Insight Communication and Research Leadership

Introduction

Market research creates value only when evidence changes a decision. A statistically significant difference, an eloquent interview quotation, or a sophisticated dashboard is not itself an insight. An insight is a defensible interpretation of evidence that explains a relevant market or consumer phenomenon and indicates what an organisation should do next.

For brand managers, the central challenge is translation: translating research questions into evidence, evidence into insight, insight into strategic choices, and strategic choices into accountable action. This chapter examines that translation process. It focuses on visualization, storytelling, dashboards, executive reporting, agency evaluation, budgeting, stakeholder management, uncertainty, and the development of an evidence-led insight function.

The chapter treats communication as an analytical activity rather than a cosmetic final step. Poor communication can exaggerate weak evidence, conceal limitations, or encourage decisions that the research cannot support. Strong communication makes the reasoning visible, distinguishes fact from interpretation, and enables decision-makers to act without confusing confidence with certainty.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Distinguish data, findings, insights, implications, and recommendations.
2. Select visualizations appropriate to different qualitative and quantitative evidence.
3. Construct persuasive but methodologically defensible research stories.
4. Design dashboards and executive reports for different decision contexts.
5. Evaluate research agencies, proposals, budgets, and deliverables.
6. Manage stakeholders, disagreement, uncertainty, and competing interpretations.
7. Establish governance principles for ethical and evidence-led insight functions.
8. Assess the role and limitations of artificial intelligence in research communication.

1. From Evidence to Action

A useful chain is:

Research question → evidence → finding → insight → implication → recommendation → decision → learning

These terms should not be treated as synonyms.

• Evidence is observed material: survey responses, behavioural data, interview transcripts, experiments, social listening records, or market statistics.
• A finding is a directly supported pattern in the evidence, such as a difference between segments or a recurring theme.
• An insight explains why the finding matters, often by connecting behaviour with motivation, context, tension, or consequence.
• An implication states what the finding means for the brand, customer experience, portfolio, or market strategy.
• A recommendation proposes an action, owner, timing, and basis for evaluation.

For example, a finding might be that repeat purchase is lower among young urban consumers. An insight might be that the brand is perceived as suitable for occasional indulgence but not sufficiently relevant to everyday routines. An implication is that increasing awareness alone is unlikely to solve the problem. A recommendation could be to redesign usage occasions and communications, then test whether consideration and repeat purchase improve.

A defensible recommendation satisfies five tests:

1. Relevance: it addresses a real strategic decision.
2. Traceability: its reasoning can be followed back to evidence.
3. Proportionality: the strength of the recommendation matches the strength of the evidence.
4. Feasibility: resources, capabilities, timing, and constraints are recognised.
5. Evaluability: success can be monitored using agreed indicators.

A recommendation should also identify what would change the decision. This is especially important when evidence is uncertain or contradictory.

2. Building a Defensible Insight

2.1 The evidence ladder

A practical evidence ladder moves from description to action:

The ladder is not a claim that every study must reach prediction or prescription. A small exploratory sample may support hypothesis generation but not population estimates. Conversely, a large survey may estimate prevalence but offer limited explanation. Researchers must communicate the boundary of inference.

2.2 Triangulation

Triangulation compares different sources, methods, analysts, or theories. It does not mean that agreement automatically proves truth. Different data sources may share the same bias. For example, survey respondents may report purchase intentions while behavioural data reveal actual purchases; the difference is informative rather than merely problematic.

Useful forms include:

• Method triangulation: qualitative interviews combined with surveys or experiments.
• Data-source triangulation: customer data, market data, and observational evidence.
• Analyst triangulation: independent coding or interpretation by more than one researcher.
• Temporal triangulation: comparison across periods, before and after an intervention, or across market cycles.

An insight is stronger when multiple sources converge or when disagreement is explained through differences in context, measurement, or population.

2.3 Claim–evidence–caveat–action structure

A concise insight statement can use four parts:

1. Claim: the central interpretation.
2. Evidence: the most relevant supporting observations.
3. Caveat: the principal limitation or uncertainty.
4. Action: the decision implication.

Example:

The brand is well known but weakly associated with convenient everyday use. This interpretation is supported by low association with convenience in the survey, repeated comments about preparation effort in interviews, and lower weekday usage in purchase data. Because the survey is cross-sectional, causality cannot be established. The immediate priority is therefore to test convenience-led propositions rather than assume that a packaging change will increase repeat purchase.

