没有合适的资源?快使用搜索试试~ 我知道了~
全球科学网:新兴的全球专业知识来源-研究论文
需积分: 0 0 下载量 42 浏览量
2021-05-20
11:31:23
上传
评论
收藏 201KB PDF 举报
温馨提示
试读
16页
此研究项目(ESRC RES-160-25-0031)的目的是评估Internet和Web是否以及在何种程度上可以改变对科学专门知识来源的访问。 在竞争激烈的理论观点上,要么是对链接最紧密和访问最多的网站的集中效应(由“权力法”驱动,要么是赢家通吃的效应),要么是在线资源更加多样化或“民主化”。 这导致了许多相关的问题,例如,互联网是否与使用更多的全球性专业知识而不是本地性专业知识有关? 各个学科之间是否存在差异? 互联网在塑造科学家对信息和专业知识的访问中起着怎样的中心作用? 除了这些实质性问题外,该项目还试图为Webmetric技术的开发和应用及其与深入访谈和焦点小组的结合做出贡献。 项目团队发现,这种三角关系使人们对互联网在获取不同专业知识来源方面的作用有了更深入的了解。
资源推荐
资源详情
资源评论
Electronic copy available at: http://ssrn.com/abstract=1328341
ESRC End of Award Report
The World Wide Web of Science: Emerging Global Sources of
Expertise
RES-160-25-0031
Ralph Schroeder, Alexandre Caldas, Shefali Virkar,
and William H. Dutton
Oxford Internet Institute, University of Oxford
Background
The World Wide Web of Science (WWWoS) project explored the degree to
which the Internet is supporting a ‘winner-take-all’ pattern of access to scientific
expertise. Rather than enabling expertise to be accessed from a more
diversified array of sources, a winner-take-all, or power law, hypothesis posits
that online access will reinforce central sources of science expertise, leading
access to expertise to become more concentrated.
This role of the Internet in ‘reconfiguring access’ to information (Dutton 2004),
including scientific expertise, is becoming increasingly important as people shift
more time and attention to the use of online resources (mainly the Web) to
access information and expertise in an expanding array of information sectors.
While there has been little research in this area, the topic of search engines,
online resources, and access to expertise has become a prominent research
issue during the course of the present project.
Objectives
The aim of this research project was to assess whether and to what extent the
Internet and the Web could be transforming access to sources of scientific
expertise. There are competing theoretical perspectives, either for a
concentration effect for the most heavily linked and accessed sites (driven by
‘power laws’ or a winner-takes-all effect) or for a greater diversification or
‘democratization’ of online resources.
1
This led to a number of related
questions, such as whether the Internet is associated with the use of more
global versus local sources of expertise? Are there differences across
disciplines? How central a role is the Internet playing in shaping access to
information and expertise among scientists?
In addition to these substantive questions, the project also sought to contribute
to the development and application of Webmetric techniques, and their
combination with in-depth interviews and focus groups. The project team found
that this triangulation yielded greater insight into the role of the Internet in
accessing different sources of expertise.
1
This literature is reviewed in the project proposal, but the classic work remains Merton (1988). An
important popularization of the ‘winner-take-all’ thesis is provided by Frank and Cook (1995), and
moves into the context of the Internet with Barabási (2003).
1
Electronic copy available at: http://ssrn.com/abstract=1328341
Methods
This was a one-year exploratory research project. It employed multiple traits
and methods to empirically examine patterns of access to scientific expertise.
We approached this by describing networks of scientific communication. Since
critical collaboration tends to be focused on particular areas of science, we
studied communication patterns surrounding specific issues. This was
accomplished by sampling key global issues that reflected a range of
challenges of world-wide importance. Topics were chosen also to avoid
favouring a winner-take-all hypothesis by selecting issues that were not
inherently more concentrated, such as in a few centres for big science.
Given the exploratory nature of the proposed research, the original proposal
identified six topics from a potentially wide range of issues. These were:
1. Climate Change,
2. Internet and Society,
3. Poverty,
4. Trade Reform,
5. Terrorism and
6. AIDS/HIV
The project was anchored in Webmetric analyses of the structure of networks of
relationships on the Web within each of these six topic areas. This involved
crawling the World Wide Web to identify and collect links which could then be
analyzed to determine the structure of online networks in each issue area.
The Webmetric results were validated and extended through two complimentary
methods. Validation of a sub-sample of issues provided added confidence in the
findings across all six issue areas:
- A set of in-depth interviews were conducted with a sample of known experts --
scientists working within four of the six issue areas. Two were anchored
primarily in the social sciences (Internet and society, terrorism) and two in the
natural sciences (climate change and HIV/AIDS). Four topics from the six were
chosen to allow for greater in-depth analysis of the Web representations and
more interviews with researchers within each topical area. The project team
interviewed 20 researchers, five from within each of the four topics. Given
limited funding for fieldwork, researchers were chosen from universities in
Oxford and Greater London, and do not represent a random sample from a well
defined population of experts. Primarily face-to-face, but also some telephone
interviews, were semi-structured and asked a series of questions to
contextualize and identify how these researchers use the Web and other offline
and online resources. All the interviews were transcribed and coded using
qualitative analysis software (e.g., NviVO) to extract relevant information from
the text of interviews.
- The results of our Webmetric analyses were validated also by two expert
panels, similar to focus groups. One looked at ‘climate change’ and another at
2
Electronic copy available at: http://ssrn.com/abstract=1328341
‘Internet and society’. Three and five participants respectively were chosen
based on accessibity to the research centre, with each focus group was
conducted in Oxford, at the OII. These enabled us to compare lists of the top
sites with the respondents’ lists of top sites, providing a means for judging the
face validity of the Webmetrics.
Changes in the Research Design
In-depth interviews replaced the originally proposed Web-based survey. The
team’s concern over problems with response rates and limited number of
questions that are practical for a Web-based survey led to abandoning this
method. The interviews provided richer information than was originally
envisaged, supporting a number of publications.
2
One further change from the original proposal was the decision not to undertake
a study of Usenet groups as an additional source of evidence on academic
networks. Usenet is one of the early systems, created in the 1970s, that
supported distributed discussions over the Internet which has continued to
enable users to post and read articles within a wide range of newsgroups. The
analysis of Usenet groups across these areas was pursued at the first WWWoS
workshop, and the OII was able to obtain permission and a license to analyse
Microsoft’s archive holdings. However, it was not taken beyond this point for two
major reasons. First, a paper had just become available (Matzat 2004) which
covered this topic well. Secondly, it became increasingly clear that the use of
Usenet had become a niche area that was increasingly unrepresentative of
online access by scientists generally. Its use was also being marginalized by
the role of search engines and other online resources, such as digital libraries
and online datasets. It was therefore decided to focus project resources on in-
depth interviews with scientists to validate, interpret and extend the Webmetric
results.
Results
A number of patterns and themes have emerged from this exploratory research,
which are substantively and methodologically useful in shaping follow-on
research. The details are developed further by the two nominated outputs, but
are summarized in this section.
First, the central finding of our study is that the winner-take-all hypothesis fails
to reflect the more complex structures of scientific networks of expertise. It is
true that a small proportion of sites capture a disproportionate share of links, but
the Webmetric analysis and interviews suggest that:
− The structure of networks was more fractal in structure than would be
expected by a simple winner-take-all hypothesis;
− Numerous clusters of institutions and websites are more prominent than
others, but there are many winners sharing the attention of more
specialized networks of researchers and other users;
2
Key early findings are provided by Fry et. al. (forthcoming), but additional publications are in progress.
3
− The type of search and the topic make a difference in the overall
structure of expertise, i.e., whether it is a directed search on a topic
where there are established sources, or if the topic requires a more
exploratory approach aimed at tapping a range of diverse sources;
− Search engines, and Google in particular, play an increasingly important
gate-keeping role – shaping winners and losers, though this function
varies between the four topics investigated and the type of search
engine, as different search engines yield different sources of expertise.
Indeed, the study provided a robust validation of the heterogeneity of
search engines, the qualitative differences among them in terms of
which functionalities they provide, and the apparently different content
and Web spaces they provide in indexing and search services.
− Researchers display significant differences in how they access online
resources (for example, if they go from search to publications, or vice
versa; go from online to offline resources, or vice versa; look for people
or institutions, etc.), which mitigates potentially more systematic impacts
of the Internet on who goes to what sources of expertise;
− There was a US bias in the search engine results, when compared to
the local UK sources that our UK experts consult, though this was more
the case where the research topic prioritised national online resources
(HIV/AIDS public health sites) than for the other three topics.
Secondly, there is an embedded social structure inherent in the distribution and
sharing of resources on the Web. Similar patterns of the centrality of key
resources, their connectivity and the way they are clustered, emerge on the
Web spaces of all six topics. These social networks can be seen as electronic
social networks, and identifying such structures will be of value in designing the
search engine technology of the future.
Thirdly, the qualitative findings of this exploratory research suggests the
potential value of embedding in future search engine technology more ‘user
oriented’ and topic specific functionalities, such as peculiar characteristics of
‘geography and locality’ relevant to the topic. Policy considerations influencing
the distribution of resources as well as the more academic nature of certain
kinds of information on the Web might explain significant differences in the
social networks embedded within these electronic networks.
Methodologically, the study supported the value of using Webmetric analyses
and interviews in a complementary way. For example, we were able to
successfully compare the most heavily linked and highly rated sites for each
topic with those that researchers regularly accessed and used. The following
sections provide a more detailed discussion of findings within and across the
two methodological approaches.
I. Results of the Webmetric Analysis
The core empirical data set was created from Webmetric research tools, which
enabled us to crawl the Web for links to and from sites to analyse patterns of
access on the World Wide Web to scientific and technical knowledge in the six
broad global issue areas of scientific research: ‘Climate change’, ‘Poverty’,
4
剩余15页未读,继续阅读
资源评论
weixin_38667403
- 粉丝: 2
- 资源: 915
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功