Qualitative Data – Types, Examples, Collection and Analysis

Qualitative data is information used to describe meanings, experiences, perceptions, qualities, actions, or contexts. It commonly appears as interview transcripts, written responses, field notes, documents, photographs, audio, video, or online interactions. Researchers analyze it systematically to identify patterns, interpretations, relationships, and explanations that cannot be understood through numerical measurement alone.

Introduction

Qualitative data helps researchers understand how people interpret events, why they behave in particular ways, how social processes develop, and what an experience means in its specific context. It is used across education, psychology, health care, sociology, anthropology, business, design, history, communication, and many other fields.

This guide explains what qualitative data is, what forms it can take, how it differs from quantitative and categorical data, how researchers collect and analyze it, and how they establish trustworthiness. It also covers sampling, ethics, data-management software, artificial intelligence, common mistakes, and reporting standards.

Key Takeaways

• Qualitative data primarily captures meaning, experience, language, behavior, representation, and context.
• Common forms include text, audio, video, images, documents, observations, and digital interactions.
• Qualitative research data is broader than nominal and ordinal categorical variables.
• Coding organizes data, but analysis also requires interpretation, comparison, and explanation.
• Sample adequacy depends on the study aim, participants, data quality, theoretical framework, and analytical approach.
• Software and AI can assist with data management, but researchers remain responsible for interpretation, ethics, and methodological decisions.

What Is Qualitative Data?

Qualitative data is descriptive or interpretive information that represents qualities, meanings, experiences, perceptions, actions, relationships, or contexts. It is usually communicated through language, images, sounds, observations, or records rather than being reduced primarily to numerical measurements.

Examples include:

• A transcript of a student describing difficulties with online learning
• Field notes documenting interactions in a hospital waiting room
• An open-ended survey response explaining why a customer cancelled a subscription
• A photograph showing how people use a public space
• A diary describing changes in a patient’s daily experience
• An audio recording of a focus-group discussion
• Social-media posts responding to a public event
• Historical letters examined for recurring ideas and social assumptions

Qualitative data is not merely “anything without numbers.” It becomes useful research evidence when it is collected, documented, interpreted, and connected to a clear research question.

A simple definition

Qualitative data describes what something is like, how it is experienced, what it means, or why it occurs in a particular context.

Quantitative data is generally used to answer questions such as “how many,” “how much,” or “how strongly.” Qualitative data is especially useful for questions such as “how,” “why,” “in what way,” and “what does this mean to the people involved?”

What Makes Data Qualitative?

Qualitative data usually has several of the following characteristics.

It preserves meaning and context

A numerical satisfaction score may show that a participant selected 3 out of 10. A follow-up interview can reveal that the low score resulted from confusing instructions, long waiting times, fear of complaining, or previous negative experiences.

The explanation supplies context that the score cannot provide by itself.

It is often detailed and open-ended

Participants can use their own language rather than selecting only from predetermined response options. This may reveal ideas that the researcher did not anticipate.

It may be produced by participants or researchers

Participants produce qualitative data through interviews, diaries, stories, photographs, drawings, messages, and written responses. Researchers also produce data through field notes, observation records, analytic memos, maps, and descriptions of interactions.

It may be structured, semi-structured, or unstructured

An interview transcript is not automatically unstructured. A researcher may use the same semi-structured interview guide with every participant. A document collection may be organized systematically by date, source, or genre.

The difference between structured and unstructured data concerns organization, not whether the material is qualitative.

It requires interpretation

Researchers make decisions about what to examine, how to segment the material, which codes to apply, what patterns matter, and how findings relate to the research question.

Interpretation does not make the process arbitrary. Good qualitative research makes these decisions transparent and supports interpretations with evidence.

It can be analyzed iteratively

Data collection and analysis often influence one another. Early observations may suggest new interview prompts, reveal missing participant perspectives, or identify issues that require further investigation.

Forms and Types of Qualitative Data

There is no single universally accepted list of qualitative-data types. A practical approach is to classify the material by its form and origin.

