The Five-Step Literature Review Development Process
Five processes in one table, one of which quietly redefines what a good topic is. The fifth cell alone contains nine separate instructions — it is the point where the table stops being a plan and becomes a method.
The five steps, as the source states them
The table is short and it repays reading exactly as written before you interpret it. Two of the five cells contain instructions that would not survive unexamined, and one contains nine instructions compressed into a paragraph.
THE DEVELOPMENT PROCESS AS TABULATED IN THE SOURCE
| Process | Task, as stated |
|---|---|
| 1) Find a topic and narrow it down | Begin with a research question |
| 2) Choose a controversial area | Identify a topic that has a great deal of difference of opinion. This will ensure that you have a great deal of published material |
| 3) Design a research question | Write a good research question that provides scope for discussion |
| 4) Find resources | Be discriminating in your selection of resources. It is best to use refereed journals. Do not choose articles that are too difficult for you to understand. Only use secondary sources if you are not able to access the primary source |
| 5) Write a summary and critique of what you read | Write a critical review for each article you read, then combine similar and related ideas — nine separate instructions, unpacked below |
Process names and tasks are the source's own. The fifth task is abbreviated here and given in full further down the page.
Note what the table is and is not. It is a development sequence for the review itself — how the document comes into existence — rather than a description of what the finished review contains. For the latter, see what is a literature review.
Steps one and three: the question appears twice
Step one says to begin with a research question. Step three, two steps later, says to design one. The table does not explain the relationship, and read literally it asks you to start with the thing you produce in the middle.
The phrase that carries the real instruction in step three is "provides scope for discussion". A question with a settled answer, or one answerable from a single document, gives you nothing to review. This is the same requirement step two states from the other direction.
Step two: choose a controversial area
This is the step to think hardest about, because it changes the selection criteria for a topic and it does so for a stated reason: volume of published material. Take the reasoning seriously — a topic where the field disagrees is a topic that has been argued about in print, and argument in print is what a review has to work with.
It is also, on its own terms, an instruction about supply rather than about value. Consider two framings of the same underlying interest in risk practice. How does my organisation number and maintain its risk registers? has a difference of opinion of approximately zero and almost no published material behind it. Does formal risk registering change delivery outcomes, or only the record of them? is contested, has been argued from several positions, and gives a review something to narrow through. The second framing is better by step two's criterion — and it is a genuinely different research project, not a rewording of the first.
What step two optimises for
- Volume of published material
- The presence of competing positions to review
- Scope for discussion in the research question
- A field with visible unresolved issues
What the earlier topic advice optimises for
- Your own interest, so the work does not become a slog
- An area where you already hold technical knowledge
- A topic that is narrow enough to be interesting and broad enough for adequate information
- Feasibility — time, access, and the topic being 'doable'
For a project manager researching alongside a delivery role, the honest resolution is usually to keep the interest and the domain knowledge fixed, and to move the framing until it lands on a contested question. That is a reframing exercise, not a change of subject — and it is what step one's "narrow it down" is asking for.
Step four: three selection rules, one of which is not about quality
Step four contains three distinct instructions bundled under "be discriminating". They do different work and they deserve to be separated.
The three rules inside step four
It is best to use refereed journals
A source-quality rule, and consistent with the rest of the week's teaching on academic value. Refereed material has been assessed by other scholars before publication, which is the basis on which the source treats it as more credible.
Do not choose articles that are too difficult for you to understand
A reading-level heuristic. It says nothing about whether an article is good, relevant or important — only whether you can currently follow it.
Only use secondary sources if you cannot access the primary source
A provenance rule. Cite the study, not somebody's account of the study, unless the study itself is genuinely unavailable to you.
Rule three is the one that quietly determines whether your reference list can be trusted. The distinction it depends on is covered on primary and secondary data sources, and the criteria for judging a source's academic value are on assessing the academic value of sources.
Step five: nine instructions, two different jobs
The fifth cell is the longest by a wide margin and it contains the entire analytical method the table offers. Separating its nine instructions makes visible something the paragraph hides: they belong to two different activities, performed at different times.
THE NINE INSTRUCTIONS IN STEP FIVE, SEPARATED BY JOB
| Instruction, as stated | Job it belongs to | What it asks of you |
|---|---|---|
| Write a critical review for each article you read | Per article | One evaluative account per work, written while the work is fresh |
| Identify main ideas | Per article | Reduce the work to its load-bearing claims |
| Select the main idea or key points when they are explicitly stated | Per article | Take the author's own statement of the point when one is offered |
| Determine the main ideas when the author is not being explicit | Per article | Infer the point when it is not stated — the hardest instruction in the cell, and the one most likely to introduce error |
| Reflect the author(s) emphasis | Per article — fidelity | Keep the author's proportions. What they treated as central stays central in your account |
| Paraphrase accurately | Per article — fidelity | Restate in your own words without shifting the meaning |
| Exclude personal opinions | Per article — fidelity | Your judgement of the work is a critique; your view of the subject is not part of the summary |
| Combine similar and related ideas | Across articles | Group by idea rather than by source — the move from a set of accounts to a review |
| Combine similar or related details into categories and provide a label or heading for ideas | Across articles | Name the groups. The labels become your section headings |
All nine instructions are the source's own wording. The grouping into per-article and across-article jobs is a reading of them, offered because the source presents all nine as a single undifferentiated paragraph.
