30

I have worked on many ideas where I found later that they were already published, which was frustrating. I usually look around before developing my idea, but sometimes the articles I read or the keywords I use to look up any related works make me believe that my work is actually original until I find later, after skimming through hundreds of papers in the field (to further check the originality), or by a response from a reviewer that it is not the case.

In short, what's an efficient way to make sure that your work is original before developing the idea?

Edit :
Thank you for your answers. As suggested in the comments, I should have included addditional information so that the answers would make more sense.

I am a PhD student in Computer Science (CS). My main research areas are : Information theory, coding and cryptography.

  • 5
    If you independently replicated someone’s work, wouldn’t replication of that work be interesting in its own right, especially given the current replicability crisis in some fields like psychology? – nick012000 Dec 24 '19 at 13:07
  • 7
    What is your discipline? It really makes a difference in answering your question because although all disciplines require originality for publication in their most reputable publication outlets, different disciplines do not necessarily have the same view of what is "original". Moreover, different disciplines have different feasible approaches to verifying originality early on. – Tripartio Dec 25 '19 at 6:49
  • @Tripartio, mine is CS. I figured that from Nick's comment! Thank you for pointing it out ! – sapienswatson Dec 25 '19 at 7:14
  • 1
    I've offered an answer, but please edit your question to add the fact that you are in computer science (and please spell it out, since abbreviations are not universally understood); this could help more people offer you helpful answers. – Tripartio Dec 27 '19 at 15:01
  • @Tripartio, Thank you for your answer. I will edit my question to include your suggestion. – sapienswatson Dec 27 '19 at 15:54
43

Ask people

Doing literature research on your own is necessary, and you should spend some time on this to make sure that the result is not already available. But there is only so much you can find, and it often happens that the result has previously been published with a different name or title, which you could not have known about. More experienced people in the field might have come across those works and know that your result is not new.

So ask people in the field that you can trust. Your supervisor/boss/similar is usually a good start. You can also ask other people in your department. As you advance in your career, you often make new contacts with whom you can discuss such things. And, as you have experienced, peer reviewers will provide an additional "safety check".

Of course, make sure you only ask people who you can trust that they will keep your work confidential. If you are at the beginning of your career, your supervisor might provide further guidance on this.

|improve this answer|||||
  • I usualy do not ask a lot. I see your point now. – sapienswatson Dec 24 '19 at 10:51
15

Talk to your supervisor

It takes experience to really know what's going on in a research field, who the major and minor players are, what other research groups are working on, what the major unsolved problems are and so forth.

This is experience new* students don't have. Can't have, aren't expected to have. But a good supervisor (who's active in that field) should have a good idea.

*new: you don't list your level of expertise, but I'm guessing it's somewhere postgrad?

|improve this answer|||||
  • 1
    yes, postgrad ! – sapienswatson Dec 24 '19 at 10:48
  • 1
    +1. Effective methods for searching the literature may vary by field. Perhaps your supervisor knows what works in your field. – GEdgar Dec 26 '19 at 13:13
3

Know the state of the art

I believe that there are no real shortcuts except in actually knowing what is the state of the art in the field, what approaches have been tried, what is being used right now, what is not being used because it's been tried in 1980s and doesn't work, what's happening right now, etc. A thorough literature search for keywords relevant to the topic, as some other answers suggest, is a necessary, but not sufficient part of that.

In addition to that, you'd need to follow up all the related approaches, perhaps try them out experimentally (depending on the field), look up if similar things have been tried for all the many, many other problems that are somewhat similar, browse the abstracts for all or most papers in journals or conferences in a subfield to see what other solutions have been tried, etc. It's not "how to do it efficiently/quickly" for a particular paper or solution - it's about doing lots and lots and lots of inefficient work to obtain lots and lots and lots of background knowledge as a table stakes of starting serious independent work in a particular subfield of science, which will then help you (and others!) for many papers in future.

I consider that researchers entering a field are expected to obtain that expertise (or most of it) during their PhD program, and that state is something that can take at least two or more years of full time work to obtain initially (i.e. it's not something that is plausible to have for your first paper in a field) and it deteriorates quickly, needing many hours each month to keep up.

