Problems with advanced publication

Now that the last of the Zhuchangtyrannus stuff is behind me, it’s worth returning to an issue that has in a way affected every aspect of the blogging and media side of things – the fact that what is actually available right now is an uncorrected proof. For those who don’t know, the basic procedure of getting a scientific paper published is that you submit your work to a journal, this goes for review, if it’s seemed suitable the reviews come back to you to make corrections based on them, this is returned to the journal editor and if deemed satisfactory are turned into a ‘proof’ version (i.e. formatted for publication), these come back to you to check and make final tweaks and then go back to the journal for final publication.

In the old days (and by this I mean as little and 5 or so years ago) unless you sent round copies to your colleagues, in general no one would see the paper until it was actually published – physically in paper form in a journal. The gap between returning your proofs and publication was generally weeks and was often months and occasionally years. As such journals soon realised that they were missing out on citations and credit for papers that were basically done and could be viewed as published and with the digital world in full flow, they could put these up in advance.

I see this as generally a good thing. It keeps things moving along and cuts down on the waiting time on work to appear and especially work that everyone already knows about and wants to cite but can’t because the journal is slow. However, many journals have now gone a step further and started putting up the proofs themselves. This is a very bad thing.

First off the more frivolous reasons, it’s annoying. I want the final version, not some pre-version. Journals still charge for these, so there’s a risk you could end up forking out twice for what is essentially the same paper. Similarly, the authors may have reasons for this not appearing before the final version for whatever reason and there’s generally no opt-out. It also makes doing media work hard (like here for me) because the paper kinda is published and kinda isn’t. It’s also a time waster – I fielded a number of e-mails from people asking for apparently ‘missing’ data in the paper which was absent only because of a formatting error, and it was, frankly, a pain as I had to keep trotting out the same reply to lots of people to give them what they wanted and suggest that I wasn’t a complete fool and did put it in there honestly.

More seriously there are taxonomic issues. I think the world’s taxonomists are just ignoring this issue from a practical point of view (which is probably for the best) but I doubt they’re happy. Zhuchengtyrannus is, right now, technically not properly published. The ICZN still require paper copies of papers to make names valid (see the PLoS One discussions of recent years). While I didn’t publish in an online only open-source journal, so far there are no paper copies of this article in libraries so they name doesn’t count. This is stupid and of course genuinely risks that some nefarious person will one day attempt to gazump a real piece of research by simply copying the diagnosis and specimen number, and self publishing a couple of pages of notes and stick them in a few locations and try to steal the name. No one wants this and it does offer unnecessary confusion.

Finally there is the real issue here. Proofs are not really, real. They can be subject to change and these can be profound. I really don’t know of anyone who hasn’t spotted a mistake in a paper they have published and most people would probably tell you there’s an error in most paper they’ve published one way or the other. It can be a simple as a missing comma, a spelling mistake or a mis-citation (I meant Smith et al. 1998 not 1999). I know of people who have spelled major things wrong (including the word ‘Cretaceous’ in the title once). Proofs are your last chance to catch these things and of course add in last minute changes (as I have to do with this one for example). Recently a co-author spotted we had a ‘not’ missing from a sentence in a paper which of course rather changed the meaning of the paragraph at hand!

While some of these (minor and major) will slip through, the vast majority won’t. That’s pretty much the point of the proofs. But if the proofs are already out there, people will see them and might think they are intended. They won’t know you went back and added in a ‘not’: Hey, Smith et al. said this IS true! Surely it’s not true!!!

Now that these are free game and not just seen by authors and editors the cat gets out of the bag in the wrong way. People will read them and cite them. But they will be reading and even citing things you might not officially have said. In the medium and long term the only record will be the real, actual, paper. A colleague told me that some time ago this almost happened to him, someone basically published a paper and in it, refereed to my colleague’s in press work (at the time a proof) and pointed out a significant anomaly in there. In the meantime, said colleague had seen it himself and fixed it. So you now have the bizarre situation that person A has a paper pointing out an error in a paper by person B. Only that error is not in the paper, and actually both papers agree on the point absolutely. This means that if you read both, A looks like a fool and if you only read A’s paper then you think B is. The situation was saved because A was polite and sensible enough to check before he went ahead.

Even so, it seems to me only a matter of time before some major eruption occurs because of these. Someone will get a major roasting in the literature for something they never officially said, or someone will leap on a point to prove their case only to see that data point vanish when the paper is properly published, or someone will be very unscrupulous with a new name or something similar. As such, I utterly hate advanced proofs and I’d love to see them return to whence the came.

23 Responses to “Problems with advanced publication”

  1. 1 David 09/04/2011 at 12:24 am

    In economics etc. we are only seeing “Corrected Proofs” going online. The only problem here is that they don’t have a volume/page number or article number. But no uncorrected proofs. We do have “working papers” but then it is in the authors control what goes up and what doesn’t. In some neighboring natural science areas, I’m seeing journals putting up a paper as “Discussion” and then later as a real published article. So all sorts of things seem to be going on in this transition phase.

