03 · What You Need to Know
Defining Documents and Documentary Evidence in Research
What Makes Something a Document?
A document is generally understood as a recorded object or material that preserves information, communication, representation, or evidence in a form that can be examined.
In social research, documents are not simply containers of information. They are also products of social, organizational, and historical processes.
For example, an institutional policy is produced through particular decision-making procedures, authorized by particular actors, and circulated to particular audiences.
Its significance may lie in what it says, how it is organized, who issued it, or how it is used.
Prior (2003) emphasizes the importance of examining documents in relation to their production and social functions rather than treating them merely as passive sources of information.
The relevant definition should therefore reflect the researcher's analytical purpose.
Documents Do Not Have to Be Printed
Digital materials can qualify as documents when they provide relevant recorded evidence.
Examples include electronic policies, PDF reports, webpages, emails, online announcements, digital correspondence, and exported records from institutional platforms.
For example, a university may communicate AI assessment requirements through an online policy page rather than a printed handbook.
The webpage can be a documentary source for research examining institutional requirements.
However, digital documents may change over time. Researchers should record relevant versions, dates, and retrieval conditions.
The fact that a document exists electronically does not make its contents permanent or stable.
Documents Do Not Have to Be Official
Official documents are only one category of documentary evidence.
Personal correspondence, informal notes, diaries, internal messages, and other unofficial materials may be relevant.
For example, a study investigating how teachers informally coordinate AI-supported instructional activities might analyze collaborative planning notes or correspondence, provided access and use are ethically appropriate.
Such materials may reveal practices that do not appear in official policies.
However, unofficial documents should not automatically be treated as more authentic than formal records.
Both types are produced within particular social circumstances and require contextual interpretation.
What Are Common Types of Research Documents?
The following categories illustrate the range of materials that may qualify as documentary data.
| Document category |
Examples |
Possible research use |
| Policy and regulatory documents |
Institutional policies, official guidelines, regulations |
Examine formal rules and responsibilities |
| Administrative records |
Meeting minutes, memoranda, reports |
Study recorded decisions and procedures |
| Educational materials |
Curricula, syllabi, assessment rubrics |
Examine instructional expectations and representations |
| Personal documents |
Diaries, letters, reflective journals |
Investigate recorded experiences and interpretations |
| Digital communication |
Emails, forum posts, online announcements |
Examine communication and interaction |
| Public communication |
Press releases, websites, public statements |
Study organizational representations |
| Visual documents |
Photographs, posters, diagrams, screenshots |
Analyze visual representations and recorded contexts |
| Historical and archival records |
Archived correspondence, historical reports |
Investigate developments over time |
These categories overlap. The appropriate classification depends on the document's origin, content, and role in the study.
Can Emails and Private Messages Be Research Documents?
Yes. Emails and messages can provide documentary evidence about communication, decisions, relationships, or organizational practices.
For example, correspondence between teachers may reveal how instructional responsibilities are negotiated.
However, private communication raises substantial ethical and privacy considerations.
Researchers should not assume that receiving access to an email account or message archive automatically authorizes the research use of every communication it contains.
Consent, confidentiality, third-party information, and applicable data-protection requirements must be considered.
The methodological relevance of a message does not remove these obligations.
Can Meeting Minutes Be Used as Documents?
Yes. Meeting minutes can provide evidence about recorded decisions, agenda items, responsibilities, and formal organizational communication.
For example, minutes from a university committee may document approval of an AI governance policy.
However, meeting minutes are not necessarily complete transcripts.
They may summarize decisions while omitting disagreements, informal exchanges, or considerations that participants did not formally record.
Researchers should therefore distinguish what the minutes document from what may have occurred during the meeting.
The original recording purpose affects the evidentiary limits of the material.
Can Curriculum Documents and Teaching Materials Be Analyzed?
Yes. Curricula, syllabi, lesson plans, learning modules, assessment rubrics, and instructional guidelines can serve as documentary data.
For example, a researcher may investigate how universities incorporate AI literacy into formal curricula.
The study could analyze course descriptions, learning outcomes, and assessment requirements.
However, the presence of AI literacy in a syllabus does not establish how the subject is taught or what students learn.
Curriculum documents provide evidence about planned or formally represented instruction.
Claims about enacted teaching or learning outcomes require appropriate additional evidence.
