Why traditional scientific journals are redundant

Was it really six years ago that I first blogged about the Academic Journal Racket which siphons off millions from hard-pressed research budgets into the coffers of profiteering publishing houses?

Change is coming much more slowly over the last few years than I had anticipated when I wrote that piece, but at least there are signs that other disciplines are finally cottoning on to the fact that the old-style model of learned journals is way past its sell-by date. This has been common knowledge in Physics and Astronomy for some time, as I’ve explained many times on this blog. But, although most wouldn’t like to admit it, academics are really a very conservative bunch.

Question: How many academics does it take to change a lightbulb?

Answer: Change!!???

Today I came across a link to a paper on the arXiv which I should have known about before; it’s as old as my first post on this subject. It’s called Citing and Reading Behaviours in High-Energy Physics. How a Community Stopped Worrying about Journals and Learned to Love Repositories, and it basically demonstrates that in High-Energy Physics there is a massive advantage in publishing papers in open repositories, specifically the arXiv.Here is the killer plot:

citations_arXivThis contains fairly old data (up to 2009) but I strongly suspect the effect is even more marked than it was six years ago.

I’d take the argument further, in fact. I’d say that journals are completely unnecessary. I find all my research papers on the arXiv and most of my colleagues do the same. We don’t need journals yet we keep paying for them. The only thing that journals provide is peer review, but that is done free of charge by academics anyway. The profits of their labour go entirely to the publishers.

Fortunately, things will start to change in my own field of astrophysics – for which the picture is very similar to high-energy physics. All we need to do is to is dispense with the old model of a journal and replace it with a reliable and efficient reviewing system that interfaces with the arXiv. Then we’d have a genuinely useful thing. And it’s not as far off as you might think.

Watch this space.

32 Responses to “Why traditional scientific journals are redundant”

  1. Phillip Helbig Says:

    I think you underestimate the fact that arXiv has reasonably high quality because most of the stuff there (apart from that which is weeded out by the moderators) is also submitted to journals. Many people might read most papers on arXiv and not at the journal (whether paper or electronic), but aiming for journal publication does increase quality.

    So, you have to keep some quality standard and, related to this, keep refereeing as well, at least in some form, if you want to make the journals obsolete. If “a reliable and efficient reviewing system that interfaces with the arXiv” can do this, fine.

    You also need to seriously address the fact that not everyone can submit to arXiv. Even their endorsement system does not give one the right to submit to arXiv. (And there is the additional problem that someone might have written a good paper but doesn’t know anyone well enough to get endorsed. The days of Ramanujan writing to Hardy are gone.) Yes, one should keep out the crackpots, but at least in some cases (I’m not referring to myself here, unless I have a problem this weekend, in which case I will report back) it appears that lack of institutional affiliation can make it difficult for some people to submit to arXiv even if there is nothing wrong with their papers. Some people have a very loose affiliation they quote, some invent their own one-man institute, some might even use a bogus affiliation, I don’t know. Take my word for it: it is a real problem. The people who have managed to avoid it, in whatever way they can, have since published several papers in the leading refereed journals, which have also appeared on arXiv. Obviously, most or all such people are hesitant to come forward.

    In some cases, acceptance by a respected journal might allow a paper to get on arXiv which otherwise wouldn’t. You have a chicken-and-egg problem here, of course, if you rely on arXiv as your sole submission path. At the very least, there needs to be an additional mechanism to bypass arXiv. (There are also institutes which have a policy of arXiv only after acceptance.) You also need a public agreement with arXiv that any paper accepted by your journal (I’m assuming that there will be some corresponding status) will automatically be allowed on arXiv.

    And, finally, unless those involved in some new arXiv-overlay electronic journal publish at least most of their own stuff there, I don’t think that it will ever get off the ground.

    • Bryn Jones Says:

      Yes.

    • telescoper Says:

      I don’t agree with your first comment. Most papers go to arXiv before, or at the same time as, submission to a journal.

      I hope we will have a better refereeing standard than most journals, actually. There’s an awful lot of dross that gets through peer review these days.

      • Phillip Helbig Says:

        “Most papers go to arXiv before, or at the same time as, submission to a journal.”

