Why is it so important? Good scientists are like poets: they HAVE TO do reseach and publish (interal motivation). If the only goal is to get citations, it is narcissism, not science. Mendel was forgotten and his findings had to be reinvented later - but that does not change the fact that he was a great scientist.

Yes. Half lives will, amongst other things, be a function of the discipline involved.

Food expiry dates are not an exact science but merely a guideline. {Otherwise I would be dead by now!} The same applies to papers. We would have to cope with egomaniac authors who predict half-lives of centuries for their pet papers. Half-live estimation is a better metric, because it is softer; it does not imply that the whole paper suddenly becomes obsolete on the 31st December of some year. Although possible, prediction will be very error-prone. Certainly, we should not destroy all copies at the end of the designated year. _Trends_ in citation counts (upward / downward) could be a useful proxy. In my own experience, the dustiest book on the shelf, and the yellowest photocopy were the most valuable to me.

Generic truths, universal models, practical circuits, and eternal truth books have the longest half-lives. Often, not the whole book, but one sentence, one paragraph, one table, one drawing. Papers similarly.

Everything is relative. It depends on the scientific evolution of reference field, the journal where is published. An article in a national language if it is not tranlated into an international language loses from the number of the possible citations. I consulted papers with an inestimable value, but were published in national journal, with limited edition and so they did not receive a big number of citations. In this situation, a further article, which quotes the previous article, can receive more citations (if it is published in a better rated journal and more widely). I observed that much more citations have the articles which open or create new horizons for research.

Why is it so important? Good scientists are like poets: they HAVE TO do reseach and publish (interal motivation). If the only goal is to get citations, it is narcissism, not science. Mendel was forgotten and his findings had to be reinvented later - but that does not change the fact that he was a great scientist.

In my opinion that there is no expiry date for a paper. Because my be someone could see the scientific content and thinks it is old fashion, however, another scientist can read between the lines and finds a great idea that could come out from that paper. I think we all agree that many developments in science came from transferring old knowledge gained in one filed to another new field. Eg. heat transfer in solids and groundwater flow in porous media.

Bean counters love citation numbers. They feel that it lets them understand how important some research is when they don't understand a thing about research.

Also, citations aren't all equal. We cite to doff the hat or give a nod to someone who had the idea in print before us, or who said something more beautifully than we could ever say it. Conversely, we give citations to authors with whom we have to take up swords and disagree with.

Sometimes we cite single phrases, single sentences and single equations. Sometimes a whole page or appendix deserves our referencing. Occasionally (horrors!) we even recommend a complete book or Web site.

The true importance of a paper is measured by how many patents it generates and how many useful products and services it makes. By how many lives of patients it saves or betters. By how many students photocopy it for their education. By how many coauthors it gets collaborating. By how many people view or download it without citing it.

Please do not use presumed expiry dates to decide when to stop storing our papers.

@ Ian , well said 👍I wonder, how the idea of expiration of papers came in ? If a paper is useful, ages later, it is still an IN paper. How can it age or expire ?

The best papers are timeless, and we are all striving to write that sort of paper. I have perhaps only one or two papers in that category. Life expectancy = infinity!

I have also researched and written some parochial papers that are closer to newspaper columns than perpetual papers. I admit that such whole papers were written for that time and era and circumstance, but they still contain a few gems worth mining, and they are proudly published on ResearchGate.

So, yes, although we could conceivably define an expiration date and useful age for our papers, these are not useful concepts. Such dangerous predictions will not appear under the keywords of any papers we write in the future.

@Ian,

"Yes. Half lives" very interesting observation. It is applicable in most cases. However, some researches never loose their novelty and are cited over centuries, in spite of the fact areas based on those researches are cultivated for indepth researches. Only a paradigm change/shift can offset them. Sometime an old research is dug up to explain new experimental results as put by someone somewhere on RG.

Packed foods expire for the sake of health hazards, due to the life span of chemicals used in preparing them. Electromagnetic materials perish in time, due to the disintegration of compounds that form the materials.

Scientific publications in the contrary, as far as my understanding is concerned, are research results which augment the knowledge bank of humans on a particular subject and in principle have to be lasting. It is true that not all results or publications are equally liked, needed and therefore equally accessed, but still by virtue of being true scientific results, they are lasting and eternal in their validity, unless there is a change in the premise and therefore on the foundations of the theory.

True knowledge, if expires, then it was not knowledge in the first place. True knowledge leads to the creation and extrapolation of new and more wider knowledge and therefore lives indefinitely.

I guess it all depends on the paper, the fundamental papers defining basic theories written in past century are still very much in citation, however the routine data reporting from one site or other sometimes are not cited even once. Thus, I would say age doesnt matter and it is teh content that matters and keeps a paper alive.