    • 2 David Hone 09/04/2011 at 8:41 am

      Yeah, corrected proofs are very common and that’s fine, it gets your work out anything up to a year in advance of formal print publication. That helps, but then the author’s done his bit and had his chance.

  2. 3 mattvr 09/04/2011 at 1:58 am

    It sounds like a formal approach needs to be taken that engages with digital publishing.
    I can only imagine that physical publishers would fight tooth and nail against a system which allowed papers to be validated simply by printing them out and storing them in the appropriate libraries?

    • 4 David Hone 09/04/2011 at 8:37 am

      Well PloS One has got round the problem easily enough, but the paper copies are another issue, it’s the release of information (which may be incorrect) before it’s ready that has me bothered.

  3. 5 Schenck 10/04/2011 at 11:38 pm

    On the bigger issue, seems like people would have to vote with their feet, or submissions rather, anything else is just discussion, no? Might not be many choices there though.

    On the smaller issue of ICZN rulings, didn’t nearly the same thing effectively happen with Epidendrosaurus / Scansoriopteryx? What was supposed to be the official publication was claim jumped by a very small circulation ‘publication’, and then there was also a weird complication of the date written on the actual publication being differnent, etc.

    As you note, these small issues get worked out and aren’t terrible important in the end. For the larger, the only way to correct them is to be even more familiar with the literature.

    • 6 David Hone 11/04/2011 at 9:00 am

      All good points. In the first case, there’s not that much choice for palaeo journals and now a couple of major publishing houses are pulling this stunt, it rather narrows the field down if you want to avoid them. I have yet to find a journal that alters you as an author to this fact wither, you have to go hunting to find that they do it.

      I don’t think anyone was claim jumping with those as there were two specimens. But yes the confusion over ‘first publihsed’ online dates vs ‘printed date’ vs ‘date which the paper was distributed’ were all in the mix. As with many things the fact that it did happen let people work out what to do for next time.

      But here I can see the problems looming (and i’m not the only one) but the journals seem unwilling to recognise that no-one apparently benefits from this (not even them as it annoys authors). One day there will be a big issue and I can’t see why we have to have it.

  4. 7 Mike Taylor 13/04/2011 at 4:58 pm

    Yes, the current culture of making uncorrected proofs available is an absolute disaster. The otherwise near-perfect journal Acta Palaeontologica Polonica goes one worse, and makes accepted manuscripts available — horrible, ugly, MS-Word files with double-spaced lines and all the figures at the end. Ridiculous.

    Journals need to sort this out. NOTHING should be available to the world until everything is. Accepted manuscripts and uncorrected proofs are private, work-in-progress documents that the world has no reason or right to see. Corrected proofs are identical to what finally gets published, so if the journal wants to make them available to moment they’re ready then it should just declare the paper published — and lots of journals do this, calling it “published ahead of print”, or e-First, or whatever.

    If you have publicity to co-ordinate, it’s essential that you have one date that is when the paper becomes available. And in any case there should OF COURSE only ever be one publicly available version — THE version — of any given paper.

    Really: what the heck are journals thinking?

    • 8 Richard 13/04/2011 at 6:18 pm

      Mike – I guess that the idea is to make papers visible as soon as possible so that other workers can already begin to cite them. This potentially increases the number of citations that a paper will receive in the all important two years after it is actually published, and thus boost the journal’s impact factor.

      Gotta love impact factors.

      • 9 David Hone 13/04/2011 at 6:28 pm

        That makes sense in terms of the the e-first etc. level, as those can come out months before papers get published, but these pre-proof version are often only around for a few days or weeks which really don;t make a huge difference. And they can still only publish what papers they have. Shifting everyhting forwards two weeks won’t chance the number of papers they have out in any one year.

      • 10 Christopher Taylor 14/04/2011 at 5:24 am

        Shifting everything forward two weeks does make a difference when you consider the incomprehensible manner in which the ISI decides the ‘two years after publication’ in which a paper can be cited. ISI only counts citations within its own two working years. So a paper published towards the end of an ‘ISI year’ (I can’t recall exactly when impact factors for the year are published) may only be eligible for just over a year’s worth of citations, not the full two years. In that case, it makes sense to try to get publications as early in the ISI year as possible.

      • 11 David Hone 14/04/2011 at 7:59 am

        I get that Chris, but surely if you move *everything* forwards two weeks then effectively you are just moving the dates for the entire journal. The March ones come out in Feb, the April ones in March etc. but you still only get X papers per month and 12X papers a year for any given 365 days. So how does that help them?

      • 12 Christopher Taylor 14/04/2011 at 8:44 am

        No, the February issue is still coming out in February, but can be potentially cited from January. You’re then sneaking an extra month into the available time frame.

