Walt at Random

Gold Open Access 2017-2022: Articles in Journals (GOA8) is now available in print book, PDF ebook, and dataset forms. The print book–a 6×9 trade paperback with color graphs–is $11.50 (or the nearest equivalent in other currencies supported by Lulu), of which I receive a stunning $0.24. The PDF ebook and dataset are both free, and all versions are CC-BY.

As usual, all links are available at https://waltcrawford.name/goaj.html, or you can use these direct links:

The Lulu trade paperback

The free PDF ebook

Figshare dataset

Same dataset on my own site.

Next, some time in July, Diamond OA 2023: The World of No-Fee Open Access Publishing, covering the same years and the no-fee portion of the dataset–but with a little lightweight new research added to give some sense of how such journals are funded (when they’re not directly from universities or societies).

Short form: The book and shared data will probably be available some time between Wednesday, October 24 and Wednesday, October 31 (2023).

That’s also the long form. I’m doing several proofing passes of the manuscript, about to start preparing the covers, and taking the time to do it right. There may be some additional checking on data, but not much–there will doubtless be a few duplicates and other minor slipups, as there have been every year.

As to the next step: Yes, I’m doing Diamond OA 2023: The World of No-Fee Open Access Journals–and I will be including some brief notes on apparent funding for journals not obviously published by universities and societies. (Guess what? They’re mostly supported by…universities and societies.) I’ve done quick checks on 1,800 of these and have about 360 more to check (interleaved with other work). Then it’s a matter of preparing an appropriate matrix and putting it all together. I’m guessing about a month’s work after GOA8 is complete, give or take a week or so. [For this first attempt. I’m including journals that appear to be temporarily diamond and the tiny number that seem to rely on Subscribe-to-Open. If there are later editions, I’ll reconsider those.]

A reminder: The spreadsheet is not intended to be an authoritative dataset; it should never be used in place of DOAJ, for example. On large-scale groupings, I’m satisfied that it’s at least 98%-99% right, but things can happen at the single-journal scale that may get missed.

The data’s as gathered as it’s going to be. 18,765 journals; 18,254 fully analyzed.

I’ve firmed up dates and ISSNs (using online ISSNs where available).

I’ve checked codes for consistency, etc.

And I’ve simplified and clarified some codes:

• Fee code is gone; just not readily available or ever called meaningful by anybody.
• Count code is gone. That was really for my own use. Of the 17,900 cases where I entered such a code, 10,081 came from DOAJ–with 2022 figures checked and frequently revised. 6,159 came from journals and were counted by using repeating metadata such as “PDF.” 1,083 came with help–e.g., article counts offered either by year or at least by issue. And 592 were damn nuisances, requiring manual counting. [These figures don’t count exclusions.]
• Code “bi” is now just “i” for “inactive” (and “a”=active.)
• Code “bx” is gone (I could probably resurrect it if needed), as are the notes.
• So the codes now are, for the main sheet, a (17,585); i (423); xm (143, up from 140 last year); and xx (102, up from 90 last uear). For the exclusions sheet: xd (defunct/duplicates, 100, up from 92 last year); xm (47, up from 12 last year); xm2 (260, DOWN from 383 last year; xn (4, down from 10 last year; xo (41l down from 119 last year); xx (40, up from 27 last year); and xx2,(20, same as last year). Total identifiable articles excluded: 3,994, almost all xm2.

As far as I can tell, of journals in the main sheet, 5,818 have fees and 12,435 don’t–but, as usual, most articles involve fees: 997,913 in 2022, compared to 440,229 without fees. Those figures could change slightly, but probably not by much. And the total is 1.438,142 2022 articles (from 16.984 journals with 2022 articles) and 1,322,021 2021 articles (from 17,344 journals with 2021 articles).

That’s it. Now I’ll go silent while IL

• Add derived data and save off the Figshare version
• Write the appendix–mostly from these reports.
• Start the book itself.

Help?

If you have OA-related contacts, let them know this is coming,

And if you know people who want to see the 12,435 “diamond” journals treated properly, let them know–and let them know I could still use feedback (it’s still an optional project, with no monetary gain). Before, say, the end of May.

I just finished the scan of xx2/xm2 journals, and–deviating from my original plan–actually did full retests in most cases.

This turned out well. At the end of the scan, I have 260 remaining xm2 cases (down from 383 last year, and that doesn’t include cases that became xm2 this year) and only 20 xx2 cases–same as last year, which means new long-term problems were offset by fixed cases. Those journals account for at most a few thousand articles–perhaps 4,000.

What stands out is that Indonesia is problematic. Last year, it had a majority of xm2 journals. This year, it has 71% of all excluded xm/xm2 journals and 83% of all xm2 journals: 218 of 260.

Right now, about 1,420,000 2022 articles are accounted for (not counting excluded journals). That figure will rise after I check 864 journals that seem reasonably likely to have more articles now than they did when first checked. These are small journals: they added up to about 37,000 articles in 2021 and, so far, 15,013 in 2022. I figure three or four days to do the retests.

[Then come overall data consistency checks, firming up ISSNs, saving the master lists, adding derived data…and, eventually, the fun part.]

Feedback on planned drop of “bx” still feasible. Feedback on “diamond OA” plan STILL REALLY WANTED. It is extra work, but if it will be useful, I’ll happily do it.

Oh, and sigh, just saw another pontification on the colors of OA that says flatly that all gold OA includes fees. THIS IS NOT HELPFUL! Diamond/platinum is a subset of gold.