3

I'm writing a review paper. I'm in the field of computer science.

There are many related papers in my research area. Some are good, some are not.

I'm planning to include only Q1-papers based on Journal Citation Reports

But, does this enough for the paper?

Should I include other papers in Q1, Q3, Q4 and conference papers too.?

If I include all, it will make my review paper really long especially when I create a table to summarize all related papers.

Improve this question
6
  • This seems like a content-of-research question.
    – Anonymous Physicist
    Mar 20, 2020 at 7:24
  • 2
    @AnonymousPhysicist nothing to with that. It's a serious question on how to choose high-quality papers for a review article.
    – Jaya A
    Mar 20, 2020 at 7:35
  • If you're intending to conduct a systematic review, PRISMA is a widely used checklist in the medical sciences which could be useful (disciplines differ of course): bmj.com/content/339/bmj.b2700
    – Rdd
    Mar 20, 2020 at 7:43
  • 2
    In a systematic review you try to find as many relevant papers as possible using an in-depth search to answer a specific research question. It takes a long time and requires at least two people to review each paper. A rapid review methodology may be more useful to you (but again this is specific to each discipline): bmcmedicine.biomedcentral.com/articles/10.1186/… . Your university library (if applicable) may also be able to help - librarians know how to do this!
    – Rdd
    Mar 20, 2020 at 8:02
  • 1
    @RuaraidhDobson I doubt computer scientists will find that familiar.
    – Anonymous Physicist
    Mar 20, 2020 at 8:06

2 Answers 2

Reset to default
8

If there was an easy algorithm to determine, which papers are important for a topic and which are not, we would not need review articles. A list of Q1 articles with short descriptions can be automatically generated from a citation database and the paper's abstracts. There would be no need to put it into a separate publication. Your primary job as author of a review article is that of curation. You should consider all (or at least as many as possible) papers on a topic, select the important ones and put them into a coherent thread based on their content alone.

It will likely be skewed towards the highly cited, because they are highly cited for a reason. But the point is, if a paper is good and advances the topic, I want a review article to tell me about it even if it only ever appeared as a preprint and hasn't really been cited by anyone. Equally, if it appeared in the best journal, is cited by hundreds, but is only tangentially related and doesn't really offer any new insight into the topic, I don't want to waste my time looking it up just because it was mentioned indiscriminately with the others.

Improve this answer
2
  • Can you summarize what you were saying in one sentence?
    – Jaya A
    Mar 20, 2020 at 10:58
  • 7
    The short one sentence summary would be: "The right papers are decided by content, not by metrics." Or, more direct: "Your approach is fundamentally wrong, abandon it."
    – mlk
    Mar 20, 2020 at 12:15
4

Different literature reviews have different purposes which my determine what kinds of articles they look for. But for most literature reviews, a systematic literature search should include ALL possible sources and there is no justification to include only articles from "top-ranked" or "elite" journals (such as the Q1 that you referred to). That would be saying that it is unlikely for articles in other journals to have anything interesting to say related to your topic, but you have no logical reason to expect that. Your reason given, "If I include all, it will make my review paper really long", is not an acceptable excuse. Any such review article would likely be rejected for publication by any journal in Q1, Q2 and even Q3.

Moreover, in computer science specifically, you absolutely must include conference articles since much of the best original research is published there.

So, what do I recommend for computer science? If you search only in the ACM Digital Library and including anything and everything you find there, then your literature search would often be considered sufficiently comprehensive for the computer science domain. However, some specialized domains might need additional database sources, even in computer science. (If you want to be really exhaustive, you could add arXiv for grey literature [preprints], but that would probably not be required.)

Improve this answer
3
  • 2
    Only on ACM Digital Library? How about other publishers?
    – Jaya A
    Mar 20, 2020 at 11:03
  • 1
    Whether the ACM Digital Library is sufficiently comprehensive is definitely subfield-specific. In my subsubfield (computational geometry/topology), the ACM DL is not even close to comprehensive.
    – JeffE
    Mar 20, 2020 at 17:03
  • @JeffE: Thanks for your comments; I have edited my answer accordingly.
    – Tripartio
    Mar 21, 2020 at 10:10

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