Rejected Papers Get More Citations When Eventually Published

scibri writes "In a study of more than 80,000 bioscience papers, researchers have illuminated the usually hidden flows of papers from journal to journal before publication. Surprisingly, they found that papers published after having first been rejected elsewhere receive significantly more citations on average than ones accepted on first submission. There were a few other surprises as well...Nature and Science publish more papers that were initially rejected elsewhere than lower-impact journals do. So there is apparently some reason to be patient with your paper's critics — they will do you good in the end."

Comments

Surprisingly?

[Anonymous Coward]

Not at all. Papers that were previously rejected benefit from additional, careful revisions by their authors, therefore they end being of higher quality than they would have.

Re:Surprisingly?

[Arker]

Not at all. Papers that were previously rejected benefit from additional, careful revisions by their authors, therefore they end being of higher quality than they would have.

The conclusion the journals would like you to reach, and it may explain some of the effect. Seems to me likely the larger effect is simply that papers which break new ground tend to be controversial to old guard, thus accumulate rejections, but when and if finally published they also tend to accumulate more citations as well, as even the old guard will then have to cite in their rebuttals.

Re:Surprisingly?

[Sir_Sri]

Reviews, in my experience in physics and computer science anyway, tend to be more about process than results. Did the work you do cite whomever the reviewer knows about in your field (usually not), is the process you used for your work valid, or importantly, is it clear that it is valid? Are your results understandable to someone who didn't do your particular experiment? Is your paper clear enough that an expert in the field but not your research can build on it? Admittedly I might be biased because I'm bad at explaining myself.

Reviewers aren't out to screw submitters for the fun of it, because they don't want to be screwed themselves and the whole point of research is to find new stuff. But internally when you're preparing a paper your supervisor, or your grad students or the like know what the fuck you're talking about, and someone on the outside can say something to the effect of 'this makes no sense'. "This makes no sense' by the way doesn't mean the work isn't valid, just that you might have done a terrible job trying to write about within whatever constraints your target journal has.

Journals are also becoming a bit overspecialized, and it makes it hard when you do something really high impact that isn't very narrow to know where to put it. My particular corner of the academic universe is broadly under the field of game development, but we really combine work in AI, computational social science, strategic studies and economics, and an AI journal may look at our work and feel (correctly) that it's not enough of an AI problem for them, the computational social science people will say the same thing and so on. The big dogs of Nature and Science publish things that can be cross disciplinary and that aren't (and shouldn't be) pigeon holed into a particular basket. The place I'm at has, or had at least, some really stellar computer vision and computer algebra researchers but odds are if you aren't specifically in those fields you couldn't care less what they do. That can be good work with few citations just because it's really really important to one very tiny problem. A journal rejecting you because you aren't doing enough pathfinding for their pathfinding edition doesn't mean you didn't do good work, it just means the work you didn't isn't as applicable to what they're doing.

I expect as time goes on we'll see more of this. New researchers aren't as biased by picking up a physical journal and reading it, they do internet and database searches (which are becoming much easier and much higher quality) and glue together concepts from multiple journals, but then their work may not really fit with what the previous ones did. It can be broadly interesting and broadly useful, but it's not the feeder disciplines.

If you want a good example take Quantum computing. The 'Quantum' mechanics side of the research fits in easily a dozen different journals (e.g. MRI, Laser spec, semiconductors etc) but very little of the work is meaningfully original, it's a new application of an old problem. The theory of computing side of things probably only belongs in one or two different theory of comp journals, but again, it's re-examining some of the fundamentals of computing theory in a different way and there's not much to do after the first guy does it. But they are, together very interesting.

Peer review doing its job

[Anonymous Coward]

Peer review *does* work. Yes, part of its job is to filter out the poor papers that don't deserve publication. That's the obvious part. But I've gotten plenty of papers back with comments like "deserves publication, but X, Y, and Z need to be fixed". Or even "rejected, but if X were addressed, should be reconsidered", and so on. So, you go off and do X, Y, and Z, resubmit, and you've got a better paper because you've addressed the critical comments. Good papers are ones that incorporate constructive criticism, so it makes sense those might eventually get cited more. Also, if it's a paper that was rejected somewhere, then it might be something controversial that people want to argue about. So, publish a paper that makes a claim some people don't agree with (hence the rejection), and those critics will publish their own paper slagging the original one. Putting it another way, in order to say someone else's paper is full of crap, you have to cite it, and if a lot of people are saying it's crap, then you'll get a lot of citations :-)

Peer review isn't perfect, but the described pattern makes sense. What I'm surprised at is their ability to statistically detect these patterns given all the other variables involved, but I guess a sample size of 80000 helps.

