03 · What You Need to Know
Trustworthiness Is Built Into the Study, Not Added at the End
Trustworthiness is often discussed through the framework associated with credibility, dependability, confirmability, and transferability. These concepts help researchers ask different questions about the quality of an inquiry.
They are most useful when treated as methodological concerns rather than four boxes requiring one technique each. The broader question of what rigor means in qualitative research depends on the qualitative tradition, research question, epistemological assumptions, and claims being made.
| Trustworthiness Concern |
Question to Ask |
Possible Strategies |
| Credibility |
Why should readers find this interpretation sufficiently plausible and well grounded? |
Appropriate sampling, sustained engagement, triangulation, participant engagement, peer dialogue, attention to contradictory evidence |
| Dependability |
Can readers understand how the study and analysis developed? |
Methodological documentation, analytic memos, decision records, audit trail, transparent reporting of changes |
| Confirmability |
Can readers see how interpretations connect to the research material and how researcher influences were considered? |
Reflexivity, evidential grounding, analytic records, consideration of alternatives, transparent links between data and claims |
| Transferability |
Can readers judge whether the findings may illuminate another context? |
Relevant contextual description, participant and setting information, boundaries of the phenomenon and conditions studied |
Begin With Methodological Coherence
Before choosing a trustworthiness strategy, establish that the design itself makes sense.
The research question should fit qualitative inquiry. The chosen qualitative approach should fit the kind of knowledge sought. Sampling should follow the logic of that approach. Data-generation procedures should create material capable of addressing the question, and the analytic method should be compatible with the methodology and epistemological position.
No amount of member checking or triangulation can compensate for a fundamental mismatch between question, data, analysis, and claims.
APA's qualitative reporting standards emphasize this coherence by asking researchers to identify their approach to inquiry, describe researcher perspectives, explain participant or data-source selection, report data-collection and analytic procedures, and describe methodological integrity.
Use Appropriate Sampling to Strengthen Credibility
Qualitative sampling is usually designed around informational relevance rather than statistical representativeness.
A researcher studying how faculty implement institutional AI policies might deliberately recruit participants from disciplines with different assessment practices, varying levels of AI use, or different institutional responsibilities. The aim is not necessarily to reproduce population proportions. It may instead be to obtain sufficiently relevant and varied perspectives to understand the phenomenon.
Sampling becomes a trustworthiness issue when important perspectives are systematically absent without justification, when participants have little experience of the phenomenon, or when recruitment procedures shape the data in ways the analysis ignores.
Researchers should therefore explain why the selected participants, cases, documents, settings, or events are capable of illuminating the research question.
Engage With the Phenomenon Deeply Enough
Some qualitative inquiries require substantial engagement with participants or settings before researchers can understand what is occurring.
Prolonged engagement is particularly associated with naturalistic and ethnographic traditions. Time in the setting may allow researchers to understand routines, develop relationships, identify unusual events, and distinguish recurring patterns from initial impressions.
More time is not automatically better, however. A focused interview study does not become rigorous merely because interviews are longer, and a researcher can spend months in a field site while still asking superficial questions.
The appropriate depth and duration of engagement depend on the phenomenon and methodology.
Use Triangulation When Multiple Perspectives Can Test or Deepen the Interpretation
Triangulation can involve different data sources, methods, researchers, theories, or perspectives.
Suppose faculty interviews suggest that an institutional AI policy is consistently implemented across departments. Researchers also examine policy documents and classroom practices. If the evidence converges, confidence in part of the interpretation may increase.
But convergence is not the only useful outcome.
If interviews describe consistent implementation while classroom observations reveal substantial variation, the disagreement may identify an important distinction between official policy, perceived practice, and enacted practice.
Triangulation therefore strengthens trustworthiness when it interrogates the phenomenon from useful vantage points. Collecting several kinds of data simply so that the study can claim “triangulation” adds complexity without necessarily adding rigor.
Use Member Checking for a Defined Purpose
Member checking can take many forms. Researchers may return transcripts, preliminary interpretations, summaries, themes, or findings to participants. These activities do not accomplish the same thing.
A participant may verify whether a transcript contains factual errors. Another may comment on whether a preliminary interpretation resonates with their experience. A group discussion may reveal perspectives missing from the analysis.
These can all be valuable.
What member checking cannot provide is an automatic certificate that an interpretation is correct. Participants may disagree with one another, change their views, misunderstand the researcher's theoretical interpretation, or reasonably object to an interpretation that extends beyond their individual perspective.
Use participant feedback when it answers a meaningful credibility question, and report what was returned, to whom, what feedback was obtained, and how that feedback affected the analysis.
Search for Evidence That Challenges Your Interpretation
Qualitative analysis becomes vulnerable when researchers notice only data that fit an emerging story.
Negative-case or deviant-case analysis deliberately examines observations that contradict, complicate, or fail to fit the developing interpretation.
Suppose most participants describe generative AI as reducing workload, but several describe it as creating additional verification work. Rather than dismissing those accounts as exceptions, researchers can ask why experiences differ. Perhaps workload effects depend on discipline, task type, institutional policy, or participants' expertise.
The contradictory cases may require refinement of the interpretation rather than abandonment of it.
