Manuel B. Garcia

Manuel B. Garcia serves as the Senior Director for Educational Technology and Digital Learning at FEU Institute of Technology, Manila, Philippines. Read More

Contact Info

1607, FEU Tech Building,
P. Paredes St, Sampaloc,
Manila, Philippines
mbgarcia@feutech.edu.ph

Follow Me

Part 1 · Guide 507 of 603

Can a Research Question Be Harmful Even When the Data Collection Itself Is Safe?

A study can collect data safely while still raising ethical concerns through the question it asks, the assumptions it embeds, the groups it characterizes, or the consequences of its findings. Participant safety during data collection is important, but it is not the whole ethical analysis.

By Manuel B. Garcia
507
Can a Research Question Be Harmful Even When Data Collection Is Safe? Guide 507 of 603
01 · The Question

What if nobody is harmed while the data are being collected?

Imagine an anonymous online survey. Participants answer ordinary questions from their own devices. No invasive procedures are involved. The survey collects no direct identifiers, participation is voluntary, and the researchers have taken appropriate steps to protect the data.

It would be reasonable to describe the data-collection process as relatively safe. But suppose the study is designed to determine whether members of a particular marginalized population are inherently less trustworthy, less capable, or more prone to an undesirable behavior. Or suppose technically sound data are later interpreted in ways that attach a negative characteristic to an identifiable community without adequate evidence.

The ethical problem may no longer lie primarily in what happened to participants while answering the survey. It may lie in the question, the assumptions built into the study, the interpretation of the evidence, or the consequences of producing and communicating the answer.

02 · The Short Answer

Safe data collection does not guarantee harmless research

In Brief

Yes. A research question can raise ethical concerns even when data collection itself is relatively safe because harm may arise from how the question frames people, how groups are categorized, what conclusions are drawn, or how findings affect participants, nonparticipants, or communities.

This does not mean researchers should avoid every question with potentially uncomfortable implications. It means ethical assessment should consider both the process of conducting research and reasonably foreseeable consequences of producing, interpreting, and communicating its findings.

03 · What You Need to Know

Research harm can arise from the process, the findings, or both

Data-collection safety addresses only one part of the research

Human-participant research ethics rightly pays considerable attention to what happens during participation. Researchers assess physical risks, psychological discomfort, privacy, confidentiality, consent, data security, and other potential harms associated with research procedures.

Those protections matter. But the ethical life of a study extends beyond the moment at which a participant answers a question, completes an assessment, or provides a sample.

Research produces claims. Those claims may characterize people, institutions, communities, or populations. They may be published, reported by news organizations, incorporated into policy discussions, repeated on social media, or used to support decisions. Consequently, the research process can be safe while the resulting representations create other forms of harm.

Process-related harm Harm arising through recruitment, consent, measurement, intervention, disclosure, data collection, or other activities involved in conducting the study.
Outcome-related harm Harm arising from what the research produces or communicates, such as stigmatizing findings, damaging group characterizations, or consequences associated with the interpretation and use of results.

Research-ethics scholarship on community-engaged research has explicitly distinguished risks arising from research processes from risks arising from research outcomes and has noted that outcome-related harms can affect both participants and people associated with the groups being studied.

The wording of a research question can embed an unsupported assumption

A question is not always a neutral container waiting to be filled with data. Its wording may define who is considered problematic, identify what requires explanation, privilege one causal interpretation, or treat a disputed characteristic as though it were already established.

Consider the difference between asking why members of a population are “academically deficient” and asking what factors are associated with an observed educational disparity. The first question assumes that the deficiency resides in the population. The second permits individual, institutional, socioeconomic, structural, measurement-related, and other explanations to be investigated.

Changing the wording does not automatically solve an ethical problem. Cosmetic euphemisms accomplish little. The substantive issue is whether the question embeds an assumption that the evidence has not established.

This is one reason ethical considerations may need to enter before the research question is finalized, rather than appearing only after the methodology has been designed.

Group-level harms can reach people who never participated

Suppose researchers collect anonymous information from members of a small identifiable community and conclude that a socially undesirable characteristic is unusually common within that community. No individual participant is identifiable.

Nevertheless, other members of the community may be associated with the published characterization simply because they belong to the group. They never consented to the study and never provided data, yet the research may affect how others perceive them.

Research-ethics literature on community-engaged research describes this distinction explicitly: risks may arise to individuals as individuals, to individuals because they are members of a group, and to communities themselves. Findings attributed to a group can therefore have implications for participating and nonparticipating members alike.

This possibility becomes particularly relevant when a study could contribute to stigmatization of the population being studied.

Safe collection cannot repair invalid interpretation

Some apparent ethical problems are also methodological problems wearing an ethical hat. A cross-sectional association may be reported as causation. Differences between groups may be described as inherent characteristics despite plausible contextual explanations. Results from a small convenience sample may be generalized to an entire population.