Can Websites Be Considered Documents?
Yes. Websites can be analyzed as documentary artifacts when their content or structure is relevant to the research question.
For example, a study may examine how universities publicly communicate their AI governance policies.
However, webpages can be updated, reorganized, or removed.
Researchers should record retrieval dates, relevant URLs, and available version information.
Where appropriate and legally permissible, archived copies or documented captures may support reproducibility.
Researchers should also distinguish the webpage's content from the wider website and explain which portions were included.
Are Social Media Posts Documents?
Social media posts may qualify as documentary data when their recorded content is relevant to the research question.
For example, a researcher might analyze public institutional statements about AI use on an official social media account.
However, social media data introduce issues involving context, platform restrictions, identifiability, consent, and the expectations of content creators.
A publicly accessible post is not necessarily free from ethical concerns.
Researchers should consider whether public social media content can be treated as documentary data and what safeguards are required.
The status of a post as a document does not establish unrestricted permission to collect or republish it.
Can Photographs and Images Count as Documents?
Yes, when the study treats visual materials as recorded evidence.
For example, institutional posters may be analyzed to investigate how universities represent academic integrity or responsible AI use.
Photographs of learning spaces may provide evidence about the visible arrangement of instructional environments.
However, visual documents require an analytical approach appropriate to their content and production context.
A photograph does not provide a complete account of everything occurring outside its frame.
Researchers should also consider consent, copyright, image manipulation, and the identifiability of people depicted.
Can Audio and Video Recordings Be Documents?
Recorded audio and video may be treated as documentary materials in some research traditions.
However, their methodological classification depends on how they were produced and used.
For example, an archived public speech may be analyzed as an existing audiovisual document.
A video recording produced specifically to observe classroom behavior may instead be described as observational data.
Both are recorded materials, but their origins and evidentiary roles differ.
Researchers should describe the recording's production circumstances rather than assume that every recording belongs to the same methodological category.
What About Automatically Generated Digital Records?
Digital systems generate records such as activity logs, timestamps, submissions, and transaction histories.
These may be analyzed as documentary or administrative data, depending on the research design.
For example, learning management system logs may record when students access particular resources.
However, a recorded access event does not establish that the student read or understood the material.
Automatically generated records reflect the system's technical definitions and recording procedures.
Researchers should examine the meaning, completeness, and reliability of these records before treating them as evidence of behavior.
Can Drafts and Unpublished Materials Be Documents?
Yes. Formal publication is not a necessary condition for documentary status.
Draft policies, internal memoranda, unpublished reports, and working documents may provide important evidence.
For example, successive drafts of an AI policy may reveal how institutional language changed during development.
However, drafts should not be confused with final authorized policies.
Researchers should record document status, version, authorship, and relevant dates.
Access and permission requirements may also differ from those applying to public documents.
What Is the Difference Between a Document and a Research Record?
A document is a broad category of recorded material.
A research record refers more specifically to documentation produced or retained in connection with conducting research.
Examples of research records include observation notes, coding decisions, consent documentation, and data-processing logs.
Such records may themselves become documentary data in a study investigating research practices.
However, their role as evidence should be distinguished from their role in documenting the conduct and integrity of a particular project.
The same recorded object can serve different methodological functions in different contexts.
Does a Document Have to Be Complete?
Not necessarily. Fragments, excerpts, and partially preserved records may provide relevant evidence.
For example, an incomplete archival memorandum may contain an important statement about an institutional decision.
However, missing sections can limit interpretation.
Researchers should distinguish what is preserved from what is unknown and avoid reconstructing absent content without evidence.
The treatment of missing and unavailable documentary material should be explicit when it affects the conclusions.
How Do You Distinguish a Document from a Dataset?
Documents and datasets can overlap, but they are not identical concepts.
A document may contain narrative, visual, or structured information, while a dataset typically organizes observations or measurements for analysis.
For example, a university annual report is a document, while a spreadsheet containing enrollment records may be treated as an administrative dataset.
However, the spreadsheet may also be examined as a documentary artifact if the research concerns how institutions record or classify information.
The distinction depends partly on whether the study analyzes the recorded content as evidence about events or examines the artifact's production, categories, and institutional functions.
How Do Documents Differ from Interview Transcripts?
Interview transcripts are recorded texts, but they are usually generated through a research interview rather than independently produced as part of ordinary social activity.
