Conversation and Connectivity in the Blogosphere

1
Conversation and Connectivity in the Blogosphere

• Blog Research on Genre (BROG) Project
• School of Library and Information Science
• Indiana University, Bloomington

2
BROG project members

• Faculty
• Susan Herring
• John Paolillo
• Students
• Ben Clark
• Inna Kouper
• Sarah Mercure
• Lois Ann Scheidt
• Sharon Stoerger
• Pete Welsch
• Elijah Wright

3
The Blogosphere

• the collective term encompassing all weblogs (cf.
  blog biosphere)
• the intellectual cyberspace inhabited by
  bloggers (Wm. Quick, 2001)
• blogs as a community blogs as a social network
  (www.samizdata.net)

4
Blogs and conversation

• Full-blown conversations are carried on
  between three or five blogs, each referencing the
  other on their argument or rebuttal of the
  others positions
• (Newsweek, 2002)

5
Blogs and conversation (cont.)

• Weblogs enable a massively distributed but
  completely connected conversation covering every
  imaginable topic of interest
• (Marlow, 2004)

6
Research questions

• How much conversation takes place in the
  blogosphere?
• What is it like?

7
Sampling procedure

• Random sampling from blo.gs site
• Selected first four blogs with links to other
  blogs in sidebars.
• Snowball sampling from random blogs
• Followed and recorded all links to blogs in
  sidebars out three degrees of separation.
• Result 5,517 unique URLs and 14,890
  source-destination pairs
• Random sampling from snowball sample
• 50 blogs with minimum 5 posts in March 2005

8
Resultant data sample

• Established, active blogs
• Links in sidebar
• Part of a loose network
• Resemble blogs most often discussed in the media
  and blog research (Herring et al., 2004b)

9
Analytical methods

• Social network analysis of links in snowball
  sample (n5,517 blogs)
• Visualization of most densely interlinked blogs
  (n254 blogs)
• Content analysis of posts and comments (n50
  blogs)
• See Herring et al. (2005). Conversations in the
  blogosphere An analysis from the bottom up.
  HICSS 38.

10
Visualization

• Pajek
• Cut-off at 10 in-degrees
• Three main clusters
• Catholicism
• Homeschooling
• A-list (political commentary humor)

11
(No Transcript)
12
Content analysis

• For each blog (A)
• Coded conversational units
• References in posts
• Comments
• Coded connections between units
• Distinguished within sample (B) and outside
  sample (non-B) references and comments
• One month (March 2005) coded
• Entirely, for B blogs
• 1st three posts 1st 10 comments per post, for
  non-B blogs

13
Example references

• Power Line has good news about Iraq a topic that
  Glenn Reynolds notes is apparently unfit to print
  in The New York Times. blog name blogger
  name
• Indeed, as the man who first linked to Frizzell
  would say. indirect nominal reference
• Get Well Soon, Blogfaddah. direct address

14
Example comments

• ...examine how National Guard deployments are
  affecting state readiness.
• readiness for what? the Parti Québécois tank
  units that are massing at the border? quotes and
  responds
• This is actually very good news for Firefox
  users. deictic reference
• Attila Girl, () I am sorry for what you went
  through. I hope you have peace with it. direct
  address 2nd person pronouns

15
A blog conversation

• (Post by Blogger A) Blogger non-B says ltlinkgt,
  and I disagree
• (Comment by Blogger non-B) Blogger A, you
  misunderstood me
• (Comment by Blogger A) Blogger non-B, thanks for
  your reply

16
Results Conversational units

• N582 units
• References in posts to other blog(ger)s N135
• Mostly blogger's name, blog name, and link to
  content
• Hyperlinked 48 Trackback 10.3
• Comments by other bloggers N447
• Mostly implicit and 2nd person address

17
Connections between units

• 0 references reply to a post in another blog in
  the sample
• Comments reply by definition
• To current post (83)
• To previous comment in thread (17)
• Longest exchange in sample (post comments) 5
  units

18
Inside vs. outside sample

• References to B 29
• Linked from sidebar to B 49
• References to non-B 71
• Linked from sidebar to non-B 31
• Comments by B 0
• Linked from sidebar to B 0
• Comments by non-B 83
• Linked from sidebar to non-B 27

19
Summary

• More comments than references in posts
  participate in conversations
• Comments address other bloggers directly
• References refer to other bloggers and/or link to
  their content
• Most comments and references involve blogs
  outside the sample
• Only 27-49 of blogs that are conversed with are
  also linked from the sidebar

20
Conclusions

• Not all blogs interconnect, although some do
• Conversing blogs dont necessarily link to each
  other
• Most blog conversations in sample take place in
  comment threads, not posts

21
Study limitations

• Sampling method creates bias towards connectivity
• Average blogs may be less conversational
• Did not track reciprocal conversations outside
  the sample
• References in posts may receive responses from
  blogs elsewhere
• 50-blog sample is small
• May not represent the blogosphere as a whole

22
Future research

• Need for systematic, large-scale study of
  exchanges among bloggers
• Need to take account of private (e.g., email, IM)
  as well as public exchanges

23
brog_at_stderr.org