I think there is no limit defined but it depends on the kind of research and the literature available on a specific topic. The more there is literature the more can be the references depending upon the knowledge and effort of the researchers.

Some time ago, I was excluded from a call for papers about Cultural Heritage, and motivation was: "Citations of studies prior to 2000."

Yes, it was true. Among the references, I forgot to exclude Marshall 1995, Kuhn 1992, Simon and Mandelbrot 1972.

I depends on your research frame and question. You have to organize and submit previous studies and show the gap in the literature and defend your research question reasonably and logically in terms of filling mentioned gap. I don't prefer to cite everything in the literature. Cite if it is necessary to accomplish your goal :)

We cannot say the exact number of references. But the important point to consider is, you should not avoid the most promising literature in your relevant field. You should know what are the top journals in your field. When you compiling a paper or specially when compiling literature reviews, literature review grid will be a good tool to use. Lets say u are drafting a literature review, past ten year literature would be enough in my point. However all depends on your research scope.. Use literature review grids for your research work..

This is a question that I am regularly asked by my students (both under and post-graduate) related to their assignments and research studies. My answer is always 'how long is a piece of string'? (hopefully that means something to those reading it; it essentially means that there is no absolute answer). I support my statement with identifying that is always 'quality over quantity'. Students often posess the misconception that the more references they include the better the awarding mark will be - and this is rarely the actual case. It's far more about how the literature is 'integrated' i.e. analytical and/or synthetical process - and that it is the most appropriate resulting from a good critical literature search i.e. the most 'seminal' pieces of work, the most current, preferably primary studies, and the ones most closely related to the topic area and findings.

I fully agree with Dean Whitehead, that there are no definite answers, but it depends mostly on the type of work done on the topic in past and there availability to the researcher. However, it has been observed that authors make a very long and exhaustive list of papers cited but it often does not show in the text at appropriate places. Moreover, people also have been seen citing there own work multiple times which often is not very relevant to the paper. I again agree with Dean, that a long list of references does not always make the paper of a high quality.

There is no specific number, but it depends on type of paper. for instance, in qualitative paper-based, 20-30 references are needed whereas in quantitative paper-based, usually, 50 is sufficient number. regards,

Maryam - how do you work that out? Methodology or paradigm approach should not have any bearing on the amount needed. it's back to my earlier point that, regardless of method, it's 'quality over quantity'. Pre-determined numbers do not come in to the equation at all.

I use saturation. When the journal articles begin repeating the same authors and offers nothing new to the literature review then it is time to stop. For some master's programs they only want new articles and set a "no older than 5 years" for the lit review. For Ph.D./Doctor programs, the saturation rule is usually used. I have spoken to Qualitative Researchers who were told they had to have a minimum of 30 references but not more than 50 references.

It depends upon the quality of the work not quantity and yes there should be no repetition and one should use new references.

My chair said that there is no set number of references, however, the ones you use should sufficiently cover your research area, be peer reviewed, and at least 85% of the references should be within the last 5 years. The other 15% can be foundational to bolster later research. Keep in mind when you will complete the study because if you use articles that are close to the 5 year mark and you have problems with your study, they may be obsolete by the time you are ready to finalize your study.

David's quoted figures would obviously not apply to people doing historical research. The suggested percentages might even be reversed.

Dear Mohammad, I wanted to check what your intended 'research paper' is since some of the posts are quite rightly making different suggestions (bigger or smaller numbers) possibly thinking about different outputs: an undergraduate assignment, a postgraduate assignment, an Honours thesis, Masters (Research) thesis or a PhD thesis...

Like Dean Whitehead I am regularly asked this by my students in relation to their assignments and my answer too is frequently 'how long is a piece of string'? ;).

I agree with the importance of 'quality over quantity' and the importance of how the literature is 'integrated' including 'seminal', current studies, and those most closely related to the topic.

If you are trying to write a journal article though, most students who have not written one before can genuinely become quite paralysed by this issue of number of references. They seem to benefit from a rough guide in terms of numbers - in the same way that journal instructions regarding number of words in an abstract and number of words in the article overall assists them. A recently set a student who asked this question the task to bring in to a supervision session 5-10 articles he was most impressed by and had learned the most from in thinking about the paper he was attempting to write. We looked at these papers together and found they had between 20-30 references. This gave him a guide, and us a chance to talk more about what made for a good article (which was probably more helpful still).