A researcher may analyze an interview transcript as qualitative textual data.
However, that does not automatically make the study documentary research.
The distinction matters because interview data arise through interaction between the researcher and participant, while many documentary sources were produced for purposes unrelated to the study.
Researchers should describe the data-generation process accurately.
The fact that evidence is written or transcribed does not, by itself, determine the research method.
How Do You Decide Which Document Types to Include?
The research question should determine documentary eligibility.
For example, a study of formal institutional AI requirements may include authorized policies and official guidance while excluding informal social media commentary.
A study of public discourse about AI governance may legitimately include both official statements and public reactions.
The inclusion criteria should specify document types, sources, relevant dates, and the relationship to the phenomenon.
Researchers should avoid expanding the corpus merely because additional materials are available.
The procedure for defining a document corpus should provide coherent boundaries for analysis.
How Do Document Boundaries Affect Analysis?
Researchers must decide what constitutes one document and what constitutes a component of a document.
For example, a university website may contain several policy pages, downloadable guidance documents, and frequently asked questions.
Should the entire website count as one document or should each policy page be treated separately?
The answer depends on the research question and unit of analysis.
A study comparing institutional policy systems may treat the collection as one institutional case, while a content analysis may code individual pages or provisions.
Researchers should define these boundaries before calculating frequencies or making comparisons.
What Is the Unit of Analysis in Documentary Research?
The unit of analysis is the entity or level about which the study intends to make conclusions.
It may be an individual document, an institution, a policy provision, a communication episode, or another defined entity.
For example, a study may collect 100 documents from 20 universities but analyze institutional policy approaches at the university level.
In that case, the number of documents is not the same as the number of institutional cases.
The researcher should distinguish the unit of collection, coding unit, and unit of analysis where necessary.
Failure to do so may produce misleading comparisons or inflated counts.
Why Does Document Provenance Matter?
Provenance concerns a document's origin and history.
Researchers should consider who created the material, when it was produced, how it was authorized, and whether it has been modified.
For example, an AI policy downloaded from an official university website may have a different evidentiary status from an unattributed copy circulated through an informal channel.
However, official origin does not guarantee substantive accuracy.
Provenance helps establish what the document is and how it should be interpreted.
Researchers should preserve relevant metadata where feasible and appropriate.
How Do You Evaluate Documentary Credibility?
Documentary credibility concerns the trustworthiness of the information for the intended research purpose.
A press release may credibly establish what an institution publicly claimed while providing incomplete evidence about the actual outcomes of a program.
Meeting minutes may accurately record formal decisions without documenting every discussion preceding them.
Researchers should therefore assess the relationship between the document's original purpose and the claim being investigated.
Bowen (2009) discusses authenticity, credibility, representativeness, and meaning as important considerations in document analysis.
These considerations help prevent researchers from treating recorded statements as transparent accounts of reality.
Can Documents Produced for Different Purposes Be Analyzed Together?
Yes, when the comparison is methodologically justified.
For example, a researcher may compare official policies, implementation guidelines, and public announcements to investigate how an institution communicates AI responsibilities.
However, these documents serve different functions.
A policy establishes formal requirements, while an announcement may summarize them for a broader audience.
Differences in wording may reflect communication purpose rather than substantive disagreement.
Researchers should consider how documents produced for different purposes should be analyzed before treating them as equivalent units.
What Ethical Considerations Apply to Documentary Materials?
Documentary research may involve privacy, confidentiality, copyright, consent, and data-protection obligations.
These concerns are particularly important for private correspondence, institutional records, identifiable photographs, and sensitive digital communications.
Public accessibility does not necessarily remove every ethical consideration.
Researchers should consider whether documents contain information about individuals who did not consent to research participation.
They should also verify applicable institutional and jurisdictional requirements.
Documentary relevance does not establish unrestricted permission to reproduce or publish the material.
How Should Document Types Be Reported?
The methods section should explain what the study treats as a document and why those materials are relevant.
Researchers should identify document categories, sources, dates, eligibility criteria, and relevant version information.
Where materials differ in format or original purpose, describe how those differences were handled analytically.
The report should also clarify whether documents were originally produced for research or existed independently of the study.
Transparent definitions allow readers to understand what evidence was included and what remains outside the corpus.