        I haven’t done the statistics, but I’ll take your word for it. But that is beside the point. What I mean is that merely the fact that it is intended for journal submission at some stage improves the quality.

      • Phillip Helbig Says:

        “I hope we will have a better refereeing standard than most journals, actually.”

        We’ll see.

        “Thereโ€™s an awful lot of dross that gets through peer review these days.”

        Occasionally I do wonder how something got through. Still, compared to completely unrefereed stuff like, say, viXra, most stuff is OK.

    • Bryn Jones Says:

      Incidentally, I hit my own ArXiv submission problem a few months ago. I wrote an article about the history of astronomy that was published last year. It seemed like a good idea to submit it to the Philosophy and History of Physics section of the ArXiv, but I found I did not have permission to do that – I’m only allowed to submit to astro-ph. I don’t know anyone else who has submitted to that category of the ArXiv, so can’t get endorsement.

      It’s a bit odd.

      It’s possible a small number of serious researchers encounter similar problems trying to submit modern research articles to the ArXiv, perhaps in some developing countries.

      • Phillip Helbig Says:

        Indeed. This is one aspect of what I mentioned. Then there are people who leave academia but still publish something, usually related to their former work, from time to time. Some of these have problems.

        In a reverse of your situation, I know a rather well known retired professor at a prestigious institute who wrote a review article of a field in which he is an acknowledged expert, and the arXiv moderators moved it to the Philosophy and History of Physics section, even though it should have gone to the “normal” area.

        When this happens, they don’t tell you why and there is little, if anything, you can do about it. Complain too loudly and you will get banned for life. I am probably taking a huge risk by just mentioning this here, but if i run into problems as a result (of course, if this happens, no reason will be given), I hope some high-profile readers here will help me out.

  2. Anton Garrett Says:

    Folks, as someone long outside the university system I would probably not be able to submit to the arXiv without at least some kind of introduction, but let’s keep perspective and support Peter’s laudable and very important aim of getting research out from under the great publishing ripoff.

  3. John Peacock Says:

    Peter, I’m a fan of the arXiv, but I’m not sure what the plot demonstrates. If you’re a strong researcher doing good stuff, you may well be more likely to have the self-confidence to stick the material out on arXiv as well – but also vice-versa. Therefore, a plausible explanation of the plot is that papers that never get put on arXiv are just less good on average (with honourable exceptions, naturally).

    As to arXiv quality staying high because most papers are also sent to journals, I disagree. arXiv quality would stay high even if journals evaporated, because most people don’t want to make idiots of themselves in public. Now that immediate arXiving is now the norm (I think – although does anyone have actual statistics on this?), you know that it’s only the arXiv version that gets read, so it needs to be good. For this reason, I take the same care with everything I post, irrespective of whether it will eventually get refereed. Sending to a journal is just a backup in case arXiv should disappear one day.

    • telescoper Says:

      I aghree that there may be a selection effect, but it does demonstrate that, statistically, papers submitted to the arXiv do attract more citations. Whether that is because people who write good papers also stick them on the arXiv is not proven.

      I agree with your last point too. Institutions would be much better off paying to ensure the arXiv keeps going than continuing to squander cash on journal subscriptions. Of course many of them are supporting arXiv already…

    • Citations are a blunt tool for measuring visibility or impact and should be treated with caution. Journals have become reliant on arxiv for visibility: I am not surprised to see higher citation rates for papers on arxiv. Arxiv still relies on journals for its quality control. It may be on the way towards becoming a journal itself, but if so it needs to find new ways to deal with quality control. We support and pay for Arxiv, but to us it hasn’t replaced journals yet.

      Longevity is an issue. If funding would stop, the Arxiv paper archive could evaporate. I hear that has already happened to one or more electronic journals, so it is not unique to Arxiv. But it does need solving.

      • telescoper Says:

        “Arxiv still relies on journals for its quality control”

        Simply untrue. One can post papers on arXiv without any journal being involved.

      • Bryn Jones Says:

        Yes, but in practice, an appreciable fraction of the papers posted on the ArXiv have been prepared for traditional journals, and many have already been accepted after refereeing. So the journal quality-control process already affects much material posted on the ArXiv.

