18th home blog_twitter_English (12OCT2010)
Download as PPTX, PDF0 likes136 views
This study analyzed the social network structures of Korean politicians on their homepages, blogs, and Twitter accounts between 2009-2010. It found that Twitter networks had higher connectivity and more evenly distributed links than homepage and blog networks. While politicians' homepage and blog networks mainly linked to others within their own party, Twitter networks showed more cross-party linking, especially between the two largest parties. The number of followers, followings, and tweets on Twitter in 2009 and 2010 were also found to be correlated. However, having more followers in 2009 did not guarantee more in 2010, showing the importance of continued online engagement.
More Related Content
Similar to 18th home blog_twitter_English (12OCT2010) (20)
18th home blog_twitter_English (12OCT2010)
- 1. A Mixed Hyperlink Network Analysis:Homepage, blog, twitter among Korean politiciansChien-Leng Hsu (Post-Doctorate Research Fellow) Han Woo Park (Associate Professor)Department of Media & Communication, WCU Webometrics Institute, YeungNam [email protected]://www.hanpark.nethttp://english-webometrics.yu.ac.krPresented at the KAJCS, 16 Oct 2010, SNU, Korea
- 2. About this studyBased on previous studies (e.g. Beer & Burrows, 2007; Hsu & Park, 2011), this research intends to investigate online communication of South Korean politicians (18th National Assembly members) and conduct comparison between the network structure of different media platforms (i.e. homepage, blog & Twitter).
- 3. (Online) Social NetworkIndividual’s behavior is related to the larger web of social connection he/she fits into and affected by the types of relations he/she has with others in the network (Freeman, 2008; Haythornthwaite, 1999)Online networks are isomorphic to traditional/offline social networksHyperlinks can be seen as mediators of a wide range of associative relations between producers of web materials (Foot, Schneider, Dougherty, Xenos, & Larsen, 2003)
- 4. Research QuestionsWhat are the respective structure of Assembly members of homepage, blog and Twitter?Are there any similarities or differences in terms of network linking patterns from homepage, blog to Twitter?Do offline connections an factor in the formation of online linkage between Assembly members?
- 5. Data collectionDate of collection – April 2010Homepage – LexiURL searcher to retrieve data from the Yahoo! databaseBlog – Manually collected by visiting Assembly members’ blog pageTwitter – An automated computer program using Twitter’s API (Application Programming Interface) to retrieve data from TwitterAnalysis & Visualization – UciNet
- 6. Analysis TechniquesBasic concepts of Social Network Analysis (SNA)Density – how network members connect with each otherDegree centrality/centralizationTo identify actors more central than others in a networkThe more neighbors one has, the more central he/she is in a network
- 7. Overall Use of Homepage, Blog & Twitter – by partyFinding: Liberal (and progressive) parties use more of alternative communication channels (e.g. Twitter)
- 13. Party Use of Twitter - Following
- 14. Party Use of Twitter - Follower
- 15. Party Use of Twitter - Tweet
- 16. Twitter 2010 – Following Mean & S.D.
- 17. Twitter 2010 – Follower Mean & SD
- 18. Twitter 2010 – Tweet Mean & SD
- 19. Comparison of Density across 3 networks Grand National Party (GNP) had a more connected network on homepage & blog, but Merged Democratic Party exploited Twitter more.
- 20. GNP policy - Every member would have a blogComparison acrossthree networksPower law distribution shows:Twitter ties are made more equally than homepages
- 21. Comparison acrossthree networks The value should fall between 0 and 1.
- 22. The average distance of Twitter 2010 was the shortest
- 23. Distance_based cohesion (larger values indicate greater cohesion): Twitter 2010 was the most cohesive network
- 24. Distance_weighted fragmentation (larger values indicate greater fragmentation): Twitter 2010 had the lowest value thus was least fragmented.Comparison acrossthree networksNumber of ties (pairs) according to path distanceResults of Path distance:Distance between any pair of ties was shorter on Twitter 2010 (the longest path distance was 3)
- 25. Cross-party linkage on homepageDLP (Democratic Liberal Party)
- 26. MDP (Merged Democratic Party)
- 27. LFP (Liberal Forward Party)
- 28. NPP (New Progressive Party)Links were mainly occurred between MDP & GNP, or between same party membersCKP (Creative Korea Party)
- 29. FHA (Future Hope Alliance)
- 30. GNP (Grand National Party)Cross-party linkage on blogDLP (Democratic Liberal Party)
- 31. MDP (Merged Democratic Party)
- 32. LFP (Liberal Forward Party)
- 33. NPP (New Progressive Party)Links were mainly created within parties (MDP, GNP), with a small portion between GNP & independent members CKP (Creative Korea Party)
- 34. FHA (Future Hope Alliance)
- 35. GNP (Grand National Party)Cross-party linkage on TwitterDLP (Democratic Liberal Party)
- 36. MDP (Merged Democratic Party)
- 37. LFP (Liberal Forward Party)
- 38. NPP (New Progressive Party)
- 39. CKP (Creative Korea Party)
- 40. FHA (Future Hope Alliance)
- 41. GNP (Grand National Party)
- 42. Similar to homepage & blog, links were mainly created within and between the two parties (MDP & GNP)
- 43. MDP created more links to other parties than GNP didPreferential attachmentTwitter 2009 & 2010 :Pearson CorrelationTwitter 2009& 2010: Spearman CorrelationBoth Pearson Correlation & Spearman Correlation show:The number of followings, and the number of followers are significant correlated
- 44. Indegree of politicians presented in both 2009 and 2010
- 45. Preferential attachmentComparison between the number of indegree of politicians appeared both in 2009 and 2010A higher number of followers a politician had in 2009 did NOTguarantee the politicians to have more number of followers in 2010.
- 46. Preferential attachmentPearson Correlation:Number of Tweet, follower & following in Twitter 2009 & 2010 were correlated.
- 47. Preferential attachmentSpearman Correlation Coefficient:Number of Tweet, follower & following in Twitter 2009 & 2010 were correlated.
- 48. ConclusionsHomepage network:The structure was sparse; few nodes had higher degree centrality and other members communicated with one another through these nodes.Blog network: Boundaries between different parties were more closely definedNodes were better connected, as compared to the homepage networkTwitter network:Nodes were much more connectedThe number of followers, followings and tweets were correlated in 2009 and 2010Small number of sample: further examinations will be required.
- 49. Twitter network (cont.):The number of followers does not grow over time:Politicians need to be active to attract more audience (e.g. keep posting)Not all politicians used Twitter limited our findingsProgressive/liberal parties were more interested in using the medium to advance the agenda, to gain support, etcParty policies – a factor in the use of the WebComparison across homepage, blog & TwitterLink distribution on Twitter was more even than homepage & blogDistance between any pair of nodes was shorter on TwitterDifferent results from other Twitter related studies (e.g. Ohmynews) reflect the need to employ different tools and methodologies for the dynamic network.
- 50. Thank you.