03 · What You Need to Know
How to Determine Which Evidence Sources Your Case Study Actually Needs
Why Are Multiple Sources Commonly Recommended in Case Study Research?
Case studies frequently investigate complex phenomena involving decisions, experiences, institutional practices, and contextual conditions.
No single evidence source necessarily captures all these dimensions adequately.
For example, interviews may reveal how faculty members understood a university's AI policy, while committee records establish which proposals were formally considered. Observations may show how the policy is enacted, and platform records may document patterns of actual use.
These sources contribute different forms of evidence about the same bounded case.
Yin emphasizes the use of multiple sources of evidence as an important principle of case study data collection. The rationale is not simply that additional sources increase the quantity of data. Rather, they may permit the development of converging lines of inquiry and a more defensible chain of evidence.
Other case study traditions may approach evidence integration differently, particularly when the research prioritizes participants' meanings or the particularity of the case.
What Are the Main Sources of Evidence in a Case Study?
Yin identifies six commonly used sources of case study evidence: documentation, archival records, interviews, direct observations, participant-observation, and physical artifacts.
These categories are useful, but they should not be treated as a mandatory checklist or an exhaustive inventory of every form of contemporary research data.
| Evidence source |
Possible educational technology example |
Important limitation |
| Documentation |
AI policy drafts, meeting minutes, implementation guidelines |
Documents may present official intentions rather than actual practices. |
| Archival records |
Historical enrollment, expenditure, training, or platform records |
Administrative measures may be incomplete or inconsistently defined. |
| Interviews |
Accounts from faculty, administrators, students, and technical personnel |
Accounts may be affected by memory, perspective, and social desirability. |
| Direct observation |
Observing policy meetings, training sessions, or classroom practices |
Observed behavior may not reflect practices at other times. |
| Participant-observation |
Researcher participation in an implementation team |
Researcher involvement may influence events and interpretation. |
| Physical artifacts |
Learning devices, instructional materials, or assessment tools |
Artifacts may not establish how or why they were used. |
Digital traces, software logs, online interactions, and platform configurations may also be relevant to contemporary case studies. Their classification depends on how the records were generated and how they function as evidence.
Does Yin Require Every Case Study to Use All Six Sources?
No. Yin's classification describes important categories of evidence rather than requiring researchers to collect all six.
The appropriate selection depends on the case and research question.
A study of historical policy development may rely heavily on documents, archival records, and interviews because the relevant events can no longer be observed directly.
A study of current classroom technology practices may place greater emphasis on observation, instructional artifacts, and teacher interviews.
Researchers should explain why the selected evidence is sufficient for their analytical objectives and acknowledge limitations created by unavailable sources.
The objective is adequate and coherent evidence, not numerical compliance with a source inventory.
Can Interviews Alone Support a Case Study?
Potentially, although the defensibility of an interview-only case study depends substantially on the research question and methodological tradition.
Suppose the study examines how faculty members interpreted a particular institutional AI policy during its first year. Interviews with appropriately selected participants may provide rich evidence about perceptions, meanings, and experiences.
However, if the study claims to explain how the university formally developed the policy, interviews alone may be insufficient to establish the chronology of decisions, actual policy revisions, or administrative procedures.
Participants may remember events differently or describe institutional processes from limited positions.
In such circumstances, documentary records or other evidence may be necessary to support the intended claims.
Researchers should also consider whether an interview-focused study genuinely investigates a bounded case or whether another qualitative approach better represents its purpose.
The distinction depends on whether case study is the appropriate research design, not simply on how many interviews are conducted.
Are Multiple Participants the Same as Multiple Sources of Evidence?
Not necessarily. Interviewing several participants can provide multiple perspectives, but all the evidence may still come from the same general method.
For example, interviewing twenty faculty members provides variation in participant accounts. Adding administrators and students may broaden the perspectives represented.
However, this is different from combining interviews with documentary records, observations, or administrative data.
Researchers should distinguish diversity of participants from diversity of evidence types.
Neither is inherently sufficient for every case study. The appropriate combination depends on the case, the questions being asked, and the claims being made.
What Is the Difference Between a Source of Evidence and a Data Collection Method?
A source of evidence is where relevant information originates. A data collection method is the procedure used to obtain or record that information.
For example, faculty members may be sources of information, while interviews are the method used to obtain their accounts.
Institutional policy documents are evidence sources that may be collected through document retrieval and analyzed through documentary or qualitative content analysis.
Classroom practices may be examined through direct observation using structured or unstructured recording procedures.
The terms overlap in ordinary methodological writing, but distinguishing them helps researchers explain what the evidence represents and how it was obtained.
Does Using Multiple Sources Automatically Establish Triangulation?
No. Collecting several sources does not automatically mean that they have been meaningfully integrated.
Triangulation involves using different perspectives, methods, investigators, or theoretical approaches to examine a phenomenon, depending on the form of triangulation intended.
For example, a researcher may compare faculty accounts of policy consultation with meeting minutes and records of policy revisions.
However, simply placing interview findings and documentary findings in separate sections does not necessarily demonstrate how they relate to each other.
The researcher should explain what each source contributes, where evidence converges or diverges, and how those relationships affect interpretation.
The detailed design of triangulation within case study research requires more than collecting different kinds of material.
Must Different Sources Agree With One Another?
No. Disagreement among evidence sources may be analytically important.