Melissa - good to see that we tend to 'work on the same page'. I like your strategy with your student. It isn't pre-determined or strict - but thought through. I often co-author with my post-grad students (but tend not to post them on RG as they tend to feel a little unsure about the 'wider exposure'). They tend to cite heavily fro their theses (and that's fine for the thesis itself) - but usually journal restrictions means that they have to heavily cull those lists to fulfill the requirements of manuscripts to journals; that's usually quite a part of my mentoring role with them.

I can understand that some type of qualitative research papers could have less references. Qualitative studies are often carried out to describe components of new phenomena. Often there are not as many research studies coming before hand so less to include in the literature review particularly.

To pick up on Melissa's advice to a student, it would be interesting to do a content analysis of different types of articles to look at average numbers of references.

The optimum number of references in the References Section of a research paper is the exact number of papers quoted or paraphrased in the body of the paper! One reference is too little, because then you are not comparing and contrasting sources. One hundred is too much for a student paper. Just try to obtain as many relevant references as is possible. Amongst other things, the mark you get will be an increasing function of the number of relevant references. Is there an optimum value for the mark you want?

Madeline - not sure that a content analysis would reveal anything - unless you did it for all disciplines across all types of popular (and unpopular) journals. I would hazard a guess and say that all you would find is a lot of 'difference'. The average might be 'skewed' because of so many 'outliers' i.e. journals that say no more than x-number - and those that do not state at all. I've seen 200+ in some articles.

What's the right length for a piece of string? Just long enough to serve the required purpose. No longer.

Mohammad speaks of an optimum range, not a minimum range or a maximum range. I believe there are physical limitations that a manuscript confronts like number of pages- ten thousand pages will definitely bear on the number of references and the evaluative capacity of the peer critics or research professor. Except perhaps for the anecdotal one-page thesis of Einstein, no other dissertation can beat that- what with the formats- title page, table of contents, rating sheet, reference list, etc. I tend to agree with Ian. There is no optimum range. There is only a range that is subject to a lot of factors. So Mohammad, I suggest that you "investigate" what your institution, your panel or the Journal, say so, It will ease down your burden of convincing them later.

The same number of references cited in the main text must be included in the final references. If you do not review a work you don't need to include it in the references. There is no limitations. It depends on relevance and saturation of important reviews on the subject.

Obviously as many have said there is no magic number. However, some publisher have been willing to give them. I have written chapters for book projects with Sage and Elsevier. Both publishers have set as a general rule one source per page average. I.e., a paper that has 20 manuscript pages might have about 20 sources. As a general rule. The number of citations is different than the number of sources. I have written papers with 75 citations (for journals that use footnotes) but still the article only had about one source per page average.

Well known publishers may give guidelines that are the average for the type of chapter that they have accepted. The writer knows that there should be a reason if he / she has a wide deviation from the average. Such deviations would apply to topic review papers, where more that the average would be expected, or really innovative research, where less than that average would be expected.

So it all depends on how heavily your research is based on prior research. You do not plan to get the national average; you only bother to work out the statistic after you have published, and someone asks you! My papers that I looked at today run from 2 references per page (for a book review) to 3 per page for papers that are heavily didactic (intended to also be used for teaching purposes).

Less than an average of 1 reference per page might indicate the possibilities of poor scholarship or plagiarism.

However, publishers cannot force you to reduce an average of 2 or 3 references per page without forcing you to be a poor scholar or a plagiarist! I know of no case in my career where this forcing has happened to any of my colleagues.

Good last few posts - especially from Ian. If the publishers want your services and they know your worth - they may recommend an average - but they are not going to say no if there is a rational variation. Michael, I've written quite a lot of book chapters for Elsevier, and my experiences have been - no more than 40 citations per chapter - so maybe they vary according to discipline, individual responses etc

Ian and Dean make some interesting points. I would reinforce Dean's answer, there is no one rule for anything. In my discipline, many scholars have just the opposite view as Ian. I am on five editorial review boards. When I see a paper with three citations per page (say 60+ for a 20 page manuscript) I start LOOKING more closely for plagiarism.

I expect a scholar to have something original to say and to have his/her own voice, not simply attribute every sentence or quote to someone else. A scholar cites quotation and ideas that are not his/her own original thinking, but just having a lot of citations means very little.

I have seen published article of high scholarly impact with six citations -- something common in the 70s and 80s., and I have seen utterly unimportant articles with 100 citations. The number of citations is not a reflection of the quality of the research ideas or arguments, so long as the scholar has conducted a proper literature review and attributed work by other scholars to the source.

Just a clarification: when I said one source per page average by some publishers, I meant "manuscript page," not published page. A ten page published manuscript with ten sources would be too few in most fields I would think, but a two-page published book review with more than three sources would be a lot.

@Michael: My fairly arbitrary threshold for looking more closely for plagiarism is stated above as less than an average figure of one reference per (understood published) page of paper. Yours is three per double spaced MS. I think that we agree! (Within experimental error of course :-)

There is a debate as to whether excessive quoting and paraphrasing should be counted as "plagiarism". I prefer to call it poor scholarship.

I wish that I had written your third paragraph.

I don't deal with double spaced MSs.

Ian and Michael - I believe we are starting to nail this. Nothing is set in stone - especially if you are an established author. Minimal or excessive citation is probably indicative of inexperience (and I like the mantra of 'poor scholarship'). Inexperience (or otherwise) is no excuse for under or over-egging anything. Michael - over-egging', I agree, is the worst scenario. To me, it demonstrates 'insecurity' that an author does not have the confidence to 'say it themselves'. That said, and looking back, I've been guilty of it myself. The problem with plagiarism (several more threads needed probably) is that it's very difficult to define, police and take punitive action - especially for publishing houses.

I find these responses most intriguing. Shall surely bear them in mind as a young researcher. The average number of citation per page appears to be a good rule of thumb, though as said by others it does vary with the publication and reviewers.

Hi Dale, The responses here have been very helpful, yes.

The average citations per page are a more helpful yardstick though I think for book chapters than journal articles.

Of course there will be an average for a journal article however it will be very uneven where the citations fall - more in the Introduction, to ground and justify the purpose of the study and research question, and in the Discussion, to explain what the results mean in the context of previous scholarly research. Some pages may well not have citations at all.

Just wanting to clarify so this does not have some newer to research hunt for a reference after 2 pages without a citation...

Dear Researchers,

Can you enlighten me about the references for a review article?? It will be appreciated!!!

I have seen some papers with as many as 150 references??

What is the best number of references for a review article on medical subject, for example an article on a small structure in the brain?

Thank you 

@Murlimanju: 

Let us say that the small structure in the brain is the "brain stem". Today Scholar gives 1.16 million results. So let us say you rather limit yourself to the "human brain stem". This give 4,640 results today. Let us say that you decide therefore to further limit your review to "human brain stem potentials". This gives 109 results today, which is manageable with today's software. Of these results, I assume that you find that half are false "hits" and you discard them. However, you also follow up on some of the "Cited by" trails, and I assume that this quickly doubles your relevant references. If you use Medline as well, the number will go up even further! You might limit your review otherwise to include either "injury", "stroke", "tumour" or "death" as keywords, depending on your personal interests.

So for a review of such a medical field, you will easily be extracting the essence of perhaps a hundred papers to do justice to the evidence. You will include each quote or your paraphrase in a sentence or in a paragraph. Each quote or paraphrase will have to be meticulously referenced, so that easily makes for a hundred references.

The fact that you are studying a small structure does not mean that your reference list will be small!

the optimum range for some good journals are 20,, it may vary based on the different format and publishing agency 

It depends. John Nash's PhD thesis (Non-Cooperative Games) contains only two references; the one his own paper and the other von Neumann's book. It is a thin document of 28 pages. But nothing to compete with. It was path breaking. In the articles that deal with literature review, references must be many, due to the nature of the write up. An article/paper has its own balance structure like a human body - not emaciated, not too corpulent or obese. References are a part of the article/paper. Must be balanced. It should be judged in that light.  No mathematical formula for a balance.

