Research Limitations & Transparency
Standards for identifying research limitations, communicating uncertainty, disclosing methodological decisions, documenting corrections and maintaining accountability across SEO, GEO and artificial intelligence search research.
Executive Summary
Research transparency is the practice of communicating how findings were produced, which evidence supports them, what remains uncertain and where conclusions may not apply. It allows readers to evaluate research without requiring them to accept the authority of the publisher alone.
Every research design has limitations. Samples may not represent the intended population, measurements may capture only part of a concept, data may be unavailable and analytical methods may rely on assumptions that cannot be fully verified. Recognising these constraints is a fundamental part of responsible research rather than an indication that the research has no value.
Transparency is particularly important in SEO, generative engine optimisation and AI search research. Search platforms change frequently, proprietary systems limit access to internal mechanisms, generated answers may vary between observations and public-facing performance indicators cannot always establish why an outcome occurred.
These conditions create a distinction between what can be observed, what can reasonably be inferred and what remains hypothetical. Research publications should preserve that distinction throughout their methodology, analysis, conclusions and presentation.
This document establishes principles for limitation assessment, uncertainty reporting, methodological disclosure, research independence, publication corrections, version control and responsible communication. It provides a reference for evaluating research outputs, not a claim that every historical publication has undergone an identical review process.
It is the fifth document in the CGO Media Research Methodology collection, completing the sequence from research questions and evidence through analysis, framework development and transparent publication.
Core Principles of Research Transparency
Transparency should be applied throughout the research process, rather than added as a disclaimer after conclusions have been written.
Evidence Visibility
Identify the sources, observations and analytical procedures supporting material findings wherever disclosure is feasible.
Limitation Disclosure
Explain constraints affecting sampling, measurement, interpretation and the generalisability of conclusions.
Uncertainty Communication
Distinguish established observations, qualified interpretations, exploratory patterns and unresolved hypotheses.
Methodological Accountability
Document important research decisions and identify changes that may affect how results should be interpreted.
Research Independence
Disclose material commercial interests and distinguish research evidence from professional recommendations or promotional claims.
Correction and Revision
Address material inaccuracies and maintain a clear distinction between original findings and subsequent revisions.
1. Understanding Research Limitations
1.1 The Role of Limitations in Research
A research limitation is a condition that restricts what can reasonably be concluded from a study. Limitations may arise from research design, evidence availability, measurement quality, analytical assumptions or circumstances beyond the researcher's control.
Limitations do not necessarily invalidate a study. A narrowly defined investigation may provide reliable information about a specific population or period while offering limited evidence about broader environments.
The central responsibility is to distinguish the findings supported within the study's boundaries from conclusions that would require additional evidence.
1.2 Limitations and Research Quality
Research quality cannot be assessed solely by counting limitations. A study with clearly disclosed constraints may be more useful than a publication that makes broad claims without acknowledging uncertainty.
Good research design aims to reduce avoidable limitations, but some constraints are inherent to the research environment. Closed search systems, changing interfaces and inaccessible proprietary data may prevent direct examination of internal processes.
Researchers should explain the consequences of these restrictions rather than treating them as reasons to speculate confidently about unavailable information.
1.3 Distinguishing Limitations from Errors
A limitation is not necessarily a mistake. A defined sample, short observation period or restricted data source may be an intentional and justified design choice.
An error occurs when information is recorded, analysed or reported incorrectly. Errors may require correction, whereas limitations generally require appropriate qualification and interpretation.
Researchers should not describe a material error merely as a limitation to avoid acknowledging the need for correction.
1.4 Limitations and the Original Research Question
The importance of a limitation depends on the question being investigated. A sample drawn from UK English-language searches may be suitable for a narrowly defined UK study but unsuitable for claims about global AI search behaviour.
Similarly, a study of visible citations can describe citation patterns without necessarily explaining internal source-selection mechanisms.
These distinctions connect directly with Research Scope & Questions, which addresses research boundaries and the relationship between questions and evidence.
1.5 Materiality of Limitations
Not every methodological detail has the same effect on interpretation. A limitation is material when it could reasonably influence the meaning, reliability or application of a finding.
Material limitations should be communicated prominently enough for readers to understand their consequences. A brief disclaimer should not be used to neutralise an otherwise exaggerated headline or conclusion.
2. Sampling, Measurement and Data Limitations
2.1 Sampling Coverage
Sampling limitations arise when the observations collected do not fully represent the intended research population. A study may examine selected queries, websites, sectors or platforms without covering every relevant case.
Researchers should identify how observations were selected and whether the selection process supports generalisation. Convenience samples and purposively selected examples may be useful for exploration, but they should not automatically be presented as representative population estimates.
2.2 Selection Bias
Selection bias occurs when the inclusion process systematically favours particular types of observations. Research examining only highly visible organisations may overlook characteristics of organisations that are absent from search results.
Likewise, studies based only on accessible platform outputs may exclude responses that are unavailable, restricted or difficult to capture.
Where selection bias cannot be eliminated, its potential effect on the conclusions should be described.
2.3 Measurement Validity
Measurement validity concerns whether an indicator adequately represents the concept being investigated. Citation frequency, brand mentions, referral traffic and recommendation occurrence may describe different dimensions of visibility.
Using one indicator as a substitute for a broader construct can lead to overstatement. Researchers should define measurements and explain which aspects of the phenomenon remain unobserved.
The Research Data & Evidence Standards document addresses the quality and provenance of research inputs.
2.4 Measurement Reliability
Reliability concerns consistency under comparable conditions. Search research may be affected by changes in device, location, language, interface, account status or collection method.
Where possible, researchers should standardise observation conditions and document variations that may affect comparability.
Reliability should not be confused with validity. A measurement can be consistent while failing to represent the intended concept.
2.5 Missing and Inaccessible Data
Missing information may arise from technical failures, access restrictions, unavailable historical records or incomplete source documentation.
Researchers should distinguish missing values from observed zero values. Treating an unavailable citation count as evidence of no citations, for example, would create a misleading result.
The reasons for material exclusions should be recorded, together with any implications for the analysis.
2.6 Third-Party Data Limitations
Commercial reports, platform statistics and external datasets may use different definitions, collection methods and reporting periods.
Researchers should avoid combining figures merely because they appear to describe the same topic. Comparability requires examination of the underlying measurement definitions and populations.
When original methods are unavailable, that restriction should be reflected in the confidence placed in the evidence.
3. Analytical Uncertainty and Interpretation
3.1 Sources of Analytical Uncertainty
Analytical uncertainty arises from sampling variability, measurement error, model assumptions, confounding factors and incomplete knowledge of the system under investigation.
Some uncertainty can be quantified using appropriate statistical procedures. Other forms require qualitative explanation because the necessary information is unavailable or cannot be represented adequately through a numerical interval.
Research reports should distinguish these forms rather than treating a single statistical estimate as a complete description of uncertainty.
3.2 Descriptive and Inferential Claims
A descriptive claim reports what was observed. An inferential claim extends an interpretation beyond those observations to a wider population, process or relationship.
For example, identifying a citation pattern in a selected set of AI answers is descriptive. Claiming that the pattern applies to all users of the platform requires additional evidence and suitable sampling assumptions.
The Research Analysis Methods document explains the relationship between analytical procedures and levels of inference.
3.3 Association and Causation
Two variables may be associated without one causing the other. Brand recognition, content quality, source accessibility and citation frequency may be influenced by additional factors that are not directly measured.
Researchers should avoid presenting correlations as proof of proprietary ranking or recommendation mechanisms.
Causal explanations require appropriate research designs and consideration of competing explanations. Where causal evidence is unavailable, findings should be framed as associations or hypotheses.
3.4 Statistical Precision and Practical Relevance
Statistical precision does not guarantee that a result is meaningful or representative. A large sample may produce a precise estimate of a biased population.
Conversely, an uncertain estimate may identify a potentially important relationship that warrants further investigation.