Re:Peer review doing its job

[WaywardGeek]

My problem with this system is that only people who get paid to publish papers wind up getting heard. For example, the entire EDA industry continues to use interconnect delay estimation algorithms that have been obsolete for 15 years, because a paper I tried to publish on how to do it right was rejected. Sure, it would be a better paper if I put in the work you talk about, but I don't get paid for that. I just get paid to deliver better implemented solutions. You could read patents I have in the area, owned by QuckLogic, but good luck finding a sensible paper in the field.

Re:Peer review doing its job

[ThreeKelvin]

Communication skills matter in science!

It doesn't matter that you have the invented the greatest algorithm since quicksort if you can't or won't tell other people about it. If you can't convince other people how great your work is, they won't use it, and therefore you won't have contributed to the field. When you die the knowledge disappears, and you might as well never have invented the algorithm in the first place.

Therefore, it is important to convince your audiance that:
- Your algorithm gets the job done. (Proofs)
- Your algorithm is better and/or just different than existing algorithms. (Extensive litterature search so that you can compare your algorithm to existing ones)

Just reporting your algorithm together with a "this is how I do it" doesn't cut it. We researchers don't have the time to examine every claim somebody makes about something in our field.

Re:Peer review doing its job

[WaywardGeek]

You are quite right. One thing I've learned over the years is that I should have learned to write well in college, rather than trying to pick it up while being paid to program. I've written SBIR grant proposals (all were funded), various statement of work proposals, and a number of patents. It was damned hard for me, but it had to be one. I'm steering my 10 year old son towards taking writing more seriously, and hopefully he'll not have the same writing handicap I had when entering the workforce.

I will dispute one point. In EDA at least, results are what counts. Once algorithms are written and benchmarks performed showing yours beats other well known algorithms in some area, IMO, you have data worth sharing. This is not how it works today. I put my own hobby research on the web, but as I've said, I gave up on dealing with the PITA journals ages ago. Here's a great algorithm for better speech frequency analysis. Here's a better speech speedup algorithm for > 2X. Promoting algorithms I develop for free is also painful. Getting sonic into Debian was not fun at all, though it seems to have been adopted in Android and several TTS engines almost magically. I believe it's now even in the Android Audible client, which is now far superior than the iOS client for high speed. I can't get either algorithm linked to on Wikipedia because my web pages don't pass their test as a credible source.

How are you supposed to share great ideas?

Another possible mechanism for this

[Brett Buck]

I haven't done a lot of publishing in open literature, but many times, the papers that fly through the vetting process with little effort are are on topics that are somewhat straightforward/trivial. And would thus not be as likely to be useful as a citation. The interesting topic raises many more questions and is more likely to require multiple tries to get through the review, but ultimately is more useful and more likely to get a citation.

    Brett

Re:Another possible mechanism for this

[WaywardGeek]

There are two kinds of papers: invited papers and papers where professors do the peer review thing. I've published some invited papers, but in my experience, there's always at least one a-hole on the review committee who will shoot down my work. The worst example is Professor Larry Pillage, who QuickLogic paid $20K to review my work which I did based on a paper called RICE. He never got it working properly like I did, but after reviewing my work, he claimed it as his own and published it at DAC the next year for a best paper in show award, and then made a mint selling it to all our competitors. That guys is a serious a-hole. He was on most committees I ever tried to publish a paper in, and while I don't get names with the reviews, the psychotic analysis I did sometimes receive seemed 100% Prof Larry Pillage. He stole ideas from great guys like Prof. Ronald Rohrer, who told me once, "We don't tell Larry anything!"

If publishing papers were more important to me, I'd do something about it, but the reality is people with ideas don't get to publish. People with the right connections and background do. This explains why Nature would do better with rejected papers.

Re:Another possible mechanism for this

[Grieviant]

I haven't done a lot of publishing in open literature, but many times, the papers that fly through the vetting process with little effort are are on topics that are somewhat straightforward/trivial. And would thus not be as likely to be useful as a citation. The interesting topic raises many more questions and is more likely to require multiple tries to get through the review, but ultimately is more useful and more likely to get a citation.

My experience is the exact opposite. Papers that address new topics or ones that are 'all the rage', even when there isn't much substance to them, get preferential treatment from reviewers. See MIMO and Cognitive Radio in the field of wireless communications. These areas are cash cows for grant money until the next flavor of the day comes along, and the fact that very little practical impact was made is quickly forgotten.

