03 · What You Need to Know
Understanding Ethical Responsibilities Beyond Individual Research Participants
What Is Community Harm in Research?
Community harm refers broadly to adverse consequences experienced by a group or community in connection with research activities, findings, or their subsequent use.
Such consequences may arise even when individual research participants experience no direct injury.
For example, a study may associate a particular community with an unfavorable characteristic. If the findings are interpreted as evidence of collective deficiency, members of that community may experience stigma or discriminatory treatment.
Community harm can also involve loss of trust, misuse of collective information, or policies that disadvantage a population.
However, the term should not be used to imply that every negative description of a group is unethical.
The relevant assessment concerns the evidence, context, foreseeable consequences, and whether the research unnecessarily contributes to unjustified harm.
How Does Community Harm Differ From Individual Participant Harm?
Individual participant protections focus on people who take part in research and the risks they may encounter through participation.
Community-level concerns extend to groups whose interests may be affected by research, including people who did not participate.
Individual participant harm
Adverse consequences affecting a person through participation, such as unauthorized disclosure of confidential information or psychological distress.
Community-level harm
Adverse consequences affecting a population through group identification, stigmatization, discriminatory applications, or other collective effects.
Overlapping harm
Situations in which research affects both participants individually and the communities with which they are associated.
These categories are not mutually exclusive. A confidentiality breach may identify a participant while also revealing sensitive information about a community.
Conversely, a publication may avoid identifying any participant yet still contribute to harmful generalizations about the population studied.
Why Does Informed Consent Not Resolve Every Community Risk?
Informed consent is an important mechanism for respecting individual autonomy. However, an individual's agreement to participate does not necessarily authorize every consequence affecting other members of a community.
For example, a participant may consent to discussing experiences of educational disadvantage without having authority to consent on behalf of an entire neighborhood or cultural group.
Similarly, a community representative may support a research project without necessarily possessing authority to provide consent for every individual participant.
Researchers should distinguish individual consent from community engagement, collective interests, and any legally recognized forms of collective authorization.
These mechanisms may complement one another, but they are not universally interchangeable.
What Types of Community Harm Can Research Create?
| Potential harm |
How it may arise |
Illustrative consequence |
| Stigmatization |
Findings are interpreted as evidence of a group's inherent deficiency. |
Students from a community are stereotyped as academically incapable. |
| Discriminatory policy |
Research findings are used to justify unequal treatment. |
Educational resources are withheld from particular populations. |
| Collective privacy harm |
Research reveals sensitive information about an identifiable group. |
A small community becomes associated with sensitive behavioral characteristics. |
| Misrepresentation |
Findings are generalized beyond the evidence. |
A nonrepresentative sample is portrayed as describing an entire community. |
| Loss of institutional trust |
Research practices or dissemination disregard legitimate community concerns. |
Community members become reluctant to participate in future studies. |
| Exploitative knowledge use |
Research outputs benefit outside organizations while creating unfair burdens for the studied community. |
Community information is commercialized without appropriate governance or benefit considerations. |
These examples describe possible pathways to harm rather than outcomes that necessarily follow from research involving particular populations.
Can Anonymous Research Still Harm a Community?
Yes. Removing individual identifiers may protect participants without eliminating the possibility that findings identify or characterize a group.
For example, a report may avoid naming individual students while identifying a small educational community and attributing sensitive characteristics to it.
Even where individual reidentification is unlikely, the findings may influence how the wider group is perceived or treated.
However, not every group-level statistic creates meaningful harm.
Researchers should examine the specificity of group identification, the sensitivity of the findings, the reliability of the evidence, and plausible consequences of dissemination.
Individual anonymity and collective protection are related but distinct ethical considerations.
What Is Group Stigmatization?
Group stigmatization occurs when a population becomes associated with characteristics that contribute to negative stereotyping, exclusion, or diminished social standing.
Research can contribute to stigmatization when findings are framed without sufficient context or interpreted as reflecting inherent group characteristics.
For example, lower average performance on a digital literacy assessment may reflect unequal access to devices, connectivity, instruction, or other educational opportunities.
Presenting the difference as evidence of innate inability would introduce an interpretation that the data may not support.
However, researchers should not conceal reliable differences simply because they may be uncomfortable.
The appropriate response is accurate interpretation, contextual explanation, and careful consideration of foreseeable misuse.
Can Accurate Research Findings Still Produce Unfair Harm?
Yes. Accuracy is necessary for credible empirical research, but it does not guarantee that every interpretation or application is justified.
A statistically supported difference between groups may be used to make decisions that are inappropriate for individuals or unrelated contexts.
