WHY ENGLISH?
Make every step from a research question to a bibliometric map visible, testable and proportionate to what the data can show
Bibliometric studies turn publication records into counts, trends, collaboration maps, citation networks and topic clusters. English matters because a beautiful map can conceal decisive choices about databases, search syntax, author names, thresholds, software and interpretation.
COMET-Bib stands for Comprehensive Methodological Transparency for Bibliometric Studies. EQUATOR registered the project on 22 January 2026 to cover the full bibliometric process, including research questions, data sources, citation and co-authorship methods, artificial intelligence, workflow visualisation and reproducibility.
As checked on 6 November 2026, COMET-Bib remained under development. Earlier initiatives already exist: GLOBAL addresses minimum reporting for bibliometric analyses, BIBLIO proposes minimum requirements for biomedical bibliometric reviews, and PRIBA proposed reporting items after examining highly cited health and medicine studies. COMET-Bib therefore should be read as a developing, cross-disciplinary workflow owner—not a claim that bibliometric guidance began in 2026.
Reproducibility and claims
Preserve versions, sensitivity tests and bounded conclusions.
Article contents
Open all chapters
- 1. Define a Bibliometric Question That Can Be Answered
- 2. Define the Unit of Analysis
- 3. Name Every Data Source and Coverage Boundary
- 4. Date and Version the Data Retrieval
- 5. Report the Search Strategy Completely
- 6. Explain Eligibility, Screening and Deduplication
- 7. Resolve Authors and Institutions Transparently
- 8. Describe Citation and Reference Matching
- 9. Define Nodes, Edges and Counting Rules
- 10. Justify Thresholds and Normalisation
- 11. Explain Clustering, Topics and Visual Layout
- 12. Report AI Use and Human Validation
- 13. Show the Complete Workflow and Provenance
- 14. Name Software, Code and Configuration
- 15. Test Coverage and Analytical Sensitivity
- 16. Conclude With a Bounded Bibliometric Claim
ORIENTATION
What this guide owns
Current fact check
The official COMET-Bib record says PRISMA cannot be applied unadapted to bibliometric work and describes a PRISMA-like workflow that makes analytical choices visible. Its scope explicitly includes data sources, search strategies, citation and co-authorship methods, AI use, workflow visualisation and reproducibility.
Canonical boundary
This article owns end-to-end reporting of bibliometric workflow under COMET-Bib. GLOBAL owns its own minimum-guidance project, BIBLIO owns biomedical bibliometric reviews, PRIBA remains a proposed set, and PRISMA-S or PRISMA-trAIce can help with specialised search or AI-assisted review reporting without replacing bibliometric methods.
Worked example
Every chapter returns to a fictional study mapping ten years of research on multilingual learning technologies across education, computer science and psychology. It is fictional, low-stakes and educational; it does not report a real school, clinical decision or guaranteed outcome.
Did You Know? For COMET-Bib, complete reporting makes appraisal possible but does not perform the appraisal, remove bias or guarantee transfer.
Warm up with eduKate’s guides to explaining scientific evidence and asking better questions in class.
CHAPTER 1 OF 16
1. Define a Bibliometric Question That Can Be Answered
Why this move matters
State the unit, field, population of documents, period, geography and whether the goal is productivity, impact, collaboration, knowledge structure or change.
Did You Know? Different questions require different records, relations and measures.
Rebuild the map before reading the colours
State the unit, field, population of documents, period, geography and whether the goal is productivity, impact, collaboration, knowledge structure or change. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: The study asks how multilingual-learning technology topics and collaborations changed from 2016 to 2025. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
A precise question can still be poorly served by available metadata. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Define a Bibliometric Question That Can Be Answered”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
CHAPTER 2 OF 16
2. Define the Unit of Analysis
Why this move matters
Say whether the unit is document, author, institution, country, journal, reference, keyword, topic or citation link and how composites are handled.
Did You Know? Changing the unit changes denominators and the meaning of prominence.