Same thing when novel but inferior algorithms are presented. For example, I once saw polynomial prediction applied to a specific problem where linear prediction had already been well studied - the results were worse in every possible way, but the novelty factor was enough to push it through.

In contrast, classical topics that are considered 'old' or 'well-studied' (but are by no means 'solved' and are still quite relevant) are poorly received. The reviewers tend to lazily dismiss the work without giving it due consideration.

maybe not editing?

[Anonymous Coward]

The summary seems to suggest that when a paper is rejected, the author edits it in hope of being less rejection-worthy the second time around.

I don't think the data provided is adequate to show that. An alternative hypothesis is that papers vary in risk and "risky papers" are more likely to both be rejected and , once approved, to be cited.

Re:maybe not editing?

[Chuckstar]

This is one of those "I came on here to say this but you said it better" posts.

The only study they site about how much editing is done between submissions seems to indicate "not much at all".

Also, this could explain the prevalence of Nature and Science in the study. Risky papers may be rejected by the orthodoxy of more specialized journals. Nature and Science may avoid that kind of orthodoxy simply by having a broader array of reviewers than a more specialized journal might.

Pick a completely zany example in a field I know nothing about: You've done cutting edge work on the transport of lipids in palm fronds. There aren't that many people in the field, and most of them have been following a line of reasoning that lipids are transported by osmosis, as there has been some evidence of that. You have the hypothesis that lipids are transported by tiny aphids. You do some interesting lab work, that seems to support the hypothesis. The results are supportable, publishable, but not entirely definitive -- you know, statistically significant, but there's really a lot more work to be done before anyone can say you've really proven your hypothesis correct.

You submit your paper to a botany journal. They have it reviewed by a bunch of palm frond experts. All of them have studied palm fronds very closely. They mostly adhere to the unproven pet theory of osmosis. They do not consciously reject your paper because it contradicts their pet theory, but between the kinda shaky results, and their unconscious bias, it gets rejected.

Now you submit to Nature. They only have a couple palm frond experts in their stable, so they have a fern expert and a deciduous leaf expert review it. Neither the fern expert nor the deciduous expert have any unconscious bias. They think the whole thing seems pretty interesting, and worthy of broader discussion. So it gets approved. Now that the whole palm frond field has seen your work, a couple guys start trying to replicate your work, adding their own twists. They come up with supporting evidence and publish, with a reference to your work. A couple more guys come up with evidence that supports a different hypothesis. Even if you end up having been wrong at this point, your work gets referenced because it is what stimulated them to do their work. Let's say they discovered a third mechanism, that seems to explain transport even better. This third mechanism is unrelated to osmosis, so few of the osmosis studies are being referenced. But because your work stimulated the whole line of inquiry, your wrong study still gets lots of references.

(I added the part at the end about the work ending up being wrong just to illustrate that risky scientific invetigation doesn't have to end up being right in order to get referenced a lot. It just has to stimulate inquiry in a different direction such that people keep referencing the original study. A paper that just advances the field in an orthodox direction may still be great science, but may get lost in a sheaf of other similar studies advancing the field in that similar direction.)

Alternate Viewpoint

[merlinokos]

"So there is apparently some reason to be patient with your paper's critics — they will do you good in the end."

I have a different possible viewpoint. The papers that are most likely to be rejected are the ones that are controversial because they challenge the status quo. But once they're accepted, they're game changers. And since they're game changers, and the first publications with the new viewpoint, they're cited disproportionately frequently by follow up work.

(formatted correctly this time)

Comment removed

[account_deleted]

Comment removed based on user account deletion

Duh

[Anonymous Coward]

Papers initially rejected are improved based upon the reviews of outside critics. It seems this means they end up being better papers overall. Who'da thunk it!

As a PhD student I was advised early on that you learn to love the rejections.

Surprising

[Joe Torres]

Something that surprised me was that "75% of all published papers appear in the journal to which they are first submitted."

I would be very interested in seeing the difference of this rate between junior faculty and senior faculty. With my limited sample size (and personal bias along with it), it has seemed that this number would be much lower for junior faculty. Possibly, junior faculty may be too eager to try to swing for the fences (Science and Nature) and miss (going down the ranks to PLOS ONE) while senior faculty already have favorite field-specific journals (where they may know editors) that will likely be accepted with revisions.

Retraction Watch -- for the details

[sillivalley]

A very good site to monitor is Retraction Watch - https://retractionwatch.wordpress.com/

They not only follow retractions in journals, but dig into them, and track them to other papers and publications by the same authors.

For those of us in industry, we forget there are areas of Academia that are dog-eat-dog, publish or perish.

