≡ Menu

Weblog conversations revisited: speed of linking

Has been playing a bit more with the data from Anjo… This we are looking at the speed of linking (=number of days between two linked posts) for links to own weblog and links to others. One of the hypotheses is that people would be more likely to link to their own “very old posts” than to those of others (simply because it’s much easier to remember relevant things in one’s own weblog. The results are below (in both cases number of links on the left is cumulative):

Speed of linking

Speed of linkingm %

Some notes:

  • There are more self-links (1086 total) than links to others (635 total), which is strange given that most of bloggers in our sample do not link much to their own content.
  • I guess we should look at the influence of outliers and to see if grouping bloggers by their self/others linking behaviour we give different picture.
  • Another reason for that could be in our dataset, since only links between bloggers in the sample are counted. So, may be there is someone who links a lot outside this group and those links are not in the picture.
  • Number of links decreases as time passes (=older posts are less likely to be linked)
    • Makes sense
    • For longer time-frames could be screwed up by the nature of dataset: if someone links to a post which is not in it those links are not counted.
  • As expected, linking to others is more “short-term” than linking to self.
    • Some numbers
    • Linking to others: 50% of links come within 3 days, 90% – within 51 day, longest time between posts is 283 days
    • Linking to self: 50% – 20 days, 90% – 169 days, longest time – 356
  • Does it mean that conversations with self are more longer-term?
  • Archived version of this entry is available at http://blog.mathemagenic.com/2007/08/23.html#a1938; comments are here.

    { 0 comments… add one }

    Leave a Comment

    This site uses Akismet to reduce spam. Learn how your comment data is processed.