This practice strengthens credibility because the analysis becomes less dependent on selective attention to confirming material.
Keep an Audit Trail When Traceability Matters
An audit trail is documentation that allows the development of the research process and analysis to be reconstructed.
Depending on the study, this might include versions of interview guides, sampling decisions, coding frameworks, analytic memos, theme-development records, methodological decisions, meeting notes, and documentation of changes made during the study.
The point is not to archive every keystroke. The useful record concerns decisions that materially affected what evidence was generated and how it was interpreted.
An audit trail can contribute to dependability and confirmability by showing how researchers moved from data toward conclusions and how the inquiry changed along the way.
Document Changes Rather Than Pretending the Study Never Evolved
Qualitative designs often develop iteratively.
Researchers may refine questions after early interviews, recruit additional participants to examine an emerging idea, reconsider an analytic category, or change the emphasis of later data collection.
Such development can be methodologically appropriate. Concealing it to make the study appear perfectly linear is less helpful.
APA's qualitative reporting standards explicitly recognize that qualitative research can evolve and recommend reporting modifications to methodological integrity and data-collection or analytic processes where relevant. Transparent documentation allows readers to distinguish principled responsiveness from arbitrary methodological drift.
Practice Reflexivity Throughout the Study
Reflexivity requires researchers to examine how their assumptions, identities, professional roles, theoretical commitments, relationships, and decisions shape the inquiry.
This is not accomplished simply by adding a positionality paragraph stating the researcher's occupation.
Imagine that a university administrator interviews faculty about institutional leadership. Participants may moderate criticism because they perceive the researcher as connected to management. Recognizing that relationship could influence recruitment procedures, confidentiality safeguards, interviewer choice, question wording, interpretation, and reporting.
Reflexivity becomes methodologically useful when awareness of researcher positioning affects how the study is designed, conducted, analyzed, and interpreted.
Make the Connection Between Data and Interpretation Visible
Readers should be able to understand how major claims arose from the research material.
Participant quotations can help, but quotations alone do not establish confirmability. Researchers choose which quotations appear and how they are contextualized.
A stronger analysis explains the interpretive pattern, illustrates it with appropriate evidence, acknowledges relevant variation, and shows how the interpretation relates to the wider dataset rather than presenting a striking quotation as though it speaks for everyone.
Confirmability is therefore strengthened by transparent analytic reasoning, not by maximizing the number of quotations in the findings section.
Peer Debriefing Can Challenge Analytic Assumptions
Discussion with colleagues can help researchers expose assumptions, consider competing explanations, examine whether claims exceed the evidence, or identify areas where the analysis remains underdeveloped.
The value comes from critical engagement rather than approval.
A colleague who simply says that the themes “look good” contributes little. A colleague who asks why contradictory cases were excluded, whether one category combines distinct phenomena, or how the researcher's theoretical commitments shaped interpretation may contribute considerably more.
Peer debriefing should therefore be understood as analytic scrutiny, not ceremonial endorsement.
Provide Contextual Detail for Transferability
Researchers cannot determine in advance whether findings will transfer to every future context. They can provide readers with enough information to make informed judgments.
Relevant description might include participant characteristics, institutional structures, geographic or cultural context, timing, professional roles, policies, social relationships, or material conditions that shaped the phenomenon.
The emphasis should be on information consequential to interpretation.
For example, a study of faculty adoption of AI tools may need to describe institutional AI policies, technological access, teaching modality, disciplinary context, and participants' roles. Elaborate description of irrelevant setting details contributes little to transferability.
Do Not Treat Saturation as a Universal Trustworthiness Test
Saturation is widely invoked to justify qualitative sample size, but the term has several meanings and is not appropriate to every qualitative approach.
Researchers may refer to data saturation, code saturation, meaning saturation, theoretical saturation, or other formulations. Grounded theory traditions may use saturation differently from thematic or phenomenological approaches, while some qualitative methodologies reject saturation as inconsistent with their analytic assumptions.
Simply stating “saturation was achieved” does not demonstrate sample adequacy. Explain what saturation meant in the study, how it was assessed, and why it fits the methodology, if saturation is used at all.
Use Multiple Researchers Only When Their Role Is Methodologically Useful
Team analysis can strengthen a study by introducing multiple perspectives, challenging assumptions, distributing analytic work, or supporting consistent coding when consistency is part of the analytic method.
It does not automatically make findings more trustworthy.
In some approaches, independent coding and agreement procedures are appropriate. In others, researchers deliberately use dialogue and reflexive interpretation rather than attempting to eliminate differences among analysts.
This is why conventional reliability procedures should not be imposed automatically on qualitative research.
Reporting Standards Help Readers Evaluate Trustworthiness
Transparent reporting makes methodological quality visible.
APA provides specific Journal Article Reporting Standards for qualitative and mixed-methods research, and its qualitative standards were designed to accommodate a range of qualitative traditions rather than prescribe one method. SRQR similarly provides standards for reporting qualitative research, while COREQ focuses specifically on interview and focus-group studies.
Reporting guidelines do not establish trustworthiness by themselves. Their value is that they prompt researchers to disclose enough about the research process for readers, reviewers, and editors to evaluate the study.
