How to Build a Literature Review Outline That Follows an Argument

A literature review outline that follows an argument is not a list of papers grouped by theme. It is a claim-by-claim structure in which every section exists to advance a position, and every cited study has a defined job in that advance. The practical search intent behind this question is usually: I have a pile of sources and a rough thesis, and I need a structure I can defend to a coauthor or reviewer. The answer is to build the outline from the argument down, not from the bibliography up — and to keep a method trail that shows how each source entered the review.

This guide gives a complete sequence: diagnosing your starting state, drafting the argument skeleton, mapping evidence to claims, running and documenting the search, building the outline artifact, and checking readiness. It uses a manuscript-revision example throughout and ends with stop and escalation rules.

Outlines fail for predictable reasons, and the fix depends on where you actually are. Classify your situation first.

• Source pile, no claim. You have PDFs and notes but cannot state what the review argues. You need to write the argument skeleton before touching the outline.
• Claim, thin evidence. You know your position but suspect gaps. You need a documented search tied to eligibility criteria before you commit to sections.
• Draft outline, reviewer objection. A coauthor or reviewer says the structure is "a list" or "doesn't build." You need to re-map each section to a claim and cut or merge sections that only summarize.
• Handoff or new project. Someone else must continue the review. You need an auditable trail: search strings, sources, screening decisions, and the link between each source and each claim.

Write one sentence naming your state and the artifact you are missing. That sentence determines which section below you start with.

An argument-following outline begins with three to five claim sentences, ordered so each one depends on the last. Draft them as plain assertions, not topics.

A workable skeleton has this shape:

1. The established position — what the current literature broadly holds.
2. The tension — where findings diverge, conditions change, or methods limit the established position.
3. The gap — what remains unresolved and why it matters.
4. Your contribution to the argument — how the review reframes or resolves the tension.

Test each claim by asking what evidence would defeat it. If no study could defeat a claim, it is probably a summary statement, not an argument step. Keep the skeleton to one page; it is the spine the outline hangs on.

Once the skeleton holds, convert each claim into one H2-level section of the review. Under each, list the evidence slots the claim requires — not the papers you happen to have.

This table is the core artifact. An empty slot is a search task, not a reason to drop the claim. A filled slot with a source that does not actually bear on the claim is a mis-mapping to fix before drafting.

The Cochrane Handbook's guidance on searching is directly useful here: plan searches from eligibility criteria, use multiple sources, combine free-text terms with subject headings, document the process, deduplicate records, screen in stages, link multiple reports of one study, and rerun searches before publication. Each of these maps onto an outline task.

• Eligibility criteria → evidence slots. Define what kind of study can fill each slot before you search.
• Multiple sources → coverage. Do not let one database define the gap.
• Free-text plus subject headings → recall. Database-specific syntax matters; PubMed supports field tags, phrase searching, filters, Advanced Search, and saved searches, and its search details show how terms were translated and combined. Google Scholar supports author and title searching, date limits, Cited by, Related articles, and alerts.
• Documentation → audit trail. Record strings, dates, sources, and counts.
• Staged screening → slot assignment. Screen titles and abstracts first, then full text, assigning each included study to a slot.
• Linking multiple reports → one study, one entry. Trials or studies reported more than once should not inflate your evidence count.
• Rerunning before publication → freshness check. Repeat searches before submission and note what changed.

If you are searching through WisPaper's scholarly interface, use the WisPaper search to run topic and keyword queries, and keep your strings and dates in the same log you use for other databases. For a public entry point to finding academic papers, see WisPaper Scholar Search.

An outline is only auditable if the sources behind it are organized. Save or upload the papers you screen in to My Library so that included studies, excluded studies, and slot assignments stay together rather than scattered across downloads.

For a review with many sources, a simple evidence log alongside the library works:

This log is what makes a lab handoff possible: a successor can see not just what you cited but why each source is in the outline.