Under such pressures, authors make up data, manipulate data and/or images, and more.

Take a look at Retraction Watch for the sordid details -- for us outsiders, it's like a soap opera for the geeky set!

not all citations are a sign of a quality paper

[ganv]

Don't underestimate the number of citations you can get by being controversial or wrong.

Re:Alternate viewpoint

[pnot]

Game-changing papers may encounter more initial resistance, but I have to tell you as a reviewer that most rejected papers are rejected because they're poor and/or trivial.

True, but remember that here we're not considering the set of all rejected papers; we're considering the set of rejected papers which were subsequently accepted. That probably removes from consideration a large chunk of the just-plain-awful ones.

Nature, Science publish papers rejected elsewhere?

[Convector]

I have never heard of a paper being rejected by a journal and then sent to Nature or Science. It's the other way around.

And especially at smaller journals,

[aussersterne]

the search for legitimacy of their own leads them to ultimately consider only papers that completely agree with conventional wisdom and support the already big names and big theories.

Not to mention that the reviewers that are willing to review for smaller journals are usually in the same boat—younger faculty trying to get a leg up—and subject to the same pressures and tendencies.

But even at the large and important journals, there is a tendency to dismiss really interesting papers unless they come from a large name / large name school. You'd better have a long track record and big names behind you or you won't get serious consideration, even if your work is sound and earth-shattering. It's just a matter of the probability of returns on the investment of labor.

I say all of this as someone that did sit as a managing editor on an academic journal and that has been a part of the review process for any number of articles.

There are serious inherent biases built into the system, both for good and for bad.

Much more important to my eye is the fact that this is all free labor but earns the publishers huge profits and costs the schools huge dollars. It's only a matter of time before the current system is overturned. Right now, schools pay money to faculty to write papers, pay money to faculty to review papers, then pay lots of money for the journals. Yet all of the authority of the paper comes from the faculty and from the institution, and circulation is limited to academics because articles run $30-$60 a pop for public access. It's only a matter of time until they cut out the middleman, save tons of costs, and grow their audience at the same time.

What about in mathematics?

[oneiros27]

I want to know how Rejecta Mathematica stacks up to the others.

(for those unfamiliar with it ... they only take papers that have already been rejected somewhere else, or when the author doesn't want to make the changes that the peer-reviewer is insisting on)

Re:Missed steps...

[blueg3]

Now, my writing does comparatively suck, and I've never had the patience to do all the leg work as you're suggesting.

So, your writing is bad and you don't have the patience for proofreading or copyediting, but you're surprised -- or rather, have come up with a near-conspiratorial excuse for the fact -- that your submissions to journals whose purpose is, ostensibly, to communicate the results of your work to others so that they may learn from it have been rejected? ...

Perhaps there's a simpler explanation here.

Re:Alternate Viewpoint

[tgv]

Perhaps 'game changers' is an exaggeration, but only papers that make minor extensions to the existing literature are accepted on first submission. In my (ex) field, papers that challenge a certain view get their share of flak from the reviewers. I've seen papers being shot down (see what I did there?) because the reviewers belonged to a different school. It's of course not always the case, but it does happen too often. One of the reasons is that such papers usually get reviewed by at least one of the opponents, or someone closely involved. Consequently, when such papers get accepted, they generate replies, and thus citations, in contrast to the papers that are in line with the main view.

I think the conclusion that the GP has a good point and that the conclusion "peer review works" cannot be drawn on.

Re:Alternate viewpoint

[mikael]

There has been a study on how research group leaders tend to cross-reference each others work. That's the only way they can keep publishing. It's known as "citation analysis". Depending on the field of science or industry these are known as "Collaboration graphs", "citation graphs" or "Hollywood graphs" (for movies actors have starred in - some actors can form a natural inspiration for each other, like Laurel and Hardy, or The Three Stooges). For academics, they co-author papers together because they are experts in the same field.

Had one of my papers blocked for publication around 2004/2005 only to see the exact same paper published from California. That sucks.

Re:Alternate Viewpoint

[mikael]

It's s a shame, but it's true. Like the invention of the combustion engine. The first working model would be considered noisy, fuel inefficient, and extremely environmentally unfriendly with soot and oil in the exhaust, but the paper published would have been considered seminal. Successive papers would document how to improve airflow, air-mixing, reduce turbulence, reduce noise, improve burn rate, improve fuel inefficiency, but they wouldn't be considered as ground-breaking.

If you had a monopoly on all the research on that field and could afford to wait 5 to 10 years, then you could put together all the improvements and that would appear to be a completely revolutionary engine.