BIBLIOMETRICS

1
BIBLIOMETRICS CYBERMETRICS

• History and Example
• David Roberts Jesse Wilbur

2
What is Bibliometrics?

• Bibliometrics is the study that uses statistical
  and mathematical methods to analyze the
  literature of a discipline as it is patterned in
  its bibliographies.
• Goals
• To describe, explain, or identify a cause, or
  interpret the meaning imparted by the
  participants
• http//alexia.lis.uiuc.edu/standrfr/beginner.html
  WRAP1054688352-68.166.22.155

3
Bibliometrics Good for What?

• Conceptual Mapping
• Delineating scholarly communities and
  intellectual structures
• Determining the "importance" of a particular
  author
• Illuminating trends in the literature of a
  particular field
• Search engine effectiveness

4
Applications

• Collection development
• Amount of interdisciplinary cross-pollination
• Delineating disciplinary boundaries in
  information space
• Evaluative link analysis research is better from
  who, what, and where?
• Tenure

5
Two Major Areas

• Citation analysis
• Bibliometric laws

6
History

• First bibliometric study
• Cole and Eales in 1917 performed a statistical
  analysis that shows fluctuation of interest in
  the field of comparative anatomy between 1660 and
  1860.
• First citation analysis by Gross and Gross in
  1927
• S. M. Lawani, Bibliometrics Its Theoretical
  Foundations, Methods, and Applications, Libri 31
  (1981) 294315

7
History

• Bibliometric Laws
• Lotka 1926 the number of authors contributing
  more than one article goes down by the square of
  the number of articles contributed.
• Zipf 1949 phonemes and principle of least
  effort. Words appear in a mathematically
  predictable curve.
• Bradford 1953 if periodicals contributing to a
  subject are ranked and then grouped in such a way
  that each group contributes the same number of
  articles, the numbers of periodicals in each
  group increases geometrically.
• http//www.gslis.utexas.edu/palmquis/courses/bibl
  io.html

8
History

• Garfield (ISI) 1963 There were three factors
  that led to the development of automated citation
  indexing back in the 1950's
• 1. Trying to find a better way to manage
  information
• 2. Growing dissatisfaction with the capacity of
  subject indexing to meet the needs of the
  active researcher
• 3. Hope that automation might hold the answers

9
History

• Indexes We Use
• Science Citation Index (SCI)
• Social Science Citation Index (SSCI)
• Arts and Humanities Citation Index (AHCI)
• Term Bibliometrics coined in 1969 by Pritchard

10
Units of Measurement

• Citation implies a relationship between a part
  or the whole of the cited document and a part or
  the whole of the citing document
• Reference is a listing in the bibliography -
  may not be as strong as a citation
• Link is a pointer or reference to - not as
  strong as citation
• Smith, 1981, p.83

11
Methods Theories

• CITATION INDEXING
• 1. Selection of a core set
• 2. Collecting the raw cocitation occurences
• 3. Raw data matrix (raw cocitation frequencies)
• 4. Profile analysis
• 5. Multivariate analysis of the correlation
  matrix
• 6. Interpreting the map

12
Cocitation raw data table
Cocitation clustering data
13
Concept map
Katherine W. McCain. Mapping authors in
intellectual space A technical overview. JASIS.
Vol 41 (6). 1990. pg.433-443
14
Bow Tie Theory

• How we get to PageRank

Broder, A. et. al. Graph Structure in the Web.
http//www9.org/w9cdrom/160/160.html
15
Impact Factor Theory

• The relative attractiveness of a individual web
  pages - how frequently a particular page gets
  linked in relation to the number of pages in its
  category

Borgman, Christine L. Furner, Jonathan.
Scholarly Communication and Bibliometrics
ARIST Vol. 36, ed. B. Cronin. 2002
16
Strengths Weaknesses

• scalable way to quantify the relationships
  between documents
• sophisticated way to measure correlation where
  other methods are impossible
• what is a citation?
• Only indexes occurences of citations - not
  necessarily meaning
• based on the assumption that citing or
  hyperlinking is a meaningful action
• exploitation of citation algorithms
  (googlebombing)

17
The Article

• Bibliometrics of the World Wide Web An
  Exploratory Analysis of the Intellectual
  Structure of Cyberspace
• Ray R. Larson
• http//sherlock.berkeley.edu/asis96/asis96.html
• THE PURPOSE
• To discover and map the intellectual structure of
  the World Wide Web, its sub-domains within
  particular disciplines using cocitation analysis

18
Problem and Rationale

• Aim
• Apply bibliometric analysis to Web.
• Main question
• Can the intellectual mapping of disciplines with
  citation indexes and cocitation techniques be
  applied to Cyberspace?

19
Methodology

• 1. Selection of the core set of items for the
  study.
• 2. Retrieval of cocitation frequency information
  for the core set.
• 3. Compilation of the raw cocitation frequency
  matrix.
• 4. Correlation analysis to convert the raw
  frequencies into correlation coefficients.
• 5. Multivariate analysis of the correlation
  matrix, using principle components analysis,
  cluster analysis or multidimensional scaling
  techniques.
• 6. Interpretation of the resulting map'' and
  validation.

20
Analysis

• Raw citation data was
• Converted to a correlation matrix
• Shows topical proximity between web sites.
• Groups sites with similar content together.
• Correlation matrix plotted on a Multidimensional
  Scale (MDS)
• Shows spatial relationship of the sites to each
  other.
• Maps cocitation data on a graph.

21
Results

• Mapping of cocitations indicate major topical
  clusters of Web sites.

22
Significance

• Bibliometric method works for the World Wide web.
• Cocitation analysis reveals reasonable clustering
  of sites with topical similarities.
• Citation analysis of hypertext links on the web
  works the same as scholarly citation analysis.
• Bibliometrics is a new tool that can be used to
  intellectually map the web. This is an exciting
  and provocative discovery.

23
Usefulness

• Bibliometrics can be applied to Web.
• Flaws
• MDS plotting of data is an arbitrary activity.
  Interpretation is at the discretion of the
  researcher.
• Used AltaVista search engine to retrieve initial
  set. Search engines are sporadic in their
  coverage of resources.

24
Summary

• My Sherbet dish likes to party!!!

Identify intellectual communitieson the web?
Bibliometrics on the web?
Study growth and usage of the web?
How to analyze the web?
Identify Core set of web sites
Cocitation analysis will produce Intellectual
map of the web
Cocitation frequency for Core set
Correlation matrix
Multidimensional scaling analysis
Borders of clusters reflect researcher bias
Mapping of cocitations indicate major topical
clusters of Web sites.
Bibliometrics can be used to analyze the Web