Rebuild the map before reading the colours
Say whether the unit is document, author, institution, country, journal, reference, keyword, topic or citation link and how composites are handled. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: One article is the document unit while authors and affiliations form separate linked entities. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
Database fields may not represent real people or organisations cleanly. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Define the Unit of Analysis”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
CHAPTER 3 OF 16
3. Name Every Data Source and Coverage Boundary
Why this move matters
Report database, collection, indexes, document types, fields, language coverage, access route and known omissions.
Did You Know? Bibliometric databases differ in journal, regional, language and citation coverage.
Rebuild the map before reading the colours
Report database, collection, indexes, document types, fields, language coverage, access route and known omissions. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: Education and computing databases are combined because neither alone covers the interdisciplinary field. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
Adding sources can add duplicates and incompatible metadata. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Name Every Data Source and Coverage Boundary”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
CHAPTER 4 OF 16
4. Date and Version the Data Retrieval
Why this move matters
Record search date, coverage end date, export date, interface or API, data release and later refreshes.
Did You Know? Citation counts and indexed records change over time.
Rebuild the map before reading the colours
Record search date, coverage end date, export date, interface or API, data release and later refreshes. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: The team freezes exports on 12 January 2026 and records each API version. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
A date cannot recreate a proprietary database snapshot by itself. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Date and Version the Data Retrieval”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
CHAPTER 5 OF 16
5. Report the Search Strategy Completely
Why this move matters
Provide concepts, controlled terms, free text, fields, operators, limits, syntax translations and full runnable strings.
Did You Know? Search wording defines which scholarly conversation becomes visible.
Rebuild the map before reading the colours
Provide concepts, controlled terms, free text, fields, operators, limits, syntax translations and full runnable strings. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: The appendix shows how multilingual-learning concepts were translated for each database. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
A reproducible search can still miss unfamiliar terminology. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Report the Search Strategy Completely”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
CHAPTER 6 OF 16
6. Explain Eligibility, Screening and Deduplication
Why this move matters
State inclusion rules, automation, reviewer roles, duplicate definitions, conflicts and record flow.
Did You Know? Bibliometric eligibility is not identical to systematic-review eligibility.
Rebuild the map before reading the colours
State inclusion rules, automation, reviewer roles, duplicate definitions, conflicts and record flow. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: Conference papers are retained but book reviews are excluded, and duplicate DOI/title matches are documented. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
Deduplication can merge distinct items or preserve duplicate versions. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Explain Eligibility, Screening and Deduplication”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
CHAPTER 7 OF 16
7. Resolve Authors and Institutions Transparently
Why this move matters
Describe name normalisation, identifiers, affiliations, mergers, transliteration, manual review and unresolved ambiguity.
Did You Know? Entity resolution can change rankings and collaboration networks.
Rebuild the map before reading the colours
Describe name normalisation, identifiers, affiliations, mergers, transliteration, manual review and unresolved ambiguity. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: ORCID, affiliation history and manual checks distinguish two authors with the same initials. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
No disambiguation method perfectly recovers identity. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Resolve Authors and Institutions Transparently”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
CHAPTER 8 OF 16
8. Describe Citation and Reference Matching
Why this move matters
Report how references are parsed, matched, normalised and treated when incomplete, duplicated or outside the database.
Did You Know? Citation networks inherit errors from reference matching.
Rebuild the map before reading the colours
Report how references are parsed, matched, normalised and treated when incomplete, duplicated or outside the database. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: DOIs are matched first, then title-year rules with an audited confidence threshold. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
Unmatched references may be systematically older or non-English. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Describe Citation and Reference Matching”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
CHAPTER 9 OF 16
9. Define Nodes, Edges and Counting Rules
Why this move matters
Explain network objects, relation direction, weights, full or fractional counting, self-links and time windows.
Did You Know? A network diagram has meaning only when nodes and edges are defined.
Rebuild the map before reading the colours
Explain network objects, relation direction, weights, full or fractional counting, self-links and time windows. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: A co-authorship edge receives fractional weight so large teams do not dominate automatically. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
One counting rule is not universally correct. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Define Nodes, Edges and Counting Rules”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
CHAPTER 10 OF 16
10. Justify Thresholds and Normalisation
Why this move matters
Report minimum counts, pruning, field and time normalisation, denominator choices and the effect of alternatives.
Did You Know? Thresholds turn continuous evidence into visible and invisible elements.
Rebuild the map before reading the colours
Report minimum counts, pruning, field and time normalisation, denominator choices and the effect of alternatives. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: The main map requires three documents, while a sensitivity map uses one and five. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
A reasonable threshold can still hide small emerging groups. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Justify Thresholds and Normalisation”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
CHAPTER 11 OF 16
11. Explain Clustering, Topics and Visual Layout
Why this move matters
Name algorithms, parameters, random seeds, text preprocessing, labels, validation and whether distance or colour has analytic meaning.
Did You Know? Visual proximity is not automatically intellectual similarity.
Rebuild the map before reading the colours
Name algorithms, parameters, random seeds, text preprocessing, labels, validation and whether distance or colour has analytic meaning. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: Cluster labels are checked against high-weight documents rather than generated from keywords alone. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
A stable-looking layout may change with a seed or parameter. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Explain Clustering, Topics and Visual Layout”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
CHAPTER 12 OF 16
12. Report AI Use and Human Validation
Why this move matters
Name model, version, prompts, inputs, outputs, privacy controls, sampling, error checks and human decisions for screening, labelling or interpretation.
Did You Know? Automation should be reported as a workflow component, not a magical shortcut.
Rebuild the map before reading the colours
Name model, version, prompts, inputs, outputs, privacy controls, sampling, error checks and human decisions for screening, labelling or interpretation. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: An AI model proposes topic labels, but two researchers inspect documents and revise unsupported labels. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
Human review can be inconsistent and AI services can change. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Report AI Use and Human Validation”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
CHAPTER 13 OF 16
13. Show the Complete Workflow and Provenance
Why this move matters
Connect acquisition, transformation, cleaning, matching, analysis and figures with files, decisions and identifiers.
Did You Know? Provenance lets readers locate where a surprising result entered the pipeline.
Rebuild the map before reading the colours
Connect acquisition, transformation, cleaning, matching, analysis and figures with files, decisions and identifiers. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: A workflow diagram links raw exports to cleaned tables, network files and each published figure. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
A diagram can omit undocumented manual edits. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Show the Complete Workflow and Provenance”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
CHAPTER 14 OF 16
14. Name Software, Code and Configuration
Why this move matters
Report packages, versions, scripts, parameters, operating environment and access to code or executable alternatives.
Did You Know? Software defaults are methodological choices when they shape outputs.
Rebuild the map before reading the colours
Report packages, versions, scripts, parameters, operating environment and access to code or executable alternatives. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: A locked environment and command list recreate each table from the archived cleaned data. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
Open code cannot supply licensed source records. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Name Software, Code and Configuration”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
CHAPTER 15 OF 16
15. Test Coverage and Analytical Sensitivity
Why this move matters
Repeat key results across sources, thresholds, counting rules, normalisation or disambiguation assumptions and explain material changes.
Did You Know? Robustness is stronger when conclusions survive defensible alternatives.
Rebuild the map before reading the colours
Repeat key results across sources, thresholds, counting rules, normalisation or disambiguation assumptions and explain material changes. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: The central topic remains across two databases, but institution rankings change after fractional counting. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
Sensitivity analysis cannot test every plausible universe. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Test Coverage and Analytical Sensitivity”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
CHAPTER 16 OF 16
16. Conclude With a Bounded Bibliometric Claim
Why this move matters
Tie conclusions to source coverage, time, search, cleaning, measures and sensitivity; avoid equating citations with quality or causation.
