Before You Draft a Single Paragraph: Mapping the Argument of Your Literature Review

Most literature reviews that lose marks do not lose them for want of citations. They lose them because the reasoning is invisible: paragraph after paragraph reports what others have written, yet nothing explains why one study outranks another, where the genuine disagreement lies, or what the present dissertation adds to the conversation.

An argument map is a modest and practical remedy. It is a single diagram — hand-drawn on A3, assembled from sticky notes, or built in any diagramming tool — that lays out the claims in your field, shows how they connect, marks where the evidence is thin, and places your research question in relation to them. Constructing one takes an afternoon. It can save several weeks of rewriting.

Why a Review Must Reason, Not Recite

Examiners read a literature review for two things: command of the field and command of the argument. The first is demonstrated by coverage; the second by structure. A review that recites is organised by author or by broad theme, and stops there. A review that argues is organised around contested propositions, and each section advances a position: this claim is well supported, that one rests on a single small sample, a third has been tested only in a different sector or country.

The distinction matters for UK master's and doctoral assessment criteria, which typically ask for critical evaluation rather than description. It matters again in the viva, where you will be asked to account for the gap you claim to fill. If you cannot point to the reasoning, you cannot defend it.

Four Symptoms of the Source Parade

• Paragraphs that begin with an author and end with a full stop: 'Smith (2019) found... Jones (2020) argues...' with no sentence linking them.
• Headings that name a topic but not a debate — 'Flexible Working', 'Leadership', 'Motivation' — rather than a proposition that could be contested.
• Citations that never reappear: sources introduced once and then abandoned, rather than used as evidence within an argument.
• A research question that appears in the introduction and the conclusion but is invisible in between.

If three of those four sound familiar, a map will do useful work.

What the Map Contains

Two kinds of element, and nothing more. Nodes are claims: propositions that someone in your field asserts and that could, in principle, be wrong. Edges are relations between claims — supports, contradicts, qualifies, extends, ignores. Attach to each node a brief note on the evidence behind it: design, sample, setting, date. Attach to each edge a word explaining the link.

Around the edge of the diagram sit three further items: the debates you have identified (each named as a proposition, not a topic), the gaps (claims nobody has tested, or tested only in conditions unlike yours), and your own research question, drawn with an arrow to the gap it addresses.

A Worked Example

Suppose the topic is the retention of newly qualified nurses in NHS trusts and the part played by flexible rostering. A first reading of the literature might produce the following.

• Claim A: flexible rostering improves retention among nurses. Supported by workforce surveys and by studies of intention to stay.
• Claim B: flexibility is constrained by staffing ratios and ward-level rostering practice. Supported by qualitative interviews with ward managers.
• Claim C: retention is often measured crudely. A methodological critique, since intention to stay is not the same as staying, and much survey evidence is cross-sectional and self-reported.

Placed on a map, the relations become visible. Claim A is supported largely by evidence that Claim C calls into question. Claim B explains why the association in Claim A may not hold in understaffed trusts — a qualifying edge rather than a contradiction. The gap sits at the intersection: little work links actual rostering patterns to actual leaving behaviour among band 5 staff in trusts with high vacancy rates. Your question follows from that gap, and so does your design — a mixed-methods study pairing trust-level rostering data with interviews with leavers.

That paragraph is very close to what the review itself should say. The map has done much of the drafting.

Building It in Four Passes

First pass — harvest claims. Read with a highlighter and extract propositions rather than summaries. One sentence per claim, in your own words, with the source attached.

Second pass — cluster. Group claims that address the same question and name each cluster as a proposition ('X improves Y'; 'X is constrained by Z'). Clusters that resist naming are usually topics rather than debates.

Third pass — draw the relations and grade the evidence. Note where a claim rests on a single study, on a non-UK sample, or on self-report. Weakness is not disqualification; unacknowledged weakness is.

Fourth pass — insert yourself. Draw the gap, then your question, then the method that answers it. If an arrow is missing, the logic has a hole, and it is far better to find it now than in the viva.

Reading the Map Back into Prose

Each section of the review should correspond to a cluster, and each paragraph should advance a single relation. Begin paragraphs with the claim rather than the author: 'Evidence that flexible rostering improves retention is largely survey-based and cross-sectional.' Then bring in sources as support, qualification or counter-evidence. Use relational verbs deliberately — corroborates, complicates, rests on, extends to — and let the final sentence of each section hand over to the next.

A Checking Routine Before You Draft

Six quick tests:

1. The fifty-word test. Can you state the argument of the review in fifty words without naming a single author? If not, there is no argument yet.
2. The contested-proposition test. Does every heading name something arguable?
3. The orphan test. Does every source appear at least twice, or once in a load-bearing position?
4. The counter-argument test. Is the strongest objection to your claimed gap represented on the map?
5. The gap-to-question test. Is there a visible line from the gap to the question to the method?
6. The honesty test. Would the authors you cite recognise their positions as you have drawn them?

Keeping the Map Ethical

A map is a thinking tool, not a shortcut. It cannot substitute for reading the sources you cite, and it must not flatten positions into convenient opposites. Where a study is contested, say so. Where evidence is thin, mark it as thin rather than overstating it, and never let the tidiness of the diagram push you into claiming more consensus than the literature supports. Accurate representation of others' work is the foundation on which your own contribution rests.

Drawn honestly, a map does something a reference list cannot: it makes the reasoning of your review visible, and therefore open to challenge, revision and defence.