Literature Review Writing: From Notes to Synthesis

A literature review becomes writable when your notes stop being a pile of summaries and start answering a question. The practical move is to convert reading into a small set of decisions: what counts as eligible evidence, how each source is tagged, what claim each source supports, and where the gaps sit. Once those decisions are visible, synthesis is mostly assembly and argument rather than a fresh act of reading.

This guide is a pillar map for students and early-career researchers finishing a course paper, a thesis chapter, or a first manuscript. It diagnoses where you are starting, names the artifacts you need at each stage, and shows how to tell when the draft is ready. A running example — a graduate student cleaning up a semester of notes over winter break before a January project start — keeps the steps concrete.

Most stalled reviews are stalled for one of four reasons, and each needs a different first move. Read the descriptions below and pick the one that matches your folder right now.

In the winter-break example, the student has roughly forty papers saved across folders and a supervisor note saying the chapter "reads like an annotated bibliography." That note is a diagnosis: the sources are organized but flat. The fix is not more reading; it is extracting claims and grouping them into themes that can carry an argument.

Synthesis depends on a stable definition of what belongs in the review. Cochrane's guidance on searching and selecting studies starts from eligibility criteria and builds the search from them, rather than searching first and deciding later what to keep. The same logic scales down to a course paper: write your inclusion and exclusion rules in a short paragraph, then apply them consistently.

A workable eligibility note answers three questions:

• Population or topic scope: which subjects, settings, or problem types are in and out.
• Source types: which study designs, publication types, or date ranges you will accept.
• Language and access: what you can realistically read and retrieve.

Cochrane also recommends combining free-text terms with subject headings and drawing on multiple sources, and documenting the process as you go. For a thesis chapter, that documentation can be a simple search log: database, date, query string, and number of records kept. It takes minutes to write and saves hours when a supervisor asks how you found your sources.

Tags are the bridge between notes and synthesis. Zotero's documentation notes that collections can place one item in multiple project groupings without duplicating it, and that tags can describe topic, method, status, or workflow. That flexibility is exactly what a review needs — but only if you choose a small, consistent vocabulary.

A practical scheme uses three tag families:

1. Topic tags — the themes your review will argue about.
2. Method tags — how each study produced its evidence.
3. Status tags — for example, to-read, key, cited-in-draft, excluded.

Keep the topic vocabulary to a handful of terms at first. If you find yourself inventing a new tag for every paper, your themes are not yet defined, and that is the real work to do next. Zotero's saved searches update automatically as items change, which means a saved search for key items can become your working set for the draft.

If you are starting from a search rather than an existing library, run queries through WisPaper search and save the papers you keep into My Library so the tagging work happens in one place.

An annotated bibliography records what each paper says. A synthesis records what each paper lets you claim. The difference is a single sentence per source, written in your own words and tied to a theme.

Use a claim card with four fields:

• Citation: author, year, and a link to the record.
• Claim: one sentence the source supports.
• Strength: what kind of evidence backs the claim — for example, a controlled study, a survey, or a conceptual argument.
• Theme: which topic tag it belongs to.

When two sources support the same claim, note the agreement. When they conflict, note the conflict and what differs — population, method, or measurement. NIH guidance on scientific rigor treats careful assessment of prior knowledge, strengths, weaknesses, and knowledge gaps as core to rigorous practice, which is precisely what a claim card captures at the sentence level.

If you have uploaded your kept papers to My Library, you can ask questions across them with Library QA to check what a set of papers actually says before you commit a claim to the draft.

This is the step where most drafts are won. Build a table with themes as rows and sources as columns, and fill each cell with the claim that source contributes to that theme. Empty cells are informative: they show where a theme rests on one study or none.

Read the table by column to see each source's contribution, and by row to see the state of each theme. A row with a single entry is a thin theme; a row with a contradiction is an argument waiting to be written; a row that is empty across all sources is a candidate gap.

This is also where a longer, multi-step task can be handed to a tool built for it. Scholar Agent is designed for longer research tasks with literature review use cases, which fits the stage where you need to survey a theme before deciding whether it is thin or well covered.