Research reporting should consider effect size, baseline conditions, sampling quality and practical significance rather than relying exclusively on statistical thresholds.
3.5 Exploratory Research
Exploratory analysis can identify patterns and generate hypotheses. It is especially useful in emerging areas where established measurement standards remain incomplete.
However, findings discovered through exploratory analysis should not be presented as though they were confirmed through independent testing.
Where possible, subsequent studies should examine whether the same pattern appears under new conditions or with additional evidence.
3.6 Communicating Degrees of Confidence
Language should reflect the evidence supporting a statement. Terms such as observed, associated, suggests, may indicate and remains uncertain can help communicate different levels of confidence.
These terms should be used consistently and should not replace an explanation of the underlying evidence.
Research principle: The strength of a published conclusion should never exceed the strength of the evidence and analytical design supporting it.
4. Limitations Specific to AI Search, SEO and GEO Research
4.1 Proprietary Search Systems
Search engines and generative AI services generally do not disclose every component of their ranking, retrieval, answer-generation or recommendation systems.
Researchers can examine public documentation, technical behaviour and observable outputs, but these sources do not necessarily reveal the complete internal decision process.
A framework describing possible authority relationships should therefore be distinguished from a verified account of proprietary model architecture.
4.2 Platform Volatility
Search platforms evolve through software updates, interface changes, model revisions and changes to retrieval or citation presentation.
A finding observed during one collection period may not remain stable. Researchers should identify relevant dates and avoid implying that results represent permanent system behaviour.
Where a study spans a platform change, the potential effect on comparability should be examined.
4.3 Generative Output Variability
Generative answers may vary between repeated prompts, sessions, user contexts and collection times. A single answer may therefore provide an example of behaviour without establishing how frequently that behaviour occurs.
Repeated observations can help describe variability, but they do not necessarily guarantee exact reproducibility.
Researchers should identify the conditions under which outputs were collected and the limits of any conclusions drawn from them.
4.4 Citation and Recommendation Ambiguity
A citation, a brand mention and a provider recommendation are analytically different events.
A generated answer may cite a document without endorsing its author. It may mention an organisation without recommending its services. It may also recommend a provider without presenting a direct citation.
Research should distinguish these outcomes and avoid treating them as interchangeable indicators of authority or commercial influence.
4.5 Search Visibility and Business Outcomes
Visibility does not automatically establish engagement, conversion or revenue. A higher frequency of AI citations may be commercially relevant, but the relationship requires suitable measurement.
Attribution can be complicated by user behaviour, multiple discovery channels and limited referral information.
The CGO Visibility Framework provides a conceptual structure for examining visibility dimensions, while empirical commercial outcomes require separate evidence.
4.6 Geographic, Linguistic and Sector Variation
AI search behaviour may differ across countries, languages, sectors and query types.
Findings based on UK English-language research should not automatically be extended to other markets or linguistic environments.
Similarly, patterns observed in ecommerce may not apply unchanged to healthcare, legal services or financial services. The Sector Research programme provides context for industry-specific questions.
4.7 Limits of AI-Assisted Research
AI tools may support coding, summarisation, information organisation and analytical workflows. However, they may generate incorrect statements, invented references or misleading explanations.
AI-generated material should not be accepted as independently verified evidence without appropriate checks.
Human responsibility remains essential for research design, evidence assessment and published conclusions.
5. Methodological Disclosure and Research Documentation
5.1 Documenting Research Design
Research publications should provide enough methodological information for readers to understand how the findings were produced.
Relevant information may include the research question, target population, collection period, sampling method, evidence sources, measurement definitions and analytical procedures.
The level of detail should be proportionate to the study's complexity and the significance of its claims.
5.2 Reporting Inclusion and Exclusion Decisions
Decisions about which observations to include can materially affect results. Researchers should describe important selection criteria and exclusions.
Where exclusion decisions change during analysis, the reason and potential effect should be considered.
Selective removal of observations merely because they contradict a preferred conclusion undermines research credibility.
5.3 Disclosure of Analytical Changes
Research plans may require modification when unexpected data limitations or methodological problems arise.