Ask people who know

Of course, it's impossible for everyone to know the state of the art for every field; and it's impossible for anyone to know the state of the art in "their" field if they're just starting. But as I said, I believe that there are no real shortcuts, except asking people who do know that. If you can't do proper independent science in a field yet, then that doesn't mean that you should be doing improper science. It means that you should be doing science that's not fully independent. In particular:

  • If you are doing a PhD in that field, then your advisor should be someone who knows that field. If it's not, you're going to have a hard time because you'll need to put in more work to get that knowledge yourself as soon as possible
  • If your research leads into another field, then it's really, really helpful to involve another researcher from that field. Interdisciplinary research is great, but barging in some field 'uninvited' and uninformed is not going to win you any favors - get a co-author to review your ideas and provide insights from that field, and it's going to be a win-win.
  • If you're experienced, and it's "your" field, and you still don't feel secure about what is original and what isn't, then you need to spend more time on reading semi-related research of others in order to keep up to date. There may be all kinds of conflicting time pressures from teaching or project work or whatever, but that's what needs to be done.

I'm aware that it's a big burden, and especially in some fast moving fields there's a real "Red Queen's race" and it takes all the running you can do to keep in the same place - but if it takes 10 hours a week of skimming papers every week to keep up with what's happening in your field, then that's what it takes to do proper research.

If someone is devoting, for example, just 20% of their working hours to science because of non-science academic duties or because they're working in industry, then that implies (at least to me) that they won't be able to do independent contributions to the subfield because that time is barely sufficient to keep up with what's the state of the art in that field, and knowing what is novel and what has been tried requires involving someone who is arms-deep in that field full-time or close to that.

It doesn't necessarily mean someone more senior - I have seen full professors who fully rely on a particular post-doc or PhD student to be up to date with all the literature on some particular topic.

I'm sorry if it looks like a rant, but to summarize I really want to make the following points:

  • a short-cut way to do it 'efficiently' seems implausible;
  • 'Ask people' and 'Talk to your supervisor' are good short-term solutions, but in the long term the goal is (or should be) to shift from 'ask people' to 'become (for your topic) the person who gets asked'. It's a lot of work, years of work, but it needs to be done.
|improve this answer|||||
2

I am doing my first paper independently (graduated with Bsc), it is in the first revision round (I pray everyday for it to pass!)

I truly understand how you feel... My suggestions goes like this:

  1. Read a lot of research papers and do not stop.
  2. Always take a quick glance at the references section if the paper is passable.
  3. Do not use the same website to search for articles, if you use google scholar a lot, try using reseachgate, academia, sciencedirect etc...
  4. Try using synonyms words for your search (i.e miss up your search by using different set of words)
  5. Do not spend all your time reading papers tightly related to your research, try to go wild a bit and read far from your current study, this will open your eyes on other possibilities.
  6. Seek other sources! I have seen multiple youtubers and science enthusiasts mention subjects, research papers and dilemmas that I have not even thought of or read about! For me as a physicist, youtube channels such as numberphile, veritasium, SmartertEveryday, PBS spacetime are insanely versatile and help me think and discover new things! Blogs are good too, I have a list of many senior physicists and mathematicians blogs, and I visit them every week or two.
|improve this answer|||||
  • +1 for keeping up with popular media. I've lost track of how many papers I've seen where the exact result was published in National Geographic or something twenty years earlier. – StackOverthrow Dec 26 '19 at 21:44
1

In addition to asking knowledgeable colleagues, which some other answers have rightly recommended, I would recommend that you target academic conferences as an intial publication target rather than targeting journals right away.

For many academic disciplines, conferences are a primary means to develop ideas before fully fleshing them out for journal articles. The dynamics in your field, computer science, are a little bit different because in computer science, quite often a highly reputed conference is the "terminal publication", that is, where the final, best developed version of a research project is published. But even in computer science, I don't think that is always the case.

Conferences often have lower standards than journal articles, which means that they often welcome work that is in a less-developed state. Practically, this means that you can spend a shorter time working on the initial idea of an article before submitting it for conference peer-review. If at that point the peer-reviewers tell you that the idea is not new, then would not you have invested as much time on the project as if you had tried to fully develop it for a journal article from the very beginning. In particular, if peer-reviewers can give you an idea of important literature that you have missed, then you can read that literature and then decide on what angle to pursue that has not been studied, and so your initial work would not be wasted at all.

The possible disadvantage of this conference strategy is that if your article is accepted, then you would need to muster the funds to attend the conference and present your work. That could be a major limiting factor, depending on your funding situation. Indeed, funding restrictions are why this strategy of conference presentation is not feasible for many academic fields (such as many of the humanities) where conference funding might not be as readily available.

|improve this answer|||||

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.