      • 13 David Hone 14/04/2011 at 8:51 am

        Yes, but then the Jan issue can be cited from December etc. So you’re still in the same position. Everything just moves.
        Like this (if the formatitng doesn’t kill it):

        J F M A M

        Cited from
        J F M A M

        J F M A M

        Cited From
        D J F M A

        But if each month only has say 10 papers, in any given month you still only have 10 new papers appearing. They just turn up early. But that’s the same as the month before and after, so it doesn’t chance the number of papers that appear and can be cited in any given month.

        Though what it does do I suppose is extend the deadline at the back end of things which will count differently. Eve so, this is still stupid for all the reasons discussed and of course most people don’t / won’t cite uncorrected proofs anyway. Ah journals, will they ever tire of screwing us around / over?

  5. 14 Mike Taylor 14/04/2011 at 8:51 am

    Christopher’s right. It’s unbearably stupid, but I bet this really is the reason journals make papers available before declaring them published: to give other workers a good run-up in the the two-year magic period.


    • 15 David Hone 14/04/2011 at 8:55 am

      In that case they can help themselves by not sitting on my papers for a year before reviewing them. I might then get them published in time with their precious citations in them….

  6. 16 Mike Taylor 14/04/2011 at 9:02 am

    Dave, I’m afraid you’re still not getting the Selfish Calculus Of Impact Factors (hereafter SCOIF).

    First, remember that you only get a two-year window after a paper comes out in which citations “count”. Then remember that in our field, getting a paper through writing, formatting, submission, review, revision, acceptance, proofing and into actual publication typically takes AT LEAST a year, which means that in any paper that I start writing in the second year of yours having been published is useless to your journal so far as impact-factor is concerned: by the time my work comes out, your paper has become obsolete in the mayfly-like world of impact factor. (I can only assume that IFs were invented by people who work in fields with much faster publication cycles). So it is very much to a journal’s IF advantage to have the paper available for citation anything up to a whole year before they declare it “published”. In that pre-publication year, few or no citations of it are going to be published anyway, because they just can’t make it through the slooow palaeo-publishing system in that time.

    Second, you suggested that journals “can help themselves by not sitting on my papers for a year before reviewing them. I might then get them published in time with their precious citations in them.” Sadly, that’s not the case. Unless your paper leans heavily towards citing only studies in the same journal, which I’m sure is a rare case, getting it out quickly will have a positive effect on the IFs of ALL the journals you cite, which cancels out from the selfish perspective of the journal in question. It’s good for the field in general (i.e. the IFs of all the journals in that field), but why would journal X care about that?

    All this constitutes just one more way in which Impact Factors screw things up. They are a complete and utter disaster.

    • 17 David Hone 14/04/2011 at 9:05 am

      “Sadly, that’s not the case. Unless your paper leans heavily towards citing only studies in the same journal, which I’m sure is a rare case, getting it out quickly will have a positive effect on the IFs of ALL the journals you cite, which cancels out from the selfish perspective of the journal in question. It’s good for the field in general (i.e. the IFs of all the journals in that field), but why would journal X care about that?”

      Well because it might not cite the *same* journal, but what about the family? Surely all journals owned by Elsevier say will benefit from me citing other them? They all go up and the net increase for the company as a whole is good. Kin selection almost…

      • 18 Mike Taylor 14/04/2011 at 10:23 am

        But you will also be citing Blackwell journals, will you not? it’s still something that is going to average out, whether across journals or publishers. From a competitive citation-scoring standpoint, there really is no benefit for a journal in getting your paper out quickly. (Of course there are lots of other benefits, such as that you will actually send stuff there in future! By contract, I know quite a lot of people now who won’t send anything to Paleobiology just because it’s so darned slow.)

    • 19 Christopher Taylor 14/04/2011 at 10:02 am

      I’m not inclined to give full credit to something quite so machiavellian. I’d be more inclined to suspect the fact that the editors are often different people from the publishers. For the publishers, their journals (and hence their journals’ circulation) are their main focus. For the editors, who are often working researchers with day-jobs beyond the journal, there are competing priorities.

      • 20 David Hone 14/04/2011 at 10:24 am

        Oh I agree, I think Mike is crediting them with too much evil genius. But when you think about it, it may well be true, in unconciously. They benefit from getting more citations if their papers are out early, but the others don’t benefit from it (or benefit far less) so there’s little motivation for them to get organised and deal with this as the authors would like.

      • 21 Mike Taylor 14/04/2011 at 10:24 am

        True, Christopher. You remind me of a very important piece of advice, known as Hanlon’s razor: “Never attribute to malice that which is adequately explained by stupidity”.

      • 22 David Hone 14/04/2011 at 10:33 am

        And I don’t think any author would disagree that there can be some monstrous stupidity when it comes to publishing.

  1. 1 An articulated alvarezsaur pes « Dave Hone's Archosaur Musings Trackback on 31/01/2012 at 8:11 am
Comments are currently closed.

@Dave_Hone on Twitter


Enter your email address to follow this blog and receive notifications of new posts by email.

Join 567 other followers

%d bloggers like this: