I am a researcher in ernest since 1988, a reviewer for scientific journals and conference proceedings volumes since 1989, a published author since 1992, an author of scientific books since 1994, and an editor of (several) journals since about the same time. I can therefore with a VERY high degree of confidence state that when most manuscripts are sent to a journal nowadays the literature review *often* is - to state it mildly - much more brief and shallow (and therefore much less useful) than it used to be.
A typical one does not even discuss an individual article in detail, but groups it together with other "related" papers, so that one gets a sweeping, all-encompassing statement on a subject followed by the phrase "see [12-18]".
The first word that pops up in my head when reflecting on this practice is: "Lazy!". Not only that: there are many negative aspects in this practice: the reader is not given any detailed information about the connections, similarities and differences among the papers, and hence a vast amount of work maintains for the reader, before being enlightened about the state of the art. In fact, providing simply a list of papers is virtually no help at all!
Do the researchers and editors at RG agree with my "analysis", or do you have a more positive view of the current state of the literature review than I have?
Michael this has been my observation/experience as well, though as you say some journals are better than others in terms of insisting that authors explain their contribution in the context of prior work-- which then requires authors to explain more about prior work.
Laziness may be a factor-- but also the overall race for the impact factor in many universities around the world (especially lower ranked ones), which is driving people to rapidly submit many papers to all sorts of journals without the scholarly due diligence of really reading, understanding and synthesizing source material. References are an afterthought, picked up from secondary sources, rather than primary ones-- which perpetuates repeating incorrect statements about others' work. It is one thing not to mention or say anything about a reference, but another to say wrong things-- often dismissingly ("an example of the old inefficient approach is ...." References are also added to placate reviewers who want their own work cited-- clearly without ever reading them.
Further contributing to this is that these same journals increasingly go to reviewers who are only marginally knowledgeable of the subject (these same authors who never bothered to read and synthesize the literature in the first place). Part of this is due to the fact that more seasoned experts are too busy, and may take too long to perform the reviews-- because they actually read the papers and try to understand and sometimes improve the work, instead of performing a perfunctory review with a verdict to reject.
There may also be a linguistic angle-- which I have observed reading students' work. Those who have not mastered the language of the journal in which they are seeking to publish (English, typically), try to take shortcuts. Synthesizing what others have said in your own words requires higher-level facility with a langguage (unlike deriving mathematical results)-- thus they either say the absolute minimum, or (worse) repeat verbatim from secondary sources citing that work, bordering on plagiarism.
The solution to all this is for reviewers and editors to insist on a framing one's contribution in a meaningful discussion of prior work. One thing that would help also is for journal editors to encourage knowledgeable authors to prepare critical surveys of key areas of importance to the journal-- this would allow authors to frame their work relative to a well thought out, well articulated and generally accepted reference.
Dear Michael, I found your question very interesting and topical. Although I am in my early career, I think that some of the reasons are (1) too much emphasis on word limit/count. I have had several problems submitting my papers online in some journals with strict word limit. The online platforms often reject papers that are longer than what the journals stated. (2) Publishers are contributing to this too. I was in a seminar organised by one of the major publishers, and the resource person strongly emphasised on the fact that background/literature review should be short, and in his word ''not history''. I supposed that this may be the outcome of the publisher's/editors' agreement to reduce page numbers and save cost of production. I have also experienced a situation where a reviewer stated that I should reduce my literature very well. In this light, what should I do? Leave the literature and get a rejection letter?
Lastly, laziness as you mentioned 'may' also be a reason.
Dear Emeka,
If reviewers, or even editors, tell you that you should not put your work in its proper historical perspective - that is, explain what those connections are to the existing literature - then you should tell the editor(s) that they are wrong, and that science will not benefit in the long run from such a practice. Otherwise, the effect will eventually be that publications become disconnected, and too many re-discoveries will be made in the future.
It appears to me that some of the most important things to look for when choosing a journal is
(a) whether the journal you submit to like the sort of field that you have written your papers on (and which can easily be checked by browsing a few recent volumes of the journal), and
(b) if they allow you to write a paper in the way YOU want to.
Those journals that have a strict limit on the number of words, and who specify too strictly the appearance of the paper, is surely much more worried about publishing MANY papers EASILY, rather than publishing GOOD papers. In some ways, the right to phrase your research findings in the way you wish to is like the freedom of speech - it must not be restricted!
Michael this has been my observation/experience as well, though as you say some journals are better than others in terms of insisting that authors explain their contribution in the context of prior work-- which then requires authors to explain more about prior work.
Laziness may be a factor-- but also the overall race for the impact factor in many universities around the world (especially lower ranked ones), which is driving people to rapidly submit many papers to all sorts of journals without the scholarly due diligence of really reading, understanding and synthesizing source material. References are an afterthought, picked up from secondary sources, rather than primary ones-- which perpetuates repeating incorrect statements about others' work. It is one thing not to mention or say anything about a reference, but another to say wrong things-- often dismissingly ("an example of the old inefficient approach is ...." References are also added to placate reviewers who want their own work cited-- clearly without ever reading them.
Further contributing to this is that these same journals increasingly go to reviewers who are only marginally knowledgeable of the subject (these same authors who never bothered to read and synthesize the literature in the first place). Part of this is due to the fact that more seasoned experts are too busy, and may take too long to perform the reviews-- because they actually read the papers and try to understand and sometimes improve the work, instead of performing a perfunctory review with a verdict to reject.
There may also be a linguistic angle-- which I have observed reading students' work. Those who have not mastered the language of the journal in which they are seeking to publish (English, typically), try to take shortcuts. Synthesizing what others have said in your own words requires higher-level facility with a langguage (unlike deriving mathematical results)-- thus they either say the absolute minimum, or (worse) repeat verbatim from secondary sources citing that work, bordering on plagiarism.