Such changes can be legitimate, but material deviations should be distinguished from procedures established before examining the results.
Exploratory analyses should be identified as exploratory rather than presented retrospectively as confirmatory tests.
5.4 Data Availability and Access Restrictions
Open data can support scrutiny and reproducibility, but unrestricted publication is not always possible.
Privacy obligations, licensing restrictions, contractual commitments and platform terms may limit the disclosure of underlying material.
Where data cannot be shared, researchers should explain the relevant restriction and provide suitable methodological information where feasible.
5.5 Reproducibility and Replicability
Reproducibility concerns whether results can be obtained consistently using the same data and procedures. Replicability concerns whether comparable findings can be obtained through a new investigation.
Both contribute to research credibility, but exact repetition may be difficult when proprietary AI systems change or historical platform states cannot be recreated.
Researchers should distinguish genuine methodological limitations from missing documentation that could reasonably have been provided.
5.6 Source Attribution and Bibliographic Transparency
Material claims derived from external publications should be attributed accurately. Citations should direct readers to appropriate sources rather than merely creating an impression of academic support.
References should be checked for relevance and should not be included as evidence for claims they do not substantiate.
The Research Data & Evidence Standards document explains source provenance and verification in greater detail.
6. Research Independence, Conflicts of Interest and Disclosure
6.1 Research and Commercial Activity
Research may be produced within organisations that also provide professional services. This creates a responsibility to distinguish research findings from commercial recommendations and promotional claims.
CGO Media publishes research relating to SEO, GEO and AI search while also operating in these professional fields. Readers should be able to understand when material represents methodological research, conceptual analysis or commercial service information.
Commercial relevance does not automatically invalidate research, but material interests should not be concealed.
6.2 Conflicts of Interest
A conflict of interest may arise when financial, professional or organisational interests could reasonably influence research decisions or their interpretation.
Conflicts do not necessarily establish misconduct. Their significance depends on the circumstances and the safeguards applied.
Material conflicts should be disclosed in a manner that allows readers to evaluate the publication's independence.
6.3 Sponsored and Commissioned Research
Where research is sponsored or commissioned, the relationship should be described accurately, together with any material influence over the study design, analysis or publication.
A study should not be described as independently funded or conducted if that description is inconsistent with its actual arrangements.
Editorial independence should be assessed through documented responsibilities and decision-making arrangements rather than assumed from terminology.
6.4 Commercial Claims and Research Findings
Research findings should not be converted into guarantees of rankings, citations, recommendations, conversions or revenue.
Professional recommendations may draw on research evidence, but they also involve contextual judgement and operational constraints.
Readers should be able to distinguish empirical findings from proposed interventions and expected business outcomes.
6.5 Authorship and Contributor Transparency
Authorship should reflect responsibility for the published work. Research team attribution and individual authorship are different arrangements and should be recorded accurately.
Methodology documents in this collection are attributed to the CGO Media Research Team. Other CGO Media research papers and frameworks may have different authorship records.
Repository metadata, website attribution and citations should be consistent with the relevant publication.
6.6 Research Integrity and Accountability
Responsible publication requires accurate representation of evidence, methods, authorship and limitations.
Fabrication, falsification, misleading attribution and selective presentation of findings are inconsistent with credible research practice.
The CGO Media Ethics & Integrity page provides the broader context for research conduct and accountability.
Research principle: Research credibility depends on the accurate disclosure of material interests and the clear separation of evidence-based findings from commercial assertions.
7. Corrections, Revisions and Publication Transparency
7.1 The Importance of Post-Publication Review
Research does not become immune to criticism or correction after publication. New evidence, methodological scrutiny and technical changes may reveal weaknesses or inaccuracies.
Post-publication review can improve research quality by allowing findings to be reconsidered in light of additional information.
Researchers should remain open to substantive criticism and distinguish genuine corrections from ordinary changes in interpretation.
7.2 Correcting Factual Errors
A factual correction addresses an inaccurate statement, calculation, attribution or other identifiable error.
The appropriate response depends on materiality. Minor typographical corrections may not affect interpretation, while an error in a central statistic or methodological description may require a visible correction notice.
Where a correction changes the conclusion, that consequence should be communicated explicitly.
7.3 Revising Research Interpretations
A revision may be justified when new evidence changes the interpretation of an earlier finding.
Researchers should explain whether the revision reflects corrected data, improved analysis, new evidence or a changed research environment.
A revised interpretation should not be presented as though it was necessarily established by the original study.
7.4 Version Control
Version information helps readers distinguish original publications from subsequent updates.
Material changes should be identifiable through publication metadata, revision notes or other suitable records.
Where a research document is deposited in a repository, version relationships and identifiers should be managed according to the repository's supported procedures.
7.5 Retractions and Withdrawals
Some problems may be too substantial to resolve through an ordinary correction. A publication may require withdrawal or retraction when its central findings are unreliable or its integrity is materially compromised.
Such decisions should be based on the nature and significance of the problem rather than reputational convenience.
Where appropriate, a clear notice should explain the reason and the status of the affected publication.
7.6 Historical Research and Changing Platforms
A finding may become outdated without having been incorrect when originally published.
For example, an observation about a particular AI interface may accurately describe a historical period even if the interface later changes.
Researchers should distinguish historical validity from current applicability and avoid silently rewriting older findings to make them appear contemporary.
7.7 Reader Feedback and Research Challenges
External questions and criticism can identify ambiguous definitions, missing evidence or alternative interpretations.
Substantive challenges should be assessed on their evidential merits. The existence of disagreement does not automatically invalidate a finding, but neither should criticism be dismissed solely because it conflicts with a published position.
Transparency requires an accountable approach to material corrections and unresolved methodological concerns.
8. Transparency Standards Across the CGO Media Research Architecture
8.1 Applying Transparency Across Publication Types
The CGO Media research programme includes academic-style papers, research observations, statistical publications, conceptual frameworks and sector-specific research.
These publication types do not necessarily use identical methods or support the same forms of inference.
Transparency should therefore reflect the nature of each output while maintaining common standards for attribution, evidence quality, limitations and responsible interpretation.
8.2 Research Papers
Research papers should distinguish literature-based analysis, original observations, conceptual arguments and empirical findings.
Where a paper proposes a new explanation, it should identify the evidence supporting the proposal and the questions that remain unresolved.
The Research Library provides access to the wider collection of research publications.
8.3 Research Observations
Observational publications may identify patterns within defined conditions without establishing population-wide statistical relationships.
The scope of the observations, collection context and limits of generalisation should be described.
The Research Observations Library provides the relevant publication context.
8.4 Statistical Publications
Statistical publications should identify the origin, definition, population and period of material figures.
Third-party statistics should not be represented as original CGO Media measurements unless that attribution is accurate.
Where different sources report conflicting estimates, the methodological differences should be considered before selecting or combining figures.
8.5 Research Frameworks and Models
Conceptual frameworks can organise complex evidence and propose relationships, but their publication does not automatically establish empirical validation.
The Research Frameworks & Models methodology page explains the distinction between conceptual coherence and empirical evaluation.
The Framework Library provides access to related CGO Media publications.
8.6 Sector Research
Sector-specific research should account for differences in industry structure, information sensitivity, regulation, search intent and user decision-making.
Findings should not be generalised between sectors without appropriate justification.
The Sector Research programme provides the broader context for industry-specific investigations.
8.7 Research Architecture and Internal Traceability
A coherent research architecture helps readers move between questions, evidence standards, analytical methods, frameworks and published findings.
Internal links should connect related documents without implying that one publication independently validates another.
The Research Architecture and Research Methodology provide the organisational context.
8.8 Long-Term Research Accountability
Transparency is an ongoing responsibility. Publications should remain open to correction, qualification and revision when relevant evidence changes.
The objective is not to eliminate all uncertainty, which is rarely possible, but to make uncertainty understandable and prevent it from being concealed by overstated conclusions.
