In many fields it is conventional to discuss related work in your paper. This lets you show where this new contribution sits in the literature, as to what novelties it adds over existing works; and to introduce works that you will later show how their techniques were adapted for the new work. One can certainly have papers rejected because the related work section is not up to scratch.

My question is can one draw a line as to what works are worth discussing based on the quality of the work in question?

  • Obviously, if a work is not closely enough related, one does not need to discuss it.
  • If one is contributing to a well established area, then there is no space to discuss every work, so many can be skipped.
  • But what if there is only a small number of works, that are related?

To use as an example what is prompting me to this question, the current work I am on is fairly new. There are only 5 related papers that I am considering discussing.

  • 2 on the inverse problem, 1 on a more general related problem, and 2 that are closely related.
  • Of these all baring 1 are published in top-venues
  • The remaining one, looks like it was an assignment.
    • I found it, via google scholar, at the URL: http:://www.<famous-university>.edu/<unitcode>/reports/<title>.pdf
    • It is formatted like a paper, which tends to make google scholar index it.
    • It may have been in preparation for a smaller workshop.
    • It was fairly well written; but lacked any kind of quantitative results, or even many qualitative results
  • However, this remaining one is one (of the two) closest related works.
  • I feel like I could include it, but if I did I would have to be quiet cutting as to its limitations.

I am in Computer Science

  • 2
    when you cite something as part of the literature review, you give it credit as worth research. I don't do this for crappy papers, no matter how related they are to mine.
    – Shake Baby
    Commented May 9, 2017 at 4:51
  • Which field are you in? Commented May 9, 2017 at 4:58
  • I recently found myself in a similar situation where in a relatively small and new sub-field (<10 relevant articles) a new paper was published in working paper/preprint form. The new paper was not particularly great in my humble opinion, and that is putting things mildly. Another group solved this by citing it in a footnote, refering to it as a related discussion.
    – mts
    Commented May 9, 2017 at 7:09
  • 2
    I recently published a paper with two co-authors, one a mathematician, one an engineer. We had quite some discussions on the literature review, as the mathematician said that the review should mention important and interesting results in the field that are related to our work and might be interesting to the reader, whereas the engineer stated that the literature review should show that we are experts on the field, that we know lots/all of the relevant publications and we can hence justify claiming that we did something new. So it seems to strongly depend on the field what to mention...
    – Dirk
    Commented May 9, 2017 at 8:22

1 Answer 1


Did the paper you're asking about contribute to your knowledge of your topic, or to the analysis or contextualisation of your results?

If yes, you should include it, wherever you found it and whatever it was originally. You might have to qualify your use of a non-conventionally or accidentally published source (if Google Scholar scooped it up from a random webpage the author might not know it's out in the wild). But attribution of sources is non-negotiable.

If no... well it might be up to your judgement. mts's comment about calling it a "related discussion" might be worth considering. If the typical practice in your broader field is to mention significant literature by way of contextualising your own work, I would probably leave this one out (due to the aforementioned questionable provenance). That particular document is unlikely to be a widely significant piece of literature. If the typical practice in your broader field is to mention as much as possible, you might be inclined to include it. A lot depends on how your field typically does things.

Which makes me think: have you been as thorough as possible in checking the author's body of published work, to make sure the same material can't be found somewhere more reputable? If it can, use that instead.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .