Co-authoring and Influential Academic Articles

(1)

Co-authoring and Influential Academic Articles

Master Thesis Strategy and Innovation

Nila Mahendra Sari 1558293

MahendrUniversity of Groningen

(2)

Summary

Most academic publication is published in journal, book or thesis form. Many academic publishing relies on some form of peer review or editorial refereeing to qualify texts for publication. The aim of every author of any academic journal is to make an influence. This study is going to be focused on influential output generated by article which is written by group of authors.

Fox and Faver (1984, p. 351) assert that “working with others, can create a social context and reality for the research which, in turn allows for better assessment of the project” and higher quality research. Many universities and grant-giving agencies promote and reward collaboration, especially interdisciplinary collaboration, because of the belief it has a positive effect on research quality. Other advantage which concluded by Williams and O’Reilly (1998) is that the teams made up of members from a variety of functional areas perform at a higher level than teams that do not have that diversity. Co-authorship can also generate disadvantage; some collaborators may supply less effort.

There are three variables discussed in this thesis paper. The first variable is communication. Some empirical studies show that the communication processes within a team are critical for team success (Keller, 2001; Lovelace, Shapiro, &Weingart, 2001). The second variable is mental models concern knowledge that involves attitudes, beliefs, ideas and expectation that is used to coordinate behavior. Yen et al. (2006) argue that “creation of a common understanding is necessary for a group performance”. The third variable is role because an effective way of dealing with the need for human resource without expanding the size of the team is through configuring team member roles.

In future research, it might be interesting to find out the elements of “trust” among co-authors.

The relationship between time consumption and producing influential article, the issue of seniority and producing influential articles and the possible conflicts which are possible to appear in co-authorship like intellectual property rights are also worth to discuss.

(3)

Chapter 1 Introduction

In order to create a publication, authors have choices whether they want to write it alone or in a group. Publication is any writing which is published; including books, magazines, articles, and newspapers. Most academic publication is published in journal, book or thesis form. Many academic publishing relies on some form of peer review or editorial refereeing to qualify texts for publication. The aim of every author of any academic journal is to make an influence. Publishing a good quality paper is already a noble aim, but if the paper has no influence, then the activity would be pointless. But then we have to classify to whom the article is giving influence to, whether to other academics, students, or practitioners. Unfortunately, it is not easy to measure the influence on practitioners and students. This study is going to be focused on influential output generated by article which is written by group of authors.

Writing articles in group, usually called as co-authorship, is believed to bring about many benefits, for example, cost savings and higher impact research (Hicks and Katz, 1997).

On his paper, Endersby (1996) marked that “the quality of research can be improved with joint participation. The strength of two or more cooperative researchers can be fused to produce improved, holistic results than each individual researcher acting independently.”

More advantages generated from co-authorship are going to be discussed further in the next chapter.

During the first half of the twentieth century, writing scientific papers with more one author were relatively rare. Scientific research was considered the work of solitary individuals, published in single-authored articles. However, the situation has changed significantly in recent decades. The trend to co-authorship was originated in and continues to be most closely associated with the natural sciences (Biagoli and Galison, 2002; Cronin, 2001; Price, 1963), but has been increasing steadily in the social sciences (Laband and Tollison, 2000; Moody, 2004). As a consequence, there has been increasing academic interest in the phenomenon of collaboration among scientists (Barabási et al., 2002; Barnett et al., 1988)

(5)

1.1 Research Outline

The first part of the research is ordered as follows; after the introduction, in the second chapter, co-authoring and impact factor are discussed to establish a solid understanding. Advantages and disadvantages of co-authoring, reasons why people choose member of co-authorship are also discussed in the second chapter. Subsequently, in the third chapter, a conceptual model is put forth. Next, still in the third chapter, the antecedents of co-authoring influential articles result are studied. When conducting research on these antecedents of co-authoring, several hypotheses are established.

In the second part of the research, starting from chapter four, research methodology is explained.

Research variables, population, sample and research approach are discussed in this chapter.

Analysis is going to be held in chapter five. Asserted hypotheses are tested empirically. Finally, conclusion will be drawn up, limitations inherent to the research are discussed and recommendations for future research are presented.

1.2 Research formulation

This thesis is an explorative research focusing on providing better understanding on under what circumstances does a co-authoring produce influential article. The involvements of multiple authors make it more complicated to produce an article. Every member of the co-authorship would affect the successfulness of producing influential article. Therefore it is interesting to discuss the implementation of co-authoring in producing articles.

1.2.1 Research objectives

The thesis paper will test hypotheses from literatures about co-authoring, number of citation received by an article, and the pattern of authorship to find out under what circumstances does co- authoring produce influential article. Influential will be defined as possessing power exerted over the minds or behavior of others. In this case, it would be measured by number of citation received by an article.

1.2.2 Research questions

By analyzing the facts mentioned above that there are some reasons why authors write articles in do co-authorship. Then, the research question is formulated:

Under what circumstances does group work produce influential article?

(6)

Chapter 2 Co-Authoring

A scientific document is called co-authored if it has more than one author. A linkage or tie between two or more individuals is established by a co-authorship. These linkages can be examined as a social network and patterns exhibited in the social network of an individual and his co-authors. The result of the examination can produce an analysis on how an author works and deals with his colleagues.

2.1 Advantages and disadvantages of co-authoring

Beaver and Rosen (1978) say that collaboration grew historically as science became

“professionalized”—taking place in dedicated institutions of science. Using an historical, nationally-based approach, they show that “collaboration becomes a mechanism for both gaining and sustaining access to recognition in the professional community” Beaver and Rosen (1978).

They refute other authors who claim that collaboration is historically recent, or that it is principally a response to specializations. Some advantages which can be generated by co- authorship has been mentioned, such as cost saving, higher impact research (Katz and Hicks, 1997), and the strength of two or more cooperative researchers can be fused to produce improved, holistic results than each individual researcher acting independently (Endersby, 1996).

There are more advantages generated from co-authoring. Fox and Faver (1984, p. 351) assert that

“working with others, can create a social context and reality for the research which, in turn allows for better assessment of the project” and higher quality research. Many universities and grant- giving agencies promote and reward collaboration, especially interdisciplinary collaboration, because of the belief it has a positive effect on research quality. So co-authorship receive grant from agencies or universities easier than sole- authorship. Other advantage which concluded by Williams and O’Reilly (1998) is that the teams made up of members from a variety of functional areas perform at a higher level than teams that do not have that diversity.