Some review articles can also have a "very long half life" as a good review assimilate understanding till that point in time on that very topic and I guess these are very important papers and will be cited for a very long time.

I really don´t understand the sense of this discussion. What shall be the reason to define halflifes of scientific results. You must not read the papers if you feel bored by them. By the hand who should define the life span of scientific results if not the scientific community by improving results, not reading the papers or contradicting to the contents? Papers are no garbadge which must be disposed of.

Hanno a paper is cited and stays relevent irrespective of how much time has elasped since it first appeared. But it is a tendency of most researchers to cite recent publications unless a paper is really a "landmark paper" without which a point of view and theory cannot be explained.

But am of the view that yes it should be there and will always be there, the only question will be how often people will use these. Take an example from earth sciences Thou once extremely popular continental drift theory is a historical lesson now when plate tectonics became widely accepted.

Maybe the key thing is not the age, but the use we can make of a paper -and, why not, a book-. Some papers are still usefull altough it´s been a long time they were published. All of us can think on a handful of old papers, basic to our discipline, that we have cited, because in them was a piece of knowledge we found basic to our research, and we must credit the autor and his contribution.

Besides that, I think that there could be another group of papers that have same kind of "U shaped" trayectory on their citations: when they are published, they receive a number of cites because they "are new", and, as been previously said, there is a tendency to cite the newest stuff that there is available. Time passes, and the number of citations diminishes, but when enough time has passed, if the contents worth it, some researches begin to cite them again, because researches have found another "use" to the content: the example of earth sciences Saif has given us is perfect to the idea I want to explain.

Scientometrics is an objective (majority-based) way to measure scientific significance, however not optimal. Besides, once other paper cites an old one (and so is basing on it in its scientific arguments) and so on, this ongoing connected chain ensures "keeping live" the old one regardless its age.

In science old beliefs and convictions were so many times turned upside down after centuries, but coming back to an old result in order to change it does not mean loss of importance for it, since we learn something from every act of scientific progress .

A important landmark research always remain live and is cited quite often. However will like to know that is it right to cite a recent article instead of the original.

@Saif: It is best to cite the original (primary) source, and not the quoting (secondary) source.

Ian I do cite original article but quite often it is so well described in secondary articles that people dont bother to get the original. Some cases are on other extremes. I will like to share an incident with a collegue who submitted his thesis in 1972 working on electron transfer, he cited a paper which was a cross reference (the original was in german), his supervisor picked this and asked him to get the original article and get it translated in english, this took him 7 months in those years. The supervisor after reading that article put a tick mark saying the statement si true and his Ph.D. was awarded.

@Raoof, good question! I do think that we can not define an expiration date for our papers! Of course, it depends on scientific area of a paper, as well as ,sometimes, of technology development in the area that paper threat!

Dear Raoof

I cite an example : an American biologist journalist wrote a false paper on the properties an anticancer molecule supposedly extracted from a lichen and submitted this paper for 304 open access journals around the world . Not only the paper was totally fabricated and obviously incorrect, but the name of the principal author

( Ocorrafoo Cobange ) and their institution ( Wassee Institute of Medicine ) were fictitious. Nevertheless , more than half of wanted journal ( 157 ) accepted the paper for publication. A scandal .

What does this mean? It means you have a lot of journal

The fact that a study was published does not mean it is correct ( worst , does not mean he is not even real ) . Science , like any other human activity , is unfortunately not free of dishonest people

On this same theme , Nature journal also recently published a report on the scandal involving four Brazilian journals that were caught practicing cross- citations - or " stacking quotes " scheme whereby a journal cites other purpose several times , in order to increase their impact factor ( and hence the prestige of researchers publishing them ) . Journal are Clinics , Journal of the Brazilian Medical Association , Brazilian Journal of Pulmonology and Acta Orthopaedic Brazilian .

The scheme was discovered by Thomson Reuters , the largest international reference in the production of statistics and scientific publication citations. As punishment , the four journal had their impact factor suspended for a year.

Caution is required !

Some examples on the pitfalls of the review process can be found here:

http://abouthydrology.blogspot.it/search/label/Reviewing

Have a nice reading!

Hi Raoof, if you choose to use citations then it is probably better to use citations over time, rather than a total number. By that I mean citations for each year. You might find that it produces a distinctive graph. For example with most citations near the time of publication rather than recent.

With the suggestion that most papers are read only by the author(s), the journal editor and the reviewer(s), there seems to be plenty of opportunity natural redundancy without the need for a "expiration date". They will simply gracefully decay as new publications are layered over the top and discovering older - uncited - papers becomes progressively harder.