These practices are scientifically problematic because the conclusions exceed the evidence. When the unsupported conclusion portrays a group negatively, the methodological error can also create social or reputational consequences.

Ethical interpretation therefore requires methodological discipline. Researchers should distinguish association from causation, sample findings from population claims, observed differences from explanations for those differences, and empirical evidence from speculation.

Anonymity protects individuals but may not protect groups

Anonymizing participant data can substantially reduce certain privacy and confidentiality risks. It does not necessarily prevent a population from being identifiable in a publication.

A paper might never identify “Participant 17” while repeatedly naming the neighborhood, profession, cultural group, institution, or other population from which participants were recruited. The relevant ethical unit is therefore not always the individual record.

CIOMS guidance for health-related research specifically recognizes that research results may stigmatize groups or expose their members to discrimination and recommends sensitivity to such risks, including respectful publication practices. The precise applicability of CIOMS depends on the research context, but the underlying issue extends beyond health research: protecting individual identities and considering group consequences are related but different tasks.

Not every possible misuse is the researcher's responsibility

There is an important limit here. Researchers cannot predict or control every possible way in which knowledge might later be interpreted, distorted, or misused. Almost any finding could conceivably be taken out of context.

If ethical responsibility required eliminating every imaginable downstream misuse, a considerable amount of legitimate research would become impossible.

The more defensible standard is reasonable foreseeability and proportionality. Researchers should pay particular attention when harmful interpretations are predictable from the topic, framing, categories, analytical choices, or communication strategy. Highly speculative scenarios deserve less weight than plausible consequences that can be anticipated during study design.

Potential harm does not mean difficult questions should disappear

Some of the questions most capable of producing uncomfortable findings are also socially important. Research on discrimination, violence, health disparities, educational inequality, corruption, harmful behavior, or social conflict may reveal patterns that communities or institutions would prefer not to see documented.

Suppressing accurate findings to prevent reputational discomfort can itself create harm. CIOMS guidance on health-related research notes that ethical assessment should also consider potential harm from forgoing research or failing to publish its results.

The ethical objective is therefore not to guarantee that research makes everyone look good. It is to ensure that the question has sufficient value, the design can answer it responsibly, and conclusions do not create avoidable harm through unsupported or unnecessarily stigmatizing claims.

Sometimes the question, rather than the data-collection method, needs redesign

If the ethical problem arises because a question presupposes an unjustified deficit, targets a population without sufficient reason, or seeks a conclusion whose likely value is trivial relative to foreseeable harm, making the survey safer will not solve the central problem.

This is precisely the kind of situation in which ethical constraints may need to change the research question rather than merely the method.

Researchers might broaden the explanatory framework, reconsider the comparison, examine structural rather than presumed intrinsic explanations, remove an unnecessary group label, or formulate a question that tests rather than presupposes the contentious claim.

Watch Out

Do not use “the data collection is anonymous” as shorthand for “the study cannot cause harm.” Anonymity may address an important individual-level risk while leaving questions about group characterization, stigmatization, interpretation, and downstream consequences unresolved.

04 · A Practical Example

A safe survey built around a potentially harmful assumption

Hypothetical Example

When the ethical concern begins with the question

A researcher observes that students from a particular low-income community have lower average participation in an optional university program. The proposed anonymous survey asks, “Why do students from Community X lack ambition to participate in university opportunities?” Completing the questionnaire presents little foreseeable risk, and no identifying information is collected.

Data-collection assessment The anonymous survey itself is relatively low burden and presents limited foreseeable risk.
Question assessment The question presumes that lack of ambition explains lower participation even though this has not been established.
Potential consequence The study could reinforce a deficit-based characterization of an identifiable community while overlooking financial, scheduling, informational, institutional, or accessibility barriers.
Reframing The researcher instead asks what individual, institutional, socioeconomic, and contextual factors are associated with participation in the program.
Result The revised question does not guarantee favorable findings, but it allows competing explanations to be investigated rather than treating one stigmatizing explanation as established fact.

The revised study is not more ethical merely because the wording sounds gentler. It is stronger because the question no longer assumes a causal explanation that the evidence has not demonstrated.

In this case, ethical reflection improves the scientific question rather than preventing the research from being conducted.

05 · What Researchers Often Get Wrong

Why safe procedures do not settle the entire ethical analysis

Misconception

If nobody is identifiable, nobody can be harmed

Anonymity can reduce individual privacy and confidentiality risks, but research may still characterize an identifiable population. Group-level social or reputational consequences can therefore remain possible.

Misconception

A question cannot be problematic until the results are known

Ethical concerns can be visible earlier. A question may embed unsupported assumptions, use unjustified group categories, or be structured around a foreseeable stigmatizing interpretation before any data are collected.

Misconception

Researchers are responsible for every possible future misuse of their findings

No researcher can anticipate every imaginable interpretation. Ethical responsibility should focus particularly on reasonably foreseeable consequences and on harms researchers can reduce through defensible design, analysis, contextualization, and communication.