Co-authorship can also generate disadvantages; the first is that some collaborators may supply less effort. Some co-author may not take responsibility for the contents of the article relating in a potential loss of quality. Co-authors may make minor or insignificant contributions in the production of research. In some cases the contributions may only be providing a data set or experience in navigating the journal publication process. This may increase publication output, but does not necessarily result in a higher quality paper. Another problem could also be generated

(7)

by doing a co authorship is regarding intellectual property rights. Sole authorship protects a researcher’s intellectual property rights for a critical or innovative scientific contribution, but not on co-authorship, authors have to really make a agreement among them how to deal with the intellectual property rights issues. When a sole author takes full responsibility for ideas presented in an article, co-authors have to divide responsibilities for ideas presented in an article.

2.2 Choosing member of a co-authorship

Having complimentary skills is important reason why authors form a certain group do a research together. One of them might be very good in statistics while the others are good with other things which they can contribute to the research articles. Hudson (1996, p. 157) maintains that “the gains from collaborative work result…….. From a sort of synergy where multiple contributors develop ideas that none would have developed on his or her own. Synergy differs from skill complementarities in the sense that it can exist between individuals with very similar skill sets.

When collaborative work draws upon such complementarities and synergies, it is most likely to represent a gain in knowledge to the economics profession.” Pelz and Andrews (1976, p. 52) contend that collaboration enhances the quality of research in a variety of ways. One is “by providing new ideas- jostling a man out of his old ways of thinking about things. Sometimes a colleague may know something another man needs to know. Then there is the possibility of a colleague catching an error which the man himself is too engrossed to see. Still another way is keeping a person on his toes- like …..running a test the way it should be done…”

In Amabile (1996) mentioned that time pressure does not fuel creativity but in co-authoring authors can remind each other to meet a certain deadline. Another dimension for analysis is to examine the roles assigned to the various group participants. The group member can be selected based on a role needed in the group, such as the ones defined by Fish et al. (1988): co-author, commenter and reader. A second example of roles is to have a single person responsible for the actual writing (scribe) and the rest of the group contributing with ideas, comments and suggestions (consultants). A third option is to have a single role of co-author taken by all participants, as is the case for the production of many scientific articles: each co-author is responsible for contributing text, commenting on the other co-authors’ work and improving the writing style.

Sometimes authors gather to write a certain article based on preferences. They write article together because they like to work together. Collaboration may be motivated by desire to alleviate

(8)

academic isolation since it offers opportunities for friendship. However, researchers are typically reluctant to co-author with those they do not get along with. Working with other researchers because of the opportunities for friendship may not be optimal method of selecting a collaborator if the desired outcome is high quality research (Medoff, 2003).

But on the other hand, contrary to Medoff (2003), Newcomb & Bagwell (1995) state that friends are often, but not always, found to outperform non friends. Theorists posit that advantages of friendship arise, in part, because friends create a psychological context that is cooperative and facilitates conflict resolution (Azmitia, 1996).

Friends and non friends do not differ in the frequency of disagreements, but friends are more concerned with resolving disagreements (Newcomb & Bagwell, 1995). Friends agree more during collaboration (Zajac & Hartup, 1997), know each other better (Ladd & Emerson, 1984), and are more similar than are non friends (Haselager, Hartup, van Lieshout, & Risken-Walraven, 1998). In friends’ interactions, one individual is less likely to dominate the other because friends engage in “balanced” rather than domineering interactions (Newcomb & Bagwell, 1995).

Authors with limited experience usually look for a chance to write articles with their senior who has more experiences. It is a moment where they can learn something from their senior. Beaver and Rosen (1978) mentioned that collaboration is intrinsically advantageous to scientists, they argue, particularly when it occurs between a “master” and an “apprentice”.

2.3 Linking the antecedents of co-authorship

This previous discussion can be rephrased in terms and known developed in the literature on teams involved in creating new knowledge.

2.3.1 Communication

In generating innovations, the knowledge and experiential bases of each team member must be activated and utilized by the team as a whole (Taggar, 2002). Most of the literatures about group working see communication as a first variable of interaction among team members (e.g.

Anderson & Narus, 1990; Jarvenpaa & Shaw, 1998; Mohr, 1999; Kratzer et al., 2004). Anderson and Narus define communication as “the number of formal as well as informal times in which meaningful and timely information is shared between persons”. Bleeke and Ernst (1993) say that

(9)

interaction among team members is the most important factor to bring successful cooperation;

“the most accurately designed relations will crumble without good, frequent interaction”.

Some empirical studies show that the communication processes within a team are critical for team success (Keller, 2001; Lovelace, Shapiro, &Weingart, 2001). A goal-oriented and well- coordinated innovation process requires that the team establishes a shared and valid task model and cooperation model (Edmondson, 1999; Gilson, Shalley, & Milne, 2002; Mathieu, Goodwin, Heffner, Salas, & Cannon-Bowers, 2000; Mohammed & Ringseis, 2001). Based on this understanding, the team can subsequently decide on which steps are to be taken in what order and how the workload should be distributed among team members. The process of team reflexivity (Schippers, Den Hartog, Koopman, & Wienk, 2003; West, 2000)—defined as the overt reflection of objectives and team processes—can be interpreted as a continuous monitoring and adjustment of the established task and cooperation model. Such adjustments become particularly important when process conflicts (Jehn, Northcraft, & Neale, 1999) arise, i.e., when team members differ with respect to the distribution of competencies and responsibilities among the team.

A group work which has a good communication among members is expected to achieve positive outcomes and likewise convey information timely and accurately. Interpersonal communication is the basic need of a good group work. A good communication can lead to get work done on time, make decisions, and coordinate teamwork. Therefore communication is, one of the primary responsibilities for every group member.

