A couple of weeks ago I posted “Sometimes it’s just a waste” about preparing a “trimmed” set of liblogs, running some metrics, and finding the results so uninteresting that I discarded the partial chapter.
I got a couple of comments on the post, one of which eventually resulted in rethinking the effort.
Most of the new study, But Still They Blog, avoids averages as being meaningless in a universe as heterogeneous as liblogs. But there may be cases where averages are still at least mildly interesting–particularly if the universe is made a little more homogeneous.
Here’s what I did:
- First, prepared a table of averages and supporting figures for all of the blogs in the study, for the cases where averages might have some use. That table fits on a single book page with room to spare.
- Second, removed all blogs that don’t have length metrics for all three years (either because there were no posts or because I couldn’t measure length).
- Third, removed eight blogs that seem atypical–most of them some form of current awareness service, one a pure link-and-headline blog committed to one very brief post per day. That’s eight of the 325 blogs that have length metrics for all three years–not many, but they make a difference.
- Then prepared a table with the same averages and totals (a shorter table, since the first one had to account for blogs with no length metrics).
- Noted the differences–to wit, a more distinct dropoff in posts from 2008 to 2009, and a more distinct, if still small, increase in average post length–and provided two more figures: The number of blogs each year with “essay length” posts, using two definitions of “essay length.” Using either definition, the number of essay-length blogs has grown substantially over the years.
It’s not a big deal–it adds four pages to what will still be the shortest chapter in the book–but it’s mildly interesting. And, given the nature of the spreadsheet, it probably took longer to describe trimming the universe than to actually do it…