aggienaut | Blogological Statistics IV

   Every year in the general area of mid-June I've examined the statistics of blogology. I hate to let traditions die, so I ran the numbers again this year.
   Additionally, it has come to my attention that many blogs see rankings on a webpage called Technorati.com as the premier measure of how successful they are. As far as I can tell these rankings are entirely based on how many blogs have linked to your own blog in the last six months. Anyway, I included technorati rankings in my latest compilation so as to compare the usefulness of the technorati rankings.

Control Group
   Since my original sampling of statistics consisted only of livejournals I thought would give interesting results, it is far from a random sample. This is unfortunate, since a random sample is exactly who you need in order to get a blogosphere-wide "average" to compare things to.
   Fortunately livejournal has a "randomly select a livejournal" feature. I recently used this to create a 20 blog sample group.
   In assembling this I came across 5 that lacked friends-of or other information in their profiles, so we can assume approximately 25% of the lj-blogosphere would fall in this group. I discarded these livejournals. These ljs looked like they probably sucked though anyway. =D
   Seven of the 20 were written entirely in cyrillic. Theoretically this would imply 35% of ljs are Russian. (I did not discard these)

Medians: The Control Group had the following medians:
Friends: 24 (average: 54)
Friend-of: 28.5 (average: 61)
Friends-of / friends: 1.00
Comments / entry: 2.39
Comments received / comments made: .79
Age of LJ: 24.0 months
Friends-of / month: 1.54 (ie growth rate)
Friends-of / entry: .178 (ie a growth rate roughly corrolated with activity on lj)
Entries / month: 7.56
Comments received / month: 14.51
Comments received / friends-of / month: .789
Technorati authority number*: 2
Technorati rank: 2,454,419th
Percentile of above: 68.4% (as in, the median is ranked better than 68.4% of other blogs on Technorati, more on this below)
Hops from me: 4.25 (as in, the median fell between a 4 and 4.5. The 4.5 comes from the fact tha lj-connect will show two paths, and if one is 4 and the next closest is 5, I coded it at 4.5)(There was only one 3 hop random person, everyone else was 4, 4.5, or 5)

   * I believe the Technorati "authority number" is how many blogs have linked to the given blog in the last six months.

   The Technorati percentile was calculated using the 0 Authority rank number of 7,758,595. In theory, it seems like the median livejournal should be right at the 50.0000th percentile, rather than high above it. I believe the difference is that Technorati of course incorporates all sorts of OTHER blogging sites ... thus we have definitive proof that everything on livejournal is 36.7% better than everything that is NOT on livejournal!!! Well not really, but I think it does make a compelling argument that livejournal is much better at getting one's blog noticed than if its floating in the dark blog-stratosphere of blogspot or such.

Specimen Group Results
   First off, one of the most alarming things I discovered was that the median for friends-of growth was 98.54% (ie -1.5%. For the sake of this discussion "100% growth" is actually no change. Follow?)-- that is, more than half the 33 people I studied had a reduction in their friends-list over the last year! By comparison, last year the median was 111.4%. I hope this is because my sample group is getting old and moving on, and not because there's a general loss of people from the lj-blogosphere. Since I can't get information on random ljs status a year ago, its going to be a year before I can see how the random sample group is faring in this regard.
   I don't know what professor_david's been up to, but he had a rather unprecedented 144% growth for the last year. For an lj thats on its third year thats pretty impressive. My illustrious sockpuppet has, however, had a growth over the last year of 309%. (= I myself increased my friends-of list by 5.68% (or 105.68% to keep consistent with terminology here).
   Two of three ljs that weren't even updated once this past year were above (had less friends-of loss than) the 25th percentile of my sample group. For at least the second year in a row, otimus loses more friends than dead people do -- his blog would be more popular if he didn't update! (And thus it is probably a complement that he thinks I have the most boring blog ever)

   Another new statistic I added was the "Comments received / friends-of / month," in order to measure the quality of one's friends list. It started out just as Comm Rec / F-of, but I quickly realized that unless I controlled for longevity there'd be more weight on ljs that simply had been accumulating comments longer.
   This statistic was interesting because the huge megablogs didn't actually do so well. Topping the list was beastmario with 6.2 comments / friends-of / month. shid did extremely good overall at 1.8, followed by pavel_lishin at 1.66, roter_terror at 1.38, apoplecticfittz at 1.25, and Technorati 99.97th percentile theferrett only at 1.23. The other top Technorati ranked blogs all came in below the Control Group median. This is not terribly suprising, however, since those blogs have huge friends-lists of passive observors. Since this statistic penalizes for having passive observors beyond a core group of active commentators its a relatively bad indicator of successful blogging. (But there's no reason not to be proud if you scored well in it!)

Rating Rating Systems
   As to the technorati rankings themselves, I suppose its a pretty good cross-blogosphere measure -- especially at the higher numbers. With the vast majority of blogs falling between 1 and 3, however, its far too imprecise for most blogs. Since the authority number is a measure of how many blogs have linked to them in the last six months, the number is well below the threshhold of statistical significance and can be bounced around wildly by such things as me randomly name dropping in this entry. For example, the fact that I mentioned Otimus will probably bounce him up an "authority point," increase his ranking by more than a thousand, and raise him some twenty percentile points. Crap on that.
   And incidently, my sockpuppet is beating most of you with an authority of 4 at the 82.7th percentile. (I myself am at 7)
   I think Comments / Entry is probably the best measure. Some high Technorati scoring blogs are still at the top, but they are decent blogs so it makes sense. Ones that Technorati ranked highly and I frankly couldn't for the life of me understand why, such as kebechet and ajat, are more reasonably placed lower on this scale, at (YTD) 9.63 and 3.5 respectively. Interestingly, ajat has a comments made / received ratio of about 5:1 -- meaning the number of her comments no one ever bothers to respond to is off the charts. High scorers from my friends list included apoplecticfittz at 27.39 and pavel_lishin at 12.32

Picture of the Day