Form	Examples	What it may reveal
Textual data	Transcripts, diaries, essays, survey comments, emails, reports	Experiences, explanations, beliefs, arguments and language use
Audio data	Interviews, focus groups, oral histories, soundscapes	Speech, pauses, tone, interaction and environmental sound
Visual data	Photographs, drawings, maps, advertisements, artwork	Representation, spatial relationships, symbolism and material conditions
Audiovisual data	Videos, recorded meetings, classroom recordings	Speech, movement, gesture, interaction and setting
Observational data	Field notes, behavior records, ethnographic journals	Practices, routines, interactions and context
Documentary data	Policies, minutes, court records, letters, archives	Institutional assumptions, historical change and official discourse
Digital data	Forum posts, social-media discussions, chat logs, website content	Online communities, interaction patterns, identity performance and public discourse
Participant-generated data	Diaries, photographs, timelines, drawings, stories	Participant priorities and personally meaningful experiences

Textual data

Text is one of the most frequently analyzed forms of qualitative data. It includes interview transcripts, focus-group transcripts, written reflections, open-ended survey responses, policy documents, clinical notes, essays, correspondence, and online comments.

Textual analysis can examine both what people say and how they say it.

Audio and spoken data

Audio recordings preserve tone, pace, pauses, emphasis, interruptions, laughter, hesitation, and other features that may disappear in a basic transcript.

Researchers should decide whether these features are relevant before choosing a transcription convention. A study of conversational interaction may require detailed transcription, while a descriptive study may use a less granular verbatim transcript.

Visual and audiovisual data

Photographs, videos, diagrams, maps, advertisements, artwork, and screen recordings can be treated as data. Researchers may analyze composition, symbols, movement, spatial arrangements, facial expressions, or relationships between speech and action.

Visual material also creates ethical challenges because removing identities from faces, voices, locations, and backgrounds may be difficult.

Observational data

Observation records what people do in a setting rather than relying only on what they later report. Field notes may include:

• Descriptions of events
• Physical settings
• Interaction patterns
• Researcher reflections
• Emerging questions
• Methodological decisions

Researchers should distinguish direct description from later interpretation in their notes.

Documentary and archival data

Documents can be studied as evidence of organizational practices, public narratives, institutional priorities, historical changes, or social assumptions.

A document was usually produced for a purpose other than the current research. Researchers therefore need to consider who created it, for whom, under what conditions, and what may have been omitted.

Qualitative Research Data and Categorical Data Are Not Identical

The word qualitative is used differently in introductory statistics and qualitative research.

Qualitative variables in statistics

A statistics textbook may use “qualitative data” to mean categorical data. Such variables place cases into labels or groups.

The main levels are:

• Nominal data: categories without an inherent order, such as blood group, department, or transport mode.
• Ordinal data: ordered categories whose intervals are not assumed to be equal, such as low, medium and high satisfaction.

These variables can be summarized through counts, percentages, modes, contingency tables, and suitable statistical tests.

Qualitative data in research methodology

Qualitative research often analyzes richer material:

• Entire interview accounts
• Stories and personal narratives
• Sequences of interaction
• Photographs or videos
• Field observations
• Institutional documents
• Online conversations

A researcher may transform part of this material into categorical codes, but the original context and meaning usually remain important.

Why the distinction matters

Consider two datasets about university students:

Dataset A:
Study mode: online, hybrid or in person.

Dataset B:
Interview transcripts explaining how students experience online, hybrid and in-person learning.

Dataset A is categorical. Dataset B is rich qualitative material. Both can be described as qualitative in certain contexts, but they support different questions and analytical procedures.

Can Qualitative Data Contain Numbers?

Yes. A qualitative study can contain numbers without becoming a quantitative study. What matters is the purpose of the data and the logic of analysis, not whether every number has been removed.

A qualitative interview transcript may contain:

• A participant’s age
• Dates and times
• A salary mentioned during an interview
• A pain rating
• The number of years someone has worked in a profession
• A participant identification number

Researchers may also count how many extracts received a code. Such counts can help manage the dataset, but frequency alone does not establish importance.

A rare comment may reveal a critical negative case. A frequently repeated statement may be superficial. Interpretation must therefore consider relevance, context, intensity, variation, and the research question.

Examples of Qualitative Data

Education example

A researcher asks first-year students:

“Describe a situation in which you felt supported or unsupported by your university.”

The responses may reveal themes concerning teacher availability, financial pressure, language barriers, peer relationships, administrative communication, or belonging.

Health-research example

Patients are interviewed about living with chronic pain. Their accounts may describe uncertainty, changes in identity, difficulties at work, family expectations, coping practices, and interactions with health professionals.