The connections among members in organizations can vary in terms of their strength. A strong tie describes a relationship that is characterized by closeness, reciprocation, and substantial time spent with the other party while a weak tie is a relationship that does not have this level of closeness and connection (Granovetter, 1973). The nature of the ties within a network leads to foretellable outcomes. Strong ties bond similar people to one another and results in mutual attachments among the network members. In contrary, networks characterized by weak ties tend to be more diffuse and have fewer reciprocal connections. Surprisingly, the most beneficial is the mix of strong and weak ties (Krackhardt, 1992). At first, we may not have a plan on with who are we going to build a strong tie nor a weak tie, but we can predict the needs of human resources to produce an articles. So, it is considerable for first author to mix the co-authors who have strong and weak tie with him/her. A first author may choose a second author who has work with him for several times and has good experiences with him. Then for the next authors, he can choose

(10)

somebody who he does not really know yet but has the capability to fulfill the needs of human resources for the article or the access to the data needed for the article.

Hypothesis 1: frequency of communication is positively related to producing influential article 2.3.2 Shared mental model

As the second variable, mental models concern knowledge that involves attitudes, beliefs, ideas and expectation that is used to coordinate behavior. Yen et al. (2006) argue that “creation of a common understanding is necessary for a group performance”. Co-authorship writing contrary to sole- authorship will not totally posses these shared perceptions or expectations. Co-authorship is heterogeneous contrary to sole authorship which is homogenous. Co-authorship members encompass members that stem from diverse organizations and cultures and moreover frequently possess dissimilar knowledge, this influencing the sharing of mental models.

Low degrees of shared attitudes, belief, ideas, and expectations among group members, the understanding level among authors will be very low whereas the novelty and newness value of ideas will be very high. Therefore, the chance to obtain influential article is higher. Conversely, High degree of shared attitudes, beliefs, ideas and expectations among team members, the understandability of members within the group is very high. The consequential organizational myopia or groupthink (Kratzer et al., 2006; Wuyts et al., 2005) will result in significantly lowered levels of influential articles.

Hypothesis 2: Degree of shared mental models will have a negative relationship with producing influential article.

2.3.3 Role

An effective way of dealing with the need for human resource without expanding the size of the team is through configuring team member roles. Team members can be assigned limited roles on a team. There are four mechanisms for doing this (Ancona and Caldwell, 1997):

- First, team can bring experts into a team for limited time or for a very specific aspect of the project. This allows the team to make use of the information or expertise of critical outsiders without having to integrate those individuals into the group.

- Second, teams can shift composition over the life cycle of the project. That is, individuals who have specific information or external contacts can be assigned to a team for a limited

(11)

tie. Some members of the team could come and go, based on the boundary activities necessary in a particular time.

- Third, some members could be assigned a part time role on a team. Individuals whose expertise or contacts are needed over the length of the project, but who may have other demands on their time or somewhat limited knowledge of broader project issues, could serve on the team on a part time basis.

- Fourth, the decision- making roles of team members could be differentiated. For teams working on highly complex, interdependent projects, the need for information and coordination with other groups may be too high to be accomplished exclusively through the boundary management or the part-time or part-cycle involvement of some members.

Such situations may require the expansion of the team. However, as the team grows in size difficulties in decision making and coordination may arise. One response this is to develop a two-tiered membership made up of relatively small member of core members- who play a major role in decision making- and a larger number of peripheral members- who play a more limited role, but are none the less full- fledged tem members.

There are also some possible configurations of team members in co-authorship relationship which are believed would increase the successfulness of producing an article. Said et.al (2007) examined a number of author–coauthor networks in order to see the role of authors and their co authors. Based on the analysis, four basic styles of authorship were identified, they are, solo, entrepreneurial, mentor, and laboratory. Solo style authorship characterizes authors who do not have any coauthors. The solo style authorship is not common and usually found among isolated scientists. Usually on the fields that are heavily mathematical are often the works of a single individual and therefore often follow the solo style.

The first style mentioned by Said et al. (2007) is called as entrepreneurial style because in essence the principal author has sought out groups with sub-specialties needed for a particular paper. In essence these specialty groups represent sub disciplines and are tightly coupled within the subspecialty. Co-authorship relation called as entrepreneurial style when authors are on the same level of expertise but in different field or sub-field.

The second style mentioned by Said et al. is called as the mentor style. This principal author tends to co-author papers with younger colleagues who were his students or other young associates. His basic strategy is to work with associates individually getting them started in scholarship. Once

(12)

they have the experience of writing a few papers, they begin to write on their own or with other colleagues in a new venue. The principal authors can have more than 10 articles in the same time with this type of co-authorship.

The third one is called the laboratory style. The main authors in this type of co-authorship have many co-authors. Usually all of the members of a particular laboratory project have their names on the article.

Hypotheses 3: Mentor style co-authorship is the best co-authorship style in producing influential article in social sciences

Based on the explanation above over some factors which can not be left out which are frequency of communication, degrees of shared mental models and role of authors, I developed a model to make everything easier to see. Below is a model which I am going to use for this thesis:

Figure 1:

Source: self developed model Frequency

of communication

Degree of Shared Mental

Model

Authors’

Role

Influential article

(13)

Chapter 3 Research Methodology

The thesis will be an exploratory, qualitative research based on primary data from semi-structured interviews and secondary data from academic literatures for theories to be applied in the analysis.

The literature study is based on the theories on publication, co-authorship, journal evaluation, and article citation.

3.1 Influential article

Sternberg (1993) gives list of what makes an article influential, they are:

- The article contains one or more surprising results that nevertheless make sense in some theoretical context.

- The results presented are of major theoretical or practical significance

- The ideas are new and exciting, perhaps presenting a new way of looking at an old problem

- The interpretation of the results is unambiguous

- The article integrates into a new, simpler framework data that had previously required a complex, possibly unwieldy framework

- The article presents an experiment with a particularly clever paradigm or experimental manipulation

- The findings or theory presented are general

Since the list which is proposed by Sternberg is complicated to measure, I have to find other indicator on how to measure how influential an article is. Cole and Cole (1973) were among the first to mention that high-quality research is that which is currently thought useful by one’s colleagues as measured by the number of citations to a particular published work. Stigler and Friedland (1975) argue that the number of citations made to a research publication is an objective measure of research quality. The number of citations of a research article’s has received reflects the degree to which others have made use of that work. Stigler and Friedland (p. 486) note that

“citations are an easy way to transfer the exposition of a theory or problem from your paper to someone else, so in the longer view citations reveal a form of intellectual collaboration. To some degree citations are influence, for they influence the reading by readers of the citing paper”.

Following what Stremerch et. al (2007) pointed out on his article that “prestigious articles are determined from the number of citation they receive”. So the more often the article is cited, the

(14)

more prestigious the article is. On the other hand, there are highly useful articles that are not cited frequently. Scientist read some journals for the same reasons people read newspapers and they may rarely or never cite such journals in their published work (Scott, 1969). There is also a more difficult way to measure degree of influential of an article which is mentioned by Gilbert (1977)

“high quality articles represent bigger breakthroughs and therefore may be path-breaking, and also they present higher reliability than low quality articles”. The point is that higher quality articles give more inspiration for further research than lower quality one.

Citations represent objective evidence that the article cited is germane to the current research frontier of knowledge and relevant to those attempting to extend this frontier. According to Medoff (2003) there are some criticisms against using citations as a measurement for an article’s quality. First, citations may merely be measuring the impact rather than the innovative quality of an article. If an article’s impact is the ability to stimulate further thought then impact (citations) and quality are positively correlated. Second, negative citations are not distinguished from positive citations. Some citations only point out errors in an article. Negative citation cases should be small, however, since once an article is refuted it garners very few citations. Third, volume of self-citation is not unusual in articles because of the continuity of a research and/or because of the uniqueness or novelty of their subject matter. Forth, citations may be used to make an article fashionable. What is fashionable or popular today may be considered out of fashion tomorrow.

Various studies have found that citation counts are a very good indicator of research quality.

Citation counts have been found to be highly correlated with numerous measures of quality such as awards of distinction, prestige of awards, and professional recognition (Cole and Cole, 1973).

Nevertheless, journal rankings are very controversial with the exception of a small number of journals that can be unambiguously regarded as top-notch. Two methods have been commonly employed for journal rankings. The most frequently adopted method is the impact factor based upon journal citations. However, drawbacks with impact factors include biases caused by self- citation and sociological and statistical factors. The other method for journal rankings is based upon surveys of faculty, department administrators, and/or journal article authors.

This study is going to use number of citation received by an article as a sign of how influential an article is, because citation frequency reflects a journal’s value and as a prove that other authors use the article for other research. There are more things which affect why authors cite other authors’ work (Scott, 1969). It is not only because of the quality of the research but also

(15)

dependant on other things. Authors’ reputation is one of the most common reasons, and then the next reasons are: attractiveness of the subject, circulation, availability and extent of library holdings, reprint dissemination, and coverage by secondary services.

Despite of the reason why author cite other author’s work, Hicks and Katz (1997) found out that from the year 1981-1991 articles which is produced by co-authorship received more citation than sole authorship article. They distinguished between co-authorship in one country and co- authorship in multiple countries. I am not going to differentiate between one country co- authorship and multiple country co-authorship.

3.2 Impact factor

Impact factor is a measure of the frequency with which the average article in a journal has been cited in a certain period of time. The impact factor for a journal is calculated based on a three- year period of time, and can be considered to be the average number of times published papers are cited up to two years after publication. Journal Impact Factor is from Journal Citation Report (JCR), a product of Thomson ISI (Institute for Scientific Information).¹

Journal Citation Report provides quantitative tools for ranking, evaluating, categorizing, and comparing journals. The annual Journal Citation Report impact factor is a ratio between citations and recent citable items published. Thus, the impact factor of a journal is calculated by dividing the number of current year citations to the source items published in that journal during the previous two years.²

3.3 Thomson scientific database

Every top journal publishers and editors are committed to provide comprehensive coverage of the world's most important and influential journals to meet its subscribers' current awareness and retrospective information retrieval needs. It would appear that an index to scientific journal literature might be expected to cover all the scientific journals published. It has been demonstrated that a relatively small number of journals publish the bulk of significant scientific results.³

1http://www.sciencegateway.org/impact/, visited on 1^st of August, 2008.

2 http://scientific.thomson.com/free/essays/journalcitationreports/impactfactor/, visited on 1^st of August 2008

3 http://scientific.thomson.com/, visited on 1^st August 2008

(16)

Thomson Sceintific began to publish Journal Citation Reports in 1975 as part of the SCI and the Social Sciences Citation Index. Each year Thomson Scientifics’ editorial staff reviews over 2,000 journals, and selects around 10-12% of the journals evaluated for inclusion in the Thomson Scientific database. Moreover, existing articles coverage in each journal products is constantly under review. Journals are monitored to ensure that they are maintaining high standards and a clear relevance to the products in which they are covered. The journal selection process described here is applied to all journals in Web of Science^®, whether covered in Science Citation Index Expanded™, Social Sciences Citation Index^®, or Arts & Humanities Citation Index^®.

The basic criterion in the evaluation process which a journal has to fulfill is timeliness of publication. It is of primary importance. Timeliness is also primary for electronic journals. If the e-journal is publishing distinct issues at a stated frequency, these issues should appear online in a timely manner. Top journals publishers also note whether or not the journal follows international editorial conventions, to get back the source of the articles. These conventions include informative journal titles, fully descriptive article titles and abstracts, complete bibliographic information for all cited references, and full address information for every author.

Usually citation studies either look at the total numbers of citations received by papers or journals, or they consider impact scores. Impact scores can be calculated in many different ways. Most studies use the Garfield impact factor produced by ISI, later the Thompson Group, which for any journal gives the average number of citations per paper received the previous two years. Whilst useful as a snapshot, such measured have several biases (Glanzel & Moed, 2002; Jennings, 1998), particularly over the width of coverage of the publication, the disciplinary area, and the short length of time within a paper’s citation life.

3.4 Citation Patterns

Every field has its own citation pattern. So a research is meaningful only in the context of article in the same general discipline. ⁴As an example, smaller fields like crystallography do not generate as many articles or citations as do larger fields such as biotechnology or genetics. It also happens in some areas, particularly in the arts and humanities, it may take a relatively long time for an article to attract a meaningful number of citations. But in other areas, such as the life sciences, it is not unusual for citations to accrue rapidly and peak after two or three years. These facts must

4 http://scientific.thomsonreuters.com/free/essays/selectionofmaterial/journalselection/, visited on 1^st of August 2008

(17)

be taken into consideration if citation data are to be used correctly. This is the reason why I am only going to use data from Human Resource and Organizational Behavior sub-field only.