For example, a group-level association between educational disadvantage and access to technology does not justify assuming that every student from the group has the same needs or capabilities.
Such reasoning may involve an ecological inference error, where characteristics observed at the group level are improperly attributed to individuals.
Researchers should distinguish what the data establish from what policymakers or institutions might infer.
The implications of accurate research findings that cause unfair population harm require particular attention when evidence informs consequential decisions.
Are Community Harms Limited to Sensitive Populations?
No. Community-level consequences can arise in research involving professional groups, institutions, neighborhoods, educational populations, or other identifiable communities.
However, the nature and severity of risks may differ depending on historical disadvantage, social power, institutional relationships, and available protections.
A study of workplace practices may affect a professional community's reputation, while research involving a marginalized population may reinforce existing patterns of exclusion.
Researchers should avoid assuming that every group experiences risk in the same way.
The relevant assessment should be sensitive to context without treating communities as homogeneous or inherently vulnerable.
What if a Community Disagrees With the Findings?
Community disagreement does not automatically establish that research is inaccurate or unethical.
Researchers have responsibilities to report defensible findings even when those findings are contested.
However, disagreement may reveal concerns about measurement, interpretation, group definitions, or the consequences of dissemination.
Researchers should examine whether the criticism identifies methodological weaknesses or contextual information that has been overlooked.
They should distinguish empirical disagreement from differences in values or policy preferences.
Community perspectives can inform interpretation without granting any stakeholder unrestricted authority to suppress valid evidence.
Should Researchers Consult Communities Before Conducting a Study?
Community engagement may be particularly valuable when research concerns identifiable groups and could produce consequential collective effects.
Consultation can help researchers understand local priorities, terminology, historical experiences, and potential consequences that may not be apparent from the research literature alone.
However, community engagement is not a single standardized procedure.
Some projects may benefit from advisory groups or participatory design, while others may require consultation through established governance arrangements.
Researchers should examine who speaks for the community and whether differing perspectives are represented.
Engagement should not become a symbolic exercise in which feedback is collected but cannot influence meaningful decisions.
Can Community Representatives Consent on Behalf of Everyone?
Not generally. Community consultation and individual informed consent serve different functions.
In some contexts, including particular Indigenous research governance arrangements, collective authority or community permission may have recognized ethical or legal significance.
However, the existence and scope of such authority depend on the relevant community, jurisdiction, and governing framework.
Researchers should not assume that a community leader can waive individual rights or that one person's agreement represents the views of all members.
Equally, individual consent should not automatically be treated as sufficient to address legitimate collective interests.
Appropriate governance may require attention to both levels.
How Does Community Harm Relate to Educational Technology Research?
Educational technology studies frequently compare populations, institutions, or learner characteristics using digital traces and assessment data.
These comparisons may reveal important inequalities and support better educational policy.
However, they may also be used to classify schools or communities as deficient without adequately considering contextual factors.
For example, a study showing lower engagement in an online learning platform may reflect unstable internet connectivity rather than lower motivation.
If researchers interpret the pattern without examining access conditions, the findings may support inappropriate judgments about students or institutions.
Methodological rigor and ethical interpretation therefore reinforce one another.
Can AI-Based Research Increase Community-Level Risks?
AI systems may produce classifications or predictions that are applied repeatedly across large populations.
When models are trained on data reflecting existing inequalities, their outputs may reproduce or amplify problematic patterns under some deployment conditions.
However, bias and harm cannot be inferred solely from the use of machine learning.
Researchers should examine the training population, measurement validity, subgroup performance, intended use, and institutional consequences of errors.
A model that performs adequately on average may still require further evaluation before being used for decisions affecting particular communities.
The NIST AI Risk Management Framework provides a general resource for considering risks associated with AI systems, although it does not prescribe a universal research ethics procedure.
What if a Community Is Identifiable but Individual Participants Are Not?
Group identifiability may remain relevant even when the dataset contains no direct personal identifiers.
For example, reporting results from a small, geographically identifiable community may expose collective characteristics that members consider sensitive.
Researchers should consider whether naming the community is necessary for scientific interpretation.
However, removing all contextual information may make findings difficult to evaluate or reproduce.
The appropriate balance depends on the research question, the significance of the information, and the likely consequences of identification.
Researchers should avoid promising complete anonymity when the published context makes group identification unavoidable.
Does Avoiding Community Harm Mean Avoiding Unfavorable Findings?
No. Suppressing valid findings solely because they are unfavorable can distort scientific knowledge and prevent recognition of important problems.
For example, evidence of unequal access to educational technology may be necessary to justify additional public investment.