No, no limit to the number but the references should be kept most relevant.

I agree with whitehead emphasizing the fact of avoiding both the extremes; least and most referencing. One should be moderate enough and balanced while preparing the list of citations depending upon the nature of work. It has been observed that unnecessary referencing may even impact in negative in some cases. The number can be optimized by citing the paper containing comprehensive contents about the research area.  That's why survey papers are somehow most cited papers in the area.

The answer is actually very simple and very exact: You should have precisely one reference per quotation or paraphrase. It really is that simple. If you have N quotations and paraphrases you will have exactly N references. (You are allowed of course to cite the same work more than once in your paper.)

Thus, if you have lots of quotations and paraphrases because it is a survey paper, you will naturally have lots of references. If it is pioneering research, you will have few references.

The condition laid down by Ian is necessary. You cannot quote/paraphrase without reference and you cannot give reference without quotation/paraphrasing.  But the real issue or the question is regarding sufficiency, not necessity.

Thanks, Sudhanshu. No paper referee or examiner ever counts the number of references. As Dean wrote above, "It's far more about how the literature is 'integrated' i.e. [the] analytical and/or synthetical process" 

I think that the students asking the question have been guided at school about sufficiency that "so many pages / so many words are needed." Later, they are admonished not to have no references, so the question arises as to how many.

Yes I agree with the views of Ian and Sudhanshu who indirectly talks about the relevance. 

@Jagdish: Let's talk directly about relevance. If a paper is not relevant, it is not cited or citable. 

A "Bibliography" section without page numbers rather than a "References" section of paraphrased or quoted sources is an open invitation to cover up plagiarism or to pretend that the writer has read more than he / she really has. 

Any attentive person, after glancing over a couple of research papers in a good journal, will know how to cite/refer books/edited volumes, journal articles, etc - styles may vary but the content remains more or less the same. What is the purpose? A reader should be able to lay his/her hands on the referred material or he/she may be able to communicate his/her requirement to a librarian exactly. The purpose is retrieving. If an author does not provide sufficient information regarding the cited/refereed work, he/she has developed sloppy habits that he/she should mend. Of course, the author may not always remember the matter that he/she wants to refer. Then, if the referees/editors permit, the author should admit the same. I will like to give an example from one of my papers published recently (Mishra, S.K. (2016). "Bihar’s Education System in Shambles: Building History with the Rubbles of My Fading Memory", The NEHU Journal, 14(2): 17-31.):

"I do not exactly remember who said it and where, but, perhaps, Bertrand Russell or Mark Twain wrote somewhere that ‘work’ and ‘recreation’ do not differ in matters of input of energy to rearrange the matter in form, space and time. The only difference is that when energy is directed to engagement at somebody else’s will or pleasure, it is ‘work’; when energy is directed to engagement at one’s own will and pleasure, it is ‘recreation’. In all likelihood, this discrimination applies to thought as well. In this sense, this writing of mine, which is an attempt based on my memory, to sketch the history of downfall of the education system in Bihar is an engagement in ‘recreation’. It should be read accordingly."

The paper referred to above with the content quoted here was acceptable to the editor. So it was published as it is. It all depends on the author, the referee and the editor. One should be bold and honest above all. 

Interesting... 

I've realised there is no clear cut as to the minimum/maximum number of papers or articles to include in your literature review.

The most important is to include current articles and most of which are directly related to your field of research.

7

Thank you so much Dean Whitehead and K.M. Singh your answers are very insightful.

I believe that in scientific research we don't have absolutes, number of references would variable according to discipline and area of knowledge; of course, it's important the papers have a considerable number of references to support a solid literature review exercise, more importantly, the type of documents which are reviewed, they must be scientific documents which contribute to quality of new knowledge is generated.

Great topic! Congrats to the Author.

No. There is not any optimum number of references in a research paper.

It really depends on how you gonna tell the story, but some journals do set upper limits for citations (e.g. 50).

This is very interesting topic to discuss because there some writing short review, comprehensive review, Systematic literature review. so at least we have some guideline on how many pages is the average number would be great.