In management articles citation pattern, typically, citations start slowly not least because the citing papers take some time to appear in print. Citations then build up as citations lead to new readers who may also cite it, then citations tail off as the material becomes dated and obsolete (Van Raan, 2004). There are several other patterns that may occur, including “shooting stars’’

and ‘‘sleeping beauties’’. “Shooting stars” refers to articles which are received many citation as soon as the article is published but then decreasing significantly. ‘‘sleeping beauties’’ refers to articles which receive almost no citation at all after it is published but then once it cited, many authors cited the article as well (Van Raan, 2004). In this paper I am going to make limitation for articles which are published between 2000-2003 to make it more specific and to avoid the effect of choosing the “shooting star” and “sleeping beauties”.

3.5 SOM University of Groningen

SOM is the research school of the faculty of Economics and Business at the University of Groningen. It combines and stimulates disciplinary and interdisciplinary research meeting international standards. It provides the training and research environment for PhD students. In 2004, approximately 125 research fellows and 70 PhD students belonged to SOM. SOM was recognized by the Royal Dutch Academy of Arts Sciences in May 1998. This accreditation has been renewed in 2003. The mission of SOM is to stimulate mono-disciplinary and interdisciplinary research related to the firm in its economic environment and to provide a PhD program with a high international standard.

SOM's research program is organized in the following seven themes:⁵

1. HRM&OB (the institute for Human Resource Management & Organizational Behavior)

2. IEB&M (the institute for International Economics, Business & Management) 3. IEEF (the institute for Economics, Econometrics and Finance)

4. Marketing

5 http://www.rug.nl/feb/onderzoek/onderzoeksinstituten/index, visited on 3^rd of September 2008

(18)

5. OPERA (the institute for Operations Management & Operations Research) 6. Management Accounting

7. B&IS (the institute for Business & Information Systems)

In this thesis I am going to take authors from Human Resource Management and Organizational Behavior institute as my population.

Why Human Resource Management and Organizational Behavior institute is so interesting to me is because Human Resource Management is a well established field with clear boundaries and clearly within management. SOM Human Resource Management and Organizational Behavior institute’s main research theme is the cooperation among employees in general and teamwork in particular. This thesis is also closely related to teamwork. Working in teams has widely been regarded as a way to enhance an organization’s capability to deal with changing circumstances, to generate more complex products and services and to handle difficult work processes more adequately. That is why the concept of team-based work has become increasingly popular, and has been adopted by a growing number of organizations. In SOM Human Resource Management and Organizational Behavior institute they study the relationship between the design, the functioning, the development and the performance of work teams. With respect to the design of teams we deal with the characteristics of the work to be done as well as the characteristics of the team members.⁶

3.6 Google scholar

The most commonly used source of bibliometric data is Thomson ISI Web of Knowledge, in particular the (Social) Science Citation Index and the Journal Citation Reports (JCR), which provide the yearly Journal Impact Factors (JIF). Nowadays the use of Google Scholar is increasing because it has a broader range of data sources. Since Google has broader range of data sources, the use of Google Scholar generally results in more comprehensive citation coverage in the area of Management and International Business (Harzing and van der Wal, 2008).

6 http://www.rug.nl/feb/onderzoek/onderzoeksinstituten/index, , visited on 3^rd of September 2008

(19)

Comparing Thomson ISI WoS and Google Scholar for citation analyses

In this section I am going to compare the respective advantages and disadvantages of Thomson ISI WoS and Google Scholar for citation analyses of individual academics. The most important advantage from Google Scholar is that it is freely available to anyone with an internet connection and is generally praised for its speed (Bosman et al. 2006). The Thomson ISI WoS is only available to academics whose institutions are able and willing to bear the (quite substantial) subscription costs of the WoS and other databases in Thomson ISI’s Web of Knowledge. Pauly &

Stergiou (2005) indicate that “free access to […] data provided by Google Scholar provides an avenue for more transparency in tenure reviews, funding and other science policy issues, as it allows citation counts, and analyses based thereon, to be performed and duplicated by anyone”.

They also mention the advantage of the no-cost Google Scholar option for research and academic institutions not only in developing countries, but also for modestly endowed institutions in developed countries.

Harzing and van der Wal (2008) proposed some disadvantages of using Google Scholar for citation analyses. The first one is that it includes some non-scholarly citations, such as student handbooks, library guides or editorial notes. The second one is that Google Scholar coverage might be very uneven across different fields of study. The social science, arts, engineering and humanities in particular seem to benefit from Google Scholar better coverage of citation in wider range of journals. The Natural and Health Sciences are generally well covered in Thomson ISI and hence Google Scholar might not provide higher citation counts (Harzing and van der Wal, 2008).

The third disadvantage is Google Scholar does not perform well for older publication. Pauly &

Stergiou (2005) found that Google Scholar had less than half of the citations of the Thomson ISI WoS for a specific set of papers published in a variety of disciplines (mostly in sciences) between 1925 and 1989. However, for papers published between 1990 and 2004 period both sources gave similar citation counts. The last disadvantage that is discussed is that Google Scholar is not updated as often as ISI WoS. Google Scholar does not provide information about its update frequency. The lack of daily updating might be problematic for accessing the latest research information in fields that change quickly.

(20)

3.7 Data source and data processing

I opened the SOM website and found the names of the research fellows of the institute for Human

Resource Management & Organizational Behavior

(http://www.rug.nl/feb/onderzoek/onderzoeksinstituten/index). There are 7 names of research fellows can be found there with their specialties. The list of the names can be found in the appendix. Then I tried to collect all of the articles produced by those research fellows on Google Scholar and collected articles which were published during 2000-2003. Then I had to choose articles which are produced by co-authoring with other author who is/ are still working in the University of Groningen. So there are at least 2 authors of the chosen paper still work in the University of Groningen. Only articles which are written in English would be counted. There are 10 articles which fulfilled all of the criteria.

