Elsevier

Journal of Informetrics

Volume 4, Issue 4, October 2010, Pages 540-553
Journal of Informetrics

The effect of scholar collaboration on impact and quality of academic papers

https://doi.org/10.1016/j.joi.2010.06.003Get rights and content

Abstract

We study how scholar collaboration varies across disciplines in science, social science, arts and humanities and the effects of author collaboration on impact and quality of co-authored papers. Impact is measured with the aid of citations collected by papers, while quality is determined by the judgements expressed by peer reviewers. To this end, we take advantage of the dataset provided by the first-ever national research assessment exercise of Italian universities, which involved 20 disciplinary areas, 102 research structures, 18,500 research products, and 6661 peer reviewers. Collaboration intensity neatly varies across disciplines: it is inescapable is most sciences and negligible in most humanities. We measured a general positive association between cardinality of the author set of a paper and citation count as well as peer quality of the contribution. The correlation is stronger when the affiliations of authors are heterogeneous. There exist, however, notable and interesting counter-examples.

Introduction

Collaboration is a fundamental and common feature in scientific research. It assumes various forms, ranging from sharing of ideas among researchers to corporate partnerships and research joint ventures. Collaboration arises at different levels within the research system: micro-level (individuals, research groups), meso-level (departments, institutions), and macro-level (institutional sectors, in particular collaborative agreements between university and industry, or regions). Collaboration is encouraged by institutions, funding bodies and policymakers for a number of positive factors that are largely discussed in literature (Bordons and Gómez, 2000, Katz and Martin, 1997, Luukkonen et al., 1992, Sonnenwald, 2007, Subramanyam, 1983) and it is frequently organized by scientists themselves.

There are both scientific and extra-scientific advantages of collaboration. Researchers can derive scientific advantages by sharing knowledge, expertise and techniques, jointly controlling the accuracy and the significance of results, restricting isolation and giving substance to the cross-fertilization of ideas. For instance, theoretical investigation, experimental analysis, and compelling and elegant writing are all assignments that require different skills that are rarely enjoyed by the same scholar. Collaboration allows to cope better with the increasing specialization in science, with multidisciplinary approaches, and with the complexity of scientific instruments. Collaborative research has been associated with higher productivity: from an economic perspective, collaboration allows the division of labor leading to reduced costs and time saving and consents the access to scientific funding, to expensive (possibly large-scale) equipment, and to unique scientific data.

Furthermore, collaboration enhances visibility of results. By means of collaborative behaviour in co-authorship, the article is brought to the attention of a larger number of researchers through personal contacts, either formally (pre-print posted on personal or institutional repositories, seminars, conferences) or by informal conversations. The visibility of the contribution is strengthened when co-authorship is carried out in more than one institution, and, in particular, international collaboration plays a more relevant role than domestic collaboration. This is in accordance with the view of co-authorship as a social relationship and with the social world described by Mark Granovetter is his famous paper The Strength of Weak Ties (Granovetter, 1973). Weak ties, that are links to a different community, play a more important role than links to our close friends, who inevitably move in the same circles and are exposed to the same information.

Collaboration, however, has not only advantages. A collaborative work needs deep integration among co-authors, since the final result should be a coherent and uniform piece of work. If integration among authors fails, the quality of the outcome definitely declines. Collaboration can entail time costs in jointly formulating the research problems, in deciding how to divide work or in keeping all the collaborators informed of the advances (Katz & Martin, 1997). Writing the results jointly may bring to disagreements among colleagues about findings and interpretation. As a consequence, time has to be spent in tuning up and gaining upon differences of opinion. Since collaboration is a social process it may imply a patient construction of personal relationships or adaptation in an unfamiliar environment, that again has a cost in terms of time. Collaboration may have operating costs if individuals have to move to different sites where parts of a research are developed or if equipments have to be transported. Furthermore, collaboration may be a restrain in career advancement for junior scientists, since their contribution to research products may be underestimated when their professional curriculum vitae is assessed, especially if they co-operate with well-known scientists (Merton, 1968, Sonnenwald, 2007).

Co-authorship in publications is widely considered as a reliable proxy for scientific collaboration. It has expanded in all academic fields in the last decades (Cronin et al., 2003, Larivière et al., 2006, Moody, 2004, Persson et al., 2004). Several bibliometric studies have explored the effect of co-authorship on the citational impact of individual articles, that is, the impact of the contribution measured by the number of citations received from other papers. They point out a general positive correlation between the number of authors and the number of citations received by a paper: the more authors a paper has, the more citations it receives. This is, however, not true under all circumstances and for all scientific domains, since some studies have provided no support for a link between collaboration and citations. Citational impact is also partially related to the heterogeneity of collaboration in terms of involved research institutes: empirical evidences show that papers with multi-institutional collaborations are generally more cited than articles made in-house (in the same research institute).

In studying the effects of scientific collaboration, we feel that a primary issue is whether the increase of citational impact observed when the number of authors or institutions grows is ascribable to a real enhancement of paper quality or it is due to various advertising factors, like the obvious greater visibility and higher amount of self-citations of papers with many authors coming from different institutions. We assume that there are two separate dimensions underlying the impact of a co-authored article:

  • A qualitative factor, that is the intrinsic value of the article in terms of originality, significance, depth, correctness, completeness, and clarity. It encapsulates the scientific acknowledgement by which an article deserves to be endorsed.