Psychology example

Participants keep diaries about their experiences of social anxiety. The researcher examines triggers, self-evaluations, avoidance, bodily sensations, and coping strategies.

Business example

Customers explain why they stopped using an application. Their comments may identify confusing navigation, hidden costs, poor onboarding, missing features, or a lack of trust.

User-experience example

A researcher observes users attempting to complete an online application. Notes record where users hesitate, misinterpret labels, abandon a step, ask for help, or create workarounds.

Humanities example

A historian examines letters, newspapers, photographs, and diaries to understand how migration was represented and experienced during a particular period.

Digital-research example

Researchers analyze discussions in an online community to understand how members develop norms, provide support, challenge misinformation, or negotiate identity.

Qualitative Data vs. Quantitative Data

Qualitative data primarily explains meaning, experience and context, whereas quantitative data represents quantities, measurements and numerical relationships. Neither is automatically superior; the appropriate choice depends on the research question.

Feature	Qualitative data	Quantitative data
Main purpose	Explore meaning, process, context and experience	Measure variables, test relationships and estimate quantities
Typical questions	How? Why? In what way? What does it mean?	How many? How much? How often? How strongly?
Common forms	Text, images, audio, video, observations and documents	Measurements, scores, counts, rates and numerical variables
Common sampling	Purposive, theoretical, criterion, maximum-variation or snowball sampling	Probability or non-probability samples designed for estimation or testing
Common collection	Interviews, observations, focus groups, documents and open-ended questions	Experiments, structured surveys, tests, sensors and administrative datasets
Analysis	Coding, comparison, interpretation, themes, narratives and discourse	Descriptive and inferential statistics, modeling and estimation
Typical output	Themes, categories, cases, narratives, models and interpretations	Tables, coefficients, estimates, distributions, tests and effect sizes
Generalization	Often analytical, theoretical, contextual or transfer-oriented	Often statistical, when design and sampling support it

When mixed methods is useful

Mixed methods combines qualitative and quantitative evidence in a planned design.

For example:

1. Survey data may identify a decline in student satisfaction.
2. Interviews may explain why satisfaction declined.
3. Findings may be integrated to develop and evaluate an intervention.

Mixed methods should involve genuine integration rather than placing unrelated numerical and textual findings in the same report.

When Should Qualitative Data Be Used?

Qualitative data is especially useful when the research aims to:

• Explore a poorly understood issue
• Understand lived experience
• Investigate a social or organizational process
• Learn how people interpret an event
• Examine language, discourse or representation
• Develop concepts or theory
• Explain unexpected quantitative findings
• Study behavior in a natural setting
• Understand implementation and context
• Generate hypotheses or improve survey instruments
• Evaluate how and why a program works differently across settings

Qualitative data is less suitable when the only objective is to estimate prevalence, calculate an average, determine a population percentage, or test a precisely specified numerical relationship.

How Is Qualitative Data Collected?

Qualitative data is commonly collected through interviews, focus groups, observations, open-ended surveys, diaries, documents, visual methods, digital sources and participant-generated materials. The method should be selected according to the research question, setting, ethical risks and analytical approach.

Interviews

Interviews allow researchers to explore individual experiences, interpretations, decisions, and explanations.

They may be:

• Structured
• Semi-structured
• Unstructured
• Narrative
• Episodic
• Life-history based
• Conducted in person, by telephone, or online

Semi-structured interviews are common because they provide consistent topic coverage while allowing follow-up questions.

A good interviewer avoids forcing participants into the researcher’s assumptions and uses prompts such as:

• “Can you tell me more about that?”
• “What happened next?”
• “What did that mean to you?”
• “Can you give an example?”
• “Was your experience always like this?”

Focus groups

Focus groups collect data through guided group discussion. They are useful for studying shared norms, disagreement, collective language, and interaction.

The resulting data is not simply a set of individual interviews conducted at the same time. Researchers should analyze how participants respond to, reinforce, challenge, or modify one another’s statements.

Focus groups may be unsuitable when participants face strong power imbalances, the topic is highly sensitive, or confidentiality cannot be reasonably protected.

Observation

Researchers may observe behavior, interactions, routines, spaces, meetings, classrooms, clinics, workplaces, or online communities.

Observation can be:

• Participant or non-participant
• Structured or open
• Overt or, in limited ethically justified contexts, unobtrusive
• Short term or ethnographically extended

Field notes should record what was observed, when and where it occurred, who was present, and which reflections are the researcher’s interpretations.

