4

What (researcher's) behaviors (desirable or not) are encouraged by an incentive system based on publication throughput?

Considering this as a research problem, I propose to list all the potential strategies that a researcher can apply to "game" various academic systems and to increase one's recognition by such systems, starting in this question with academic systems which evaluation is based on the number of academic publications. Later objectives could be to design mechanisms to detect and measure tendencies to follow such strategies, and to identify other set of strategies potentially used to maximize one's number of citations, or more interestingly one's h-index and i10-index, but those later objectives are not part of the discussion here.

To make the discussion cleaner, I propose to remove all moral judgments about the strategies, and to merely list all that could be applied by an academic sociopath in order to maximize his/her success. The goal is not to encourage such behaviors (obviously?), nor to criticize institutions which embrace some of those strategies, but rather to identify clearly the consequences (desirable or not) of an incentive system based on publication throughput.

Feel free to add other strategies in your own answer below, or to edit my own answer below, collaborative Q/A style.

  • 2
    If the asker knows the answer to the question, a blog is a more appropriate medium of expression. I downvote. – mkc Apr 12 '15 at 15:14
  • 7
    @Ketan Your downvote is misplaced. Stack Exchange explicitly encourages self-answering. – David Richerby Apr 12 '15 at 15:38
  • 3
    If you want to increase your academic recognition, focusing on your number of publications is utterly counterproductive. A small number of high-impact publications is not just more desirable in principle than piles of salami-sliced incremental papers that nobody reads, it's more a significantly more successful strategy in practice. – JeffE Apr 12 '15 at 17:20
  • 2
    @JeffE, I know (my publication record should tell anyone that I am definitely not maximizing my number of publications). Yet many institutions have incentives toward maximizing publication throughput: e.g. Chilean universities pay research AND salary money for each separate publication, require "at least two publications in international conference" for each PhD students, Chilean funding institutions require "at least one publication per year", etc... The answer to the question above would show them what they are pushing for with such incentives. I will edit the question accordingly. – Jeremy Apr 12 '15 at 17:44
  • 8
    To everyone: this is not the horrible question that it may first appear to be. I first read it as the OP trying to game the system. Rather, I think he's trying to record the absurd conclusion that current employment conditions in some parts of the world imply. – Pete L. Clark Apr 13 '15 at 0:27
4

This is a terrible question, because it presumes that publications are the goal of science, as opposed to an indicator of actual meaningful intellectual contribution.

Therefore, I will provide a solution that proves the metrics are meaningless per se:

  1. Generate a sequence of N meaningless papers using Sci-Gen.
  2. Have each of the meaningless papers cite K of the other meaningless papers in the sequence.
  3. Identify a set of crap journals that do not perform meaningful peer review with the aid of Beall's list.
  4. Publish the meaningless papers in series in the crap journals, thereby obtaining an i10-index of N and an H-index of K (assuming N>=K>=10).

If self-citations are excluded, then enlist (or fake) a colleague to do the same and cite each other's meaningless papers rather than one's own. If Journal Impact Factor matters, then sequence the papers over three years.

Citation metrics are the shadows on the wall of Plato's cave.

  • 2
    I love your algorithm! Note that the question does NOT presume "that publications are the goal of science", on the contrary: the goal is stated as "to identify clearly the consequences (desirable or not) of an incentive system based on publication throughput". Your algorithm would contribute to a list of strategies encouraged by an incentive system based on one's h-index and i10-index. I fully agree with your metaphore on Plato's cave. – Jeremy Apr 12 '15 at 16:48
  • 3
    I am very tempted to try to drive a stake through the heart of publication metrics in this manner. – jakebeal Apr 12 '15 at 17:36
  • So, a metric is meaningless because there is noise? – JiK Apr 13 '15 at 8:36
  • 1
    @JiK The metrics can be easily gamed, meaning that any meaningful judgement based on them is not really based on the metric, but on something not so readily quantified, i.e., scientific contribution. – jakebeal Apr 13 '15 at 11:37
3

If Academia was a game, and if the winning rule was to maximize one's number of publications (which I don't think it is, but which the administrators of some institutions seem to believe), I think that the following strategies would make sense:

