Yon Soo Lim, Ph.D. Yon Soo Lim, Ph.D. Research Fellow Research Fellow WCU Webometrics Institue WCU Webometrics Institue Yeungnam University Yeungnam University [email protected] [email protected]
Nov 19, 2014
Yon Soo Lim, Ph.D.Yon Soo Lim, Ph.D.Research FellowResearch Fellow
WCU Webometrics InstitueWCU Webometrics InstitueYeungnam UniversityYeungnam University
[email protected]@gmail.com
Network AnalysisNetwork analysis is a set of research
procedures used for identifying structures in social systems based on the relations among a system’s components (Richards & Barnett, 1993; Rogers & Kincaid, 1981).
Network Analysis
Network AnalysisNode
Unit of analysis (individual or higher-level component)
Other terms: actors, points or vertices
LinkRelationship between nodes Other terms: lines, edges or geodesics
Network AnalysisSocial/communication network analysis
Semantic network analysis
Neural network analysis
Social Network AnalysisSocial network analysis is a set of research
procedure identifying the structure of communication relationships and flows among components in a social system, such as people, groups, organizations, nations, and other information processing entities.
SoftwareUCINET (Borgatti, Everett & Freeman)MULTINET (Richards & Seary)NETMINER (Cyram)PAJEK (Batagelij & Mrvar)
Network Survey QuestionnaireWhich of these people do you communicate with on a weekly basis? If you communicate with them, please place a one (1) by their name and if you do not communicate with them weekly, place a zero (0) by their name for both business and social non-work issues. If you do speak with them, please indicate whether face-to-face, phone, email, etc.
Do you communicate with this person:
Business communication Social Non-work Resources Used
Ed Peet
Karen Shed
Carlton Smith
Tracy Pope
Dr. Todd McCune
Dr. Linda Clark
Jennifer Heisner
Melanie Lowe
Cindy VanMeter
Tracy Gilmore
Beth Murnane
Bill Morrison
Tom Redmond
Rene Kunego
Joanne McIllwaine
Frank Blasioli
Sue Ann Burley
Vivian Bulriss
Results of Social Network Analysis
Business Communication Social Communication
A Cross-Cultural Comparison of Online Forum Structures on MMOGs
Power:Power: RedRed > Yellow > > Yellow > GreenGreen
Korea United States
Number of Nodes = 4,384 Number of Links = 9,304 Density = 0.000951 Centralization = 24.8%
Number of Nodes = 10,966 Number of Links = 28,268 Density = 0.000458 Centralization = 11.4%
International Aid Network
Int’ Telecommunication Network
Semantic Network AnalysisSemantic network analysis is a systematic
technique of content analysis to identify the meaning structure of symbols or concepts in a set of documents by using network analysis (Monge & Contractor, 2003; Monge & Eisenberg, 1987).
Semantic network represents a relationship of shared understanding of cultural products among members in a social system (Monge & Contractor, 2003).
CATPACA popular software for semantic network analysis
A self-organizing artificial neural networkKeyword identificationCluster analysis (dendogram)MDS
Explore the pattern of interrelationship among words based on the probability of their co-occurrence
Semantic networks(Blogs vs. Newspapers)
Centralization = 19.6%Centralization = 19.6%
Blogs NewspapersCentralization = 44.2%Centralization = 44.2%
Semantic networks(Progressive vs. Conservative)
Centralization = 20.0%Centralization = 20.0%
Progressive News Conservative NewsCentralization = 46.6%Centralization = 46.6%
Semantic Network of IDCs
Neural Network AnalysisNeural network analysis is modeling dynamic
attitudinal structures and cognitive processes based on natural neural networks in human brains.
Defining AttitudeAttitudes can be defined as associations, associations,
represented in memory, between attitude represented in memory, between attitude objects and their evaluations objects and their evaluations (Fazio, 1990; 1995).Attitude objects: anything that can be
designatedEvaluations: summary assessments considering
cognitive, affective, and behavioral associations AttitudeObject
Evaluation
AttitudeAttitude
Interattitude StructureA simple example
ChickenChicken DuckDuck
Evaluation
“I don’t like chicken because I am allergic to chicken.”
“I don’t like duck either because it is a kind of chicken.”
Galileo Spatial-linkage ModelNeural networks in human brains
The model represents interattitude structure within a spatial coordinate system.
Theoretical propertiesCategorizationCognitive mapSelf-concept
CategorizationCategorization is a basic cognitive process to
understand our world. Similarity / Dissimilarity
≠=
Cognitive MapMapping spatial mental structure of human
brain
Mental Structure
Galileo Space
Self-ConceptConscious representation of a personAn attitude can be defined as the distance
between the self-concept and the designated concept.
Galileo Space
Myself
ExampleGalileo survey
Participants: 218 UB undergraduate studentsA paper-based survey12 concepts for the study
A complete list of 66 pair comparisons
PovertyEducationHealthHuman RightsHuman ResourcesNatural Resources
Social SafetyGovernmental LeadershipGlobal CooperationGlobal ConflictInternational Aid
SelfSelf
Example of Galileo SurveyIf COOPERATION and CONFLICT are 100
units apart, how different or how far apart is each word or phrase from the other in the pair?
POVERTY and EDUCATION are __ units apart.
Mean Distance Matrix
Neural NetworkGalileo space
Message Strategy for IDCsHealth
Internationalaid
HumanRights
EducationTarget (Self)
Resultant
14.41 units
72.67 units
MessageINTERNATIONAL AID is closely related to EDUCATION, HEALTH, and HUMAN RIGHTS.
International aid promotes human rights for everyone throughout the world by providing health care and education. We want you to pledge your support. Join the growing global movement for international aid. Your support will help to achieve an increase in international aid, promoting human rights to provide health care and education for everyone in the world.
29
Comparison of the Galileo spaces
Comparison of Concept Distances
Edu & Health62.16(7.02) 59.07(6.92) 3.09 45.33(4.63)16.83Edu & HR 53.34(6.66) 47.95(6.29) 5.39 46.00(6.37) 7.34Edu & Aid 68.52(7.04) 55.76(7.41)12.75 47.88(4.57)20.64Health & HR 65.02(7.33) 50.32(6.53)14.69 45.64(4.79)19.38Health & Aid 64.67(6.80) 55.64(6.12) 9.03 50.93(5.33)13.75HR & Aid 60.28(5.94) 49.36(6.41)10.92 55.65(5.50) 4.63Aid & Self 104.94(9.44) 76.93(7.18)28.01 80.70(7.03)24.24
ΔDistance ΔDistanceDistanceControl (n=65) One-time (n=59) Two-time (n=94)
Note. () standard error
Kruskal-Wallis Test: χ2 = 7.043, df = 2, p < .05Mann-Whitney U test: Control > One-time = Two-time (p < .05)
DiscussionNetwork analysis has been applied for
diverse social scientific research.
Its practical application has a large scope from individual to international level.
However, network analysis has several limitations.Modeling dynamic processes over timeSampling & gathering dataMeasures (e.g. centrality and density)
ResourcesInternational Network for Social Network
Analysis (INSNA)
http://www.insna.org
SOCNET - a LISTSERV list, allows network researchers worldwide to discuss research and professional issues, make announcements, and request help from each other.
http://www.insna.org/INSNA/socnet.html