«Andrés Gregor Zelman The University of Amsterdam 2002 ii Mediated Communication and the Evolving Science System: Mapping the Network Architecture of ...»
“can parallels be found in journal-journal distributions over time.” In accordance with the expectations of the analysis, the relationships between the identified clusters of journals provided an overview of the changing disciplinary emphases in the science system relevant to the SOEIS during the time period 1996 thorough 2000. Research Policy and Scientometrics were shown to have a unique relation vis-à-vis each other.
On the side of Scientometrics the two journals appeared to grow closer together over the five years of the analysis; on the side of Research Policy: the two journals grew apart, with the exception of the year 2000 in which they appeared closer. The systemic analysis of the SOEIS publication environment has shown a cognitive orientation of the SOEIS group toward policy and informatics journals and has reinforced the finding of the network analysis that revealed a unidirectional citation behaviour: Scientometrics relies upon Research Policy for the legitimization and authentication of published analyses, but Research Policy was not shown to cite Scientometrics with the same fervour.
The systemic analysis of the SOEIS mailing list environment aimed to find selforganizational properties of the project and field level Internet mailing lists under analysis (the intermediary or national level lists were left out of the analysis, as it was shown before that they do not operate in terms of threads). When observed from this macro perspective it was expected that some mailing lists would exhibit selforganization while some would not, thereby enabling a distinction. When compared for thread size and frequency none of the mailing lists were found to exhibit these properties. It was concluded that the datasets as compiled were not suited to the analysis since it became evident that much larger datasets would be needed to find statistically significant evidence of Self-Organization.
Finally, the concluding section of this chapter will highlight the central findings of the dissertation. Here the core findings are listed in rank order with the most significant findings down to those of lesser significance. The findings are also divided into those which concern the internal communicative dynamics of the SOEIS and those which concern the external.
1. Internal: Print and electronic writing differed in their architectural make-up;
the SOEIS print communications proved to be heavily codified and aggregative, whereas electronic communications appeared resistant to codification. These findings were argued in the architecture and system 137
analyses in Chapter IV: Analysis of Print Communication and Chapter V:
Analysis of Electronic Communication.
2. Internal: Print and electronic keyword distributions were different in their respective emphases; the SOEIS print communications were shown to bias function oriented words, which were shown to increase over the dataset, in contrast to the electronic communications which were shown have a decreasing occurrence of function words in favour of words which contributed to the activity of the SOEIS project; email communications were found to supplement project activity. These results were shown in the network analyses of Chapter IV: Analysis of Print Communication and Chapter V: Analysis of Electronic Communication.
3. External: SOEIS publications were shown to bias Policy and Informatics oriented journals as evidenced by the strong predominance of Scientometrics and Research Policy. The SOEIS group was shown to cite Research Policy articles in an effort to authenticate the policy relevance of scientometric research; the group was in turn cited by journals published in Scientometrics thereby revealing a one way publication flow. This result was shown in the network and system analyses in Chapter VI: Analysis of Journal Publication.
4. External: SOEIS mailing list environment revealed that email does foster unique network relations between researchers, but that email serves to supplement the print medium as evidence was found to support the expectation that the informal / informal distinction associated with print was found to be imported into this new medium as project, intermediary and field level lists. The project mailing list EuroCon-Knowflow was found to behave like field level lists, as revealed by its high level of mail activity and thread participation as shown in the architecture and network analyses of Chapter VII: Analysis of Mailing List Environment.
Overall it can be argued that the internal dynamics of the SOEIS as contained in the print and electronic communications conformed to the expected distinction between Mode I and Mode II processes of knowledge production, respectively. However, in the external communicative domains it was found that although predominantly Mode I, journal publication did exhibit Mode II characteristics (Innovation literature) and the mailing lists were shown to exhibit characteristics associated with the print medium such as the distinction between formal and informal communications, and was therefore also Mode I oriented. The distinction between Mode I and Mode II was clear in the analyses of the internal communicative domains of the SOEIS, but was less so in the analysis of the external domains thereby revealing the intricacy of the overlap.
This exploratory analysis has highlighted some specific ways in which we can examine these phenomena into the future. In the last and final chapter the limitations identified in the course of this research will be discussed, suggestions for further analyses provided, and finally it will deliver some design specifications for an envisaged software program to aid in the performance of similar analyses into the future.
Discussion The preceding chapters have addressed the key problematic identified in the introduction: understanding the role and impact of print and electronic media use in processes of knowledge production. The central research question was asked: “do media foster unique types of information environment, and are different modes of knowledge production and meaningful exchange thereby implied with each medium and its use?” Given the conclusions outlined in the final section of the previous chapter, it is now possible to answer this question in light of the knowledge gained through this study. Different media were shown to foster different types of information exchange, and different modes of knowledge production were found with respect to each medium and its designated function in the SOEIS.
By establishing new exploratory research methods using a range of computer assisted research software, this analysis has contributed to the fields of Science & Technology Studies and Communications Studies. This contribution was achieved by formalizing a research methodology that was theoretically informed using both traditions.
