1. Not finding help here? Sign up for a free 30min tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Difficulty Making Citations in a Research Paper

  1. Oct 30, 2008 #1
    I get the impression that people who review research papers want to see recent papers listed in the bibliography. To me, this implies that I would have to list papers that are not neither the biggest landmarks nor the most relevant.

    When I do research, it's mostly my own thing. It's interdisciplinary and it is not a Frankenstein paper that was made by stitching together other papers to make some unique creation. This creates a problems when I make citations. The primary source is seemingly incoherent scriblings in my notebook.

    Anyway, I can make citations, they just don't always seem totally relevant to me. I guess there is the issue of demonstrating sufficient knowledge in the subject area, but this requires a lot of effort, as my field hardly has an established body of research to begin with...which means reading papers from other fields and contributes to neither the originality nor the usefulness of the research.

    Do I sound lazy?
     
    Last edited: Oct 30, 2008
  2. jcsd
  3. Oct 30, 2008 #2

    tmc

    User Avatar

    Don't make irrelevant citations.

    If the reviewer feels that you left out something important, he'll ask you to add it. He won't simply outright reject paper just because it lacked recent citations.
     
  4. Oct 30, 2008 #3

    Moonbear

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Yes. It IS a lot of work to read the relevant literature, along with all the irrelevant literature that you need to read to determine that it IS irrelevant. If you haven't actually searched the literature, how can you even know if you're working on something novel or that hasn't been completely dismissed long ago? And, no, you should not just focus on the most recent literature. A thorough literature search includes both recent and older papers; whatever is relevant.
     
  5. Oct 30, 2008 #4

    Mapes

    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    If you don't want to position your work correctly in the field (which includes a literature review that informs the reader of previous relevant work), why are you publishing?

    You don't need to cite anything except the ideas, methods, results, and interpretations that aren't yours. And ignorance of previous work is no excuse; it's your responsibility to be very confident that your work is original.

    Leaving aside the question of laziness, I honestly can't see why time in the library would be considered a waste. Think about the fantastic ideas and approaches that could be transferred from established fields to your new field. Unfortunately, I see a lot of papers that reinvent well-known techniques because the authors feel the way you do and the reviewers don't know any better.
     
  6. Oct 31, 2008 #5
    The way my references tend to work (example, my graduate work in experimental non-linear optics):

    Cite the main texts in the field where the process is first developed or where the experiment is first observed (often in my case these works were from the 60's). To fail to do this is bad and shows you have no sense of history. Yes, these works have hundreds of citations... but this is why these people are absolutely famous (in my case, this would include N. Bloembergen, who won the Buckley Prize and the Nobel prize, and Franken et. al.'s first observation of second harmonic generation, where the dot from the SHG was famously removed from their photo by the editor, who thought it was a spot of ink!).

    I'd suggest looking then looking at works that relate to the motivation of your work. If your work is in a field where the motivation is really there -- aren't other people working on it? How is your work similar to theirs, and how is your approach different (showing you've considered their techniques)? Are you using someone else's idea, but with new materials? Who made your materials? Did anyone previously examine this material using a different technique? All these modern sources need credit and help position your work in the field.

    But of course don't make irrelevant citations. You don't need fifty sources about those using similar techniques, just find the most relevant/related ones. And you don't need sources for EVERY technique (ex. I did some AFM to characterize surfaces of materials... I didn't modify the AFM technique, I used commercial equipment, so I didn't cite anything about AFM.... although when I got results that related to the surface differences, I backed up those results by theory and other experimental observations: those references where it was first predicted and first seen or explained).

    I agree with Moonbear and Mapes: You're being lazy, and time spent in the library isn't a waste of time.... it's a glorious luxury that I try to do once a week.
     
  7. Oct 31, 2008 #6

    Choppy

    User Avatar
    Science Advisor
    Education Advisor

    When I read a paper and up-to-date, relevant work is not referenced, it immediately raises a red flag.

    As a reviewer, one has to assess the originality of the work presented, the interest and relevance of the material to the journal's audience, as well as the context in which the work was performed - all of which require a review of the relevant literature.

    That being said you don't want to go overboard either. Most journals don't want you to perform an exhaustive review of your subject. Naturally the best way to learn how to find the right balance is by reading how others do it.
     
  8. Nov 1, 2008 #7
    What about highly relevant papers that are old? Would they be deemed to be not "up to date" and should they be omitted as references?
     
  9. Nov 1, 2008 #8

    Moonbear

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    No. If they are highly relevant, they should be included. If the science was good and they haven't since been refuted, they absolutely should be included.
     
  10. Nov 1, 2008 #9

    Choppy

    User Avatar
    Science Advisor
    Education Advisor

    If the material is relevant, the age of the paper shouldn't really matter. "Out of date" to me means that the results or methods have since been improved on and thus more recent work should be referenced in its place.
     
  11. Nov 29, 2008 #10
    I've got a rough draft now. I also seem to have an abundance of references too, and they span across a very comfortable time range...Unfortunately, my rough draft needs to be tripled or maybe even quadrupled in length.
     