      • Judging whether an Arxiv paper is worth reading does involve, amongst others, where it will be published or has been submitted. Remove the journal information, and the judgement becomes more (even more) biassed by who the authors are and where they work. Yes, you can submit papers to Arxiv which will never see a journal (and one person does list papers as ‘accepted by astro-ph’). Would you read them?

        Peer review widens participation by validating papers by unknowns, it prevents errors creeping in to the literature, and forces authors to consider viewpoints other than their own. Peer review may be imperfect, but it is a solution to a problem. Without journals, that problem will still need solving.

      • Phillip Helbig Says:

        “Judging whether an Arxiv paper is worth reading does involve, amongst others, where it will be published or has been submitted. Remove the journal information, and the judgement becomes more (even more) biassed by who the authors are and where they work.”

        One of my points, exactly.

        “Peer review widens participation by validating papers by unknowns, it prevents errors creeping in to the literature, and forces authors to consider viewpoints other than their own. Peer review may be imperfect, but it is a solution to a problem. Without journals, that problem will still need solving.”

        Indeed. What we don’t want with regard to some sort of arXiv review is the equivalent of the Facebook “like” thumbs up. ๐Ÿ˜

  4. John Peacock Says:

    Bryn, Albert: I disagree in several respects. I select papers for further scrutiny in my regular arXiv trawls purely based on title followed by an iteration based on abstract. I assume that most will go to a journal but never check this. It’s a mistake to focus on known authors since great young people are entering the literature all the time. Of course you miss good things in this first pass; the papers you pick up subsequently are the ones you see cited or that get mentioned in conferences. This is the ultimate form of peer review: good papers get noticed, but bad ones are ignored. This is much more powerful than one referee sending a personal view.

    As for whether arXiv papers are improved by piggy-backing on journal peer review, the question is how often the refereed paper is strongly improved by refereeing. My impression is that 90% of papers are altered so little that you’d be hard-pressed to spot it. And perhaps this figure is biased low as not everyone uploads the final version of their paper. I presume this is because the differences are really tiny – and if this is not the case, then the authors concerned deserve all they get.

    So broadly speaking, authors get it right on first posting. To repeat: I’m sure this is so because they want to avoid public ridicule, not because they’re worried about what a single referee might say. If the journals had all packed up 10 years ago, the contents of arXiv would still be pretty much as it is.

    • Phillip Helbig Says:

      Of course, you can check this only by comparing arXiv papers before and after refereeing. Some don’t update after acceptance, some wait until acceptance to post to arXiv. I’m sure that you get it right the first time, but I don’t know how true this is in general.

    • John, I think without the referee step we’d see a lot more inflated claims being made in arxiv-only papers (they’d risk becoming press release like). I’ve reviewed papers where I had to push them to put in the caveats, that they were willing to state in response to a Referee, but forgot to add to the paper.

      So I do think that having journals/refereeing does provide some upward pressure on quality – but I guess there are a variety of approaches which would yield the same effect through some other less onerous/exploitative process.

      Finally, I have a real worry about relying on a group like arxiv, where its not clear who is chosen as a moderator and why. Having some oversight from the IAU would seem essential if we really were going to become reliant on it as a sole method of archiving and communicating (rather than its current use as a bulletin board).

      • Phillip Helbig Says:

        “Finally, I have a real worry about relying on a group like arxiv, where its not clear who is chosen as a moderator and why. Having some oversight from the IAU would seem essential if we really were going to become reliant on it as a sole method of archiving and communicating (rather than its current use as a bulletin board).”

        Good idea.

        Note that it appears that A HREF=”http://arxiv.org/help/moderation”>some appeals process now exists, though this appears to be solely within arXiv.

      • Phillip Helbig Says:

        Here is the proper link.

        With all the money saved when journal subscriptions are abolished, we should buy Peter a WordPress upgrade to allow commentators to edit their comments within 5 minutes or whatever. ๐Ÿ˜

      • Phillip Helbig Says:

        “Finally, I have a real worry about relying on a group like arxiv, where its not clear who is chosen as a moderator and why.”