  • An advertising factor, consistent with the view of co-authorship as a social relationship. As Goldfinch, Dale, and DeRouen (2003) noticed, “it may be that citation rates for an article are not simply a reflection of quality, but to some extent reflect the access to greater social networks the co-publication can allow”.

The main purpose of the present investigation is to study how collaboration, at the micro-level of paper co-authorship, influences both the impact and the quality of academic contributions. More precisely, our investigation is shaped by the following research questions:

  • 1.

    How does scholar collaboration vary across disciplines?

  • 2.

    What is the effect of scholar collaboration on citational impact and quality of papers?

  • 3.

    Does heterogeneous collaboration enhance citational impact and quality of papers?

We took advantage of the comprehensive dataset of the first Italian research assessment exercise (VTR), which covered about 18,500 research products published from 2001 to 2003 in scientific-disciplinary areas covering science, social science, arts and humanities (Abramo et al., 2009, Franceschet and Costantini, 2010, Reale et al., 2007). VTR was a peer review exercise: each submitted product was judged by at least two referees who assigned it a quality rating. Furthermore, for scientific areas in which the standard publication is a journal article (mainly sciences and a limited part of social sciences) most of the submitted products are indexed in Thomson Reuters Web of Science (WoS) database. Thus, for all submitted products we have access to the qualitative judgement given by peer reviewers, which is a direct indicator of quality, and for a significant subset we have at disposal the number of citations they received from other papers in the WoS database, which can be regarded as an indicator of impact. This provides the possibility to analyse, on a large range of disciplines, both the relationship between co-authorship and paper citational impact and that between co-authorship and paper quality.

Section snippets

Dataset and methodology

The first Italian research assessment exercise, VTR, was managed by the Committee for the Evaluation of Research (CIVR) and was designed as an ex post assessment exercise based on peer review. CIVR divided the national research system into 20 scientific-disciplinary areas, 6 of which were interdisciplinary sectors, and set up an evaluation panel responsible for the assessment of each area. The exercise was then articulated in three phases, that were in charge of research structures, panels and

How does scholar collaboration vary across disciplines?

In this section we study how collaboration, at the author level, varies across disciplines. We adopt two attributes generally used in collaboration studies (Levitt & Thelwall, 2009). The first attribute, the collaborative rate (CR), is the proportion of papers with more than one author. The second, called collaborative level (CL), is the average number of authors per paper (CL). Since CL may be biased because of distortions caused by highly collaborative articles, we introduce two further

Related work

Moody (2004) observes that the propensity of scholars to collaborate widely differs among disciplines, and it is more common in natural sciences than in social sciences. A clear dichotomy between sciences and humanities pertaining to papers collaborative rates has been detected by Larivière et al. (2006). They computed the collaborative rate for different scientific fields, using the whole set of Canadian papers that were published during period 1980–2002. They noticed that almost all articles

Conclusion and future work

With the aid of a large-scale dataset taken from the first national research assessment exercise in Italy, we have investigated the effect of scholar collaboration in different disciplines of science, social science, arts and humanities. Our main conclusions can be summarized as follows.

The intensity of research collaboration is negligible in arts and humanities: the set of paper co-authors is frequently a singleton. Social science researchers often work in team, sharing competencies and other

Acknowledgements

The authors would like to thank CIVR and its President Franco Cuccurullo for making available data used in this paper, through the agreement protocol between CIVR and PhD course in “Strumenti e metodi per la valutazione della Ricerca” of the University of Chieti-Pescara. The first author is partially supported by PRIN 2008 project “Innovative and multi-disciplinary approaches for constraint and preference reasoning” (20089M932N). The second author thanks the staff of Evaluation and Control

References (28)

  • G. Abramo et al.

    Allocative efficiency in public research funding: Can bibliometrics help?

    Research Policy

    (2009)
  • R.L. Hart

    Collaboration and article quality in the literature of academic librarianship

    The Journal of Academic Librarianship

    (2007)
  • J.S. Katz et al.

    What is research collaboration?

    Research Policy

    (1997)
  • M. Medoff

    Collaboration and the quality of economics research

    Labour Economics

    (2003)
  • D.W. Asknes

    Characteristics of highly cited papers

    Research Evaluation

    (2003)
  • M. Bordons et al.

    Collaboration networks in science

  • B. Cronin

    Hyperauthorship: A postmodern perversion or evidence of a structural shift in scholarly communication practices

    Journal of the American Society for Information Science and Technology

    (2001)
  • B. Cronin et al.

    A cast of thousands: Coauthorship and subauthorship collaboration in the 20th century as manifested in the scholarly journal literature of psychology and philosophy

    Journal of the American Society for Information Science and Technology

    (2003)
  • Franceschet, M., & Costantini, A. (2010). The first Italian research assessment exercise: A bibliometric perspective....
  • W. Glänzel

    Coauthorship patterns and trends in the sciences (1980–1998): A bibliometric study with implications for database indexing and search strategies

    Library Trends

    (2002)
  • W. Glänzel et al.

    Analyzing scientific networks through co-authorship

  • S. Goldfinch et al.

    Science from the periphery: Publication, collaboration and ‘periphery effects’ in article citation rates of the New Zealand Crown Research Institutes 1995–2000

    Scientometrics

    (2003)
  • M.S. Granovetter

    The strength of weak ties

    American Journal of Sociology

    (1973)
  • H. Herbertz

    Does it pay to cooperate? A bibliometric case study in molecular biology

    Scientometrics

    (1995)
  • Cited by (205)

    View all citing articles on Scopus
    View full text