This structure prevents a common error: presenting an interpretation as if it were a direct observation.

3. Visualization as Analytical Reasoning

Visualization is not simply decoration. Choices about scale, ordering, aggregation, colour, and annotation influence what the audience notices and concludes.

3.1 Matching visual form to question

A table is often superior to a chart when exact values, small differences, or many categories matter. A chart is superior when the audience must recognise a pattern quickly.

3.2 Quantitative integrity

Every quantitative visual should make the population, measure, time period, sample size, and denominator clear. Percentages without denominators are especially dangerous. If 60% of ten respondents express a preference, the result should not be visually presented as equivalent in reliability to 60% of 2,000 respondents.

When comparing proportions, uncertainty may be represented with confidence intervals. For an estimated proportion p̂ from a simple random sample of size n, a commonly used approximate standard error is:

Here, p̂ is the sample proportion expressed between 0 and 1, n is the sample size, and SE is the standard error of the estimate. The approximation assumes independent observations and is less reliable when n is small or p̂ is very close to 0 or 1. A confidence interval is constructed using a suitable interval method; the familiar normal approximation is not always appropriate, particularly for small samples.

For a difference between two independent proportions, p̂₁ − p̂₂, the uncertainty depends on both sample sizes and proportions. The groups must be defined clearly, and weighting, clustering, or repeated measures may require more advanced variance estimation. A visual should not display a difference as meaningful merely because one bar is taller.

Axis design also matters. Truncating a bar-chart axis can exaggerate differences because bar length no longer begins at a common zero. A line chart may use a non-zero axis when the purpose is to show variation, but the scale must be explicit and should not mislead.

3.3 Qualitative visualization

Qualitative analysis can be visualized through:

• thematic maps showing relationships among themes;
• customer journey maps showing stages, needs, emotions, and friction;
• laddering diagrams connecting attributes to functional and symbolic consequences;
• evidence matrices linking themes to respondent types and verbatim excerpts;
• tension maps showing competing motivations.

A quotation is evidence of a participant’s expression, not proof of population prevalence. Labels such as “a recurring theme” or “illustrative quotation” are more accurate than “what consumers think” unless supported by appropriate prevalence evidence.

4. Research Storytelling

Research storytelling is the disciplined arrangement of evidence around a decision. It is not an invitation to dramatize results or remove inconvenient evidence.

4.1 The executive narrative

A strong executive narrative usually answers six questions:

1. What decision is required?
2. What is the most important thing we learned?
3. What evidence supports it?
4. Why does it matter commercially or strategically?
5. What should be done now?
6. What should be monitored or tested next?

A useful report may therefore begin with the decision and conclusion rather than with a long description of methods. Methodological detail remains essential, but it can be placed in a transparent technical section or appendix.

4.2 Narrative structures

Different situations call for different structures:

• Answer-first: conclusion, evidence, implication, action. Useful for senior executives.
• Problem–tension–resolution: current situation, unresolved conflict, proposed response. Useful for brand strategy.
• Journey narrative: consumer context, experience, friction, opportunity. Useful for customer experience.
• Hypothesis–test–learning: proposition, test, result, decision. Useful for innovation and experimentation.
• Scenario narrative: plausible futures, triggers, consequences, strategic options. Useful under uncertainty.

The “pyramid principle” is particularly useful: state the governing answer first, then group supporting arguments, then provide evidence and detail. Each assertion should be supported by evidence at an appropriate level of specificity.

4.3 Avoiding narrative bias

Stories can create false coherence. Researchers should actively test for:

• disconfirming evidence;
• minority or marginalized perspectives;
• alternative explanations;
• changes over time;
• differences between stated attitudes and observed behaviour;
• sampling and measurement limitations.

A balanced report may include a “what would make us wrong?” box. This strengthens rather than weakens credibility because it identifies the conditions under which the recommendation should be revisited.

5. Dashboards and Executive Reports

A dashboard is an information system for recurring monitoring; an executive report is usually a structured interpretation of a specific question or decision. Confusing the two produces dashboards overloaded with measures and reports that cannot be updated.

5.1 Dashboard design

A useful dashboard begins with decisions, not available data. Each metric should have:

• a precise definition;
• a named owner;
• a source and refresh schedule;
• a target or benchmark where appropriate;
• an interpretation rule;
• a documented limitation;
• a specified escalation or action.