Worldwide there is no clear cut of number of references. Though, depends upon the universities and the program. Always abide by the rules and regulations of any programs or institutions you in

Some journals limit the number of references while others dont put any limit. I think it is quality of referencs that you r citing in yours article that mattrs....

Dear sir, regarding number of references..there is no particular rule but you please include references which have maximum no of citations

The number of references depends on several factors such as novelty of the research, number of earlier studies that are accessible. However, many Journals advocate a minimum of 40 referencess or above

There is no rule that says the number of references should be limited. but to my own view, the whole of your article should not be fill with people ideas alone. at least you should have a significant contribution to your study and not only copying from others

Many researchers increase the number of references in their papers to be more visible and then more citable. I think Statistics could answer this question!

My view of it is if you self-cite more than 10% it becomes a bit of a joke and you are just pumping your stats to appear more important than you actually are (https://www.nature.com/articles/d41586-019-02479-7). But it does happen in small fields. If every (1) last (2) immortal (3) point (4) is cited? It is actually rather distracting to the reader and I will find someone else to cite. As a reviewer, I start looking pretty hard at someone who has more than 60 in a standard article and usually ask them to 'prioritize' and ask them to write it in a more focused way. Reviews are different.

Please refer to our paper titled "The nearly universal link between the age of past knowledge and tomorrow’s breakthroughs in science and technology: The hotspot ": https://advances.sciencemag.org/content/3/4/e1601315.full

Sure: not too few - as we need to refer backwards to where the topic began, and how we now have ended up here, where new material/perspectives will be offered. We also are forced to discuss what others have done and perhaps are doing, in order to balance your new contribution with the past, and the current status quo. It is definitely unfair to throw every paper you have written into the paper, but I have seen it a few times. Don't.

The number of references depends on several factors, but, in general, no limit but the references should to be up to date. Some Journals advocate a minimum of 30 references or above.

I hope you don't mean "Up to date" as in "recent". We need for the authors of the paper to trace some of the paths from the original question(s) to the state-of-the-art, and that might take a rather a lot of space. I have never seen a journal specifying limits, as it must be possible to trace parts of the original work through to today - so that there is a natural thread, or even more than one in a multiple-disciplinary work within the scope of the paper.

Beware of instructions from journals with shady publishers that state that you should try to find articles from their journal to cite (I guess even when they are irrelevant) - because that is both fraudulent and dishonest. By the way - don't publish in those journal, period! :-)

Dear Prof. Michael I mean must include some latest journal article references published from 2016-2020.

Th references should to be enough to coverage of other relevant work. Including a wider overview of the most recent literature would significantly contribute to the quality of the manuscript and would strengthen its academic perspective and scientific foundation.

Right Prof Michael Patriksson many journals do that

"Beware of instructions from journals with shady publishers that state that you should try to find articles from their journal to cite"

Good! Keep up the good work! :-)

This question was posted in 2013, seven years ago. For no special reason, I googled the references question because my current book, much smaller and more punctual than my dissertation 26 yrs ago, is already at 285 references.

I have a hypothesis that will remain so, I guess (I don't see how it can be tested or falsified). The more controversial is the original argument I am making, the more I feel the need for the most thorough literature review possible. Also, it never ends.

Two things seem to be associated with the number of references: the controversial nature of the argument and maybe a little bit of depression, just enough to empower the superego.

I don't think there is a standard number.

Some very good answers in this thread. In light of my recent experiences and what I follow I would like to share my views:

1. There is no limit. But it has to be enough and thorough.

2. For a paper, any number of arguments that are important to support, contradict or prove novelty of the paper is necessary. Just giving the references that support your hypothesis is in my opinion not correct. Contradictory findings often help us locate exact areas where the experiment differed and can actually sometimes help in making stronger arguments in support of your hypothesis.

3. Unnecessary references for flexing that do not add value to the paper or do not make vital points to the central line of thought should be minimised.

4. Already well established popular facts need not be referenced.

5. Also for a single argument, I like to put maximum of 2-3 most relevant references and not all references since the beginning of time that said the same thing.

6. Also being timely is important so I like to put references as recent as possible. This establishes that the point is still valid. Of course for fundamental ideas..the original paper should be cited which can be old.

