Walt at Random

It’s September 30–the last day of the month, when I fetch usage statistics for my websites (as always, omitting about 18.5 hours of that last day), so here’s an update on GOAJ–just the total numbers this time:

• Paperbacks: Still just my own copies.
• Dataset: 876 views, 430 downloads.
• GOAJ: 36 Lulu copies, 6,912 copies from my site: total 6,948. Since the stats miss a bit less than 3% (that last day), it’s fair to assume that more than 7,000 copies have been downloaded. That’s remarkable,
• Subjects: 17 Lulu copies, 171 other copies, 188 total.
• Countries: 8 Lulu copies, 907 other copies, 915 total.
• C&I: To date, 994 copies of the excerpted GOAJ version (16.5) and 3,724 copies of “APCLand and OAWorld” (16.4.)

Just a quick informal update on my “gray OA” research, having achieved two milestones:

1. Getting past row 655 of the publisher spreadsheet, the Publisher Who Shall Not Be Named (actually four imprints). It didn’t show up as 700+ full-OA journals when I checked it out (at least not from the APC table, the easiest way to deal with it), but as 618 journals, 54 of which were hybrid “OA” and 19 empty, leaving 545 with at least one article between January 2012 and June 30, 2016. Articles (using the publisher’s DOI scheme, which leaves out editorials in most cases): 9,041 for the first half of 2016; 14,198 for 2015; 10,798 (in 337 journals) for 2014. That’s about 20% of all the active journals for the first 655 rows and 15% of the 2016 articles, but only 13% of active-in-2014 journals and 10% of 2014 articles.
2. Getting to row 701 (that is, 700 publishers of 1,027 total).

Roughly, the first 700 show 2,718 journals with at least one 2016 article, for slightly under 60,000 articles January-June–and 2,723 journals with at least one 2014 article, for slightly under 105,000 articles for the full year.

It is, of course, possible that the remaining 327 publishers and 1,000+ “independent” journals will provide the 315,000 articles for 2014 you’d need to get to the frequently-flouted peer-reviewed number of articles in 2014 in “predatory” journals–but it’s a tad unlikely. (This also assumes that all gray OA is predatory, despite the lack of any evidence at all in 85+% of cases. That’s another discussion.)

Now for the rest of them…

As noted in August, I’m trying to determine the actual size and flavor of “gray OA”–that is, gold OA journals that weren’t in DOAJ as of 12/31/15.

I’ve now reached roughly the halfway point: 500 publishers and “publishers” (and ones that are neither) out of slightly more than 1,000.

Here’s what I find:

• There are 10,044 journal names so far–excluding ones that are in DOAJ, but…
• Only 2,655 of those have published at least one article between 1/1/2011 and 6/30/2016, and can be properly analyzed.
• 6,402 “journals” are entirely empty–and most of those consist of nothing more than template-generated webpages.
• 587 might have articles, but don’t state APCs–and most that I’ve checked have very few articles.
• 402 either can’t be reached, don’t work, aren’t OA at all, are “hybrid” or otherwise don’t belong.

You may notice that those numbers are nowhere near being 25% larger than in the August update, although they include 25% more “publishers.”

Article counts? 38.666 for the first half of2016; 79,787 for 2015; 75,608 (approximately) for 2014; 56.040 (approximately) for 2013; 40,593 (approximately) for 2012.

Are there some paper mills here? Yes, I think so, although not many (and I’m guessing most of the authors know exactly what they’re doing). To wit, the ten journals with the most 2015 articles (that’s ten of 2,655, or less than 0.5%) have between one-sixth and one-fifth of all the articles–e.g., 15,870 for 2015 and 14,678 for 2014.

Still not willing to suggest what the final numbers will be.

And now setting it aside to write the big essay for the next C&I…and take care of other business.