Open-ended survey questions

Open-ended questions allow respondents to answer in their own words. They can collect useful qualitative data at scale, but short survey comments usually provide less depth than interviews.

A weak question asks:

“Any comments?”

A stronger question asks:

“What was the most difficult part of the application process, and what change would have improved it?”

Diaries, journals and experience sampling

Participants can record events, emotions, decisions, or activities close to the time they occur. Diaries reduce reliance on distant recall but require sustained participant effort.

Documents and existing records

Researchers may analyze:

• Policies
• Meeting minutes
• Reports
• court judgments
• News articles
• Personal correspondence
• Historical archives
• Clinical records
• Curricula
• Advertisements
• Organizational communications

The researcher should evaluate authenticity, provenance, purpose, audience, missing material, and access conditions.

Visual and creative methods

Participants may be invited to create photographs, drawings, maps, collages, timelines, or videos. These methods can help participants communicate experiences that are difficult to express verbally.

Researchers must not assume that an image has one self-evident meaning. Participant explanation, cultural context, production conditions, and researcher interpretation all matter.

Digital and online data

Digital qualitative data may include forum discussions, comments, chat logs, websites, livestreams, online meetings, virtual-world interactions, and social-media posts.

Public accessibility does not automatically eliminate ethical obligations. Researchers should consider platform expectations, vulnerable users, identifiability, quotation searchability, consent, terms of service, and institutional review requirements.

Sampling and Sample Adequacy

Qualitative sample size should be justified by the study’s purpose and the adequacy of the resulting information, not by one universal numerical rule.

Common qualitative sampling strategies

Purposive sampling

Participants or cases are selected because they can provide relevant information about the phenomenon.

Criterion sampling

All selected cases meet a defined criterion, such as having completed a particular program or experienced a particular event.

Maximum-variation sampling

Researchers deliberately include diverse cases to examine commonalities and contrasts across relevant characteristics.

Homogeneous sampling

A relatively specific group is selected when the study requires detailed examination of a shared experience.

Theoretical sampling

In grounded theory, sampling decisions develop alongside analysis to elaborate emerging categories and theoretical relationships.

Snowball sampling

Participants help identify other eligible participants. This can support access to difficult-to-reach groups but may reproduce existing social-network biases.

Information power

Malterud, Siersma, and Guassora (2016) proposed that a sample may require fewer participants when it has greater information power. Relevant considerations include:

• A narrow, clearly defined study aim
• A highly specific sample
• Strong use of relevant theory
• Rich, high-quality dialogue
• An intensive case-oriented analysis

A broad aim, heterogeneous sample, weak interviews, or extensive cross-group comparison may require more data.

Saturation

Saturation is often described as the point at which further collection produces little relevant new information. However, researchers should specify what they mean.

• Code saturation: few or no new codes are appearing.
• Meaning saturation: the researcher has developed a sufficiently rich understanding of the codes and their variation.
• Theoretical saturation: emerging theoretical categories and relationships have been sufficiently elaborated for the chosen grounded-theory approach.

Saturation is not appropriate for every qualitative methodology. Researchers using reflexive thematic analysis, narrative inquiry, case study, or phenomenological approaches should follow guidance appropriate to that method rather than automatically claiming saturation.

How to Analyze Qualitative Data

Qualitative data analysis generally involves preparing and becoming familiar with the material, coding relevant features, developing categories or patterns, testing interpretations against the dataset, and presenting an evidence-based explanation. The exact process depends on the chosen methodology and research question.

Step 1: Clarify the research question and analytical approach

Before coding, specify what the study is trying to understand.

Compare:

• “What difficulties do students report?”
• “How do international students construct a sense of belonging?”
• “How does institutional language assign responsibility for student failure?”
• “How do students narrate changes in identity across their first year?”

These questions require different analytical attention.

Step 2: Prepare and document the data

Preparation may include:

• Transcribing recordings
• Checking transcripts against audio
• Correcting obvious errors
• Standardizing filenames
• Removing or replacing identifiers
• Recording source metadata
• Creating a data inventory
• Documenting exclusions
• Importing files into analysis software
• Keeping protected original files separate from working copies

Transcription is itself an interpretive decision. Researchers should document whether pauses, nonverbal sounds, overlapping speech, dialect, and grammatical features were retained or simplified.

Step 3: Become familiar with the dataset

Read, watch, or listen to the material more than once. Record early observations without treating them as final findings.