Symbolic and modelling approaches were integrated using the Architecture – Network – System heuristic model; the theoretic triad thereby enabled the development of a standardized methodology. Both Science & Technology Studies and Communications Studies have been enhanced by using this combinatory framework, as each emphasizes the importance of understanding the relationship between social processes of knowledge production and the technological means through which communication is made possible. The analysis has shown that these processes are inextricably linked.
By providing a distinction between communications that share similar content but use distinctly different media this study has isolated the dominant features of print and electronic communication and has highlighted their centrality in processes of academic communication.
The central problematic identified in the introduction was addressed. Basic differences were found between print and electronic media in the context of the shared communications of the SOEIS research project. In short, print writing was shown to exhibit both a priori and processual codification, and networks of keyword distribution that emphasized words associated with the functioning of the research project. Electronic writing was shown to resist the codification evident in the set of print communications, and to have keyword distributions which emphasized the activities of the SOEIS group such as arranging meetings. SOEIS print publications identified a unidirectional flow of citation from Scientometrics to Research Policy.
Finally, the EuroCon-Knowflow mailing list of the SOEIS research project was shown to behave like field level lists.
It can be stated that this study rests in an steadfastly emerging academic discipline.
The Internet is growing exponentially.1 Accordingly, the amount of studies 1 Internet growth and usage statistics: http://cyberatlas.internet.com/big_picture/stats_toolbox/article 139 concerning the relationship between academe and the Internet is likewise growing dramatically, as are other Internet related analysis which monitor user behaviour in a myriad of different ways. Given this, it is striking that the difference and relationship between print and electronic media remain sorely under-represented issues in the field. This study was motivated in part by this apparent lack, and has established a general methodology through which issues of media overlap could be addressed. It is expected that increased understanding of the overlap between media in academic and other social environments will encourage, or indeed demand, similar types of analyses into the future.
This final chapter will reflect upon the challenges encountered and limitations identified through the course of performing this research. The aim is to provide relevant guidelines for those interested in pursuing similar analyses in Science & Technology Studies and Communication Science. In the following section, I will reflect upon the challenges and limitations involved with maintaining an empirical – theoretical balance and the process of model building involved therein. The section will also provide suggestions for future research. It is also important in this reflection to acknowledge the limitations implied with the manual collection and analysis of data using such a diverse range of software programs employed throughout this study.
In response to the constraints identified in the course of performing such exploratory research, the concluding section will introduce the Media Analysis Toolbox (MAT).
The MAT is a proposed software toolbox comprised of a modularized set of tools to aid in the collection, collation, and analysis of ‘media generated’ traces of communication.
For the benefit of those who may pursue argumentation along these lines in the future, it is important to emphasize that the analysis of media overlap should not be restricted to the observation of academic environments. However, while other environments may exhibit the richness of media overlap, it can be problematic to find enough data to make the analysis viable. Academe does provide an ideal set of communications to study the phenomenon of mediation, in the sense that it is rich with traces of human communication. In part, this is why the SOEIS research project was selected – within the scope of the project we have print and electronic communication about the same subject matter, an identifiable time line, and a relatively stable body of participants.
Arguably, a similar analysis could be performed whereby the oral communications of a group could be recorded, transcribed and then compared with the group’s print and (or) electronic communication. There are also a range of other salient environments one could study to gain insight into the impact of media overlap. Examples include the comparison of traditional print oriented citation patterns with the emerging phenomenon of online citations as hyperlinks (Rousseau, 1999), the comparison of related websites on the basis of their textual content, the comparison of on and offline newspapers, and the comparison of oral and print processes of knowledge production in research environments such as the SOEIS.
One central challenge encountered in the course of this research was that of model building. Once the research group had been selected and its respective 140 communications assessed for analytic potential, it became clear that a stable framework would need to be established in order to contain the diversity of materials and methods. It simply would not have been possible to co-ordinate all of the heterogeneous research foci, research questions and expectations, much less compare two distinctly different media, were it not for the grounding of the work into a theoretical context using what was termed the theoretic triad of Architecture – Network – System. By developing this heuristic model it was possible to coordinate these otherwise disparate approaches into a workable framework.
It is necessary to realize that the development of new analytic techniques using computer assisted research tools also introduces new analytic challenges. Importantly, using techniques of this sort provides the researcher with entirely new subject matter to interpret. The research output becomes, in essence, another body of information that should be read through the lens of abstraction. That is, the researcher must appreciate that the information has been filtered using certain logics, and one must therefore interpret these new research objects in that light. Making the research object-oriented makes the simulated object a ‘real thing’ that can be manipulated.
Thus the graphic image becomes a key instrument for research. The rationale behind a visually oriented representation is that everyone viewing an image will agree on its meaning, provided they each understand the signs. The researcher must be aware how the use of such programs affects the research process. S/he must appreciate the difference between the interpretation of a text, and the interpretation of an interpretation. A text is an interpretation by definition; the author interprets the world and inscribes it. With hermeneutic examination of a text, the researcher is interpreting the interpretation. This double-bind is referred to by Giddens as the Double Hermeneutic. (1991) With the increased use of visual representations of data, researchers must acknowledge this double bind, lest their research lead to spurious conclusions.