The lists of articles, from the one which got most citations until the one which got least citations are:

• Gerben Van Der Vegt, Ben Emans and Evert Van De Vliert (2000)

Team Members’ Affective Responses to Patterns of Intergroup Interdependence and Job Complexity (cited 45 times)

Journal of Management, Elsevier, 2000.

• Eric Molleman and Manda Broekhuis (2001)

Sociotechnical systems: towards an organizational learning approach (cited 25 times)

Journal of Technology Management, Elsevier, 2001.

• Nico van Yperen and Onne Janssen (2002)

Fatigued and dissatisfied or fatigued but satisfied? Goal orientations and responses to high job demands (cited 25 times)

Academy of Management Journal, Elsevier, 2002.

(21)

• Gerben S. Van der Vegt and Onne Janssen (2003)

Joint Impact of Interdependence and Group Diversity on Innovation (cited 23 times)

Journal of Management, Sage Publication, 2003.

• Jannes Slomp and Eric Molleman (2002)

Cross-training policies and team performance (cited 20 times)

International Journal of Product Research, Taylor and Francis LTD, 2003.

• Evert van de Vliert and Onne Janssen (2002)

"Better than" Performance Motives as Roots of Satisfaction across more and Less Developed Countries (cited 16 times)

Journal of Cross Cultural Psychology, Sage Publications, 2002.

• Eric Molleman, Jannes Slomp and Samantha Rolefes (2002)

The evolution of a cellular manufacturing system – a longitudinal case study (cited 10 times)

International Journal of Production Economics, 2002.

• Eric Molleman & Jannes Slomp (2001)

The Impact of Technological Innovations on Work Design in a Cellular Manufacturing Environment (cited 4 times)

Blackwell Publication, 2001.

• Eric Molleman, Bas van Delft and Jannes Slomp (2001) The application of an empowerment model (cited 3 times)

Human Factors and Ergonomics in Manufacturing, John Wiley and Sons Inc., 2001

(22)

• Evert Van de Vliert and Onne Janssen (2002)

Competitive Societies Are Happy If the Women Are Less Competitive Than the Men (cited 1 time)

Sage Publications 2002.

I subtracted names which are still working for University of Groningen from the list above and came out with 8 names to interview. They are:

• Evert Van de Vliert (for 3 articles)

• Gerben S. Van der Vegt (for 2 articles)

• Onne Janssen (for 3 articles)

• Eric Molleman (for 5 articles)

• Manda Broekhuis (for 1 articles)

• Nico van Yperen (for 1 articles)

• Jannes Slomp (for 4 articles)

• Ben Emans (for 1 articles)

The interviews will be in a form of e-mail, direct interview, or phone conversations. Various online publications will also used as additional information or examples to strengthen the information gathered from the interviewee.

The next chapters will be the second part of the thesis that confronts the interview results and the theories. After the discussion chapter, a final chapter will follow to conclude the discussion and suggestion for future research.

(23)

Chapter 4 Discussion

I had interviews with the authors of the articles on sample and concluded the information gathered. There are some new interesting findings which could not be found on literatures. Van de Vliert (2008) said that according to his experiences, one of the advantages in doing co- authorship is that you get more and better feedback without having to enlist a friendly reviewer.

Another advantage which is mentioned by Slomp is that by doing co-authorship, the heavy workload can be shared among authors. Slomp (2008) prefers to do co-authorship than sole authorship because in co-authorship the expertise is complementary among authors which can be very helpful in writing an article. Janssen mentioned that the discussion happened on the time of the writing is very valuable to do. The only disadvantage of doing co-authorship was mentioned by van der Vliert during the interviews was when the co-author does not live up to the time schedule.

4.1 Communication

Frequency of communication is hypothesized to be positively related to producing influential article. This is due to the fact that some empirical studies show that the communication processes within a team are critical for team success (Keller, 2001; Lovelace, Shapiro, & Weingart, 2001).

According to the information gathered from the interviews, there are several important findings which are worth to mention here:

1. Frequency

The average frequency of authors’ communication about the article they were writing was 15 times per article. Authors do not need to communicate very often if the role for every author is clear. Just like on van de Vliert and Onne Janssen’s cases when Janssen did the calculations and van de Vliert did the writing. They did not have to communicate much although they lived in the same city because they did their tasks rather independently. It happened for their two articles Competitive Societies Are Happy If the Women Are Less Competitive than the Men (2002) and "Better than" Performance Motives as Roots of Satisfaction across more and Less Developed Countries (2002). On Slomp and Molleman’s case, they needed to meet 20 times for an article because of the separation of the task is not very strict. Slomp did the mathematical models and the writing when Molleman did the statistics calculation and the writing. Slomp’s and Molleman’s wrote 4 articles which are included on sample. Cross-training policies and

(24)

team performance (2002), The evolution of a cellular manufacturing system – a longitudinal case study (2002), The Impact of Technological Innovations on Work Design in a Cellular Manufacturing Environment (2001), and The application of an empowerment model (2001). They have different expertise. Slomp’s expertise is in the field of Operation Management and Molleman’s expertise is in the field of Human Resource Management. That was why they had to talk more to make the article as a whole piece. They met 20 times per article until the articles were done. On Janssen and Yperen case when they were writing the article Fatigued and dissatisfied or fatigued but satisfied? Goal orientations and responses to high job demands (2002) they needed to meet 10 times to finished the paper.

Authors in the sample admitted that during producing the articles, they were also busy writing another article with the same or other authors. It is not very easy to manage the time to finish the work and keeping contact with other authors. But since they thought communication is important for producing a co-authorship article, they had to meet the other authors, which is in line with what Bleeke and Ernst (1993) say that interaction among team members is the most important factor to bring successful cooperation; “the most accurately designed relations will crumble without good, frequent interaction”.

2. Means of communication

My sample is authors who worked (by the time they wrote the article) and still working for University of Groningen. So most of the case is they worked in a same group, or at least they lived in the same city. So they did not have problem to meet face to face. They also preferred face to face meeting instead of using any other means of communication.

They only used other means of communication such as e mail and telephone to make appointment to meet face to face. On the final stages authors usually changes manuscripts.

