After reviewing the numbers in The Gold OA Landscape 2011-2014 and considering what I can and, more significantly, cannot reasonably ascertain and judge in non-English journals and in short visits to websites, and in consultation with SPARC contacts, I made a number of changes in grades and, as a result, in exclusions.
I did not change the list of subjects and areas, although a few journals may have been assigned new subjects—and, as in the previous study, PLOS One is omitted from subject and area figures but included in overall discussions.
The fundamental meaning of Grade B has changed from “deserves attention” to “might be excluded from DOAJ or in some versions of Open Access.”
Changes in Grade A Subgrades
All subgrades for Grade A have been eliminated. Subgrade C (ceased) is now a subgrade for Grade B. Subgrades D, E, H, O and S—all cases where some year other than the first had fewer than five articles—have been collapsed into Grade B, Subgrade F (few or no 2015 articles) if the article count for 2015 is less than 5 and simply Grade A otherwise.
Changes in Grade B Subgrades
Grade B consists of journals that may or may not belong, either in DOAJ or in a study of open access, depending on your definitions. The old subgrades all have to do with mild visual or editorial issues that now seem as though they’re imposing my own values inappropriately.
There are four new subgrades—two from Grade A and two from Grade X, albeit with different letters.
- C: Ceased—journals that published at least one article later than 2010 but explicitly ceased during or before 2015, have merged with other journals, or show no articles more recent than 2012.
- F: Few or no 2015 articles—journals that published at least one article later than 2012 and published fewer than five articles in 2015. (By current DOAJ rules, these are subject to delisting.)
- R: Conference and other reports—journals consisting entirely or primarily of conference papers and other reports. These were previously excluded, in subgrade XN, as not OA.
- S: Sign-in or registration required—journals that require some form of registration before reading articles. These were previously excluded, also in subgrade XN, as not OA.
Changes in Grade C Subgrades
Grade C, “avoid this journal,” has been narrowed somewhat, specifically to eliminate subgrades that involve personal judgment or have so few journals that they’re hardly worth noting. Specifically, subgrades E (very bad English), S (incoherent site) and T (absurd article titles—there were almost none of these) have been eliminated, leaving subgrades A (APC missing), F (clear falsehoods), O (mix of problems) and P (implausible peer review turnaround). Briefly, clear falsehoods are statements such as “the leading journal in this field” for a brand-new journal; implausible peer-review turnaround involves promises to complete all peer reviews in a couple of days.
Changes in Grade X Subgrades
Grade X, excluded journals, retains the same subgrades—but the two largest categories within subgrade N (not OA) have been moved to subgrades BR and BS.
A Partial Checkpoint
What are the consequences of these changes? In general, and combined with more exhaustive checking of some difficult situations, they should mean that more journals will be included in the full analysis. As for specific results, those won’t be clear until the project is complete.
I thought it would be worth offering some glimpses into what might be happening at a natural breakpoint: essentially halfway through the first pass of data gathering (actually 5,500 of 10,948).
First pass? Yes indeed. There will be a second pass, beginning no earlier than April 1, 2016, for quite a few of the journals, for various reasons:
- Many smaller journals, especially in the humanities and social sciences, post online articles and issues with significant delays. In practice, even waiting a year won’t get them all. I’m rechecking all journals that appear to be missing final issues for 2015; this gives them at least three months to get the articles posted.
- I’m rechecking all journals that couldn’t be reached or that showed signs of malware, as well as those that showed as parking or ad pages or were unworkable.
- I’ll take a second look at journals excluded for various reasons, trying harder to make sense of opaque cases and translation difficulties, looking more closely for apparently-missing APCs, rechecking whether certain journals are OA or not.
So far, it looks as though I’ll need to recheck about one-fifth of the journals: 1,047 of the first 5,500. I’d be delighted if that percentage goes down in the second half—but I’d also be surprised.
All the rest of these numbers are truly tentative, since review of the journals may change their categorization.
Free and Pay
Some journals started imposing APCs that didn’t have them previously (one large publisher dropped all of its free introductory periods); some (fewer) drop APCs; and some clarify the nature of their charges.
Overall, the percentage of no-APC journals (among journals where it’s clear) among the first half dropped from 64.9% to 59.8%: there are more no-fee journals than in the previous study, but there are a lot more APC-charging journals. (There are also, to be sure, more journals in general: about 412 so far.) There are fewer journals (so far) where there is an APC but it’s hidden.
Most journals that weren’t in the 2014 study are simply A (that is, “nothing special here one way or the other”), but 30 have fewer than five articles in 2015, a few couldn’t be contacted or were unworkable, a handful fall into various other categories—and, unfortunately, nine showed signs of malware.
Some changes in grade and subgrade are neutral: they’re just redefinitions. That’s true for the journals that changed from various A grades to BC (ceased explicitly or with no articles later than 2012): there are some 218 BC so far. It’s also true for the various A subgrades that are now simply A (around 230 of them) and for a number of other changes including quite a few moving from B subgrades to A.
Some 300 journals had five or more articles in 2014 but not in 2015, moving them all to BF: some of those will add articles in a recheck.
Changes for the Good
Some 27 journals previously graded CA (APC missing or hidden) now have more clarity (and four changed to various X subgrades).
Quite a few journals with explicit falsehoods on their homepages have been cleaned up—at least 80 of them.
Half a dozen journals flagged for malware no longer seem to have that problem (but see later!).
Most “not OA” entries in the first half have moved elsewhere on re-examination or redefinition, including 35 journals oriented to conference programs (another seven that had been “A” appear to be predominantly conferences and have been moved here) and ten that require registration to read articles. Some two dozen moved elsewhere, including 17 that now appear to be proper OA journals.
Most journals that I previously found too difficult to count (XO) are now handled, and I hope to reduce the number (70 for this half in the previous study is currently down to 28) even further.
Roughly half of the XT (couldn’t understand the site well enough to measure it) cases have been cleared up: so far, there are only three such journals in the first half, and I’ll try all of them again.
Changes for the Bad
A few journals have changed home pages such that I can no longer find an APC (but am sure they have one), but it’s a tiny number.
Some 70 journals that were reachable the last time around are either unreachable or unworkable when I checked this time; they’ll all be rechecked, but it’s unfortunate that there are so many.
Finally there’s the most unfortunate group, in my opinion: journals that now show signs of malware—frequently, I suspect, because they include ad networks that don’t have proper standards. A journal gets flagged for malware if Malwarebytes or McAfee Site Advisor or Windows Defender flags it or some of its components as malware; cases include phishing attempts and deliberate malware downloads. There are now twice as many of these as there were (for this subset of journals) in the previous study, and that’s about 72 too many.
Hundreds of new journals; a much shorter and simpler set of grades; adding literally thousands of peer-reviewed articles that were given as conference papers.
Far fewer journals falling by the wayside because I only read English (thanks, Google!) or because I can’t or am unwilling to count them (with true broadband, I’m willing to open up a dozen PDFs a year to see how many articles there are).
There will still be some approximate counts, but fewer (and better approximations) than last time around.
And, of course, the results will be freely available to everybody. In a few months.