Process

The Internet gives you easy access to thousands of high-quality academic papers on any subject you care to search for. In a literature review for a typical undergraduate or Masters project, you should be aiming to use around 30. Of those, only a handful should be treated as key references - those which have the most relevance for your project. The rest will be used to support minor points, provide definitions, support the methodology, etc. Finding that small number of references can be a difficult and time-consuming process and using them to create the literature review can be equally onerous. Both activities can be made easier with a systematic approach.

Building up a literature review is best treated as an iterative process as shown in Fig. 1.

Figure 1: Literature review process

• Literature search: Filtering the available material for potentially relevant items
• Triage: Using the abstract and/or conclusions to reject unwanted items
• Development: Reading remaining items in detail and extracting important points
• Writing up Turning rough notes into well-constructed text
• Redrafting: Reviewing and improving the text

During the iterative part of the process, you should be working entirely with notes. If you were to write up sections too early, there is a good chance that they will need to be updated as you read more material. To avoid this re-work, simply leave writing up until later.

To keep control over the process, you can think of each iteration corresponding to a particular period of time such as a week. This helps you keep up a good pace and means that you will see progress on a regular basis.

Tools

There are three main tools you need when using this process for your literature review:

• A site to search for academic material (journal articles and conference papers) - Google Scholar is the easiest to use.
• A writing application - Microsoft Word is recommended for its outline view.
• A reference manager - Mendeley is recommended for its Word plugin and the watch folder feature.

Please note that other options are available for each of the three tools. You may have good reasons for choosing something different. The main thing is to make sure that you know how to use your tools effectively. You should be in control of the software rather than letting the software be in control of you.

Literature search

There are several web services that allow you to search for academic articles. The simplest on to use is Google Scholar, but you could also use ACM, Science Direct, IEEE and others. If you already have one or more references for your chosen topic area, then you can use them as a starting point. Otherwise, you will need to experiment with some search terms to narrow the search hits down to the ones you are interested in.

Google Scholar allows you to manipulate the search terms you use in quite sophisticated ways to zero in on the papers of most value. For example, you can search for multi-word phrases instead of the individual words by enclosing them in quotation marks. You can also exclude terms by preceding them with a minus sign. This blog post summarises all of the search operators available.

Once you have one relevant paper, you can find older related papers by looking at its list of references. You can also use Google Scholar to find more recent papers use it as a reference by using the cited by link as illustrated in Fig. 2.

Figure 2: Google Scholar

With each iteration of the literature search step, you should aim for a shortlist of around 10 - 20 papers which appear to be relevant and current.

Triage

There is not enough time to read every potentially useful article in detail. The goal of the triage step is to narrow down your shortlist further by weeding out articles that seemed useful at first glance, but which ar not worth reading in full. The abstract is useful for this purpose: it summarises the entire paper including aim, methods and results. You will be able to reject several items this way. If you need a little more information before deciding whether or not to keep a paper, you can look at the conclusions section.

Out of an initial shortlist of 10-20 papers, you should aim to keep 2 - 5 after the triage step.

Development

This is the most demanding stage in the process because you need to do two things at the same time. You need to read each retained paper in detail, and you need to make notes as you go along. This is difficult because academic papers typically require a great deal of concentration to follow, and it you interrupt the flow of the reading to make a note, it may be hard to pick up the thread again.

Microsoft Word has a little-known feature which can help with the development stage. The Outline view shown in Fig. 3 presents your document as a collapsible tree structure based on the heading styles. The outline view is very useful for maintaining an overview of your document's structure as it grows. It also supports drag-and-drop so that it is easy to reorder sections.

Figure 3: Word outline view

Rough notes can be built up in the outline view by copying a section of the reference text and pasting it into the appropriate section of your Word document being careful to enclose it in quotation marks to show that it is verbatim. You also need to show where it came from by including an in-text citation. If you have installed a reference manager, this is a very simple procedure. This process minimises the time needed to take notes and quickly lets you go back to reading the reference paper without a big interruption. Of course, you can also add your own comments along with the direct quote if that would be useful.

As you read more articles, you gradually build up a series of notes in each section of your document. As time goes on, you will be able to see which sections have a good amount of content and which need more. This gives you some parameters to use on the next iteration of the literature search step.

A further advantage of working this way is that it allows you to place notes from a reference into the appropriate place in your document. Thus the procedure distributes multiple notes from the same source around the document rather than keeping them all together.

Writing up

Once a section of your document has a sufficient quantity of notes, you can turn them into full sentences and paragraphs. Because you have collected related notes from different sources together, it is a relatively straightforward process to rewrite the points in your own words eliminating the direct quotations as you go along. The references are already in place, and can be used to go back to a source if you need to refresh your memory of the context for a particular note. Because you have captured verbatim quotes, you can also use the wording to go quickly to the same text in the source document.

When turning your notes into text, you may discover that two authors have made very similar points. If that happens, you only need to make the point once and provide an in-text citation that contains both references. An example is shown below, and your reference manager should make this very easy.

Several authors identified cultural inertia as the main source of resistance to new software tools (Brown, 2023; Jones, 2022; Smith, 2019)

Sometimes the best way to communicate your message to the Reader is to mention the author of a paper directly. This is known as an integrated citation:

The main proponent of this theory in the early part of this century was Jackson (2006) who developed the central principles.

Redrafting

You should always plan to review and redraft your literature review to ensure that it has a logical flow and that there are no gaps. You should do this as if from the perspective of the Reader. It can sometimes help to ask someone with no knowledge of your topic to read your literature review and highlight any parts that are not clear (If you use a proofreader, please be sure to mention them in the acknowledgements section).

If you are going to do the reviewing and redrafting yourself. it is often a good idea to put the document aside for a couple of days before re-reading. That will put a little bit of distance between you and the text making it easier to spot things that might not be clear to the Reader.

Restructuring

This activity is not captured in Figure 1, but will probably be essential at some point. You can introduce new subsections, split existing ones and reorder them at any time during the development phase or the writing up phase. It will take a little time because if you are splitting a subsection into two, for example, you will have to decide which material will go into each of the new subsections. Your main guiding principle, as always, is whether the current structure communicates the content clearly to the Reader or not. If not, you should modify it.