8 from the 10 articles from the sample are using the system which every author had their own task and their own deadline. So they just meet to discuss the topic, then worked on they own until the deadline to change manuscripts. There are 2 articles which had different methods of communication during the production. They are articles from van Yperen and Janssen “Fatigued and dissatisfied or fatigued but satisfied? Goal orientations and responses to high job demands” (2002) and article written by van der Vegt and Janssen Joint Impact of Interdependence and Group Diversity on Innovation

(25)

(2003). Janssen was the 2^nd author who gave ideas to 1^st author. Janssen only gave active discussion during the pre-writing with giving ideas then waited passively for the drafts from the 1^st author. After receiving the draft, Janssen as the second author would come out with ideas to develop the article further which they discussed during face to face meeting.

3. Good communications and positive outcomes

All of the interviewees said that with a good communication workload of producing an article can be distributed according individual expertise among authors. This is in line with theories from Edmondson (1999), Gilson et al. (2002), Mathieu et al. (2000) and Mohammed & Ringseis (2001) who say that team can subsequently decide on which steps are to be taken in what order and how the workload should be distributed among team members. Good communication does not always mean continuous monitoring. 1^st author did not have to all the time control what the other authors were doing. Most of the case in the sample research shows that every author got very clear role what they have to do. So they work rather independently. Continuous monitoring is not needed. It is contrary to what Schippers, et.al (2003) and West (2000) say that continuous monitoring is important.

Communication is not one of the primary responsibilities for every group member because it is dependent on what role an author has in the co-authorship. The 1^st author has to lead and guide the 2^nd author so, 1^st author has the main responsibility; it means that the 2^nd author only has secondary responsibility.

Conclusively, based on the analysis from the interviews, I can see the importance of communication to make the tasks of every author clearer. Good communication gives boundaries on what an author should do and not to do, so continuous monitoring is not needed. Good communication leads to producing influential article. Therefore, the first hypothesis is confirmed.

Frequency of communication is positively related to producing influential article.

4.2 Shared Mental Model

Degree of shared mental model is hypothesized to have a negative relationship with producing influential article. Low degrees of shared attitudes, belief, ideas, expectations among group

(26)

members, and the understanding level among authors will be very low whereas the novelty and newness value of ideas will be very high. Therefore, the chance to obtain influential article is higher. Conversely, high degree of shared attitudes, beliefs, ideas and expectations among team members, the understandability of members within the group is very high. The consequential organizational myopia or groupthink (Kratzer et al., 2006; Wuyts et al., 2005) will result in significantly lowered levels of influential articles.

From the interview van de Vliert (2007) said that he did not select co-authors based on their shared attitudes, belief, ideas and expectations among group member. At the beginning of co- authorship authors did not have to share the same mental model, but then when the project has begun every author found out whether they have high degree of shared mental model or not. This is contrary to what Yen et al. (2006) argue that “creation of a common understanding is necessary for a group performance”. On van de Vliert and Janssen’s case, they shared a high shared mental model. They had the same shared attitudes, beliefs, ideas and expectations. Looking back on the number of citation they get for the 2 articles, one influential and the other one is less influential then the rest of the articles in the sample. The "Better than" Performance Motives as Roots of Satisfaction across more and Less Developed Countries (2002) was cited 16 times and Competitive Societies Are Happy If the Women Are Less Competitive than the Men (2002) was cited once.

On the influential article Cross-training policies and team performance (2002) which got cited 20 times Slomp and Molleman did not have high shared mental model. Even, the topic was changed during the production of the article. By not having a high shared mental model authors are freer to develop or change the topic of the article but of course with the agreement of all of the authors.

Yperen and Janssen when writing Fatigued and dissatisfied or fatigued but satisfied? Goal orientations and responses to high job demands (2002) which got cited 25 times, they had the same ideas and expectation before writing the article that they were going to write an influential article since both of them thought that their ideas was very valuable to develop. In short they shared a high mental model and produced an influential article.

In the article produced by van der Vegt, Emans and van de Vliert (2000) Team Members’

Affective Responses to Patterns of Intergroup Interdependence and Job Complexity which got cited 45 times authors did not have a high shared mental model. They said that they just put their maximum effort for the article without thinking about the end result. Eric Molleman also

(27)

confirmed that giving the best effort is more beneficial than sharing the same expectation on producing an article. Molleman said that referring to the article which he wrote with Broekhuis in 2001 Sociotechnical systems: towards an organizational learning approach which got cited 25 times.

Conclusively, based on the analysis from the interviews, I can see that high level of shared mental model, which are included attitudes, beliefs, ideas and expectations, does not result in an influential article. Therefore, the second hypothesis, degree of shared mental models will have a negative relationship with producing influential article, can be confirmed.

4.3 Role

Referring to Said et al. (2007) a co-authorship is called as entrepreneurial style when authors are on the same level of expertise but in different field or sub-field. A co-authorship is called as the mentor style when an author writes an article with younger colleagues who were his students or other young associates. Laboratory style happens when the main author has many co-authors.

Usually all of the members of a particular laboratory project have their names on the article.

Mentor style co-authorship is hypothesized to be the best co-authorship style in social sciences.

From the interviews done I can conclude that the most popular co-authorship style in social sciences is mentor style. 1^st author does co-authorship with younger colleagues who were his students or young associates. Strict role is applied on the most co-authorship styles. 1^st author is the most important author who organizes who does what. Like the case of van de Vliert and Janssen, they wrote 2 articles: "Better than" Performance Motives as Roots of Satisfaction across more and Less Developed Countries (2002) and Competitive Societies Are Happy If the Women Are Less Competitive than the Men (2002) which are included on my sample and both of them were written with the same strict role. Van de Vliert did the writing part and Janssen did the calculations part. Authors who did the mentor styled co-authorship on the sample are: van der Vegt, van de Vliert and Janssen, Molleman and Broekhuis, Molleman, Slomp, and Roelefes, and the last one is Molleman, Slomp, and van Delft. On Molleman, Slomp,and van Delft case when they wrote The application of an empowerment model (2001), van Delft task was to gather data in a company. He was a master student in Groningen University. He did not have to write anything for the article but his name is on the article as one of the authors.

The second popular co-authorship style is entrepreneur style, where authors are on the same level of expertise. This entrepreneurial style is done by Molleman and Slomp when they wrote The

(28)

Impact of Technological Innovations on Work Design in a Cellular Manufacturing Environment (2001) and Cross-training policies and team performance (2001), Yperen and Janssen when they wrote Fatigued and dissatisfied or fatigued but satisfied? Goal orientations and responses to high job demands (2002) , and van der Vegt and Janssen when they wrote Joint Impact of Interdependence and Group Diversity on Innovation (2003) . They are all on the same level of expertise in different sub specialties. So they compliment each other in producing an article.

From the sample, we can see that all of the articles are basically influential articles since they got cited more that once. In this case I see it as the more citation the article got, the more influential the article is. If we analyze the style of co-authoring and the number of citation the article gets, then the best style of co-authorship from the sample is mentor style. The most influential articles which are produced with mentor style co-authorship are: van der Vegt, Emans and van de Vliert (2000) with their article Team Members’ Affective Responses to Patterns of Intergroup Interdependence and Job Complexity which got cited 45 times. The second one which is also written with mentor style co-authorship is article from Eric Molleman and Manda Broekhuis (2001) with their article Sociotechnical systems: towards an organizational learning approach which got cited 25 times.

On articles Fatigued and dissatisfied or fatigued but satisfied? Goal orientations and responses to high job demands (2002) which was written by van Yperen and Janssen and Joint Impact of Interdependence and Group Diversity on Innovation (2003) which was written by van der Vegt and Janssen, Janssen’s contribution was giving ideas for developing the articles. The 1^st authors (van Yperen and van der Vegt) did all of the writing, and Janssen had to read the drafts and gave comments.

Conclusively, based on the analysis from the interviews, I can see the most popular co-authorship style is mentor style. This style is also proven effective to produce article which received many citations. Therefore, the third hypothesis is confirmed, mentor style co-authorship is the best co- authorship style in producing influential article.

4.4 Result discussion

The average frequency of authors’ communication about the article they were writing was 15 times per article. Authors do not need to communicate very often if the role for every author is clear. Authors did their tasks rather independently and continuous monitoring is not needed. This

(29)

is the opposite to what Schippers, Den Hartog, Koopman, & Wienk, 2003 and West, 2000 said that team processes can be interpreted as a continuous monitoring and adjustment of the established task and cooperation model. My sample is authors who worked (by the time they wrote the article) and still working for University of Groningen. So they did not have problem to meet face to face.

Edmondson, 1999; Gilson, Shalley, & Milne, 2002; Mathieu, Goodwin, Heffner, Salas, &

Cannon-Bowers, 2000 and Mohammed & Ringseis, 2001 said that a goal-oriented and well- coordinated innovation process requires that the team establishes a shared and valid task model and cooperation model this is one line with the result of this thesis paper since all of the interviewees said that with a good communication workload of producing an article can be distributed according individual expertise among authors. For the theory which said that communication is one of the primary responsibilities for every group member is not proven since it is dependent on what role an author has in the co-authorship.

At the beginning of co-authorship authors did not have to share the same mental model this is the opposite if what Yen et al. (2006) argue that “creation of a common understanding is necessary for a group performance”. Most of the influential articles on the sample shared a low mental model, authors just tried to give their best to produce an article. Co-authorship members encompass members that stem from diverse organizations and cultures and moreover frequently possess dissimilar knowledge, this influencing the sharing of mental models.

From the interviews done I can conclude that the most popular co-authorship style in social sciences is mentor style. 1^st author is the most important author who organizes who does what.

This principal author tends to co-author papers with younger colleagues who were his students or other young associates. The principal authors can have more than 10 articles in the same time with this type of co-authorship (Said et al. 2007).

The second popular co-authorship style is entrepreneur style, where authors are on the same level of expertise. Said et al. (2007) call is as entrepreneurial style because in essence the principal author has sought out groups with sub-specialties needed for a particular paper. In essence these specialty groups represent sub disciplines and are tightly coupled within the subspecialty. They compliment each other in producing an article. Then if we analyze the style of co-authoring and

(30)

the number of citation the article gets, then the best style of co-authorship from the sample is mentor style.

(31)

Chapter 5 Conclusion

I found out that there are some more advantages and disadvantages of co-authorship which I could not find in the literatures before. The advantages are that that you get more and better feedback without having to enlist a friendly reviewer and authors have always other authors to share if problems arise. The disadvantages which are mentioned by the interviewees are not on schedule co-authors and conflict may arise during producing articles.

In the end of this thesis, I have proven based on the interviews that communication, shared mental models and role are important in co-authoring. Coming is the detail conclusion per element.

1. Conclusively, based on the analysis from the interviews, I can see the importance of communication to make the tasks of every author clearer. Continuous monitoring is not needed. Good communication leads to producing influential article. Authors preferred face to face meeting and on the final stages authors usually changes manuscripts. With a good communication workload of co-authorship can be well-distributed among members.

Communication is not primary responsibilities for every group member.

2. At the beginning of co-authorship authors did not have to share the same attitudes, belief, ideas, and expectation, but then when the project has begun every author found out whether they had high degree of shared mental model or not. Everything could change during the production of the articles. Based on the analysis from the interviews, I can see that high level of shared mental model, which are included attitudes, beliefs, ideas and expectations, does not result in an influential article.

3. From the interviews done I can conclude that the most popular co-authorship style is mentor style, is that when an author writes an article with younger colleagues who were his students or other young associates. Second popular is entrepreneurial style is that when the principal author has sought out groups with sub-specialties needed for a particular paper. Strict role is applied on the most co-authorship. This style is also proven effective to produce article which received more citations than the other.

(32)

Figure 2:

Source: self developed model

With the use of the model we can see that frequency of communication as one of the elements in producing influential article is also influencing each author’s role in producing an article. With a good communication co-authorship would run smoothly because everybody knows what they have to do and what they do not have to do. Degree of shared mental model was not important at the beginning process of producing an article. Later on authors would found out the shared beliefs, expectation, attitudes and ideas. It is also depend on the role of each author contributing on the article. It is important to decide who does the 1^st author role on each article because 1^st author has the most contribution on organizing the co-authorship. With a good leader and clear role of each author influential article would be produced.

Frequency of communication

Degree of Shared Mental

Model

Authors’

Role

Influential article

Co-authoring and Influential Academic Articles