboration in Italy: A bibliometric examination

University–industry collaboration in Italy: A bibliometric examination

Please download to get full document.

View again

of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.

Cognitive Psychology

Publish on:

Views: 6 | Pages: 10

Extension: PDF | Download: 0

  Technovation 29 (2009) 498–507 University–industry collaboration in Italy: A bibliometric examination Giovanni Abramo a,b , Ciriaco Andrea D’Angelo a,  , Flavia Di Costa a , Marco Solazzi a a Laboratory for Studies of Research and Technology Transfer, School of Engineering, Department of Management, University of Rome ‘‘Tor Vergata’’, Italy b Italian National Research Council, Italy Abstract This work investigates public–private research collaboration between Italian universities and domestic industry, applying abibliometric type of approach.The study is based on an exhaustive listing of all co-authored publications in international journals that are jointly realized by Italianuniversity scientists and researchers in the private sector; this listing permits the development of a national mapping system forpublic–private collaboration that is unique for its extensive and representative character. It is shown that, in absolute terms, mostcollaborations occur in medicine and chemistry, while it is industrial and information engineering that shows the highest percentage of co-authored articles out of all articles in the field.In addition, the investigation empirically examines and tests several hypotheses concerning the qualitative–quantitative impact of collaboration on the scientific production of individual university researchers. The analyses demonstrate that university researchers whocollaborate with those in the private sector show research performance that is superior to that of colleagues who are not involved in suchcollaboration. But the impact factor of journals publishing academic articles co-authored by industry is generally lower than thatconcerning co-authorships with other entities. Finally, a further specific elaboration also reveals that publications with public–private co-authorship do not show a level of multidisciplinarity that is significantly different from that of other publications. r 2008 Elsevier Ltd. All rights reserved. Keywords:  University–industry collaboration; R&D cooperation; Bibliometrics; Multidisciplinarity; Italy 1. Introduction The capacity of a nation to produce wealth dependsincreasingly on the investment it undertakes in strengthen-ing the so-called ‘‘triangle of knowledge’’, which iscomposed of research, education and innovation. In thisregard, European nations, in accepting the Lisbon 2000agenda, assumed an ambitious objective: to make Europethe most competitive and dynamic knowledge-basedeconomic system in the world. The strategy, as furtherconsolidated in the Barcelona summit, set the objective of EU member states assigning 3% of GDP to research by theyear 2010. These directives indicate the desire to remedyEurope’s competitive weaknesses at the international level.As a vehicle for action, both the attention of policymakers and the accompanying debate seem heavilyfocused on the research apex of the research–innovation– development triad, with the Barcelona summit objectives(though recognized as a difficult target for most EUnations) clearly emphasizing the provision of resourcesfor research. And yet the existence of a ‘‘Europeanparadox’’ is well known, meaning that there is anincapacity to translate the excellent results from Europeanresearch into innovations that are successfully destined forthe marketplace (EC, 1995). Many analyses, comparing tothe reference experience of North America, show that thegreater competitive capacity of the nations there has clearlybeen favored by policies and legislation (such as theBayh–Dole Act in the United States), which havestimulated technological transfer and provided incentivefor osmosis between the worlds of public research andindustry (Shane 2004; Thursby and Thursby, 2003; Mowery et al., 2001). ARTICLE IN PRESS$-see front matter r 2008 Elsevier Ltd. All rights reserved.doi:10.1016/j.technovation.2008.11.003  Corresponding author. Dipartimento di Ingegneria dell’Impresa,Universit a`  degli Studi di Roma ‘‘Tor Vergata’’, Via del Politecnico 1,00133 Rome, Italy. Tel.: +39672597362; fax: +39672597305. E-mail address: (C.A. D’Angelo).  Within Europe then, the difference between the levels of scientific performance and technological and industrialcompetitiveness is most pronounced in nations such asItaly, where the government’s expenditure in research ishigher (50.7%) than the private sector’s; industry isprimarily specialized in low and middle-low technology,and the industrial structure composes a disproportionatenumber of micro- and small enterprises. In the Italiancontext it is even more urgent that the nation promotecollaboration between the public research system andindustry, thus creating favorable conditions for commercialexploitation of the research results from universities andpublic research laboratories (Grandi and Sobrero, 2005).But, in observation, Italy registers a low propensity tocapitalize on the results of public research. In 2001 (afterthat date, the introduction of