Describe what experts agree on and where they differ
Avoids a false binary of settled versus disputed.
[QUESTION][EXPERT SOURCES][REVIEW DATE]
Topic · 27 prompts
Research & analysis
Explore research & analysis prompts with clear inputs, complete instructions and ways to adapt the result. Explore “Describe what experts agree on and where they differ” and “Select case studies without cherry-picking”.
Select case studies without cherry-picking
Makes inclusion rules visible before choosing examples.
[RESEARCH QUESTION][CANDIDATE CASES][SELECTION CONSTRAINTS]
Trace a reported number to its dataset
Checks denominators, revisions and definitions.
[NEWS CLAIM][LINKED REPORT][DATA TABLE]
Translate uncertainty intervals carefully
Explains what an interval does and does not mean.
[RESULT TABLE][METHOD NOTE][AUDIENCE]
Separate original findings from commentary
Stops secondary summaries from becoming evidence.
[CLAIM][SOURCE LIST][PRIMARY DOCUMENTS]
Document an archival item's provenance
Preserves uncertainty around origin and custody.
[ITEM DESCRIPTION][CATALOGUE RECORD][PROVENANCE NOTES]
Design a fair before-and-after comparison
Names confounders before interpreting change.
[INTERVENTION][OUTCOME][AVAILABLE BASELINE]
Use a translated source without flattening meaning
Tracks translation uncertainty in research notes.
[ORIGINAL EXCERPT][TRANSLATION][RESEARCH CLAIM]
Check whether a citation supports its sentence
Reads the cited passage in context.
[DRAFT SENTENCES][SOURCE EXCERPTS][SOURCE LINKS]
Test whether a proxy measures the claimed concept
Separates convenient metrics from valid evidence.
[CLAIMED CONCEPT][PROXY MEASURE][STUDY CONTEXT]
Find who a study could not observe
Examines the sample frame before generalising.
[STUDY METHODS][TARGET POPULATION][RECRUITMENT NOTES]
Audit the date of official guidance
Avoids relying on superseded instructions.
[QUESTION][OFFICIAL SOURCE SET][AS OF DATE]
Trace how a survey question changed across waves
Tests whether wording and response-option changes make survey results incomparable over time.
[EARLIER QUESTIONNAIRE][LATER QUESTIONNAIRE][MEASURE OF INTEREST][CODEBOOK NOTES]
Design a fair test for a product claim
Converts marketing wording into measurable conditions and failure criteria.
[PRODUCT CLAIM][AVAILABLE RESOURCES]
Turn a research method into a replication card
Extracts inputs, steps and missing details needed to repeat a study.
[PAPER OR PROTOCOL][RESEARCH QUESTION]
Check a project against a funding call before writing
Tests eligibility and fit with direct evidence and open questions.
[FUNDING CALL][PROJECT FACTS]
Plan a local archive search before visiting
Builds specific catalogue queries and a provenance capture sheet.
[RESEARCH QUESTION][KNOWN NAMES][ARCHIVE CATALOGUE]
Compare two versions of a standard by changed obligations
Finds substantive changes and marks interpretation for expert review.
[OLD STANDARD][NEW STANDARD][USE CASE]
Prepare expert questions from evidence gaps
Turns a research hole into precise questions.
[CURRENT FINDINGS][UNRESOLVED GAPS][EXPERT ROLE]
Find blind spots in a source list
Audits provenance, geography and incentives without tokenism.
[SOURCE LIST][TOPIC]
Explain a failed hypothesis usefully
Turns a null result into a better next experiment.
[HYPOTHESIS][METHOD][OBSERVED RESULT]
Design interviews that discover rather than sell
Produces neutral prompts and a respectful analysis plan.
[RESEARCH QUESTION][PARTICIPANT TYPE]
Track how a claim changed over time
Tests whether a statement survived newer evidence.
[CLAIM][DATED SOURCE SET]
Trace a viral claim to its first source
Maps quotation chains instead of counting copied articles.
[CLAIM][AVAILABLE LINKS OR CLIPPINGS]