An application specific step-by-step guide to doing a literature review/search
Why write this guide?
Why did I write this blog/guide on doing a literature search (aka. lit reviews)? The main reason I have for writing this guide is rather selfish… I want to externalize my memory of my own workflow for doing literature searches. There are many detailed steps involved in preparing for the literature search, using the academic search engines, extracting the articles of interest, downloading the pdf, inputting them into the reference manager, cleaning the references up in the manager, and storing notes as I read through these articles. This is actually a lot more tedious and detailed than doing a “simple” literature search may seem… Especially if you want to be systematic about it.
Initially, I had worked out (over many trials and errors) a workflow for doing lit reviews. But! I didn’t write the workflow down (I know, tsk tsk)… And because I don’t do lit reviews/searches that frequently, I forgot how I set up my workflow from the last time. Which is why I decided that I needed to write it down. This blog post is an expanded and elaborated explanation of my workflow.
The second reason why I am writing this guide is for those interested in the precise workflow a researcher (myself!) may take. This is because when I first started doing literature reviews in my MSc, I searched online for how other researchers do it. However, I found next to nothing on the detailed process for the nitty gritty of doing lit reviews. While there were several that I found helpful for the general concept of it (see here, here, here, or here), I wanted to know exactly how some people did it. So I am writing this to give others an idea of how lit reviews may be done.
Caveat: This is an application specific detailed guide… As would be expected given that this is a detailed guide.
Applications used in this guide
As the main title states, this is an application specific guide. I use the following applications for this guide and for my own lit searches:
JabRef is a reference manager that uses
the BibTeX format to store references. I use LaTeX and pandoc to
create my manuscripts and/or reports so I need a program that can read
.bib files. The nice thing about JabRef is that all the
customizations and editting of the references are done to the
reference file, which means that edits and formatting I do to the
.bib file is independent of the computer. Meaning I can just take
the single, plain text file with me wherever and I have the entire
bibliography and formats in one place. JabRef is also available on
Linux, which I use.
Docear is a great, fairly recent application that combines the power of mindmaps with the power of pdf annotating (i.e. highlighting and commenting). It imports annotations from article pdfs into the mindmap, allowing a visual representation of your ideas and how research articles fit into your ideas and research. Docear is also Java-based and uses JabRef as a backend for the references.
I have structured this guide into two parts:
- A brief, concise guide.
- A detailed, descriptive guide.
Brief step-by-step guide
- Develop search terms, using PubMed and MeSH to refine
- Use the terms in the “Web Search” field in JabRef
- Compare and record number of hits between PubMed and JabRef
- If JabRef’s Web Search complains or the hits are different, use
PubMed to output to XML
and import into JabRef (
- If JabRef’s Web Search complains or the hits are different, use PubMed to output to XML and import into JabRef (
- Sort hits by year
- Search through and select relevant articles, checking for duplicates
- Import into database
- Select all imported articles and add relevant keywords using “Edit -> Manage keywords”
- Run “Tools -> Clean up entries”
Month, Note, File, Institutionfields using “Search -> Set/clear/rename fields”
- Mark entries based on priority and relevance to project (
Ctrl-M; darker = most relevant)
- Download PDF for the marked articles using
Christoph Lehner’s plugin for JabRef
- If it doesn’t work, manually download using the DOI link or
Google, renaming the pdfs by the BibKey and move them into the
article/folder (in the folder or up one from the
- If it doesn’t work, manually download using the DOI link or Google, renaming the pdfs by the BibKey and move them into the
- Run “Search -> Synchronize file links”
- Run “Tools -> Clean up entries”
- Confirm rename worked; if not, fix and re-do steps 22 and 23
Filefield (as in step 9)
- Run “Search -> Synchronize file links”
- Import the pdfs into Docear
- Move articles into their relevant branches/mindmap
- Read the articles and make comments
- Import the annotations
- Move the read article into the read branch
- Repeat steps 19-21
Detailed step-by-step guide
A. Firefox (or other browser)
- Develop a priori search terms for searching PubMed or other article search engine.
- Use PubMed and MeSH to refine the terms to capture an appropriate and relevant number of articles. - Obviously too many or too few articles are never a good thing.
- Using the search terms, run a search of PubMed and check the number of hits the search terms retrieved.
- Record the number of hits in a plain text file or some other location, preferably in the same parent directory as your research project for a manuscript.
B. JabRef — Use your master
.bib file, or create a new reference database
— Article search and selection
- Paste the search terms developed above into the Web Search
plugin/utility (usually on the sidepane), using Medline as the
search engine, and hit the
- Check the number of hits and compare to the hits from the PubMed search.
- Record date of search and number of hits somewhere, preferably in
the same directory as your research project.
- Sometimes, the Web Search doesn’t like some of the search
terms and just doesn’t search using them. This can result in
differences in the number of hits. In this case, search
using PubMed (more reliable) and output the search results
into XML via the “Send to: -> File -> XML” option below/near
the “Search” and “Filters:” buttons. Import the XML into
- Sort results by year.
- Search through articles, selecting (checking the check box) only relevant onces to be imported; check for duplicates.
- Import into the current database (or a new one).
— Cleaning the imported files
- Select all the imported references, go to the
Editmenu at the top, and click
- Add appropriate keywords, for instance topics that the articles represent (e.g. Diabetes), the relevant project that the articles are linked to (e.g. course assignment or manuscript), or any other descriptive keywords.
- Under the
Clean up entries. Sometimes the articles have been messily imported and this will clean it up.
- In the
Set/clear/rename fields; type in
Month, Note, File, Institutionat the top of the dialog box, select the
Clear fieldsoption, and click OK.
— Marking by importance and downloading pdf
- Search through articles and based on their title and abstract, mark
Ctrl-Min JabRef) the articles according to how relevant the article is and how soon the article should be read. The darker the colour, the more relevant it is.
- Use Christoph Lehner’s excellent plugin
to download the pdf from within JabRef, rename the file, and save
it in the pdf directory set in the JabRef settings automatically,
all by just clicking the
Journal pdfbutton! - Sometimes journals are behind a paywall (damn closed access), so if you have institutional access, just download the pdf manually, rename it to match the BibKey, and move it into the
- Once all your relevant pdfs have been downloaded, run
Synchronize file linksin the
Clean up entriesin the
- Confirm that the rename worked. If not, it is likely a file name issue, so fix that and do the above again.
- Delete the
Filefield in the selected articles using the
Set/clear/rename fieldsoptions in the
Searchmenu to fix the file link in the articles. Sometimes JabRef doesn’t rename the file link, but does rename the file itself.
Synchronize file links.
B. Docear — Use your main mindmap to import new pdf files
- In Docear’s main
literature_and_annotations.mmfile, use the “Reload monitored folder” button to search the
articles/folder as set in the options and import new pdfs.
- Move the imported articles into a branch that is relevant to the project. At this point, you have a few options. One is to use the main mindmap to import the annotations into, or another option is to make a new mindmap specific to your project, move the articles over into that mindmap and read them from there (this is the option I use).
- Read the articles, make comments, highlight text and so on. Once
done, right-click the article, select
PDF -> Import all annotations.
- Move the read article into a new branch called
Read, or something similar.
- Repeat until all articles are read.
That is all for now. I know this is fairly detailed, but it helps me remember my workflow and I hope that in some way helps you, the reader, to make your own workflow!