Did You Know? COMET-Bib addresses methodological transparency, not proof of research value.
Rebuild the map before reading the colours
Tie conclusions to source coverage, time, search, cleaning, measures and sensitivity; avoid equating citations with quality or causation. Trace the question, source, search, cleaning, entities, relations, measures, thresholds and visualisation in order.
Follow one record through the workflow
Worked example: The map describes indexed scholarly attention and collaboration, not educational effectiveness. Keep the raw record, deduplication, author matching, citation links, network position and plotted label connected.
Separate measurement from meaning
A database count, cluster or centrality score is a constructed measurement. Explain coverage, time, language and modelling choices before interpreting influence or fields.
Respect the development boundary
Transparent reporting cannot remove database bias. COMET-Bib is under development. This guide uses its documented scope and neighbouring official guidance without inventing final items.
Practical reader check
- Point to the exact source or method sentence supporting the claim.
- Name the person, setting, date, input and decision rule that could change its meaning.
- Label the remaining uncertainty reported, inferred, contested or unknown.
Student, parent and career transfer
Make one evidence card for “Conclude With a Bounded Bibliometric Claim”: source location, plain-language meaning, limit and next question. A parent or teacher can ask for one everyday explanation and one specialist explanation; both must preserve the same conditions. This habit supports library science, research evaluation, education, medicine, data science, science policy, publishing, information retrieval, open science and research communication without promising a grade, admission, funding decision or job.
PRACTICE LAB
Build a COMET-Bib decision trail
Student route
- Choose one original report inside this guide’s scope.
- Copy its objective and underline actor, action, context, evidence and intended decision.
- Draw the specific reasoning chain used in this article.
- Mark each step reported, inferred, contested or unknown.
- Rewrite the conclusion so it preserves the narrowest important limitation.
Parent or teacher route
Ask the learner to explain how English helps us make every step from a research question to a bibliometric map visible, testable and proportionate to what the data can show twice: once to a classmate and once to a method specialist. Praise a precise correction and a well-stated unknown; both show disciplined attention to evidence.
Early-career route
Create a one-page COMET-Bib audit with source link, date checked, scope, method logic, decisive wording, contrary evidence, limit and next action. A colleague should be able to continue without guessing.
Decision rule
If a shorter explanation removes the condition that made the original claim true, it is not clearer yet. Restore the condition.
Continue with eduKate’s guides to reading course and career information and giving useful feedback in group projects.
USEFUL NEXT READING
Connect this narrow owner to its neighbour
Read the PRISMA-trAIce AI-assisted review guide for a related but different reader job. Return here when the question is specifically how to make every step from a research question to a bibliometric map visible, testable and proportionate to what the data can show. This boundary keeps one canonical owner per intent.
For English practice across research, school and career communication, visit the English at work and in public life guide.
FREQUENTLY ASKED QUESTIONS
Questions students and parents often ask
What does COMET-Bib stand for?
Comprehensive Methodological Transparency for Bibliometric Studies.
Is COMET-Bib final?
No. EQUATOR lists it as under development.
Why not use PRISMA unchanged?
Bibliometric studies have distinct data, networks, cleaning and analytical workflows.
Do citations prove quality?
No. Citation counts reflect attention shaped by field, time, database and social processes.
Why report thresholds?
They decide which nodes and relations become visible.
Can AI label clusters automatically?
It can assist, but model details, prompts, errors and human validation should be reported.
PRIMARY SOURCES
Read the original COMET-Bib guidance
Facts were checked against original publications and official records on 6 November 2026. Follow the live source for later corrections, completed updates or companion material.
- COMET-Bib official record, EQUATOR Network
- GLOBAL official record, EQUATOR Network
- BIBLIO official record, EQUATOR Network
- PRIBA original paper
A cheerful final note
Precise English helps us make every step from a research question to a bibliometric map visible, testable and proportionate to what the data can show more generous, not less. People can see how their evidence and values entered the work, where disagreement remains and what a responsible next step looks like.