Is there a way to find review articles by the keyword?

Question

When I search for papers using some keywords in academic databases (like Google Scholar), there are lots of papers, but I would like to read literature reviews first because I’m new in that area. (By literature reviews, I mean article-length papers that analyze existing work on a specific topic.)

So I think it would be helpful for me if there were a way to find literature reviews using the keyword or using the paper’s title. I want to categorize papers I’ve found by its document type (research, survey, review, etc).

What I've tried is using a specific academic database (ScienceDirect), which provides an advanced search, in which one can filter the paper by its type (original research article, review article, short survey, etc) – but I feel some limitations of this method. I think I can also do this by adding the word review to my keywords for searching or by reading a paper’s abstract, number of references and guessing whether it is review article or not.

Is there a more efficient way to find review articles by the keyword?

(My research area is computer science.)

UPDATE

"feel some limitations" in my post, I'm not sure but it seems that ScienceDirect only searches for literatures that belongs to ElseVier company. I think It doesn't cover the journal articles well, comparing to e.g) Google Scholar so I said I feel some limitations of it

Answer 1

Google Scholar does not offer as concise query commands, as for example, Scopus and ProQuest. Note that there is a bug in Scopus. When I use commands such as the following, they don't work unless, I add an extra space " ", at the end of each line. Here is one of my Scopus queries, which shows you an example, of exactly what you are looking for. The keyword fields are ar:artical, cp:conference contribution, comp:computer science.

(  
    ( 
        ( DOCTYPE(ar) AND PUBYEAR AFT 2008 ) OR 
        ( DOCTYPE(cp) AND PUBYEAR AFT 2010 ) 
    ) AND 

    TITLE-ABS-KEY 
        ( 
            "design science" and "literature review"
        ) AND 

    ( 
        SUBJAREA(COMP)
    ) AND

    LANGUAGE(English) 

)

Answer 2

To me, using academic citation databases (such as Scopus or Web of Knowledge) is the obvious choice for these types of searches, since they have option for filtering on article type (which Google Scholar doesn't have). You haven't explained why you "... feel some limitations.." when using ScienceDirect, so it is hard to know exactly what you feel is lacking. If your issue is that it is a closed subscription platform, you can also use PubMed, which can also filter search results on article type (as review[Publication Type] in advanced search or by clicking on "Review" in the left panel when viewing search results). The main issue with PubMed is that it has a smaller coverage than the other databases, at least in some fields of science.

Answer 3

You can simply search for "review*" as a required keyword in the title, abstract or keywords in a subscription scholarly database. There are three important aspects to my answer:

• Search for "review*" as a required keyword: In addition to all your keywords of interest, you should connect them to "AND review*". This will get you all articles related to your topic that also mention something about "review". A literature review might be called "a literature review", "a systematic review", "reviews the literature", or other related terms, so "review*" (don't forget the "*") would be needed to capture all these.

• Search in the title, abstract or keywords: Do not do a full text search, or else you will return far too many irrelevant results, since most articles probably mention "review*" somewhere. However, I expect that the vast majority of literature reviews would mention that keyword in their abstracts, if not in their title or keyword list. Such a search would return some articles that are not standalone literature reviews, but you always want a search strategy that returns a little too much, and then you can manually filter the results. That way, you don't miss anything.

• Search a subscription scholarly database, not Google Scholar: Google Scholar is great for many things (I use it almost every day), but absolutely not for this kind of search. Google Scholar does not allow you to search only in the title, abstract or keywords. It only allows title searches, or abstract searches for articles in the last 12 months (or something like that). I believe that this is because publishers allow Google to index their texts only on the condition that it is not allowed to be as useful as paid subscription databases. So, it can only return thousands of results for this kind of search; you can't use it for the higher precision strategy I describe here.

Answer 4

There are a couple of methods to do this. I detail all 4 methods here

https://medium.com/a-academic-librarians-thoughts-on-open-access/4-ways-to-find-review-papers-systematic-reviews-meta-analysis-and-other-rich-sources-of-82898aebb6e7

1. Use built-in filters that exist in databases like Pubmed, Scopus, Web of Science. However these tend to have somewhat selective in coverage, and as I write this in 2020, Microsoft Academic (Microsoft's answer to Google and almost as large), has a autogenerated vocab that includes labels for "Review Articles", Systematic Reviews and Meta-analysis which can be used for filtering.

2. Do a keyword search and hope to match title, abstracts etc. This is the main method used for Google Scholar. As noted this can be quite inaccurate beyond the first few results, because Google Scholar has limited search synatx to do field searching.

3. Advanced boolean searching - using a complicated and advanced boolean search query on a syntax capable search engine, one can create a search with high precision and high recall to find such articles. I detail a method here.

4. Using bibliometric citation mapping to identify review papers - The idea here is that if intutively if you see a lot of papers that you are interested in are cited by this one paper, it is likely this paper could be a review paper (and even if not it is probably relevant). Tools like CoCites exploit this to try to detect such papers.