academic privilege in Italianpatent legislation would make international comparisonsuneven), for example, the number of patents by the entiretyof Italian universities was roughly equal to that of theUniversity of Wisconsin alone, while that of the totalityof all universities plus all public research laboratorieswas inferior to that of the Massachusetts Institute of Technology (Abramo and Pugini, 2005). As context forthe comparisons, it may be worth knowing that theR&D expenditures in 2001 were around 486 million euroat MIT, 675 million euro at the University of Wisconsin,and 4418 million euro at Italian universities (NSF, 2003;Istat, 2003).To this must be added the limited capacity for transfer of patents to the productive system; the National ResearchCouncil, the major Italian research institution, awardslicenses for the actual use of less than 20% of the patents itfiles annually; Italian universities award licenses for anaverage of 13% of patents, compared to 60% foruniversities in the United States and the UK (Abramo,2007). Yet a 2005 study by Abramo and D’Angelo thatexamined the alignment of public research supply withItalian industrial demand, through a survey of leadingpublic research scientists in high-tech sectors, found thatmost research project results do seem to have immediateindustrial applicability, even if in one third of the casesthere are no Italian companies able to exploit the results.The ensemble of these observations points to the clearnecessity of fine-tuning the match between research policyand industrial policy, with greater attention to allinitiatives that may foster the transfer of public researchresults to domestic industry.The relations between universities and industry presentlytake form in various modes, variable in the extent to whichthey are codified and formalized. Typical modalitiesinclude joint research projects, awarding of researchcontracts, awarding of know-how and patents underlicense, consulting, training services and personnel mobi-lity. The observation of such modalities, their empiricalstudy and the analysis of their underlying determinants canfurnish useful cognitive bases for the policy maker called tostimulate them. In this regard, the present study proposesto investigate research collaboration between universitiesand domestic industry through a bibliometric approach, inwhich ‘‘collaboration’’ is represented by ‘‘co-authorship of scientific articles’’, and to develop a mapping system ableto identify the technical-scientific fields in which alignmentbetween private demand and public offer of knowledge isrealized with greater frequency. In a complementarymanner, such mapping will highlight those sectors in whichthe connection between academic and productive systems isweak or completely absent. This information may result asuseful for the policy maker, both for choosing thedevelopment directions and aims for programming sector-ial priorities, as well as in monitoring results from previousinterventions with similar objectives (Owen-Smith et al.,2002; Veugelers and Cassiman, 2005). Another relevant aspect of the proposed investigationsconcerns the analysis of the determinants of collaborationfor public–private research: these can be considered asexchange relationship in which both parts obtain benefits(Meyer-Krahmer and Schmoch, 1998). In particular, onthe part of university researchers, collaboration withprivate business guarantees access to additional financingfor research and/or to complementary assets. It shouldbe noted that the correlation between defined objectives ex ante  and benefits obtained  ex post  is not always linearand that many benefits are obtained in an unexpectedmanner (Lee, 2000). All this should have a significantimpact on qualitative–quantitative productivity of scien-tists (Balconi and Laboranti, 2006; Barnes et al., 2002; Van Looy et al., 2004). However, since collaboration involvesinteraction between individuals, and in the case of public–private cooperation, between individuals appertain-ing to systems that are very different in their identity andmission, it brings about transaction costs. These are costsresulting from needs to negotiate and mediate objectives,choose methodologies, deal with results, manage logisticsfor communications, manage gatherings and face-to-facemeetings, and for further coordination needs, and they arecosts that would logically create disincentives towardscollaboration (Belkhodja and Landry, 2005; Drejer and Jorgensen, 2005). In effect, a vast survey conducted inGreat Britain by D’Este and Patel (2007) showed that thedeterminants of the variety and frequency of public–privateinteractions depend above all on the individual character-istics of the researchers involved, more so than thecharacteristics of their home organizations. Most impor-tantly, it seems there is little evidence of conflict betweeninteractions with industry and more traditional academicroles (Boardman and Ponomariov, 2008).In view of these potential benefits and transaction costs,another objective of this study is to test whethercollaboration with the private sector actually producesscientific results that are qualitatively better (from theviewpoint of publication placement) and if academicscientists who collaborate with