Some journals that puts a maximum cap on the number of references..I try to follow these points more stringently.

I am yet to come across any journal that puts a minimum cap.

Wonderful submission Biswajoy Ghosh , I must confess. Especially in the area where you said 2-3 to support an argument. I like working with a maximum of five citations for an argument unless they are less.

References are one of the areas the strength, coverage,. currency of any research paper is measured. As long as they are relevant and have covered recent studies from wider perspective to the context (topic and location), the length or number does not matter as said by Biswajoy Ghosh. It should not be haphazardly used too as every cited article must complement the arguments of prepositions of the writer. This is my opinion anyway..

In my opinion, the range is not the important thing, but the references should be adequate, that is to say, they should provide important contributions that support the discussions and conclusions of a scientific work.

It is important to note that papers vary by types. Reviews,opinion, empirical etc. The deeper the study, especially an empirical paper, the more the references. That does not put a rule on other types of article.

There is no absolute answer or number for references to be included in a research paper. However, number of references should not be the focus while writing a research paper. Instead focus on the most recent literature and most relevant literature to your topic. However, if you are looking to publish a paper in a specific journal, then you should check with its policies and guidelines for authors to see if they have any specific requirements about it.

Sometimes I wrote between 7 or more references for 2,500 words. It depends on your writing.

@Sukri, that is ok. But I feel it depends more on the type of research. I believe a good or empirical paper should cover at least 50% of current sources. However, this is not possible in places with open access problem. Like Africa.

There is no optimum range of references to writing a research paper. It depends on the type of research, for example - references are expected to be more in a review paper than a research paper. It is also guided by the area of research and works carried out on the topic so far. If the area of research is new, the references may be less and if there are more works available in the area of research then references got multiplied.

I liked Michael Patriksson's answer from May 9, 2020 regarding the need to have

references included to establish 'where the topic began, and how we now have ended up here, where new material/perspectives will be offered'. You do indeed need to compare and contrast the new findings from your study too to existing relevant literature.

Don't forget also it is a good practice to include references to justify your methodology selection as robust, and also applicable to the field or specialisation you are writing about. You may need references to establish the validity of tools or measures you have employed to collect data or an approach to data analysis.

The reference range depends upon the type of the article. Following range is commonly observed in journals.

Letter and Short Communications - 5 to 15

Regular Articles - 20 to 40

Review Article in regular journals - 50 to150

Review Article in Journals dedicated to publishing Review Articles - Unlimited

for research article mostly from 35 to 50, and maximum up to 60

- The average number of references per article is the highest for the social sciences, physics, and astronomy, and arts & humanities (roughly 54 references per article).

- On the other hand, health professions and earth and planetary sciences are the fewest references per article at an average of 8 and 17 references, respectively.

- Math and engineering averaged at roughly 29 references per article.

- Biochemistry, genetics and molecular and other biological sciences averaged at 51.

- Hard and natural sciences more frequently cited recent literature while social sciences and math were likely to include older sources.

It depends on the area of research and literatures available. Because if someone working in new area where the available literature is few, then even less than 10 references are enough. if more literatures are available means then we have to include more numbers of references at suitable places to justify our new results / findings. In this case, 40 and above references could be used for regular research article.

I think there is no optimum number. When one is writing a review paper he/she should try to include as many as research items possible. In the case of a research paper, one should determine the subject clearly and mention state of the art no matter how many references are required.

If we take the concept of other author(s) on to our present work, we need to cite. In this case, we have to limit the number of pages we have to write, the important concepts and notes we have to refer (take from others) and so on.These may limit the number of references we use. In fact, even though I am not sure, most things have limits.

Thanks!

If you are going to show off by citing as many references, you may not be able to prove that yours is a noble research. Next, if you have made a real research, you will get only a few previous studies to cite; this is what you are always expected. How many now? 12/15.

I would think of it in terms of the total length of your paper (assuming it is original research), I would generally suggest 1 ref for every 100 words eg 4000 words total = 40 refs. I would allow 10% either way ie 40 refs = 36 - 44.

This would be different, of course, for review and so on.

Of course, some journals might already specify a limit - eg max 30 - for a certain type of article.

for research article mostly from 35 to 50, and maximum up to 60