        I think at some point it was unclear who the moderators were, at least in some fields, but there is now a public list of moderators.

      • Phillip Helbig Says:

        According to the proper link above, “arXiv moderators are volunteer subject specialists who have been approved by their discipline-level advisory committees and by arXiv staff”. The advisory-board page has links to the advisory committees in the various fields.

        Of course, the moderators are not referees.

  5. Anton Garrett Says:

    Out of (mostly) idle interest, plus issues relating to the cost of maintenance, just how big is the arXiv currently in GBytes?

    • Phillip Helbig Says:

      From the main page: “Open access to 1,065,858 e-prints in Physics, Mathematics, Computer Science, Quantitative Biology, Quantitative Finance and Statistics”. 1 Mb per e-print is probably a gross overestimate, so probably way less than a terabyte. You could fit it all on a modern USB stick on your key chain. ๐Ÿ™‚

      Even in the early days (back when it was xxx.lanl.gov), it was pointed out that disk space is not an issue, though bandwidth can be.

      • Phillip Helbig Says:

        1 MB is probably an overestimate for PDF files of -eprints, which are of course compressed. LaTeX source doesn’t take up much space, but of course PostScript files do. So, allow another factor of 10. Still, not that much in modern terms.

      • Anton Garrett Says:

        I’m sure that’s true of the text, but the photos accompanying one million papers?

      • Phillip Helbig Says:

        Do you mean figures in the papers? If these are PostScript plots, that is taken into account in my “another factor of 10”. Huge plots are bitmapped, so there is an upper limit to the size of a plot. I guess some observational papers contain “plates” which might be higher resolution, but presumably they are JPEG files which of course are compressed. I’m sure that the papers themselves don’t contain full-resolution photos from gigapixel cameras or whatever. Ditto for data cubes from simulations and so on; presumably stuff like this is accessed via a URL in the arXiv “abstract” and is not stored at arXiv.

        Note, however, that times are changing: “MNRAS now supports embedded video figures”. ๐Ÿ˜

        All well and good, I suppose, but every couple of months I do take pleasure in reading The Observatory with the old-style numerals. ๐Ÿ™‚

  6. Phillip Helbig Says:

    Let me float my suggestion again for the New Journal to have some sort of formal acceptance associated with a timestamp. Otherwise, papers become moving targets, and authors can revise them to take criticism into account after the paper was “published”.

    There are papers (I’ve written a few myself) which point out some flaw in another paper. There are many such papers, debunking Penrose’s rings in the CMB, cosmic birefringence, claims that observational data support the Einstein-de Sitter universe, and a host of other things. I would even claim that such papers, pointing out flaws in others, are essential for the progress of science. It is self-correction in action.

    A journal needs to be more than a list of links to arXiv papers. It also needs to be more than a list of links to arXiv papers with some sort of rating system. Since arXiv allows revision, it might not always be clear which version of the paper is being discussed. I don’t think readers should have to compare multiple versions of a paper to see what was changed when, possibly in response to which comment.

    There definitely needs to be refereeing in some form, and some flag to indicate that the editors have accepted the paper. After that, I don’t think it should change, except perhaps for updating references and so on.

  7. Anton Garrett Says:

    “Otherwise, papers become moving targets, and authors can revise them to take criticism into account after the paper was โ€œpublishedโ€.”

    They should be able to, to improve quality; the problem is about citing something that might have changed. Perhaps you should be allowed one change at one week, another at one month, and another at one year, each given different reference tags.

    • Phillip Helbig Says:

      “the problem is about citing something that might have changed”

      Right.

      Another issue: Suppose one starts with a paper which is completely wrong, but revises it as new information comes in, ultimately taking credit for the input from those who did the real work. I don’t see the point of this.

      When something is wrong, one can publish an erratum. This is nothing new, of course, and covers the cases where something really was wrong. But what about stuff like including more relevant references, adding plots which show the essential conclusions in a better manner, doing additional checks to improve the original idea? This would make sense for a review paper, or something like the particle-data book, but not really for the way original research is done.

      Not only for historians of science it is also important to have a snapshot of what people, or at least the authors, thought at the time the paper was written.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: