This slide presents practical ways to conduct bibliometric analysis. Based on my own experience on bibliometric analysis research, I wrote this step-by-step guide to help students and professionals who will use citation, co-citation, biliographic coupling techniques in their research. Visit my service research blog for more updates: http://servicesresearch.blogspot.com/
2. Bibliometric Analysis (1)
Definition
The application of mathematical and statistical techniques to analyze the
patterns that appear in publications and documents (Groos and Pritchard,
1969).
Rationale
The collective action of contributors from the discipline ultimately
determines its identity and direction (Banville and Landry, 1989).
Advantage
- Quantifiable and objective.
- Unveil research topics currently undetected by expert evaluations.
- Examine the exchange of knowledge and scientific communication.
3. Bibliometric Analysis (2)
Techniques
- Citation analysis
- Co-citation analysis
- Bibliographic coupling
Unit of analysis
Article (author/journal)
Application
- Most often used in the field of library and information science.
- Many research fields use bibliometric methods to explore the impact of
their field, a set of researchers, or a particular journal.
4. Analysis Procedure Overview
1. Journal selection: choose representative journals in most relevant
disciplines.
2. Sample preparation: decide keywords and timeframe.
3. Sample refinement: manually refine the article sample.
4. Coding and purification: eliminate inconsistencies.
5. Data analysis: conduct bibliometric analyses.
5. Analysis Procedure (1)
Journal Selection
Objective
To choose representative journals in the most relevant disciplines.
Instructions
(1) Select relevant disciplines.
(2) Refer to journal ranking articles in each discipline, e.g., innovation
management (Linton and Thongpapanl, 2004), marketing (Baumgartner and
Pieters, 2003).
(3) Select representative journals.
Tips
- Notice the change of journal title.
- Same journal may be included more than once.
6. Analysis Procedure (2)
Sample Preparation
Objective
To use effective keywords to retrieve relevant articles within a specific time period.
Instructions^
(1) Decide search keywords and timeframe.
(2) Run search in the online database, such as WoS and Scopus.
(3) Save results into plain text.
(4) Process *.txt file with Bibexcel (steps) and generate *.doc file.
(5) Compile the full *.doc file from *.doc files of different databases.
Tips
- Tutorials of WoS search can be found here.
- Bibexcel can be downloaded here and its manual can be found here.
- One database may not have the full coverage of a journal. Search priority: WoS >
Scopus > Publisher’s own website.
7. Analysis Procedure (3)
Sample Refinement
Objective
To manually exclude irrelevant articles.
Instructions^
(1) Use Bibexcel to export *.doc file to an Excel table file (steps).
(2) Review article title and abstract to determine its relevance.
(3) Mark relevant articles and irrelevant ones.
(4) Remove irrelevant articles in the original *.doc file.
Tips
- Engage independent researchers who do not participate in the project.
- Solve disagreement through discussions and review of the whole article.
8. Analysis Procedure (4)
Coding and Purification
Objective
To supplement missing contents and to eliminate inconsistencies.
Instructions
(1) Supplement *.doc file with missing contents so that each article has key fields:
authors, title, abstract, keywords, year of publication, journal name, and a list of cited
references.
(2) Correct mistakes in *.doc file, such as wrong names and publication years.
(3) The updated *.doc file is ready for data analysis.
Tips
- Common mistakes of cited references in journal style: wrong or missing volume and
page numbers and inconsistent journal name abbreviations.
- Common mistakes of cited references in book style: multiple editions and
inconsistent title abbreviations.
9. Analysis Procedure (5.1)
Data Analysis – Citation Analysis
Definition
A procedure to examine the exchange of knowledge (Garfield, 1979).
Rationale
Authors cite papers which they consider to be important to the development of
their research.
Advantage
It is amenable for providing insights into the influences of research that prevails
within its own field and across other academic disciplines (Hoffman and Holbrook,
1993).
10. Analysis Procedure (5.1)
Data Analysis – Citation Analysis
Instructions for basic citation analysis
(1) Convert updated *.doc file to Excel table file with key fields: author, title,
journal name, publication year.
(2) Use Excel to conduct data analysis.
Instructions for forward citation analysis^
(1) Convert updated *.doc file to *.out file in Bibexcel by extracting article title.
(2) Use Excel to open *.out file and copy article titles.
(3) Search in WoS for the whole set of articles.
(4) Click
and then
(5) Save all records into plain text.
(6) Convert *.txt file to Excel table file with essential fields.
(7) Use Excel to conduct data analysis.
11. Analysis Procedure (5.2)
Data Analysis – Bibliographic Coupling
Definition
A technique to cluster source articles that refer to similar references (Kessler,
1963).
Rationale
Studies from the same research stream are more likely to cite similar references
than studies with different origins.
Advantage
It is of particular use for mapping the full coverage of the literature in one
research field and providing a valid representation of the underlying structure
(Persson, 1994).
12. Analysis Procedure (5.2)
Data Analysis – Bibliographic Coupling
Instructions^
(1) Convert updated *.doc file to *.out file in Bibexcel by extracting cited
references.
(2) Select *.out file and indicate coupling strength in Bibexcel
.
(3) Click Analyze Shared units/coupling Shared units *.cou file.
(4) Select *.cou file and click Analyze List units in pairs.
(5) Click
and highlight all items in “The List” box.
(6) Select *.cou file and click Analyze Make a matrix for MDS ect *.ma2 file.
(7) Use Excel to copy contents of *.ma2 to UCINET spreadsheet editor
.
(8) Remove all the value on the diagonal and save the file.
(9) Click Tools Correlate columns across datasets.
(10) Select the previous file and define the output file.
13. Analysis Procedure (5.2)
Data Analysis – Bibliographic Coupling
Instructions^
(11) Copy the correlation table into SPSS.
(12) Click Analyze Scale Multidimensional Scaling (PROXSCAL).
(13) Use default setting in “Data Format” window and click Define.
(14) Select all variables and move them into Proximities box.
(15) Model setting: Lower-triangular matrix, Similarities, Ordinal (check untie
option), dimensions 2x2.
(16) Options setting: Torgerson.
(17) Click OK and save common space coordinates.
(18) Plot MDS map.
* See also here.
14. Analysis Procedure (5.3)
Data Analysis – Co-citation Analysis
Definition
A method to reveal the knowledge structure in a field according to the frequency
with which two references are cited together (Small, 1973).
Rationale
References represent concept surrogates and a group of closely cited references
comprises the consensual structure of concepts in a field.
Advantages
Able to objectively reflect change in a field, despite scholars’ subjective views of a
semi-permanent disciplinary structure (White and McCain, 1998).
15. Analysis Procedure (5.3)
Data Analysis – Co-citation Analysis
Instructions^
(1) Convert updated *.doc file to *.out file in Bibexcel by extracting cited
references.
(2) Select “whole string” under
.
*.cit file.
(3) Select *.out file and click
(4) Click
and highlight the items in “The List” box whose co-citation
counts are above pre-defined threshold.
(5) Click Analyze Co-occurrence Select units via listbox.
(6) Select *.out file and click Analyze Co-occurrence Make pairs via listbox
*.coc file.
(7) Click
and highlight the pairs in “The List” box whose co-citation
count is above 2.
(8) Select *.coc file and click Analyze List units in pairs *.ccc file.
16. Analysis Procedure (5.3)
Data Analysis – Co-citation Analysis
Instructions^
(9) Highlight all items in “The List” box.
(10) Select *.coc file and click Analyze Make a matrix for MDS ect *.ma2 file.
(11) Use Excel to copy contents of *.ma2 to SPSS.
(12) Remove all the value on the diagonal and make sure there is no column with
all zeros.
(13) Click Analyze Data Reduction Factor…
(14) Select all variables and move them into Variables box.
(15) Rotation setting: Varimax.
(16) Options setting: check “replace with mean”, “sorted by size”, and “suppress
absolute values less than 0.x”.
(17) Use excel to analyze results.
* See also here.