Misconception

A harmful implication means the finding should not be published

Not necessarily. Suppressing valid and important findings can itself be harmful. Researchers should report evidence accurately while avoiding unsupported generalizations, sensationalism, and claims the study cannot establish.

Misconception

Group harm is simply the sum of harms experienced by individual participants

Research-ethics scholarship distinguishes harms to individual participants from harms arising through group membership and harms to communities themselves. People who never participated may sometimes be affected by findings attributed to a group.

06 · What This Means for You

Evaluate what the study does with knowledge, not only how it collects data

A simple decision framework

If data collection is safe but the question contains an assumption about a population
Ask whether the assumption is supported by existing evidence or should instead become something the study tests.
If findings could characterize an identifiable group
Consider reasonably foreseeable consequences for participants, nonparticipants, and the wider group alongside individual protections.
If a potentially harmful conclusion depends on causal or population-level claims
Confirm that the design and sample genuinely support those claims before making them.
If harm arises mainly from unnecessary framing or categorization
Redesign the question or analysis without concealing scientifically relevant differences.
If substantial foreseeable harm remains despite reasonable safeguards
Assess whether the expected scientific or social value is sufficient to justify proceeding or whether the research idea should be substantially changed.

The central discipline is to widen the ethical lens without making it limitless. Researchers should not be expected to control every future interpretation of knowledge. They should, however, take responsibility for the assumptions they build into questions, the categories they create, the conclusions they draw, and foreseeable harms that reasonable changes could reduce.

In some projects, this broader assessment will also require examining consequences for communities before settling on the research question. Safe participation and responsible knowledge production are complementary goals, not interchangeable ones.

07 · A Quick Checklist

Look beyond the safety of the data-collection procedure

Before finalizing the study, check:
The research question does not treat an unsupported assumption about a person or population as established fact.
Group categories and comparisons are scientifically necessary rather than included merely because the data are available.
Potential harms to identifiable groups have been considered separately from individual privacy and confidentiality risks.
The design can support the causal, comparative, or population-level claims you intend to make.
Plausible contextual and alternative explanations will be considered where relevant.
Foreseeable downstream harms have been considered without treating every speculative misuse as equally probable.
Potentially uncomfortable findings can still be reported accurately rather than suppressed or cosmetically softened.
The expected value of answering the question remains proportionate to significant foreseeable harms that cannot be reduced.
08 · Frequently Asked Questions

Questions about harms beyond data collection

Can an anonymous study still harm a community?

Potentially. Individual participants may remain anonymous while findings identify or characterize a recognizable group. People associated with that group may then experience social, reputational, or discriminatory consequences even if they never participated.

Does this mean researchers should avoid controversial questions?

No. Controversial or uncomfortable questions may have substantial scientific and social value. Researchers should distinguish legitimate inquiry from unnecessary stigmatization, unsupported assumptions, and interpretations that exceed the evidence.

Is a research question harmful simply because someone finds it offensive?

Not necessarily. Offense can be ethically relevant in context, but it is not by itself a complete test of research harm. Researchers should consider the question's value, framing, scientific justification, foreseeable consequences, and whether less harmful approaches can achieve the same legitimate objective.

Are researchers responsible if the media misrepresent their findings?

Researchers cannot control every third-party interpretation. They can reduce foreseeable misunderstanding by reporting limitations clearly, avoiding sensational claims, providing appropriate context, and ensuring that their own conclusions do not exceed the evidence.

Can accurate research findings still cause harm?

Yes. Accurate information can have adverse consequences. That does not automatically mean it should be withheld. Ethical judgment requires considering the value of the knowledge, foreseeable harms, available safeguards, and possible harms from not conducting or communicating the research.

What if the ethical problem is built into the research question?

Reconsider the question itself. A change in wording may be insufficient if the underlying assumptions or objectives remain problematic. The question may need substantive reframing, a different comparison, or, in some circumstances, abandonment.

09 · The Bottom Line

Safe participation is necessary, but research can have consequences beyond participation

The Bottom Line

A research question can raise serious ethical concerns even when data collection itself is safe because harms may arise from the assumptions built into the question, the characterization of groups, unsupported interpretations, or reasonably foreseeable consequences of the findings.

Do not treat this as a reason to avoid every difficult question. Instead, assess both how evidence is collected and what the study is positioned to claim, consider individual and group-level consequences where relevant, and redesign questions that create avoidable harm without sacrificing legitimate scientific inquiry.

10 · Sources and Further Reading

Authoritative sources on research risks and group-level harms

11 · Cite this Guide

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.

Has the Field Guide helped your research?

If a guide helped clarify a question, inform a research decision, or move your work forward, I would love to hear about your experience. Your story may also help other researchers discover the Field Guide.

Share Your Experience
Takes only a few minutes