The solution to all this is for reviewers and editors to insist on a framing one's contribution in a meaningful discussion of prior work. One thing that would help also is for journal editors to encourage knowledgeable authors to prepare critical surveys of key areas of importance to the journal-- this would allow authors to frame their work relative to a well thought out, well articulated and generally accepted reference.
Dear Hani, I think I agree with your views, especially on the fact that impact factor issue drives scholars to specific journals. The emphasis on impact factor and publishing in specific journals by universities and grant organisations may not permit scholars in some fields to do what Michael suggested.
I am all for the writing of review papers - I've even done so myself, with great pleasure even! :-)
I don't think that it is a problem of laziness: it is one of the many consequences of the profound changes that have characterized the research world in the last decades. In order to increase the education level of the populations, the number of universities rose in all countries of the world, accompanied by a corresponding high increase in the number of teachers, which are also researchers. The latter need to publish papers for their careers, and this has given rise to a progressively increasing number of journals. Both researchers and journals' publishers like that the papers have many bibliographic references, irrespective whether the quoted literature has been read or not by the papers' authors. In fact a great number of bibliographic references increases the probability that the researchers' papers are quoted - which is useful for their careers - and at the same time raises the impact factors of the journals, and this is useful to publishers, whose number of subscribers increases correspondingly.
A consequence of the increase in the number of journals is that publishers encounter difficulties in finding editors and reviewers that are really expert in the research fields of their journals, which therefore publish few high level original papers, along with many others that are elementary applications of established methodologies, or trivial refinements of the latter, with bibliographic references that in general are not discussed, because they were never read by the papers' authors.
I think that this evolution cannot be stopped, as it is a consequence - negative but unavoidable - of the desire to increase the education level of populations. A good idea could be to convince the best publishers to create - for any research field - a very limited number of journals, under the responsibility of really expert editors and reviewers, which should publish only high level original papers, leaving to the other journals the task of fulfilling the career's expectations of researchers.
My opinion is that a high level original paper should have few references, those very important with respect to which the paper represents a real innovation, and they should be deeply discussed as suggested by M. Patriksson. As an example, "si parva licet componere magnis", the Einstein's paper on the special relativity has no bibliographic reference.
Unfortunately, most of the papers being dicussed here are nowhere near the significance or originality of Einstein's paper on special relativity-- representing fairly small incremental contributions that border on "recreational mathematics". In those cases, placing one's work relative to the existing literature is even more important.
The (more than small) hints from some journal editors asking authors/reviewers to ensure that more papers from that journal are added in the reference list also pisses me off.
I tend to agree with your analysis, in the sense that, at least in my experience, literature reviews are often simply lacking or strongly autoreferential or too biased towards the scientific production of the authors' group or scientific 'party'. On the other hand, in some cases journal editors are responsible for that as well. I always tend to include rather long literature review sections that try to put my work into context and more than once I have been asked by the reviewers/editors to reduce significantly the size of the literature review and of the reference list, which I find rather absurd, especially since there are no real length constraints nowadays while there is such an abundance of published papers that providing some orientation for the reader should be seen as a positive aspect of a paper.
To very mildly contradict you, Hani, on your language point - "Synthesizing what others have said in your own words requires higher-level facility with a language (unlike deriving mathematical results)".
I do not fully agree with you there: even deriving a mathematical result may quite often require skills in other fields than the math itself. A proof has not only a logical structure (and by the way, also in the logical structure in the derivation there may be several optional avenues to take, some being natural, some others being less so) - there is storytelling in there, too, when the proof is a complex one, perhaps combining several smaller results (lemmata) before collecting the story, leading to the climax (the THEOREM!).
There may be cliff hangers, around which a detailed discussion might be needed, concerning the options possible, or small climaxes within the main storyline - like smaller propositions along the way (a hold-up at gun point in a dark alley). It could be like one of those last chapters in a crime novel featuring Sir Gideon Fell, Sir Henry Merrivale, Ellery Queen, or Miss Marple, in which the loose ends are tied together in a description of an extremely complex chain of plots. And as classic crime novel lovers know - there were some pretty dull novelists out there around 1900-1940.
I for one have shot down a few manuscripts that have had potentially very nice results in there, for the reason that the authors have simply not been able - or not been willing! - to adjust their storytelling so that it becomes coherent.
Let's agree then that the first thing a beginner PhD student in the mathematical sciences and related fields should do is to get a copy of Nick Higham's book on Mathematical Writing, which describes in detail the principles and processes of writing scientific articles.
Michael-- not disagreeing at all with you about the art and drama in mathematical proofs, and their presentation. I was mostly trying to convey that one reason otherwise competent researchers may not do a good job in synthesizing the literature is due to being less comfortable expressing ideas in non-native language-- while they're perfectly able to derive and present the KKT conditions for a math program (not suggesting it is easy or trivial to convey clearly).
Hani: it would be quite interesting to hear/read what your experience is as an editor - incredibly experienced as you are - whether you feel that items such as the "survey" is getting better (as in being complete, detailed, and useful) or worse (as in getting more shallow, seemingly randomly gathered from the literature, and less useful) over the years. As an editor and reviewer, that is my experience.
I can actually make a statement myself about the increased shallowness of the literature review: after reading carefully the journal Operations Research I have learnt that the literature review DOES shrink over time. This is sad news for those who come after us, and who need to spend much more effort to find the good stuff - as we apparently are not willing anymore to help out.
Michael, one more observation from biology field: some manuscripts have shorter literature analysis to increase importance of own findings and avoid critics in non-original ideas and research results. Unfortunately, even some researcher with high reputation do it.
- Badges
- Science topic
- Similar topics
- Academic Journals