Now assemble the outline itself. For each section, record four things: the claim it advances, the evidence assigned to it, the function it serves in the overall argument (establish, complicate, bridge, resolve), and the transition to the next section.

A worked example, from a manuscript revision:

• Section 1 — Established position. Claim: the dominant account treats the phenomenon as stable under standard conditions. Evidence: two foundational studies. Function: establish. Transition: "But this account assumes conditions that later work varied."
• Section 2 — Tension. Claim: findings diverge when conditions change. Evidence: studies with conflicting results, linked so multiple reports of one study count once. Function: complicate. Transition: "The divergence is not random; it clusters around one unresolved question."
• Section 3 — Gap. Claim: the unresolved question has not been tested directly. Evidence: reviews and primary studies naming the limitation. Function: bridge. Transition: "Reframing the question makes the gap tractable."
• Section 4 — Contribution. Claim: the reframing reorganizes the evidence and identifies the next test. Evidence: studies supporting the reframing. Function: resolve.

If a section cannot state its function in one word, it is probably a summary section. Merge it into a neighboring claim or cut it.

Tool use should follow the outline, not replace it. A few placements that fit this workflow:

• When a claim needs a broader sweep than your current strings produce, Scholar Agent supports longer research tasks in WisPaper's literature review, inspiration, and reproduction use cases.
• When you need to interrogate the papers you have already collected — for example, to check whether a slot is genuinely filled — Library QA lets you ask questions based on papers uploaded to My Library or added from search results.
• Before submission, when the reference list is final, TrueCite screens BibTeX against real academic databases to flag hallucinated references.
• If the review is part of a larger effort with coauthors, Research Projects gives the work a project home.

Each of these sits at a defined checkpoint: broad sweep, slot verification, reference check, project organization. None of them substitutes for the argument skeleton or the evidence log.

Use explicit exit criteria so you do not draft on an unstable outline.

1. Skeleton checkpoint. Three to five ordered claims, each falsifiable in principle. Done when a coauthor can restate your argument without reading the sources.
2. Slot checkpoint. Every claim has evidence slots, and every slot is marked filled, partial, or empty. Done when no slot is silently missing.
3. Search checkpoint. Strings, sources, dates, and counts are logged; duplicates removed; multiple reports linked. Done when a second person could rerun the search.
4. Mapping checkpoint. Every included source is assigned to a slot, and every slot's sources actually bear on its claim. Done when no source is cited only because it exists.
5. Outline checkpoint. Each section states claim, evidence, function, and transition. Done when removing any section breaks the argument.
6. Freshness checkpoint. Searches rerun before submission, with changes noted. Done when the log shows the rerun date.

Clear rules prevent an outline from quietly becoming a bibliography.

• Stop and rewrite the skeleton if two sections make the same claim, or if a section's function cannot be named.
• Stop and search if a gap slot is empty; do not soften the claim to fit available sources.
• Stop and re-map if a source appears in more than one slot without a stated reason, or if a source's conditions do not match the claim it supports.
• Escalate to a coauthor when the tension claim depends on studies with incompatible methods and you cannot state the comparison fairly.
• Escalate to a methodologist or librarian when eligibility criteria or search syntax are contested, or when a database's translation of your terms is unclear.
• Escalate before submission if the freshness rerun surfaces studies that change a claim; revise the outline rather than the citation list alone.

The outline is ready when four tests pass. First, the argument test: reading only the claim sentences in order produces a coherent argument with no gaps. Second, the evidence test: every claim has at least one assigned source whose conditions match the claim, and no source is unassigned. Third, the audit test: a colleague can trace any cited study back through the log to the search that found it and the slot it fills. Fourth, the freshness test: the log shows a rerun before submission and records what changed.

If any test fails, return to the corresponding checkpoint rather than drafting around the weakness. An outline that passes all four is one you can hand to a coauthor, defend in peer review, or pass to the next lab member with the method trail intact.