Frontiers: Get the numbers right

On October 13, 2015, Frontiers posted a piece on its blog, “Frontiers’ financial commitment to open access publishing.” It’s sort of an effort to be transparent about the OA publisher’s finances, although “sort of” may be the right qualifier, as it lumps all the publishing-related stuff into one $6.8 million chunk (only 34% of total spending).

But I’m not commenting on the piece in general; at least, it’s more data than we have for a lot of other players in scholarly journals. Instead, I’m concerned about the second paragraph, because I believe it at tends to undermine the remainder of the post.

I have a slight concern about the very first sentence in the post as well, to wit “In 2014, the annual cost of traditional, subscription-based scholarly journal publishing was $14 Billion” Suddenly seeing total spending on traditional scholarly journals (this isn’t cost, this is price) jump by 40% is rather startling. But at least that figure is sourced, sort of: apparently it involves combining two different reports, one of them published before the year it supposedly covers. But I’ll leave that for somebody else to deal with.

Here’s the paragraph in question:

Open Access does away with subscriptions to allow any reader in the world unrestricted access to scholarly articles. To provide this option, Open Access publishers directly charge the authors an Article Publishing Charge (APC), which authors typically pay from their grants or receive institutional support to cover the cost. The APC generally ranges from $500 to $6,000 with an industry average of around $3,000. Often people wonder “Where does this money go?”

The first sentence is just fine. The last sentence is just fine.

But those middle two sentences:

To provide this option, Open Access publishers directly charge the authors an Article Publishing Charge (APC), which authors typically pay from their grants or receive institutional support to cover the cost.

Can y’all repeat with me the old refrain? Most OA journals do not have Article Publishing Charges. And yet this sentence doesn’t say “most Open Access publishers” (which would be false) or “the OA publishers publishing the most articles” (which would be true for some fields and as a whole, but false for others). Nope. It’s an unqualified “Open Access publishers.” Which is convenient, of course, if you’re an APC-charging OA publisher…

I’ll give you two real numbers–one that’s Published by a Reputable Publisher but covers only about two-thirds of serious gold OA journals (that is, the journals in DOAJ), another that covers nearly all serious gold OA journals but is not Published by a Reputable Publisher, instead being self-published by, well, me.

Published by a Reputable Publisher*:

67% of the gold OA journals that are accessible to English-speaking readers, reachable on the web, and actually published articles between January 1, 2011 and June 30, 2014 do not charge article processing charges. (Biomed is the only broad segment in which a slight majority of journals–53%–do have APCs.)

Published by Walt Crawford with transparent methodology:

Of the 9,512 gold OA journals that don’t raise warning flags and published articles between 2011 and 2014, 74% do not charge APCs. (A higher percentage of non-English journals are free to authors.) Of the 8,760 that actually published articles in 2014, 72.9% do not charge APCs. In 2014, 42.8% of articles in gold OA journals were in journals that don’t charge APCs. Even in biomed, a majority of journals (56.2%) do not charge APCs.

Now, let’s look at the second of the two offending sentences:

The APC generally ranges from $500 to $6,000 with an industry average of around $3,000.

As with the first sentence, this one’s not sourced, but here are some real numbers–again, one covering two-thirds of serious gold OA journals (those at least partly accessible to English-speaking people) and Published by a Reputable Publisher, and one covering nearly all serious gold OA journals:

Published by a Reputable Publisher*:

Fees range from $8 to $5,000. The average is such a silly figure that it wasn’t published, but only 12% of APC-charging journals charge $2,000 or more and only 16 out of 2,064 charge $3,000 or more. So an “industry average” of $3,000 must define the “industry” to include only the 30 most expensive journals. (42% of the journals charge less than $450, for what that’s worth.)

Based on article counts, the average APC per article in APC-charging journals is $1,045; the average across all articles is $630. Even for biomed, the average across APC-charging journals is $1,460, a far cry from $3,000.

Even in biomed, a minority of APC-charging journals charge $1,451 or more.

Published by Walt Crawford with transparent methodology:

I didn’t provide top and bottom figures because they’re not very meaningful, but the top quadrant of APC-charging journals (that is, the 25% with highest APCs) begins at $1,420 and the second quadrant begins at $600, so the average is somewhere in the $600 range. (Even in biomed, only 520 out of 1,365 APC-charging journals charges $1,420 or more.)

Based on article counts, the average charge per article for APC-charging journals in biomed was $949 in 2014.

Let’s look directly at the spreadsheet (not published, and in this case I’ll even include a few hundred journals that seem to be sketchy–those graded C–for a total of 9,824 journals):

  • The highest APC in 2014 was still $5,000, with only one journal at that level, but among this broader group of journals, there are 28 charging $3,000 or more–that is, 28 out of 2,619 APC-charging journals.
  • The low is now $2 (in U.S. dollars), not $8.
  • The average APC is $830.
  • The median APC is $600, as you’d expect.

Ah, but those numbers include a few journals that I regard as sketchy. So let’s look at just the 9,512 journals that appear to be good:

  • The high is still $5,000, with 28 journals out of 2,470 charging $3,000 or more. (That’s just a little over 1%–an odd version of “average.”)
  • The low is still $2.
  • The average APC per journal, a very silly figure, is now $842.
  • The median APC is…still $600.

I suppose you could inflate that “average APC” a lot by including so-called “hybrid” journals, which tend to charge extremely high fees for the handful of suckers wealthy authors/funding agencies that pay to (possibly) make their articles open while shoring up the $10 (or $14?) billion subscription marketplace. But by my reading, it’s a bad set of numbers (with no sources provided).

*Open-Access Journals: Idealism and Opportunism, published as the August/September 2015 issue of Library Technology Reports, an imprint of the American Library Association.

2 Responses to “Frontiers: Get the numbers right”

  1. Obinna Ojemeni says:

    So 67% of Gold OA journals do not charge author-side fees, this is quite surprising because it is usually believed that APC is a must for OAJs. Considering that Beall is of the opinion that APC is what introduces the conflict of interest that encourges predators. However, he described those OAJs that do not charge APCs as Platinum OA.

  2. Walt Crawford says:

    Yes, a majority of gold OA journals have been funded through other means for a long time.

    As for Beall’s assertions, it’s odd that he never mentions that many, perhaps most, subscription journals also impose author-side charges (“page charges”), and that subscription journals justify price increases on the basis of publishing more papers–thus setting up precisely the same conflict of interest.

    As for terminology, Beall is neither authoritative nor particularly interesting.