The last two rows are the hinge of the whole table. Everything above them produces a set of per-article critiques — which, stopped there, is an annotated bibliography. The combination step is what converts that set into a review, and the instruction to label each category is what gives the review its structure. That distinction is the subject of what a literature review is not.
Three of the nine are fidelity rules — reflect the author's emphasis, paraphrase accurately, exclude personal opinions. They constrain how you write about a work you have decided to include, and they are the ones an examiner can check against the original. The techniques for satisfying them are on critical summary and paraphrasing.
Working the table without over-reading it
Two practical framings, offered as synthesis. First, run steps four and five as a loop rather than as phases — the categories you create in step five tell you what to search for next, and new resources change the categories. Second, expect step three to be revisited: a question written before the reading rarely survives it intact, and the table gives you no reason to defend one that should change.
Working through the five steps
- The topic has been narrowed, and you can state what it excludes as well as what it includes
- You can name at least two positions in the literature that genuinely disagree with each other
- The research question, as currently written, could not be answered from a single source
- Your resource set is predominantly refereed, and any secondary source in it is there because the primary was inaccessible
- Every article you have read has an evaluative account written against it, not just a highlighted PDF
- Similar ideas have been grouped, and each group carries a label you would be willing to use as a section heading
- You could point to any sentence in your summaries and show which part of the original it came from
What to carry forward
- Five steps: find and narrow a topic, choose a controversial area, design a research question with scope for discussion, find resources discriminatingly, then summarise and critique.
- Step two optimises for volume of published material. The earlier topic advice optimises for interest, technical knowledge and feasibility. Both are in the source; neither is reconciled with the other.
- "Do not choose articles that are too difficult for you to understand" is a reading-level heuristic. It is not a judgement about the article, and excluding hard papers systematically skews a review.
- Step five contains nine instructions covering two jobs — critiquing each article, and combining across articles. The combination step is what makes it a review rather than a bibliography.
- Three of the nine are fidelity rules: reflect the author's emphasis, paraphrase accurately, exclude personal opinions. They are the ones a reader can check against the original.
Frequently asked questions
Do I have to choose a controversial topic?
The source's table says to, and gives a specific reason — difference of opinion ensures a great deal of published material. That is a supply argument rather than a claim that contested topics are more worthwhile. The same teaching series elsewhere tells you to choose a topic you are interested in and that is doable, and the material never reconciles the two, so the judgement is yours to make and record.
Why does the question appear in both step one and step three?
The table does not say. The workable reading is that step one's question is a provisional working question good enough to direct a first search, and step three's is the designed question written after you have seen where the field disagrees. The source labels neither as provisional, so treat this as an interpretation rather than a stated distinction.
Should I really skip articles I find too difficult?
The source says not to choose them, and states it without qualification. Read it as a warning about summarising accurately — a paper you cannot follow cannot be represented faithfully — rather than as a filter on quality. The safer response to a difficult but central paper is to get help with it, and to declare the exclusion if you do leave it out.
What is the difference between step five's per-article and across-article work?
Writing a critical review for each article, identifying its main ideas and paraphrasing it accurately are all done one work at a time. Combining similar and related ideas into labelled categories is done across the whole set. Stopping after the first job leaves you with an annotated bibliography; the second job is what produces a review.
How many sources does the process require?
None are specified. The table gives no volume, no minimum and no stopping rule, and neither does the surrounding teaching material. Step two's logic implies that a contested area will supply enough, but that is an implication rather than a stated threshold.
When do I stop searching and start writing?
The table implies a sequence — find resources, then summarise and critique — but the two work better as a loop. The categories you create while summarising tell you what is still missing, and writing is usually what exposes the hole. The source sets no rule either way.
References and source attribution
- The supplied teaching source: master's-level teaching material on reviewing the literature, supplied as a teaching document, accompanying slides and printed study notes. The five-step table and all nine step-five instructions are quoted from that material; the earlier topic-selection criteria are quoted from the same series' material on developing a research topic.
- O'Leary, Z. 2017, The Essential Guide to Doing Your Research Project, 3rd ed., Sage Publications, London — listed among the reference works cited by the supplied source.
- Bryman, A. 2016, Social Research Methods, 5th ed., Oxford University Press, Oxford — listed among the reference works cited by the supplied source.
- Naoum, S. G. 2013, Dissertation Research & Writing for Construction Students, 3rd ed., Routledge — set as reading in the supplied source's material on topic selection.
Suggested questions for Ask KEVOS
- What are the five steps in the literature review development process?
- Why does the source tell me to choose a controversial research area?
- What are the nine instructions inside step five of the process?
- Is 'too difficult to understand' a legitimate reason to exclude an article?
- How do I turn per-article critiques into a synthesised literature review?