GuidesPlaybook

Literature-review map

A review map with clusters, anchor papers, disagreement lines, and gaps to investigate.

Intermediate
Playbook

Literature-review map

A review map with clusters, anchor papers, disagreement lines, and gaps to investigate.

Steps5Time~155 minArtifactReview mapLevelintermediate
Before you start
  • Familiarity with academic databases
  • Basic understanding of citation management
0%0/5 steps

Review builder · 2–3 hours

A literature review is a map, not a summary queue.

This runbook turns search results into themes, tensions, and next reads before you draft prose.

Review map

Review map

One page with 4–6 clusters, two anchor papers per cluster, disagreement notes, and missing-evidence questions.

Cluster: AI feedback in writing courses. Tension: speed and revision volume vs shallow acceptance of generated phrasing.

Set the review boundary

15 minElicit

Write the population, intervention, context, and years you will include — plus one tempting topic you will exclude.

A boundary statement for the review map.
Pitfall
Letting an AI search broaden the topic until every paper looks relevant.
Verification
You can reject a plausible paper without improvising the reason.

Build the source graph

30 minConnected Papers

Seed Connected Papers or Litmaps with two known-good papers and export the candidates worth opening.

A candidate graph with visible seed papers.
Pitfall
Treating graph proximity as quality rather than a lead to inspect.
Verification
Each kept candidate has a reason beyond 'the graph suggested it'.

Cluster by claim, not keyword

45 minZotero

Group opened papers by what they claim, what they measure, and what they cannot answer.

4–6 clusters with two anchor papers each.
Pitfall
Letting abstracts stand in for methods and limitations.
Verification
Every cluster has at least one methods note and one limitation note.

Find disagreement

35 minConsensus

Use Consensus or Elicit to surface where evidence points in different directions, then verify the cited papers yourself.

Three disagreement lines with source pairs.
Pitfall
Flattening disagreement into a safe average sentence.
Verification
You can name which design choices explain the conflict.

Outline from gaps

30 minClaude

Write section headings that answer: what is established, what is contested, and what your next search must resolve.

A review outline led by gaps, not tool summaries.
Pitfall
Asking a model to write the review before you know the argument structure.
Verification
Each heading points to papers and a reason the section exists.
Finish line

You should now have: Review map.

  • Every cluster contains opened papers, not just saved search results.
  • Disagreement lines cite the paper designs that create the tension.
  • The outline starts from evidence gaps, not AI-generated section names.

PrivacyDo not upload licensed course packs or unpublished manuscripts to consumer tools unless your institution permits it. Keep bibliographic metadata separate from private annotations when in doubt.