A literature review gets difficult when several papers seem to answer the same question in different ways. One reports an association, another tests an intervention, and a third uses the same term for something else. A folder of summaries keeps the papers together. A synthesis board helps you work out what they establish together.
This guide uses Acres for that board. We make Acres, and recommend it for moving between PDF passages, Markdown notes and a visual argument. Bibliographic management is a separate job.
UNB Libraries recommends synthesis charts and matrices for comparing sources. The workflow below adapts that idea to a canvas: keep sources, evidence and your interpretation distinct, then connect them.
Start with a question you can inspect
Create a card at the top of the board with the question and scope. “Remote work” is too broad to guide a comparison. “How do teams preserve the reasoning behind decisions when communication is asynchronous?” gives you something to look for.
Record the population or setting you care about, the terms you’re using, and the limits of the review. Keep a separate search log with databases, queries, dates and inclusion decisions. Those records explain how you found the material; the board explains how you understand it.
If you’re conducting a systematic or scoping review, follow the process appropriate to that review. For systematic reviews, PRISMA 2020 provides reporting checklists and flow diagrams; arranging papers on a canvas doesn’t satisfy those requirements by itself.
Give sources and claims different cards
Use three kinds of cards. They serve different purposes, even when they sit side by side.
| Card | What to record | What it lets you check |
|---|---|---|
| Source | Full title, authors, year, DOI or URL, study design, setting and citation-manager record | Which work you’re using and what kind of evidence it offers |
| Evidence | Exact passage or a clearly labeled paraphrase, page or section, source link and relevant limitations | Whether the paper supports your reading |
| Synthesis | Your claim, supporting sources, disagreement, limits and open questions | What follows from comparing the evidence |
Write the source metadata yourself or bring it from a checked bibliographic record. A filename is rarely enough: paper-final-2.pdf won’t help you distinguish editions or prepare a citation.
In Acres, text PDF highlights can become cards with a backlink to the source page. Follow the PDF highlights workflow for the exact steps and the difference between text and area selections.
Keep a quoted excerpt intact. Write your explanation below it, or in a separate synthesis card. A reader should be able to tell where the author’s words end and yours begin.
Arrange the board around the argument
Create named sections for Sources, Claims, Disagreements and Questions to resolve. The names can change as the review develops. Place the question where it stays visible.
Put a synthesis card in Claims, with the relevant evidence nearby. Add connectors when their meaning is useful: “supports,” “different setting,” “challenges the definition.” Explain the relationship in the card body too, so it survives outside this particular board.
A source may contribute to several themes. Reuse its card on the relevant Acres boards rather than making another copy of the same note. A card is shared across its placements, so a correction to its body is available in each view.
Start with a few papers. Once a section gets crowded, make a separate board for that question and keep the overview focused on the main claims.
A worked example: disagreement or different questions?
The entries below are fictional source notes, included to demonstrate a comparison. They are not papers you can cite or research findings.
Suppose your question concerns decision records in asynchronous teams:
| Example source | What the note records | What it doesn’t establish |
|---|---|---|
| Source A: interviews | Participants describe losing the reason behind older decisions | Whether a particular documentation format fixes the problem |
| Source B: short experiment | A structured record helps participants find an earlier decision in a controlled task | Whether people maintain those records over months |
| Source C: team case study | A shared log still needs someone responsible for updating it | Whether the same arrangement suits other teams |
The synthesis card might read:
Example interpretation: Finding a decision and keeping its explanation current are separate problems. Source B addresses retrieval; Sources A and C raise questions about what is recorded and maintained. These notes don’t yet establish a durable process for a working team.
That is more useful than “A says this, B says that, C says the other.” It identifies the difference between the sources and the next question.
Before calling two papers contradictory, compare their definitions, methods, settings and timeframes. A reported experience and a measured task outcome may be complementary. Put a real disagreement in its own card, with the passage on each side and your explanation of the conflict.
Turn a cluster into a paragraph
Choose one synthesis card and draft a paragraph with a claim, evidence, qualification and implication. Return to the originals as you write. Check that a paraphrase hasn’t quietly become stronger than the source.
Leave unresolved questions visible. “The included papers use incompatible definitions” can be part of the argument. It also gives you a precise search task: find work that defines or measures the missing concept.
Use your citation manager when moving into the manuscript. Zotero’s Word, LibreOffice and Google Docs plugins maintain document citations and dynamic bibliographies. An Acres deep link returns you to your working material; it doesn’t replace that citation.
Use AI for a bounded comparison
An agent can help compare selected notes or identify places where your draft lacks a source. It should work from material you’ve actually provided.
Read these source and evidence cards. Compare their definitions and methods. For each suggested synthesis, identify the supporting cards and explain its limits. Mark gaps rather than supplying missing findings. Show the draft before creating a card.
Read the output against the papers. A plausible explanation may still combine incompatible results or attach the wrong source. The local MCP guide covers retrieval, drafting and board layout as separate steps.
A local note library doesn’t determine where AI processing happens. Check the model route before sharing material; the offline AI guide explains the local-model option.
Before drafting the review, inspect the board for uncited claims, evidence with missing locations, apparent contradictions and unanswered scope questions. Preserve useful synthesis cards beyond this project with a visual Zettelkasten. The test is whether you can follow a sentence back to its evidence and explain why you included it.