Suppose administrators report that faculty members were extensively consulted during AI policy development, while faculty interviews indicate that important decisions had already been finalized.
Meeting minutes may confirm that consultation occurred but reveal limited opportunities to revise substantive provisions.
The discrepancy could illuminate differences between formal participation and perceived influence.
Researchers should not automatically treat one source as correct or force all sources into a single consistent narrative.
Instead, they should examine why the accounts differ, what each source can establish, and whether the disagreement reveals an important feature of the case.
Can Institutional Documents Be Treated as Objective Evidence?
Documents may provide valuable evidence about formal decisions, stated policies, and recorded events, but they are not automatically neutral representations of institutional reality.
Meeting minutes may omit disagreements. Official reports may emphasize successful implementation. Policy documents may describe intended procedures rather than actual practice.
Researchers should consider who created the document, for what purpose, when it was produced, and whether it was revised.
For example, a university's AI policy may require faculty consultation, but the policy document alone does not establish that consultation was meaningful or consistently implemented.
Documentary evidence should be interpreted in relation to its provenance, institutional purpose, and other available information.
What Can Observations Establish That Interviews May Not?
Observations can provide evidence about practices, interactions, and events as they occur.
For example, faculty members may report that an AI assessment platform is integrated into everyday teaching. Classroom observations may reveal when and how it is actually used.
However, observation has limitations. Researchers may observe only selected sessions, and participants may alter their behavior when they know they are being observed.
Observations also do not automatically reveal participants' intentions or interpretations.
Interviews and observations may therefore provide complementary evidence, but neither should be assumed to offer unrestricted access to the phenomenon.
How Can Digital Platform Logs Contribute to Case Study Evidence?
Educational technology research increasingly involves digital records such as learning management system activity logs, software usage reports, assessment submissions, and platform configuration histories.
These records may help researchers establish when systems were accessed, how frequently particular functions were used, or how implementation changed over time.
However, digital traces are not transparent measures of learning, engagement, or meaningful technology adoption.
For example, frequent platform logins do not necessarily indicate effective instructional use. Low recorded activity may reflect offline learning, shared accounts, or limitations in logging procedures.
Researchers should examine how the data were generated, whether definitions changed, and which interpretations the records can legitimately support.
Privacy, consent, institutional permissions, and data protection requirements should also be addressed according to the applicable research setting.
What Is a Chain of Evidence?
A chain of evidence allows readers to understand how the researcher's conclusions relate to the research questions, collected data, and analytical procedures.
In Yin's case study approach, maintaining a chain of evidence is an important principle for strengthening the traceability of the inquiry.
For example, a conclusion that faculty consultation influenced AI policy revisions should be traceable to relevant consultation records, policy drafts, participant accounts, and the analysis connecting them.
This does not mean every claim must be supported by every available source.
It means the researcher should preserve a coherent evidentiary basis for the claims and make analytical decisions sufficiently transparent.
A case study database, organized document identifiers, interview records, analytical memos, and decision logs may help maintain this traceability.
How Do You Decide Whether Additional Sources Are Necessary?
Begin by identifying the claims the study intends to make and the evidence needed to support them.
If the question concerns how participants experienced a policy, interviews may be central. If it concerns what decisions were formally made, documentary records may be essential. If it concerns actual classroom practices, observations or other practice-based evidence may be particularly informative.
Additional sources are valuable when they address an important evidentiary gap, examine competing explanations, or provide information unavailable from existing material.
They are less useful when they merely repeat information without improving the analysis.
The decision should consider evidence adequacy, ethical access, resources, and the risk that collecting additional data will expand the case beyond its intended boundaries.
What If Important Sources Are Unavailable?
Case study researchers may encounter restricted institutional records, unavailable observations, incomplete archives, or participants who decline interviews.
These limitations do not automatically invalidate the study, but they may constrain its conclusions.
Suppose confidential committee minutes are unavailable. Researchers may still examine public policy drafts, interviews, and implementation reports.
However, they should avoid claiming to have reconstructed the complete decision-making process when consequential deliberations remain undocumented.
The researcher may need to narrow the research question, qualify explanatory claims, or explicitly identify unresolved interpretations.
Missing evidence should not be replaced with assumptions presented as established facts.
Can Too Many Sources Make a Case Study Weaker?
Potentially. Collecting more evidence than the research question requires can increase analytical complexity without improving the study.
For example, a researcher investigating one university's AI policy development might collect thousands of unrelated platform records, student surveys, and administrative reports.
If these materials are not connected to the unit of analysis, they may distract from the central inquiry.
Large datasets can also create additional privacy, storage, and interpretation responsibilities.
The relevant standard is not maximum data accumulation but sufficient, appropriate, and well-integrated evidence.
Researchers should preserve the boundaries of the case and avoid treating methodological complexity as a substitute for analytical clarity.
How Should Multiple Evidence Sources Be Reported?
The methods section should identify each evidence source, explain why it was selected, and describe how the information was collected and analyzed.
Researchers should clarify the period covered, the relevant participants or records, and any important limitations.
They should also explain how evidence from different sources was related during analysis.
For example, policy documents may have been used to reconstruct the implementation timeline, while interviews examined how stakeholders interpreted the decisions and observations documented actual practices.
The reporting should make clear which conclusions are supported by which evidence and where unresolved disagreements remain.
Multiple sources strengthen a study only when their contributions to the inquiry are analytically meaningful.