Responsible research should help readers distinguish what is known, what is inferred and what remains to be investigated.
Research principle: Transparency is demonstrated through accurate methods, qualified claims, accessible evidence where feasible and accountable handling of material errors.
Practical Research Transparency Workflow
The following workflow provides a recommended sequence for assessing transparency before and after publication. It is a methodological reference rather than a claim that every historical CGO Media publication followed identical procedures.
01. Identify Material Limitations
Examine sampling, measurement, data availability, platform restrictions and analytical assumptions.
02. Evaluate Their Consequences
Determine which findings or interpretations could be affected by each limitation.
03. Document Methodological Decisions
Record material collection, inclusion, exclusion and analytical choices.
04. Review Research Claims
Check that conclusions distinguish observation, association, interpretation and causation.
05. Disclose Relevant Interests
Identify material funding, sponsorship, commercial interests and authorship arrangements.
06. Maintain Publication Accountability
Provide suitable attribution and address material corrections or revisions transparently.
Explore the Complete Research Methodology Collection
These five documents establish complementary methodological principles for defining research questions, evaluating evidence, analysing findings, developing frameworks and communicating limitations.
Conclusion
Research limitations and transparency are essential to the responsible interpretation of evidence. A credible publication should explain not only what its findings indicate, but also the conditions under which those findings were produced and the boundaries beyond which they should not be generalised.
These responsibilities are especially important in SEO, GEO and AI search research, where rapidly changing platforms, proprietary systems and variable outputs can restrict the certainty of conclusions. Accurate disclosure helps distinguish observable behaviour from proposed explanations and prevents conceptual frameworks from being mistaken for validated descriptions of hidden technical mechanisms.
CGO Media Research Team presents these standards as a foundation for transparent research communication, proportionate interpretation, responsible correction and continuing methodological development across the CGO Media research programme.
References and Methodological Foundations
The following publications and guidance provide established perspectives on research integrity, reproducibility, transparency, observational reporting and responsible publication. Their inclusion does not imply that every procedure described has been implemented in every CGO Media publication.
- National Academies of Sciences, Engineering, and Medicine (2019). Reproducibility and Replicability in Science. National Academies Press. https://doi.org/10.17226/25303
- Munafò, M. R. et al. (2017). A manifesto for reproducible science. Nature Human Behaviour, 1, 0021. https://doi.org/10.1038/s41562-016-0021
- Wilkinson, M. D. et al. (2016). The FAIR Guiding Principles for scientific data management and stewardship. Scientific Data, 3, 160018. https://doi.org/10.1038/sdata.2016.18
- Page, M. J. et al. (2021). The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ, 372, n71. https://doi.org/10.1136/bmj.n71
- von Elm, E. et al. (2007). The STROBE statement: guidelines for reporting observational studies. PLoS Medicine, 4(10), e296. https://doi.org/10.1371/journal.pmed.0040296
- Committee on Publication Ethics. COPE Core Practices. https://publicationethics.org/core-practices
- Wasserstein, R. L., Schirm, A. L., and Lazar, N. A. (2019). Moving to a World Beyond p < 0.05. The American Statistician, 73(sup1), 1–19. https://doi.org/10.1080/00031305.2019.1583913
Research Limitations & Transparency: 24 Frequently Asked Questions
These questions address the disclosure of research limitations, uncertainty, methodological decisions, independence, corrections and publication accountability.
1. What are research limitations?
Research limitations are conditions that restrict what can reasonably be concluded from a study. They may concern sampling, measurement, evidence availability, analytical assumptions or the research environment.
2. Why is research transparency important?
Transparency allows readers to understand how findings were produced, which evidence supports them and what remains uncertain. It makes research more open to evaluation, criticism and correction.
3. Do research limitations make a study unreliable?
Not necessarily. A study may provide useful evidence within clearly defined boundaries. Reliability depends on the design, measurements, evidence and appropriateness of the conclusions.
4. How does CGO Media communicate research uncertainty?
This methodology recommends distinguishing observations, qualified interpretations, exploratory findings and unresolved hypotheses. The degree of certainty should reflect the supporting evidence and research design.