A metric catalogue prevents “metric drift,” in which the meaning of a measure changes over time. It should record numerator, denominator, population, time period, filters, weighting, and treatment of missing data.

Dashboard layers can include:

1. Strategic layer: brand health, penetration, retention, equity, and commercial outcomes.
2. Diagnostic layer: drivers, segments, touchpoints, and customer experience measures.
3. Operational layer: campaign delivery, service performance, response, and issue logs.
4. Learning layer: experiments, interventions, confidence, and accumulated knowledge.

A dashboard should distinguish leading indicators from lagging indicators. Awareness or search behaviour may change before sales, but neither is automatically a valid causal predictor. Indicators should be connected through an explicit theory of change.

5.2 Executive report architecture

A concise executive report can contain:

1. Decision required.
2. Executive conclusion.
3. Three to five key findings.
4. Evidence and uncertainty.
5. Strategic implications.
6. Recommended actions, owners, timing, and measures.
7. Risks, caveats, and unresolved questions.
8. Technical appendix.

The main body should not hide critical limitations. A limitation that could change the decision belongs near the relevant claim, not only in an appendix.

Accessibility is part of quality. Use sufficient contrast, readable type, direct labels, text alternatives for charts, and colour schemes that remain interpretable for people with colour-vision differences. Confidentiality rules should be applied to small cells and quotations that might identify participants.

6. Communicating Uncertainty

Uncertainty is not a defect to be concealed; it is information about the reliability and scope of a conclusion. It may arise from sampling error, nonresponse, measurement error, model assumptions, coding interpretation, market volatility, or unknown confounding factors.

Researchers should distinguish:

• Sampling uncertainty: variation arising because a sample rather than a census was observed.
• Measurement uncertainty: imperfect wording, scales, recall, or instruments.
• Inferential uncertainty: limits on generalization or causal interpretation.
• Scenario uncertainty: genuine unpredictability in future conditions.
• Model uncertainty: dependence on analytical specifications or assumptions.

Probability language should be calibrated. “The data suggest” is not equivalent to “the data prove.” “No statistically significant difference” does not prove that groups are identical; it may reflect limited power, noisy measurement, or a genuinely small effect. Practical significance must be considered alongside statistical significance.

For forecasts or scenario work, ranges and assumptions are usually more informative than a single point estimate. Decision-makers should be told what is known, what is estimated, what is assumed, and what is unknown.

7. Evaluating Agencies and Research Proposals

Agency selection is a research-governance decision, not merely a procurement exercise. The lowest fee may represent an under-specified design, weak quality control, or substantial change-order exposure.

7.1 Proposal evaluation criteria

A weighted scoring model can support transparent comparison. If criterion i has weight wᵢ and proposal j receives score sᵢⱼ, the weighted score is:

Weights should sum to 1 (or 100%), and all proposals must use the same scoring scale. The formula supports structured judgment; it does not eliminate judgment. A proposal with a high total score may still fail a mandatory requirement, so minimum thresholds should be defined in advance.

7.2 Red flags

Red flags include guaranteed outcomes, vague samples, unexplained “nationally representative” claims, excessive reliance on a single method, no pilot plan, quotations presented as prevalence, unclear data ownership, unrealistic timelines, and an inability to explain how contradictory evidence will be handled.

A good commissioning brief specifies the decision, population, context, methodological constraints, required outputs, governance standards, timeline, budget range, and evaluation criteria while leaving appropriate room for professional design.

8. Managing Budgets and Research Portfolios

Research budgeting should be based on decision value rather than methodological prestige. Costs commonly include design, recruitment, incentives, fieldwork, technology, translation, analysis, visualization, reporting, data storage, and project management.

A portfolio view distinguishes:

• Foundational research: enduring understanding of people, category, and brand.
• Diagnostic research: explanation of a current problem.
• Evaluative research: assessment of an intervention.
• Exploratory research: discovery of emerging needs or opportunities.
• Monitoring: repeated measures for trend detection.

A budget should protect activities that affect validity, such as sample quality, piloting, translation, quality control, and analysis time. Cutting these may reduce cost while increasing decision risk.

Stage gates can control expenditure: approve the research question, then design, pilot, fieldwork, analysis, and dissemination. However, governance should not become bureaucracy that delays time-sensitive decisions. Proportionate documentation is the objective.