You might also like to think of publication as a cycle with the conference/journal paper at the start and through various degrees of aggregation, reviews, systematic reviews, handbooks and textbooks and so on. Ideas/research that has credibility or traction absorbed into the literature. I do recognise this is an idea more prevalent in the social sciences. BW Matt

We cannot define the expiry date of our research paper until we can define the expiry dates of theories and equations used for completing that paper.

Regards,

Nitish

I do not think you can apply your principle to a single publication.

That is, what makes it impossible [or extremely complex, therefore impracticable] to predict how long the value of "citability" of a single publication, and when it will expire, is the certainty that there will exist n other publications. Not only on the same field, theme, but also in similar themes, objects — and gradually more and more distant from the 1st.

So much, so that it will be difficult to predict not only the number of such other publications, and their content, but even if one of them [or a different number of them] will cause a *shift* of your current / prevailing scientific paradigm [Kuhn], enough to change the relevance of that 1st publication you would predict the "duration".

However - like all hyper-complexities - it is a challenge that may seem fascinating.

g

Dear Raoof,

Let me repeat your question again here:

"Can we define an expiration date or useful age for our papers?"

I think we need some distinction concerning the topic of the question separating the "expiration date" and the "useful age".

(A) papers that get citation for long time;

(B) papers that contains information/concept that still holds;

(C) seminal publications that influenced many scientists.

Please note that Case C may differ from Case B. An example for Case C might be Aristotle geocentric cosmology that is still being taught even if we know that it is outdated. The influence of his concept is so important that everyone should be aware of.

In this context (Case B) there are some papers/publications that are not cited, though they are the basics of our scientific models/thinking: these are the results of scientist whom you learn about in secondary school. There are many such scientist, I just mention Einstein here, his contributions are still being cited, but NOT in all cases the actual contribution is given, only contextual reference is mentioned.

Let me cite another scientist, this time a geoscientist, namely Wegener, who not just OBSERVED an important fact, but also tried to put the puzzle together, despite the environment that (mostly) denied the concept. Today Wegener's works are cited very few times (precisely with bibliographic data), but I can hardly imagine any textbook on geoscience without mentioning his name or even recreating his paleogeographic map.

Such citations are often not given credit in the databases. (It is not needed, as Gyorgy mentioned above, since they are anyhow in the Hall of Fame of scientists.)

So the MEASURING of the half-life/expiry date of papers is, at least at the moment, somewhat biased, maybe this will improve in the future. Since the original, seminal works are more and more available on the net nowadays, I see a weak, but strengthening tendency to cite the original works as Ian and Saif mentioned above, even if it is not in English. (On the other hand once a reviewer requested removal of original Hungarian references from our manuscript "because they are not available and in Hungarian" that we wanted to cite in our paper on the evolution of the Danube Bend. I cannot agree with that, since these great researchers in the beginning of the 20th century recognized important facts that we do not want to claim as our idea! At that time it was obvious to publish it in Hungarian, sometimes in German...)

This leads me to another type of papers in Case B: there are some observations that are not repeatable anymore, especially in geoscience: the phenomenon was ephemeral (e.g., a landslide, or, yes, a glacier...), the outcrop is not existing anymore (the mining activity destroyed it), it is not accessible anymore (built-in or under water in the mine), etc. Those papers describe the observation, and it is very important. These could be among "a few gems worth mining" that Ian formulated so nicely above.

Again, these papers will not fade out (see Amro's and Victor's contributions above), even the implications published in them would not stand the test of the time.

In summary, I would make a clear distiction between expiration date and useful age.

Thanks for all contributions above, I enjoyed them. Regards, Balázs

Whenever a new research surpasses the benchmark defined by the earlier paper, then automatically the earlier one finds no place... However for BASIC research the earlier paper still is a foundation for the particular research.

Fantastic question! A paper is a novelty in the history of science - approx. since the 19th Century, when paper started to take the lead over books. The question hence can also be addressed to books - but I do understand and completely agree with the sound sense of the question raised by Raoof.

Science, so to speak, is a bet for history and farther on, if you wish to eternity. Kind of platonism, if you wish. But that is indeed the case. Fundamental papers are authentic jewels of the mind and the spirit.

Intelligent question. I think old knowledge or papers serve as foundation for generating new knowledge/papers. So we should not mention expiration date or useful age for our papers, it would be reader who decides.

Outdated papers do not exist in historical sciences...

I learned a lot on this thread. During my Masters and PhD, my lecturer insisted that I must have recent papers, she defined as within the last 5 years. I think that many of us in that Research Design Course got the idea somehow that that's the expiration period! Mistake again!

Hi Raoof,

You have raised an interesting question, but I should say that the answer is very relative to your discipline.

Popper suggests a never-ending process of conjectures and refutations. Likewise, Thomas Kuhn believed that a shift in professional commitments to shared assumptions takes place when an anomaly subverts the existing tradition of scientific practice.

