20

For starting a new problem I generally need to look for work already published through Google Scholar. For example, recently I came across an equation, the Pochhammer-Chree equation. When I type this equation name in the Google Scholar tab, a large number of publications on this equations piles up. It is not possible to look for every single publication to check what type of work has already been done on this equation.

Let me elaborate a little more. Suppose I need to check if the Lie symmetry analysis on this equation has been carried out or not. Then I have certain options in advanced search, that to include the words symmetry or symmetry analysis in the title, abstract or in the whole article along with the word Pochhammer-Chree. Here of course the time range can also be given. Unfortunately, even by this type of search I can miss some of the article that have been already published on symmetry analysis of Pochhammer-Chree.

Is my way of search correct? If not, how should I search through the web by not missing a single article on the topic I am looking for?

  • 1
    It is not totally clear to me what's the actual question here. In the title it seems like you are searching some specific article. The question body seems to indicate that you would only like to find any work that could be relevant. It may also be that you want to ask "How can I find out if something has already been treated?". Please clarify. – Dirk Jul 8 '16 at 9:43
  • See How to do a literature search – D.W. Jul 8 '16 at 17:35
  • @D.W. Is there a similar question on Academia.SE? That one is specific (with specific journals) to the cryptography field. – J. Roibal - BlockchainEng Jul 8 '16 at 18:09
  • @J.Roibal, not that I know of. Feel free to ask such a question here, if you'd like to see a similar answer here. However, my answer there is not at all specific to cryptography. The main elements remain the same in any other field -- I would make only extermely small changes for non-cryptographers ( remove recommendations to look at IACR or Crypto.SE; those are the only crypto-specific parts of my answer). – D.W. Jul 8 '16 at 18:14
6

Just to elaborate on what the other posters have mentioned: your first job is to find at least one paper that is very relevant to the specific topic that you are interested in, e.g. symmetry analysis of Pochhammer-Chree equations. Possibly just searching on Google Scholar and browsing the results is your best option here. The number of citations that a specific paper has is an indication of whether that paper was important/influential within your field (i.e., these are the papers your reviewers will know about), so you should make sure to study these papers in particular. Once you have a good reference paper, you can use Google Scholar to click on the blue 'Cited by ##' hyperlink that is on the bottom left of every search result on Scholar. You can use this to find other relevant papers going forward in time. As you collect relevant papers, you can also look at their references (and read the 'related work' sections of the paper) and find other papers that you may have missed.

In most cases, it is impossible to 'not miss a single article' on the topic that you are looking for, and I don't believe this is a useful goal anyways. Of course, in some fields, a result may be binary: either you have proven/shown something or you have not, and once it's done, there is no point doing it again. But in most fields, if two people independently pursue the same topic, they will approach it in different ways, and confirming each other's findings in this way provides a lot of value. When your paper is reviewed, it is essential that you demonstrate an intellectual heritage to your work - that you care about prior work, and have used it to guide you, and you are building on a tradition. If you miss something, the reviewers will be happy to point it out, and in my experience this is rarely the reason a paper gets rejected. And even if a paper is rejected for inadequate understanding of prior work, this is an easy problem to fix for resubmission elsewhere.

So in summary: you can only do your best. The process outlined above is the process used by most people these days. If you follow this approach, and put in the proper amount of effort, then if you miss something anyways, I don't think anyone will hold it against you (and most likely no one, including you, will ever know).

  • 5
    I think the key takeaway is that in research, it is far more likely that you will need to read many papers over weeks or months to get a feel for the state and texture of human knowledge in an area. It is very unlikely that one paper was written that answers your question specifically, even if you had the perfect search string, and fills in just the gaps you need filled. Expect to invest substantial time and intellectual effort. – mightypile Jul 8 '16 at 16:46
6

It is true; there are too many papers out there to even read the abstracts of all of them. That is why many research papers contain phrases like "to the best of our knowledge", etc.

My approach to minimize the risk of missing out publications is as follows: I start with reading review papers. These will sum up the research up to a certain point and will show you what the seminal papers in the field are. Then I look up the latest papers that cite the reviews and/or the seminal papers. This approach works very well with Google Scholar.

  • IMO, the phrase "to the best of our knowledge" should be acronym-ify as "TTBOOK" for short. – Ooker Jul 8 '16 at 17:46
  • 1
    Isn't it better to simply assume that the authors looked back in time "to the best of their abilities"? If someone wants to correct them, they can correct them directly. – Utkarsh Sinha Jul 8 '16 at 18:56
  • @ooker - this acronym does not sound common at all to me so I do not see the need to change it in the post – magnon2020 Jul 10 '16 at 16:26
  • @magnon2020 I just try to have some fun here... – Ooker Jul 11 '16 at 17:48
3

One trick that I find useful, when researching an area that I'm relatively new to, is to sort by the number of citations. Among the top-cited papers from the search will often be the original paper that introduced a concept, and/or a useful review article. Both of those provide excellent starting points to follow citations forwards and backwards through time. If neither of those appear, then out of the top few papers, pick the one with the most relevant sounding title, and with any luck it will, at the very least, provide references within the text that will form a useful starting point.

1

The search feature on Google Scholar is a fantastic starting point to begin your research, but there are also other methods which can be used with Google to provide highly relevant results.

Once you have found a paper in your general topic and have determined to be of high-value (usually the first result in Google Scholar Search) you can then view which articles cite this article, and which articles are cited by this article. That is a fantastic way to develop an interconnected web following the same line of thought through time. You can learn who researched what when, and what advancements were made to that theory or reasoning and how long each development took.

For Example, when searching google scholar for "CRISPR Cas9" (A new methodology for genetic engineering) A number of results is displayed, and underneath each result, statistics are available. Under the first result, it says "Cited By 850", by clicking this link, you are able to see what recent publications on the topic are covering. These can also be sorted by citations, date published, authors, etc. This is one of the most effective methods for performing google scholar searches and finding relevant information.

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.