However, the same evidence can be communicated in ways that either clarify structural disadvantage or encourage unsupported stereotypes.
Researchers should accurately report findings while distinguishing observed differences from speculative explanations.
The ethical obligation is not to guarantee favorable portrayals. It is to avoid unnecessary misrepresentation and consider foreseeable harmful uses.
How Should Researchers Communicate Group Differences?
Researchers should identify the population studied, measurement procedures, uncertainty, and limits of generalization.
Where relevant, they should discuss contextual factors that may help explain observed differences without asserting causal explanations unsupported by the design.
For example, differences in digital literacy scores should not automatically be attributed to motivation, intelligence, or cultural characteristics.
Researchers should also avoid treating average differences as descriptions of every individual within a group.
The principles of communicating sensitive group differences without reinforcing stigma are particularly relevant when findings concern historically disadvantaged populations.
What if Policymakers Misuse the Findings?
Researchers may have ethical reasons to clarify when their findings are materially misrepresented or applied beyond the evidence.
For example, if a policymaker uses a descriptive study to claim that a population is inherently incapable of benefiting from technology, the researcher may explain that the study did not establish that conclusion.
However, researchers should distinguish empirical correction from disagreement over legitimate policy choices.
They may not possess authority to determine every policy decision informed by their work.
The appropriate response depends on the seriousness of the misinterpretation, the researcher's influence, and available channels for clarification.
The broader question of responsibility for misuse of research findings helps define these limits.
How Should Researchers Balance Individual and Community Interests?
Individual and community interests may align, but conflicts can arise.
For example, publishing detailed contextual information may strengthen transparency while increasing the risk of identifying a small community.
Restricting such information may reduce exposure but also limit reproducibility and the usefulness of the findings.
Researchers should consider which interests are affected, whose rights are involved, and whether less restrictive alternatives are available.
No universal rule establishes that community interests always override individual autonomy or that individual consent resolves every collective concern.
The appropriate balance depends on the context and applicable ethical obligations.
The question of balancing individual and community interests warrants careful attention when different groups may bear unequal consequences.
Can Community Engagement Prevent Every Harm?
No. Engagement may improve research design and interpretation, but it cannot guarantee that every stakeholder will agree or that all downstream uses will be beneficial.
Communities are rarely homogeneous. Members may differ in their priorities, experiences, and views about acceptable research.
Researchers should avoid treating consultation with a single representative as evidence of universal community endorsement.
Engagement is most useful when it informs decisions and when disagreements are documented rather than concealed.
Researchers should also consider whether participation in consultation creates additional burdens or expectations.
What Should Ethics Committees Consider About Community Harm?
Where community-level consequences are relevant to the study and within the committee's remit, ethics review may examine foreseeable group risks alongside individual participant protections.
Relevant questions may concern recruitment, group identification, data governance, dissemination, and the potential use of findings.
However, review bodies differ in their jurisdiction and expertise.
Some projects may benefit from additional community consultation or specialized advice beyond conventional participant-focused review.
Researchers should not assume that approval eliminates every possibility of collective harm.
Nor should they impose additional procedures without considering their relevance, feasibility, and effects on the research.
When Might Community Harm Justify Changing the Research?
Changes may be warranted when serious foreseeable harms are closely connected to the research design, group identification, or intended use of the findings.
Possible responses include revising recruitment or reporting procedures, modifying how groups are defined, restricting sensitive data access, or reconsidering the research question.
However, changes should preserve scientific validity and avoid concealing meaningful differences.
For example, suppressing all results concerning disadvantaged communities could undermine evidence needed to address educational inequality.
The objective is to reduce unnecessary harm while retaining an accurate account of the phenomenon studied.
When the risk is inherent in the intended capability, the principles concerning misuse risks and the choice of research question may be relevant.
How Can Researchers Document Community-Level Ethical Decisions?
Researchers should record the groups potentially affected, foreseeable risks, consultation undertaken, relevant methodological decisions, and reasons for selected safeguards.
Where concerns are raised, the record should distinguish empirical evidence from stakeholder judgments and unresolved uncertainty.
For example, a research team may document why it reported regional rather than institution-level results to reduce unnecessary identification risks.
Such decisions should also consider whether aggregation changes the meaning or usefulness of the findings.
Documentation supports accountability without implying that every ethical trade-off has one objectively correct solution.
Watch Out
Do not assume that individual consent and anonymization eliminate every ethical concern. Research may affect communities through group identification, interpretation, or policy use. At the same time, concern about stigma should not become a justification for concealing valid evidence or preventing communities from being represented in research.