A gap statement is a claim about the literature, so it needs the same support as any other claim. Write it as a sentence with a scope: Among the studies in this review, few examine [population] in [setting], and none combine [method A] with [method B].

Test each gap statement against your synthesis table:

• Does the table actually show the absence, or does it show that you have not searched for it?
• Is the gap about evidence, method, population, or theory?
• Would a reader accept the gap from the sources you cite?

Cochrane's recommendation to rerun searches before publication is a useful habit here: a gap that survives a fresh search is stronger than one that reflects an old query. If a gap points toward a new direction rather than a missing study, Idea Discovery is built to take research topics, background, references, and experimental constraints and generate research ideas — a natural next step once the gap is written down.

With themes and gaps in hand, the writing order matters less than the logic. A reliable sequence is:

1. Frame the theme in one or two sentences.
2. State the shared finding and cite the sources that agree.
3. Introduce the tension where sources diverge, naming what differs.
4. Close with the gap that the theme exposes.

This structure keeps every paragraph doing synthesis work rather than summary work. If a paragraph only describes one paper, it belongs in your notes, not the draft.

For a thesis chapter, keep a running list of claims you have made and the sources behind each. That list becomes your revision checklist when supervisor feedback arrives, and it makes citation cleanup a mechanical task rather than a hunt.

Citation errors are the most avoidable reason a draft comes back. Before you send a chapter or submit a manuscript, screen your BibTeX against real academic databases. TrueCite is built for exactly this: it checks BibTeX entries and flags hallucinated references, so you can fix them before a reader finds them.

Pair that with a manual pass: confirm that every in-text citation appears in the reference list, that every reference is cited, and that the claim attached to each citation matches what the source supports. If your review is a systematic review rather than a narrative one, PRISMA 2020 provides a reporting checklist and flow-diagram templates that cover search, selection, and synthesis methods — worth consulting before submission so your reporting is complete.

The draft is ready when it passes a short set of checks, not when it feels finished.

If a row fails, you know exactly which artifact to fix. That is the advantage of building the review as a set of decisions rather than a single long document.

Not everyone starts at the same point. Use the route that matches your situation.

You have notes but no structure. Start with eligibility criteria, then build the tagging scheme. Move to claim cards only after the vocabulary is stable.

You have a tagged library but no themes. Build the synthesis table directly from your tags. Let empty and contradictory cells tell you where the argument is.

You have themes but thin gaps. Rerun searches for each candidate gap and write the gap statement with a scope. If the gap suggests a new study, move to idea generation.

You have supervisor feedback. Convert each comment into a checklist item tied to a specific artifact — a claim card, a theme row, or a citation. Then revise in that order.

You are starting a January project. Set up the project structure first: a project area, a library, and a tagging scheme. Research Projects is where you can create a new project, and WisPaper Scholar Search is a public entry point for finding papers to seed it.

Return to the graduate student with forty papers and a supervisor note. Over two weeks, the work looks like this:

• Days 1–2: Write eligibility criteria and a search log. Exclude off-topic papers and record why.
• Days 3–4: Apply three tag families in the library. Save a search for key items.
• Days 5–7: Write one claim card per kept source. Flag agreements and conflicts.
• Days 8–9: Build the synthesis table. Identify thin themes and candidate gaps.
• Days 10–11: Draft theme paragraphs in the fixed order. Write gap statements with scope.
• Day 12: Screen BibTeX and reconcile in-text citations with the reference list.
• Day 13: Run the readiness checklist and send the revised chapter to the supervisor.

The output is not a longer document. It is a chapter where every paragraph makes a claim the sources support, and every gap is one the table can defend.

Pick the starting state that matches your folder, then do the first move for that state today. If you are in the scattered-notes stage, write your eligibility criteria in one paragraph before you open another PDF. If you are in the flat-notes stage, write ten claim cards and build the first version of the synthesis table. Either way, you will end the session with an artifact that moves the review forward — and a clear signal of what to do next.