  12. Nov 29, 2008 #11

    f95toli

    User Avatar
    Science Advisor
    Gold Member

    Just to add a comment to what Moonbear wrote.
    If the results are so well established that they are part of the standard literature there is no need for a reference in the first place, there is e.g. no need to refer to Schrodingers original paper just because you are using the SE somewhere.
    If you referring to a whole sub-field or topic it is usually better to refer to a book or a review paper. One example would be if you are writing an introduction to a paper in condensed matter physics and have a sentence like "these methods are well known in the field of quantum optics", here you should refer to a book on QO which contains a description of the methods in question (preferably the one you used).

    The phrase "see [x] and references therein" is also very useful, especially when you are writing a letter and need to avoid a long list of references (letters are usually 3 or 4 pages, space is always a problem).
     
  13. Dec 20, 2008 #12
    I'm still a little confused about paper writing. Sometimes a paper will present some new application of some older theory. This involves presenting old theory in the context of a new problem. I was thinking about doing this, but I'd worry that the old theory might be perceived as not having sufficient complexity. On the other hand, I've certainly seen many other papers, if not numerous papers, do this.
     
  14. Dec 20, 2008 #13

    Mapes

    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    Try imagining yourself as a novice reading your paper. If you wanted to learn more, are the references sufficient to position the paper correctly in the field and guide you in learning more?

    Or try handing the paper to a colleague and asking them what they think.

    Your reviewers will likely work in the same field. Will they be upset because their relevant work isn't cited? Because you haven't found the previous papers that addressed (maybe even solved) the same problem?

    Or look at it this way: it's hard to go overboard, except in citing your own or your colleague's work excessively. That's irritating.
     
  15. Jan 17, 2009 #14
    I'm trying to figure out the structure of all this. It doesn't make sense to write it like a news story with the most important stuff first. I've also heard that papers give mathematical results to problems, or they have a complete derivation. I don't think I'll do derivations, but to state my results by themselves doesn't make sense because that would make the paper too short. Obviously, I could write some background. I could write it such that I unravel a puzzle, but that approach seems a little abnormal to me.
     
  16. Jan 17, 2009 #15

    Choppy

    User Avatar
    Science Advisor
    Education Advisor

    These seem to be questions for your thesis advisor (I'm assuming you're a grad student).

    You should write the paper following the format used by all the other papers in the journal. In general it should be broken up as: introduction, theory, methods, results, discussion and conclusions - with a succinct abstract at the start. It's not a newspaper article or an essay or a novel - it's a scientific paper and as such is supposed to convey relevant information in an organized and efficient manner.

    And what's the issue with length? There's nothing wrong with a short, to-the-point paper so long as it has all the necessary information. You don't need to show every little step in your work (especially when it can be referenced), but you do need to provide a clear picture of how your arrived at your results so that someone else can verify your work.
     
  17. Jan 19, 2009 #16
    I don't think that the format is standardized. A lot of papers also lack a clear hypothesis, so the results can be a little bit fuzzy. I've also noticed that many many papers end by saying some to the likes of "In conclusion, we have written a paper."

    There's not really an issue with length. Although once I struggled to comprehend long paper. Then I had a professor describe it in six words, and everything just, all of the sudden, clicked. It makes me wonder if the paper really needed to be so long in the first place.
     
  18. Jan 19, 2009 #17
    I hate papers that are written for the sole purpose of the author writing a paper. You see this a lot with people that conduct simulations using commercially available software and then publishing their results. Whats the point? To me, this is the most important reason why recent and relevant citations are needed for a paper. They are not necessarily there to provide a source of information or methods of how experiments are performed, but to show that the work you are publishing has meaning and purpose. Papers should be published because work has been performed that is relevant to solving a present problem. Not because publishing papers is just something that people in research do.
     
  19. Feb 14, 2009 #18
    Another thing I've noticed is that research papers rarely cite anything other than other research papers, books, lecture notes, and dissertations. Why does seem that patents and other non-academic-related materials are never cited?
     
  20. Feb 15, 2009 #19

    MATLABdude

    User Avatar
    Science Advisor

    Just my opinion, but I think it's just that that's where the information is most rigorous, complete, and well-vetted. I'm not sure if I've ever seen tech or application notes referenced, but if they were, I'd imagine it'd be in a technological journal (say, IEEE or ASME as opposed to Science). I'm not sure if anybody's cited work found out on the Internet (I have seen people refer to unpublished work--but this usually comes from others who come from an acknowledged institution), though, if you found someone's grad school coursework mini-review paper, or somebody's coursework, you can probably find a better one (from which their own derives).

    I have yet to see patents referenced, whether by authors who later received a patent on the subject matter in the article, or other people referencing the first guys. Probably because you can make all kinds of claims, which, as long as they don't appear, at first glance, to violate fundamental principles (or those that the examiner is aware of) may pass muster. Also, publications and patents are at somewhat of a cross-purpose: publications share your discoveries with the rest of the scientific community, allowing them to build off of your work and further the field. Patents let you make money off of your hard work, but also keep others from making a quick buck (or, forces them to get creative enough to get around your particular patent).

    I think the closest I've ever seen was a review paper which mentioned that some of the pioneering work in a particular field had been commercialized, but that none of the add-on projects had in the dozen or so years afterwards.
     
  21. Feb 15, 2009 #20

    Choppy

    User Avatar
    Science Advisor
    Education Advisor

    I have two words for you: peer review.

    The bottom line in any scientific manuscript is that if you reference an accepted fact, the justification of that acceptance comes in the form of a peer-reviewed publication.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?



Similar Discussions: Difficulty Making Citations in a Research Paper
Loading...