9. Stakeholder and Research Leadership

Insight leaders operate across marketing, finance, sales, product, data science, legal, agencies, and senior management. Their authority often depends more on credibility and relationship quality than on formal hierarchy.

Effective stakeholder management includes:

• agreeing the decision before commissioning research;
• identifying supporters, skeptics, users, and affected groups;
• clarifying roles and approval rights;
• involving decision-makers early enough to build ownership;
• separating evidence review from advocacy;
• documenting disagreements and unresolved assumptions;
• creating feedback loops after implementation.

A stakeholder disagreement is not necessarily a failure. Different functions may value different outcomes or hold different risk tolerances. The insight leader should make the disagreement explicit: Is it about the evidence, the interpretation, the objective, the feasibility of action, or the distribution of benefits and costs?

An evidence-led function also requires institutional memory. Repositories should preserve research questions, instruments, datasets where permitted, codebooks, analysis decisions, limitations, presentations, and subsequent outcomes. Without this record, organisations repeatedly fund similar studies and cannot learn whether recommendations worked.

10. Ethics, Governance, and Responsible Technology

Research communication must protect participants and preserve analytical integrity. Important principles include informed consent, voluntary participation, privacy, confidentiality, minimization of collected data, secure handling, fair inclusion, and honest reporting.

Quotations should be anonymized where identity could be inferred. Small subgroup results require suppression or careful aggregation. Sensitive attributes should not be used casually for targeting or segmentation, particularly where decisions could create exclusion or discrimination.

Artificial intelligence can assist with transcription, translation, coding suggestions, summarization, anomaly detection, and draft visualization. Its use requires human review because automated systems may hallucinate, reproduce bias, flatten context, or expose confidential data. An AI-generated theme is a hypothesis for analyst verification, not an independent finding.

Governance should document the tool used, data supplied, purpose, review process, known limitations, and final human accountability. Synthetic respondents or generated quotations must never be presented as real participant evidence. Automated sentiment scores should be validated for the language, category, and population in which they are used.

11. Developing an Evidence-Led Insight Function

An insight function becomes strategic when it moves beyond producing reports to shaping better decisions and organisational learning. Its maturity can be considered across five dimensions:

1. Ad hoc: isolated projects, inconsistent standards, limited reuse.
2. Repeatable: common briefs, methods, templates, and quality checks.
3. Integrated: research linked to brand planning, analytics, innovation, and performance management.
4. Strategic: insight informs resource allocation, portfolio choices, and capability development.
5. Learning organisation: decisions are tracked, outcomes are measured, and evidence is continuously improved.

Capability building involves methodological expertise, commercial understanding, communication, data literacy, facilitation, and ethical judgment. The function should measure not only activity—such as number of projects—but also usefulness: decisions influenced, uncertainty reduced, actions implemented, learning captured, and outcomes evaluated.

An insight agenda should include recurring strategic questions, a research calendar, a knowledge repository, agreed standards, supplier relationships, and a mechanism for escalating important evidence. The goal is not to make every decision data-driven in a narrow sense. It is to make assumptions visible, evidence proportionate, and learning cumulative.

Applied Case: Repositioning a Ready-to-Drink Beverage

This is an illustrative case.

Background

A national beverage brand had high aided awareness but declining repeat purchase among consumers aged 25–34. The marketing team proposed increasing digital advertising because brand tracking showed strong awareness but weak consideration.

Problem

The insight team had to determine whether the problem was inadequate mental availability, an unattractive proposition, poor availability, or a mismatch between brand meaning and consumption occasions. The decision involved a substantial media budget, so a recommendation based only on awareness data would be insufficient.

Analysis

The team combined four sources: a brand-tracking survey, retail purchase data, online reviews, and depth interviews. The survey showed high awareness but low association with “fits my everyday routine.” Purchase data indicated that trial was relatively strong while repeat purchase was weaker on weekdays. Interviews revealed that consumers liked the taste but viewed the product as an occasional treat requiring deliberate selection. Reviews also contained recurring complaints about portability and pack size.

The team created an evidence matrix separating observed findings from interpretations. It found convergence around a usage-occasion problem, but not proof that packaging alone caused low repeat purchase. The research therefore rejected the simplistic conclusion that more advertising would solve the issue.