those in the private sectordemonstrate superior performance with respect to collea-gues who are not involved in such collaboration. ARTICLE IN PRESS G. Abramo et al. / Technovation 29 (2009) 498–507   499  The last aspect for exploration concerns the multi-disciplinarity of projects undertaken in cooperation be-tween academic scientists and private sector researchers. Itcan be hypothesized that projects that interest privatecompanies necessarily imply, by their nature, a call forheterogeneous and varied competencies. This weighs ontransaction costs, which will clearly be variable with thelevel of heterogeneity among the members of a mixedresearch team. Thus, the last objective to be pursued in thisstudy is to examine and verify whether there is a higherlevel of multidisciplinarity in public–private researchprojects, which could determine a surplus transaction costfor this type of project and a consequent disincentive forthe actors in play, in particular for academic scientists.From a methodological perspective, this study, althoughlimited to observation of the Italian situation, is character-ized by its very ample field of analysis, both for the numberof academic institutions (all 68 Italian universities) and forthe scientific sectors analyzed (the full 183 sectors of theeight technical-scientific areas of the Italian academicsystem). This constitutes an innovative aspect with respectto preceding studies, which have generally been based onpartial measures of one or a few universities, and/or havefocused on single scientific sectors.The authors are aware that co-authorship-based indica-tors should be handled with care as a source of evidence fortrue scientific collaboration, as has been cautioned bymany bibliometricians (Melin and Persson, 1996; Laudel, 2002; Tijssen, 2004; Lundberg et al., 2006). As Katz and Martin (1997) stated, some forms of collaboration do notgenerate co-authored articles (university researchers mightfor example publish without mentioning the direct involve-ment of industrial researchers) and some co-authoredarticles do not reflect actual collaboration (a publicationcould suggest an inter-institutional collaboration that hasnot taken place, for example if an author has moved from auniversity to industry and in his/her publication lists boththe prior and current affiliation).However it is incontestable that, in the literature,analysis of co-authorship has become one of the standardways of measuring research collaborations between orga-nizations, evidently because it offers notable advantages incounterpoint to the limitations noted above. Co-authoredpublications indicate the achievement of access to an ofteninformal network, and can be viewed as successful scientificcollaboration in themselves, while also indicating diffusionof knowledge and skills. Moreover the indicator isquantifiable and invariant, while measurement is notinvasive and analysis is relatively inexpensive. Finally, withreference to the specific character of the study proposed,the numerous cases observable as proxy (more than 1500publications, for a total of almost 2000 collaborations inthe 2000–2003 triennium under examination) certainlyguarantee a level of significance that could not be reachedthrough alternative approaches, for example those basedon listings of patents authored by academic scientists butowned by private firms, or on sample-based surveys.The next section of this report presents the data set usedin the study, while Section 3 depicts the mapping of collaboration, by area and disciplinary sector. Section 4, inreference to the second objective of the study, presents theanalysis of the qualitative–quantitative impact of colla-boration with private sector colleagues on the researchperformance of academic scientists. Section 5 explores thelevel of multidisciplinarity of research projects in private– public co-authorship, while the last section closes the workwith a brief synthesis and the final thoughts of the authors. 2. Data set As noted in the introduction, the investigation of thephenomenon of collaboration in academic research typi-cally considers scientific publications in international journals that are co-authored by universities with anyother type of organization: other universities, publicresearch laboratories, domestic companies, organizationsfrom other nations, etc. However, for the objectives of thisparticular study, the data set under specific investigationconsists of publications in co-authorship with domesticindustry.The source of reference is the Observatory of PublicResearch ( Osservatorio sulla Ricerca Pubblica , or ORP)which registers, for the 2001–2003 triennium, the interna-tional scientific production of all Italian universities. TheORP is in turn based on the data of the Thomson ScientificSCI TM , Cd-Rom version. In order to assess to what extentthe SCI TM and consequently ORP can serve as representa-tive of the academic research outputs in the ‘‘hard’’sciences, a verification was made by Abramo et al.