  • Focus on Research :: Abandon all attempts to personal life (and good teaching) and dedicate your life to research. You might manage to produce more publications than fellow researchers who have a personal life and/or spend a large proportion of their time teaching, but never more than fellow researchers applying a combination of the following strategies.
  • Focus on low hanging fruits :: Choose a finite set of conferences. While attending or when the proceedings are published, review it quickly to mark publications overlapping with your sphere of expertise. Assign a limited time-length (e.g. two days or two weeks) to each problem, and see if you can produce some quick results on it. Spend more time on it ONLY if you have had results within the allotted time, otherwise move to the next topic. Aim to delegate the writing of the results to a junior colleague or to a student in order to optimize your time. Notice how attacking a "long-standing open problem" will NOT help you to maximize your publication throughput.
  • Minimal Publishable Unit :: adjust the content of a submission to the minimum required, so that to maximize the number of publications. NEVER join two conference papers into a single journal publication, even when they largely overlap, as this would reduce your number of journal publications. Aim to publish the journal version of each conference article as soon as possible so that, if the results of your next conference publication overlaps with the last, the journal version is reviewed BEFORE the second conference publication is submitted.
  • Self-Plagiarism :: when a partial result of independent interest can be used in various more important results, avoid creating an easy to reference lemma in one publication and to refer to this lemma in the others, but rather reproduce its proof in each of the parallel submissions, in order to make each submission look more technical and lengthy. Aim to submit the various results in parallel so that to further justify this multiplication with the excuse to make each submission independent from the other. Aim to write minor variants of the result in each submission in order to further confuse the situation.
  • Publishing Clusters :: collaborate with a finite set of x colleagues of your field who have had in recent years at least y yearly publications per year at a given level agreed upon. Agree on a number z (less than y) of research themes per year that each of the members of the cluster is to propose to the whole cluster, to research them and describe their results on their own, and to submit it to the whole cluster for proof-reading. Aim to increase by a term of roughly zx publications (minus potential rejections) your number of publications per year. Aim to participate in at least two distinct clusters in order to confuse things, and optimally to y distinct clusters, contributing exactly one publication to each cluster. As your career advances, rotate clusters to further confuse things, at the cost of a more complex management. Participate in various clusters at various levels of publications to further confuse things.
  • Pyramidal Scheme :: Maximize the number of doctoral students, postdoctoral fellows and junior faculties in your laboratory or research group. Minimize your involvement in each project, but insure co-authorship of each report/publication. Focus on acquiring funding to attract and recruit more people. Pay attention to your people's need in term of future careers only when they are relevant to attracting more people and to maximizing your output. When you reach the maximal size at one level, apply to the next level.
  • Falsify experimental results :: Producing fake results, presented in a difficult to reproduce fashion, in a field where experiments are costly and lengthy, will give you an edge over fellow researchers who labor to gather the funds to perform those same experiments, and make them waste time trying to reproduce your experiments. Aim to recruit junior researchers and/or students and to push them in non-explicit ways (or at least not in writing, only orally) to fake their results so that you can put the blame back on them if/when discovered.
  • Focusing on the research and low hanging fruit is not cheating as the tag in the question seems to be suggesting. – mkc Apr 12 '15 at 15:13
  • 1
    @ketan: I removed the tag #cheating given that you find it misleading: the theme of cheating is related but not central to this question. Note that the contribution to human knowledge of incentives encouraging scientists to "focus on research" or on "low hanging fruits" is debatable (but that this debate is, once again, out of the scope of this question). – Jeremy Apr 12 '15 at 16:51
  • 1
    @ketan: note that beyond "Falsify Experimental Results", it is not clear of the other strategies are universally considered as "cheating": for instance, the "pyramidal scheme" seems a good description of laboratory work in Chemistry and Biology, and maybe of the general structure of research institutes in Germany. – Jeremy Apr 12 '15 at 18:43
  • @Jeremy A few points: (1) faking results is a high risk strategy in that it could end an academic's career; (2) there are a range of bodies that allocate rewards (e.g., grant bodies, promotional bodies; societies; etc.): these bodies typically use a different mix of priorities - some emphasise quantity, but others focus more on true quality (actual reading of papers; grant proposals; top papers) or imperfect proxies for quality (e.g., impact factor, citations, etc.). In summary, some of these strategies would work for some audiences, but not others. – Jeromy Anglim Apr 13 '15 at 9:01
  • @JeromyAnglim: I am not advocating faking results, and I am not recommending the other strategies (so it does not matter in which communities people could get away with them). I am just arguing that incentives based on publication throughput are pushing people toward such strategies. The recent scandals about faking data in biology and physics might be the top of the iceberg in this regard. – Jeremy Apr 13 '15 at 17:26
2

There is one behavior that is encouraged: vociferously advocate to change the mainstream publishing model to cater the needs of authors to publish more stuff instead of providing the readership with quality, curated scientific content. Subscription-base journals, editorial rejection and scrupulous peer review have to be made obsolete because they only prevent scholars form lengthening their publication list.

  • Encouraged by whom? Or do you mean which should be encouraged? – Jeremy Apr 13 '15 at 17:26
  • 1
    @Jeremy encouraged by "an incentive system based on publication throughput" as per your question. In my opinion it certainly should not be encouraged, it's a regrettable consequence, but you said you were not interested in opinions. – Cape Code Apr 13 '15 at 18:33
  • oh, ok. I guess I should be more specific. I was aiming for researcher's behaviors rather than publisher's behaviors. I will amend the question accordingly, sorry for not being clear :/ – Jeremy Apr 13 '15 at 18:43
  • 1
    @Jeremy I have adapted my answer according to your comment. – Cape Code Apr 13 '15 at 19:01
  • 1
    @Jeremy yes, it's swarm-sociopathy. – Cape Code Apr 13 '15 at 20:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.