Approach and solution

The executive report used an answer-first narrative:

• Claim: awareness is not the principal constraint; everyday relevance is.
• Evidence: high awareness, weak routine association, weekday repeat-purchase weakness, and recurring occasion barriers.
• Caveat: the evidence is correlational and does not isolate the effect of packaging, price, distribution, or competitive activity.
• Action: test a convenience-led proposition across communication, pack format, and selected retail occasions.

The team recommended a staged programme: prototype alternative pack formats, conduct qualitative usability work, run a controlled market test where feasible, and monitor repeat purchase rather than awareness alone. A dashboard connected leading indicators—convenience association and intended weekday use—to lagging indicators such as repeat purchase and value per buyer.

Outcome

The organisation redirected part of the proposed media increase toward proposition and packaging tests. The first test improved reported fit with weekday use, but the commercial effect was modest in stores with weak distribution. This led to a second learning: the proposition required both consumer relevance and physical availability.

Learning points

1. Awareness is not equivalent to demand.
2. Multiple methods can reveal different parts of the same problem.
3. A limitation can guide the next test rather than invalidate the current decision.
4. Dashboards are useful when measures are tied to a theory of change.
5. Research leadership includes resisting an attractive but weakly supported solution.

Challenges and Limitations

Insight communication faces persistent tensions. Compression can make complex evidence usable but may remove important qualifications. Visualization can clarify patterns but can also magnify trivial differences. Storytelling can create shared meaning but can encourage premature closure. Dashboards can improve accountability but can reward only what is easy to measure.

Qualitative evidence is vulnerable to selective quotation and overgeneralization. Quantitative evidence is vulnerable to false precision, poor sampling, measurement artefacts, and inappropriate causal claims. Mixed-method research does not automatically solve these problems; integration must be designed explicitly.

Organisational politics also matter. Senior audiences may prefer certainty, agencies may defend their methods, and teams may interpret findings through existing incentives. An evidence-led leader cannot remove politics but can make claims, assumptions, uncertainty, and decision rules more transparent.

Best-Practice Checklist

Before communicating research, ask:

• Is the decision clearly stated?
• Can every major claim be traced to evidence?
• Are fact, interpretation, implication, and recommendation distinguished?
• Is the visualization appropriate and quantitatively honest?
• Are denominators, sample sizes, and uncertainty visible?
• Have alternative explanations and disconfirming evidence been considered?
• Is the recommendation feasible, owned, timed, and evaluable?
• Are participant privacy and accessibility protected?
• Are AI tools governed and human-reviewed?
• Will the organisation capture what happens after the recommendation?

Chapter Summary

Insight communication is the disciplined conversion of research evidence into decisions and organisational learning. Findings describe patterns; insights explain their significance; implications connect them to strategy; and recommendations specify defensible action. Visualization must match the analytical question and preserve quantitative integrity. Storytelling should organise evidence around a decision without exaggerating certainty or suppressing contradictory evidence.

Dashboards support recurring monitoring when metrics have precise definitions, owners, sources, and action rules. Executive reports should communicate conclusions, evidence, implications, uncertainty, and next steps in a decision-friendly structure. Agency evaluation requires attention to methodological fit, sampling, quality assurance, governance, deliverables, capability, and commercial clarity. Research leaders manage budgets and stakeholders by aligning evidence with decisions, protecting validity, documenting disagreement, and building institutional memory.

An evidence-led insight function is ethical, transparent, accessible, and capable of learning from outcomes. Artificial intelligence may increase efficiency, but it does not replace methodological judgment, contextual understanding, or human accountability.

Key Terms

Actionability: The extent to which evidence supports a feasible decision or intervention.

Dashboard: A recurring information display designed to monitor defined measures and support decisions.

Evidence: Observed information used to support or challenge a claim.

Evidence ladder: A progression from description through explanation and prediction to prescription and evaluation.

Executive report: A structured communication designed to inform a specific management decision.

Finding: A directly supported pattern or result in research evidence.

Insight: An explanatory interpretation that makes a finding strategically meaningful.

Insight function: The organisational capability responsible for generating, communicating, applying, and learning from research-based understanding.

Leading indicator: A measure expected to change before a later outcome.

Metric catalogue: Documentation defining measures, denominators, sources, owners, and limitations.

Recommendation: A proposed action grounded in evidence and linked to a decision.

Sampling uncertainty: Uncertainty arising because conclusions are based on a sample rather than a census.

Stakeholder: A person or group that influences, uses, funds, or is affected by research.

Storytelling: The structured communication of evidence, interpretation, consequence, and action.

Triangulation: Comparison of multiple methods, sources, analysts, or time periods to strengthen interpretation.

Uncertainty: The degree to which a conclusion is limited by sampling, measurement, inference, models, or future unpredictability.

Visual integrity: The faithful representation of quantitative or qualitative evidence through design choices.

Theory of change: An explicit account of how an intervention is expected to produce outcomes.

Review Questions

Short Answer Questions

1. Distinguish a finding from an insight and a recommendation.
2. Why should a dashboard begin with decisions rather than available data?
3. Name three types of uncertainty relevant to market research.
4. What is the purpose of a metric catalogue?
5. Why is a participant quotation not automatically evidence of population prevalence?

Long Answer Questions

1. Explain how the claim–evidence–caveat–action structure improves research defensibility.
2. Compare the purposes and design requirements of an executive report and a dashboard.
3. Discuss how visual design choices can mislead decision-makers even when the underlying data are correct.
4. Explain the criteria that should be used to evaluate a research agency proposal.
5. Discuss the capabilities, governance practices, and measures required to build an evidence-led insight function.

Analytical / Application Questions

1. A survey reports that 58% of respondents prefer Brand A and 54% prefer Brand B. Design a communication that avoids overstating the difference. Identify the additional information needed before recommending a strategic change.
2. An agency proposes a large sample, rapid fieldwork, and guaranteed “actionable insights,” but provides little information about recruitment, quality control, or analysis. Evaluate the proposal and specify the questions you would ask before approval.
3. A dashboard shows rising awareness, stable consideration, and falling repeat purchase. Develop two alternative explanations, identify evidence needed to distinguish them, and propose an executive recommendation that communicates uncertainty appropriately.

Multiple Choice Questions with Answers

Question 1

Which sequence most accurately distinguishes research outputs?

A. Recommendation → finding → evidence → insight

B. Evidence → finding → insight → recommendation

C. Insight → evidence → recommendation → finding

D. Finding → recommendation → evidence → insight

Correct Answer: B. Evidence → finding → insight → recommendation

Question 2

What is the principal purpose of a dashboard?

A. To replace all qualitative research

B. To display as many available metrics as possible

C. To support recurring monitoring and decisions

D. To guarantee causal explanations

Correct Answer: C. To support recurring monitoring and decisions

Question 3

Which visualization is generally most suitable for showing change over time?

A. Line chart

B. Pie chart with many categories

C. Word cloud

D. Unordered table

Correct Answer: A. Line chart

Question 4

A quotation from one interview participant should normally be described as:

A. Proof of population prevalence

B. An illustrative expression of a participant’s view

C. A statistically significant result

D. A representative market estimate

Correct Answer: B. An illustrative expression of a participant’s view

Question 5

In a weighted proposal score, the weights should normally:

A. Be different for every proposal

B. Sum to a common total such as 1 or 100%

C. Be assigned after seeing the preferred result

D. Apply only to price

Correct Answer: B. Sum to a common total such as 1 or 100%

Question 6

Which statement about a non-significant statistical result is most accurate?

A. It proves the groups are identical

B. It proves the research was invalid

C. It may reflect a small effect, noisy data, or limited power

D. It proves the alternative hypothesis

Correct Answer: C. It may reflect a small effect, noisy data, or limited power

Question 7

Which is a leading indicator in the beverage case?

A. Historical repeat purchase

B. Past annual revenue

C. Convenience association before a market test

D. Previous-year distribution

Correct Answer: C. Convenience association before a market test

Question 8

What is triangulation?

A. Using only the largest available sample

B. Comparing multiple sources, methods, analysts, or time periods

C. Removing contradictory evidence

D. Replacing analysis with visualization

Correct Answer: B. Comparing multiple sources, methods, analysts, or time periods

Question 9

Which is an appropriate governance practice when using AI for qualitative coding?

A. Presenting generated themes without review

B. Uploading confidential data to any public tool

C. Treating automated themes as hypotheses requiring human verification

D. Using generated quotations as participant evidence

Correct Answer: C. Treating automated themes as hypotheses requiring human verification

Question 10

Which characteristic makes a recommendation defensible?

A. It is certain and requires no monitoring

B. It is based only on senior preference

C. It is traceable, feasible, proportionate, and evaluable

D. It uses the most complex available method

Correct Answer: C. It is traceable, feasible, proportionate, and evaluable