(2009). The articles in international journals indexed inORP amount to an average of 95% of the total outputssubmitted by the Italian universities in the first and onlyItalian research evaluation exercise. Selecting every listedpublication with at least one address corresponding to anItalian university, the ORP then applies a disambiguationalgorithm to attribute the publication to its respectiveacademic authors. For details see Abramo et al. (2009).Since Italian university research persons are subdivided byscientific disciplinary sector (SDS), it is possible to linkeach publication (and each collaboration) to the SDSs towhich the university authors appertain. The SDSs aregrouped in macro-university disciplinary areas (UDAs).The field of observation for the present analysis considerseight technical-scientific UDAs (mathematics and compu-ter sciences, physics, chemistry, earth sciences, biology,medicine, agricultural and veterinary sciences, industrialand information engineering) including 183 SDSs. Thislevel of detail permits overcoming several distortionstypical of aggregate analyses that do not give dueconsideration to the different ‘‘fertility’’ of scientificdisciplines and the different representivity by disciplinewithin the journals that are listed in the source database(Abramo et al., 2008). ARTICLE IN PRESS G. Abramo et al. / Technovation 29 (2009) 498–507  500  The task of listing the publications of interest, i.e. thoseco-authored by universities and domestic companies, alsoimposed the identification and consistent rendition of allthe possible names of domestic firms present in the addressfield of publications listed in the ORP.The work here is unique with respect to the internationalstate of the art for at least two features, firstly for its broadfield of observation; studies in the previous literature haveonly been based on limited samples of the population of interest, and have tended to focus on restricted disciplinarysectors or single institutions. Instead of these approaches,the study proposed here refers to the entire population of all academic research scientists from all technological-scientific fields, being a total of 33,000 scientists. Secondly,the study is unique for the method used, of categorizingeach collaboration and comparing individual performance;each scientist has been individually identified, thenclassified and grouped by role and scientific field of specialization. This permits the limitation of otherwiseinevitable distortions in productivity measurement due tonon-homogeneity of units under comparison (see Abramoand D’Angelo, 2009). The analysis is based on the entirepopulation of Italian university research staff and thusavoids problems in robustness and significance of infer-ential analyses. It further presents an undeniable advantageof objectivity and homogeneity in the source data, notalways found in examinations based on questionnaires. 3. Sectorial mapping of university–industry collaboration inresearch Overall, there were 791 domestic companies (the legalentities considered are private companies located in Italianterritory. The following have been excluded: publiclyowned organizations, mixed public–private consortiumsand foundations) in the 2001–2003 triennium, whichrealized at least one international scientific publicationlisting in the ORP. Of these, 483 collaborated at least oncewith an Italian university. On the other side, 63 out of 68universities collaborated with industry, in the areas underexamination. Such collaboration resulted in 1534 articles,approximately 3% of the over 52,000 articles bearing thenames of university researchers. Each article can indicatemore than one collaboration, in function of the number of universities and private firms present in the address field of the article itself. There are four possible cases:   one university, one corporation  ¼  one collaboration;   m  universities, one corporation  ¼  m  collaborations;   one university,  n  private firms  ¼  n  collaborations and   m  universities,  n  corporations  ¼  mn  collaborations.As a whole the 1534 co-authored articles embed 1983collaborations, of which 1195 (60%) are of the first type,646 (33%) of the second type, 92 (5%) of the third type and50 (2%) of the fourth type.To quantify the level of intensity of collaborationbetween universities and private companies in the variousscientific sectors, four types of indicators were taken intoconsideration:   The number of university articles in co-authorship withprivate researchers, in a given SDS/UDA.   The percentage of articles in co-authorship with privateresearchers, out of the total of articles realized in thespecific SDS/UDA.   The percentage of articles in co-authorship with privateresearchers, out of the total articles realized in co-authorship, in the specific SDS/UDA. By this indicatorwe can see to what extent public–private collaboration issector specific.   The number of articles in co-authorship with industryper researcher in the specific SDS.Table 1 presents the data relative to the analysis bydisciplinary area. Double counting of articles may occurhere because an article may fall in more than onedisciplinary area. In terms of mass (number of articles inco-authorship), the medicine and chemistry areas dom-inate. Referring to the other two (normalized) indicators, itis industrial and information engineering that leads, andthe ranking of the first four disciplinary areas is invariant:for industrial and information engineering, over 6% of publications bear the joint signature of university scientists ARTICLE IN PRESS Table 1Ranking of the top four university disciplinary areas (UDA) by university–industry collaboration.UDA 1 UDA 2 UDA 3 UDA 4Number of articles inuniversity–corporation co-authorshipMedicine (416) Chemistry (415) Industrial and (358)Biology (308)Percentage of articles in co-authorship withprivate sector out of the total UDA articlesIndustrial and (6.4%)Chemistry (3.9%) Agricultural and veterinarysciences (2.8%)Biology (2.8%)Percentage of articles in co-authorship withprivate sector out of the total UDA articleswith co-authorshipIndustrial and (10.6%)Chemistry (5.7%) Agricultural and veterinarysciences (4.4%)Biology (3.9%) G. Abramo et al. / Technovation 29 (2009) 498–507   501  and researchers from private firms. In quite distant secondplace arrives chemistry (3.9%) and in sequence, agriculturaland veterinary sciences and biology (2.8%). Medicine,which leads the rankings for number of articles inuniversity–industry co-authorship is placed below fourthposition in the normalized ranking for overall scientificproduction.The details by individual scientific disciplinary sector arepresented in Tables 2–4. Ahead of all others, electronics isthe sector with the most articles with co-authorshipbetween universities and corporations, numbering a full114 (Table 2), followed next by internal medicine (109),pharmacology (94) and biochemistry (89). Below the fourthposition we find five disciplinary sectors from the chemistryarea and one from physics (experimental physics).Considering, in each SDS, the rating for incidence of articles in co-authorship with the private researchers as apercentage of total scientific production by all universityscientists in the same SDS (Table 3), electronics recedes tothe third place (12.6%), overtaken by the sector of energyand environmental systems (15.0%) and by polymermaterials science and technology (13.3%). In the first 10positions, we find eight sectors from the industrial andinformation engineering disciplinary area and two fromchemistry.The domination of the industrial and informationengineering area is also observed in the rankings forpercentage of articles realized in co-authorship with privatesector researchers out of the total of articles with co-authorship (Table 4). In the first 10 positions, a full nine,and among these the first seven, are occupied by SDSs fromthis area. In general, it is possible to observe that thesectors concerned, as could be expected, are those directedtowards applied science. This can probably be retraced tothe structure of the Italian productive system, whichprimarily articulates around small and medium enterprises,operating for the most part in ‘‘non-high-tech’’ areas andtherefore more inclined to collaborate with universities if this involves research projects with a prevalently practicalapplication.Relating the scientific production realized in co-author-ship with private sector researchers to the number of  ARTICLE IN PRESS Table 2Ranking of top-ten scientific disciplinary sectors (SDS) by number of articles in university–industry co-authorship; the UDA for each sector isindicated in parentheses.SDS UDA Number of articlesElectronics Industrial andinf. engineering114Internal medicine Medicine 109Pharmacology Biology 94Biochemistry Biology 89Industrial chemistry Chemistry 78Organic chemistry Chemistry 76Experimental physics Physics 74General and inorganic chemistry Chemistry 73Pharmaceutical chemistry Chemistry 69Physical chemistry Chemistry 60Table 3Top-ten SDS ranking by percentage of university–industry co-authoredarticles out of total SDS articles.SDS UDA Academic articlesin co-authorshipwith industry, outof total SDSarticles (%)Energy and environmentalsystemsIndustrial and inf.engineering15.0Polymer materials scienceand technologyChemistry 13.3Electronics Industrial and inf.engineering12.6Aerospace installations andsystemsIndustrial and inf.engineering11.1Primary materialsengineeringIndustrial and inf.engineering11.1Industrial chemistry Chemistry 10.9Electrical systems forenergyIndustrial and inf.engineering10.9Hydrocarbons and groundfluidsIndustrial and inf.engineering10.0Applied physical chemistry Industrial and inf.engineering9.8Electrical and electronicmeasurementIndustrial and inf.engineering9.7Table 4Top-ten SDS ranking according to the percentage of university–industryco-authored articles out of total SDS co-authored articles.SDS UDA Academic articlesin co-authorshipwith industry, outof total SDS co-authored articles(%)Energy and environmentalsystemsIndustrial and inf.engineering30.0Manufacturing technologyand systemsIndustrial and inf.engineering21.4Electrical energy systems Industrial and inf.engineering21.2Commodities engineering Industrial and inf.engineering20.0Applied physical chemistry Industrial and inf.engineering17.9Electronics Industrial and inf.engineering17.8Electrical and electronicmeasurementIndustrial and inf.engineering17.7Environmental chemistry Chemistry 17.1Aerospace constructionand structuresIndustrial and inf.engineering16.7Aerospace systems andplantsIndustrial and inf.engineering16.7 G. Abramo et al. / Technovation 29 (2009) 498–507  502
Related Search
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks