Bibs & Blather
Books and Blogs
Three items, some of which you may have seen on Walt at random. One of the items has been updated from the earlier post.
I may be a slow learner, but I can take a hint.
There have been no sales of Public Library Blogs: 252 Examples since June—and two sales of Academic Library Blogs: 231 Examples since June.
Barring a significant increase in sales, I’ll accept that these books didn’t serve a need and remove them from sale—probably around the beginning of 2009. That may make the (so far) 35 copies of Academic Library Blogs collector’s items…or not.
I might be interested in doing a better, deeper, lateral version of the two studies as a single study—but that could only happen with sponsorship.
I’m still open to The Liblog Landscape 2007-2008 having a sponsor, at the right price. Shoot me a line.
And I’d be delighted to work with the right agency on a combination of writing, data analysis/research and editing. Again, shoot me a line (waltcrawford at gmail.com).
Here’s what I said on September 24, 2008:
Phase 1 of the book is complete, resulting in a long chapter containing 607 brief blog profiles (to be revised later on) and one big Excel2007 workbook:
· A base sheet 607 rows deep and “S”–I guess that’s 19–columns wide
· A derived sheet (references to some of those 19 columns, plus a bunch of derived numbers) also 607 rows deep but “AD”–I guess that’s 30–columns wide
· A “working” sheet that starts each subphase as a values-plus-formats copy of the derived sheet, then gets manipulated to look at quintiles, build charts, create pivot charts, etc.
· A separate results workbook for results I know I’ll need separately–because it’s easier to have one workbook on one screen and the other on a second, in separate instances of Excel.
When I wrote that, I didn’t realize I should also make a special spreadsheet for each graph that goes in the book, since Word wants to store the spreadsheet to support proper editing. (Yes, I lost one graph in the first five chapters because it was linked back to a spreadsheet that I then deleted. Ah well: Ten minutes to recreate the graph and save it properly.)
So far? It’s turning out to be quite interesting, but I’m not ready to start publishing intermediate results. Here’s what I have so far in polished-draft form:
What this is all about, how I built the universe, some special notes (e.g., blog software used–and yes, there are still a few MovableType liblogs out there, although 18 doesn’t compare to 230 WordPress and 222 Blogger liblogs), notes about authorship and affiliation, some graphs related to age of liblogs.
Looking at the number of posts in each blog in the two study quarters (March-May 2007 and 2008), including quintiles for each year, quintiles for the change from 2007 to 2008, graphs as appropriate, a list of blogs in the first quintile for 2008 and a few notes about “subsets” (e.g., blogs whose authors are affiliated with public libraries, blogs affiliated with medical institutions, etc.–I think there are six or seven subsets large enough to evaluate).
A similar analysis–but this time there are two sets of metrics: Overall blog length and words per post, in some ways a more interesting figure.
Since September 24, I’ve also done:
Similar analysis of overall comments for a blog and the number of comments per post—and year-to-year changes in both metrics.
A brief chapter dealing with figures (illustrations, pictures, graphs) in each blog and figures per post—secondary metrics but still interesting.
This is the chapter I talked about (when it was Chapter 7) on September 24, as follows:
I’m particularly looking forward to Chapter 7, a three-factor analysis that will actually show whether a significant proportion of liblogs have “fewer posts, longer posts, more comments per post” or whether that’s a false or skewed supposition. (Well, it should show a lot more than that–but there are 45 different models for the three factors, even assuming only three values per factor: Significant growth, significant reduction, about the same. Yep, 45: 3×3x3=27, plus 3×3=9 (posts and post length) for blogs without comments and 3×3=9 (posts and comments) for blogs where length couldn’t be measured. Fortunately, that should work out to three tables and a whole bunch of commentary…
I miscounted. There are 48 different patterns—including three where neither comments nor length could be measured.
The chapter now has two parts. The first takes a simplistic view (each factor up or down, with no “about the same”), portrays it in one table and discusses that table and its meaning. The second has four tables showing the 48 patterns and some of their combinations, discusses some of the meaning and lists blogs that fit into most of the patterns.
Since the first draft of that lengthy chapter is complete, I can now say my naïve hypothesis on liblog changes is correct. And that it’s wrong. Well, you’ll have to read the book…
This chapter looks at 45 correlations between pairs of metrics, using a straightforward test to find cases where there’s a medium or strong correlation. For the seven cases where such a correlation exists, I provide the numbers, a scatterplot (cleared of extreme cases, with the modified correlation coefficient provided) and some comments about the correlation.
Chapter 8 looks at liblogs discussed in 2006 that are still around, considering changes from 2006 to 2008.
Chapter 9 discusses the “visibility” issue. No visibility or reach measures appear anywhere else in this book. I’ll explain why, what’s happened and consider a new minimal approach.
Chapter 10 may summarize conclusions and next steps, if there’s enough there.
Chapter 11 (or several chapters beginning with 11) will include individual blog profiles, all of which exist but need updating.
The book should be available from Lulu by the start of the ALA 2009 Midwinter Meeting—that is, January 23, 2009. I hope to have it available before then.
Yes, I will mount a list of the blogs somewhere on the web, either as an Excel spreadsheet or as a web page with links for each blog. No, barring sponsorship, I will not make the full metrics spreadsheet available for free. Even (most) OA advocates don’t argue that people should carry out hundreds of hours of unpaid work for nothing, giving away the results because it’s the right thing to do.
I’ve seen a number of comments on Technorati’s recent State of the Blogosphere / 2008 (technorati.com/ blogging/state-of-the-blogosphere/). This year’s report goes beyond most earlier ones (quarterly in some years, annual recently), with lots of analysis based on a survey of 1,000 bloggers. I’m going to ignore all that because it’s not relevant to my own interests—that is, liblogs and library blogs.
Most observers seem to focus on polls demonstrating how “mainstream” blogs have become (which I don’t doubt) and the growth in blogging–and ignore history, even though Technorati provides a direct link to the 2007 report and earlier reports.
Here, then, a few facts about blogs and related facts about liblogs. I assume Technorati’s actual numbers are factual; I see no reason to assume otherwise.
The “blogosphere” is much like Second Life: If you compare actual residents (active blogs) to counted residents (started blogs), it’s mostly a ghost town.
What’s “mostly”? 94% or more, depending on how you measure:
· The 2007 report said there were 70 million blogs as of April 2007, with 120,000 new ones emerging each day. If that 120,000 rate continued, there would be (or have been) about 120 million in June 2008, when the new study was done. The new study does not state the number of blogs or the number of new blogs each day–although it says “133 million blog records” since 2002, which presumably means 133 million blogs at some point. Technorati also quotes Universal McCann as saying that 184 million blogs have started as of March 2008. So let’s say there are (or have been) somewhere between 133 and 180 million blogs.
· Meanwhile, Technorati says 7.4 million blogs had at least one new post within 120 days—a pretty modest measure of “active”—and just over 5 million posted in June. But if 120,000 new blogs were being created each day (each with at least one post), you could reduce that 5 million to a mere 1.4 million ongoing blogs. Of course, on that basis, 7.4 million is smaller than the number of new blogs during a 120-day period.
· Those figures make no sense, so let’s be as charitable as possible and say between 5 and 7.4 million blogs are active, not just one-shot wonders. That’s somewhere between 5.5% of 133 million and 3% of 180 million.
A liblog comparison? Excluding “friends and family” blogs, I come up with 533 liblogs (not library blogs) that were active (using a 90-day cutoff) in 2008—and at least 90% of those had at least one post a month. Have there been 9,700 English-language liblogs since 2001 (or 17,800 worst case)? It’s possible—but I doubt it. Roughly 10% of the “visible” English-language liblogs that were active in 2007 had no posts during the 90-day study period in 2008.
That’s the truly impressive figure–and it does appear to be a direct comparison:
· In April 2007, Technorati counted an average of 1.5 million posts per day.
· In June 2008, Technorati counted an average of 900,000 posts per day: 40% fewer.
· Looking back, Technorati reported 1.2 million posts per day in April 2006–and 900,000 in August 2005.
· In August 2005, Technorati reported 14.2 million blogs and said 55% of them–or 7.8 million–were active. If that’s right, the active blogosphere is basically as active as four years ago, with a lot of churn in between.
· That also comes out to about one post every eight days for the active blogosphere (although of course the average doesn’t exist–only 1.5 million blogs had posts in a seven-day period)
Compare that with liblogs:
· For March-May 2007, 523 blogs had countable posts, for a total of 22,969 posts.
· For March-May 2008, 533 blogs had countable posts, for a total of 19,616 posts. (That 533 doesn’t include 54 blogs with posts in 2007 but not 2008–and includes 64 new blogs and blogs without posts in the 2007 quarter).
· That’s a drop—but a drop of 8.5%, which is a whole lot better than 40%!
· Those 533 blogs averaged about 213 posts per day as a whole, or about one post every 2.5 days per blog.
Overall, liblog posting declined at a much slower rate than blogs as a whole–and active liblogs are about three times as active as blogs as a whole.
· One-third of bloggers in general operate anonymously or with pseudonyms. That compares with 18% for liblogs.
· Roughly 85% of blogs as a whole have comment systems. Roughly 20% of liblogs didn’t have any comments in 2008–but that includes blogs that don’t allow them and blogs that just didn’t have them.
While liblogs have fewer posts now than a year ago—and for both liblogs and blogs in general, it appears that the peak was probably early 2007—liblogs are doing much better than blogs as a whole.
And when somebody blathers about hundreds of millions of blogs or says “everybody will blog in the future,” feel free to ignore them. It’s trivially easy to start a blog—but a lot of people (95%? 97%?) find, sooner or later, that they really don’t have much to say that belongs in a blog. Why should that be a surprise?
Cites & Insights is sponsored by YBP Library Services, http://www.ybp.com.
Opinions herein may not represent those of PALINET or YBP Library Services.
Comments should be sent to email@example.com. Cites & Insights: Crawford at Large is copyright © 2008 by Walt Crawford: Some rights reserved.
All original material in this work is licensed under the Creative Commons Attribution-NonCommercial License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc/1.0 or send a letter to Creative Commons, 559 Nathan Abbott Way, Stanford, California 94305, USA.