For instance, in  astronomy and physics we have various paradigm shifts such as:

Claudius Ptolemy

Copernicus, Kepler, Galileo

Newton

Einstein

Nevertheless, each and every of these scientists are regarded as a prominent figure in their field of study, yet, time has proved that their statements are not really accurate today.

Paradigms have priority over their rules; because of difficulty in discovering the rules and also because application matters.

Yet again, it very much depends on your field of research, and time-sensitivity of your topic.

A given paper may be worth citing for many years, particularly if it contains something that breaks new ground in a subject.

But in constructing a literature review for a new paper, reviewers will want you to cite recent work in your field.  When I perform peer review on a paper, I look for roughly half or more of the citations to be within the past five years.  This way you demonstrate that you based your research on the most recent work in the field, not on older ideas.

Alireza makes a hugely important point here about research.  And Michael points out the other essential matter--literature reviews usually demand more current research, at least, something new coming out of an older idea.  Once can certainly cite a landmark work--one that sets a standard for the field.  I like to remind my literature students that we professors of literature are still reading Gilgamesh (2740 BCE) so we may be not be so involved in the date (we don't cite the date in parenthetical citations as the APA format does).  But I also tell them I don't want to see literary research with no citations within the last 10 years. For instance, in ancient Germanic studies, Friedrich Klaeber is always cited as THE landmark study of Beowulf.  But the literature review must move on since Klaeber studied in the 1890s and early decades of the 20th century.  There was no Sutton Hoo finding to back up the weapons descriptions in the text of Beowulf.  There was no longboat dug up in Scandinavia.  If there is not much current research available on a piece of literature, it may simply indicate that the text is not currently taught or studied.  When I was in graduate school in the 1980s, one of my favorite "reads" was Gothic fiction, usually dominated in the West by women.  My professors considered that writing "trash" and not worthy of academic ranking. Nowadays, the Golthic is everywhere (good and bad), but women like Ann Radcliffe and Jane Austen ("Northanger Abbey") are regularly studied and appreciated.

 The short story "The Yellow Wallpaper" written in the 1890s was always interpreted by the critics (all male, of course) as a piece of Gothic fiction--until the 1970s when the feminists began to write critical reviews (and found a publishing outlet).  I like to have my students do comparative papers of the critical viewpoints just to get them acquainted with how dramatically views can change from one period to another.  A paper citing only those males from 1920 would fail in any literature class--because it is not taking into account the social changes that have altered our viewpoints, our world view even.

I think in science dates are even more essential. I also teach academic writing with many of my students in the sciences. For instance, if one reads medical journals about heart attacks in women in the 1950s, the data basically says  women do not suffer much from this situation.  There seemed to be no Alpha Females in the 1950s.  If you read medical journals now, rates of heart attacks in women have been on the rise--quickly.  And the presence of Alpha Females is an accepted fact.  Similar situations exist in the treatment of diseases--like systemic lupus, a chronic illness whose recent flare-up  would have cost me my life 30-40 years ago.  But now with CT scans, ultrasounds, fMRIs, doctors do not have to engage in invasive procedures just to see what's going on inside the body (not subject to X-rays). Reading a journal from 1980 on auto-immune disease would give a totally different view than the one present now. As Alireza makes clear, the rules of citation and the preparation of literature reviews is highly dependent on the discipline.

This depends on the contents of the paper and relevant findings. For example, a research on a dynamic subject may have some validity period and in general a review paper has a long expiry date than a research paper. But there is no such rules and necessity at all.

If you are doing agronomic field research, some of the methodological sound and statistical procedure papers are older. Surely, recent papers are great to cite, but older ones with grounded knowledge always have their unique role to cement your point. In my opinion, saying that papers expires makes no sense or else literature review would be totally worthless.

There is one case in which a positive answer can be provided wit jut hesitation namely case studies. As such, a case study is a present-focused study aimed at a number of goals. However, it almost never is addressed as a historical consideration. Case-studies can become outdated as strategic, financial or organizational considerations. 

I think it depends on the field  of the paper. If it is a scientific paper, then it should not have such an expiration date, any thing scientific and valid is always valid and is a knowledge. Having expiration date implies it is more of a danger to use the material than its use beyond that date, but scientific results will never have damaging effects what so ever unless they are found false which in that case will be invalid and discarded. In other fields, anything established should be preserved for use either as a resource of knowledge or as a historical reference when case arises.

"Journal articles are an outdated way of sharing scientific research, says open-access advocate." This is fine article!

Research results stay for ever, the way of their dissemination change!

http://www.timeshighereducation.co.uk/414579.article

I think, "No, Expiration date is not required for any paper"...