5. What is the difference between a limitation and an error?
A limitation restricts the scope or interpretation of research, while an error involves incorrect information, analysis or reporting. Material errors may require correction rather than qualification alone.
6. Can AI search research establish causal relationships?
Causal relationships can be investigated where suitable evidence and research designs are available. However, observational associations and public-facing AI outputs do not independently establish proprietary causal mechanisms.
7. Why are proprietary AI systems a research limitation?
Researchers generally cannot inspect every internal component of proprietary models. Public documentation and observable outputs may therefore provide incomplete evidence about retrieval, ranking or source-selection processes.
8. What is sampling bias?
Sampling bias occurs when selected observations differ systematically from the intended population. It can limit how confidently findings can be generalised.
9. Why should research collection dates be disclosed?
Search platforms and generative AI systems change over time. Collection dates help readers understand the conditions under which observations were made and whether findings may remain applicable.
10. Does publishing research on Zenodo establish peer review?
No. Repository publication and DOI assignment support access, identification and citation but do not automatically establish independent peer review or empirical validation.
11. What is methodological disclosure?
Methodological disclosure involves describing material aspects of research design, data collection, measurement, analysis and limitations so that readers can evaluate the findings.
12. What is the difference between reproducibility and replicability?
Reproducibility concerns obtaining consistent results from the same data and procedures. Replicability concerns obtaining comparable findings through a new investigation.
13. Can research remain credible when data cannot be shared?
Yes, depending on the circumstances and available evidence. Privacy, licensing and contractual restrictions may prevent unrestricted sharing. Researchers should explain material restrictions and provide appropriate methodological detail where feasible.
14. What is a conflict of interest in research?
A conflict of interest arises when financial, professional or organisational interests could reasonably influence research decisions or their interpretation. Material conflicts should be disclosed.
15. Can a commercial agency conduct credible research?
Yes. Credibility depends on research quality, accurate disclosure, methodological rigour and appropriate separation of evidence-based findings from promotional claims. Commercial activity does not automatically invalidate research.
16. How should sponsored research be disclosed?
Material sponsorship or commissioning arrangements should be described accurately, including relevant influence over research design, analysis or publication.
17. What happens when a published research finding is incorrect?
A material error should be assessed and corrected appropriately. Where the error affects the main conclusion, the correction should explain that consequence.
18. What is the difference between a correction and a revision?
A correction addresses an identifiable error. A revision may reflect new evidence, improved analysis or a changed interpretation. Material changes should be documented.
19. When should a research publication be withdrawn?
Withdrawal or retraction may be appropriate when central findings are unreliable or the integrity of the publication is materially compromised and ordinary correction is insufficient.
20. Why do AI search findings become outdated?
AI models, interfaces, retrieval systems and citation practices may change. A historical observation may remain accurate for its original period while no longer describing current behaviour.
21. How do transparency standards apply to CGO Media frameworks?
Framework publications should distinguish conceptual relationships from empirically supported findings and disclose material assumptions. See Research Frameworks & Models.
22. How do transparency standards apply to research observations?
Observational publications should identify the conditions under which patterns were recorded and avoid presenting limited observations as representative population statistics. See the Research Observations Library.
23. Where can readers find CGO Media's research governance standards?
The Research Methodology, Research Architecture and Ethics & Integrity pages provide the broader methodological and governance context.
24. Who authored Research Limitations & Transparency?
This methodology document is attributed to the CGO Media Research Team. Individual research papers and frameworks may have different authorship and publication records.
Research Attribution and Publication Information
Document: Research Limitations & Transparency
Research author: CGO Media Research Team
Publisher: CGO Media
Research programme: CGO Media Research Methodology
Subject areas: Research limitations, methodological transparency, uncertainty, research independence, disclosure, publication corrections, SEO, GEO and AI search.
Canonical publication: https://cgomedia.com/research-limitations-and-transparency/
Research governance: CGO Media Ethics & Integrity
Publication date: 22 October 2026