Questions may include:

• What appears important to participants?
• Where do accounts agree or differ?
• What assumptions are being made?
• What is surprising?
• Which contextual details affect interpretation?
• What is absent or difficult to discuss?
• How might my expectations influence what I notice?

Step 4: Code relevant material

A code is a concise label attached to a segment of data because that segment is relevant to the research question.

Codes may describe:

• An action
• Experience
• Belief
• Emotion
• Process
• Barrier
• Strategy
• Interaction
• Contradiction
• Assumption
• Contextual condition

Coding may be:

• Inductive: developed primarily through engagement with the dataset.
• Deductive: informed by existing theory, prior research, an evaluation framework, or predefined questions.
• Semantic: focused on explicit meaning.
• Latent: focused on underlying assumptions, concepts, or structures.
• Descriptive: summarizes what is being discussed.
• Interpretive: proposes what the extract indicates or accomplishes.

Most real projects combine these tendencies to some degree. Researchers should describe their actual process rather than applying a label inaccurately.

Step 5: Compare codes and cases

Examine:

• Similarities and differences
• Repeated processes
• Contradictory accounts
• Relationships between conditions and outcomes
• Variation across participants or settings
• Changes over time
• Negative or deviant cases
• Connections to theory

This stage moves beyond isolated labels toward an analytical understanding.

Step 6: Develop categories, themes, narratives or explanations

The appropriate output depends on the method.

A theme should usually communicate a meaningful pattern or central organizing idea, not merely name a broad interview topic.

For example:

• Weak topic label: University support
• Stronger analytical theme: Support became visible only after students reached a crisis

The second communicates an interpretation that can be examined against the data.

Step 7: Review the developing analysis

Ask:

• Is each finding supported by sufficient evidence?
• Have contradictory cases been considered?
• Are themes distinct but related?
• Does the analysis answer the research question?
• Have important contextual differences been flattened?
• Are claims broader than the data can support?
• Does the analysis reflect the chosen methodology?

Step 8: Keep reflexive and methodological records

Maintain records of:

• Coding decisions
• Changes to definitions
• Theme revisions
• Sampling decisions
• Questions for further collection
• Disagreements among team members
• Researcher assumptions
• Software or AI use
• Reasons for excluding material

These records strengthen transparency and help researchers explain how conclusions developed.

Step 9: Write the findings as an argument

A qualitative findings section should not be a collection of quotations.

A strong section:

1. States the analytical claim.
2. Explains what the claim means.
3. Provides carefully selected evidence.
4. Interprets that evidence.
5. Shows variation or exceptions.
6. Connects the finding to the research question.
7. Avoids claiming that one quotation represents every participant.

Major Qualitative Analysis Methods

Method	Main purpose	Suitable data and questions	Typical output
Thematic analysis	Identify and interpret patterned meaning	Broad range of textual, visual or multimodal data	Themes explaining important patterns
Qualitative content analysis	Systematically classify and interpret content	Documents, responses, media, transcripts and records	Categories, meanings and sometimes descriptive counts
Grounded theory	Develop theory grounded in iterative data collection and comparison	Interviews, observations, documents and process-focused data	Conceptual categories and an explanatory theory
Framework analysis	Organize data in a structured matrix for applied questions	Policy, health, evaluation and multidisciplinary research	Cases-by-themes matrix and applied findings
Narrative analysis	Examine how stories are constructed and what they accomplish	Life stories, interviews, diaries, testimonies and biographies	Narrative structures, identities, turning points and plots
Discourse analysis	Examine how language produces meanings, identities and social realities	Speeches, media, policy, interviews and institutional talk	Discourses, rhetorical practices and power relations
Conversation analysis	Study the detailed organization of naturally occurring interaction	Audio or video of real conversations	Turn-taking, repair, sequencing and interactional practices
Phenomenological analysis	Explore how a phenomenon is experienced and understood	Detailed first-person accounts	Interpretations of lived experience
Case-study analysis	Develop an in-depth understanding of a bounded case	Multiple sources concerning one or more cases	Integrated case descriptions and cross-case explanations

Thematic analysis

Thematic analysis identifies and interprets patterned meaning across a dataset. Researchers should state which form of thematic analysis they used because coding-reliability, codebook, and reflexive approaches rely on different assumptions and quality practices.

Braun and Clarke’s reflexive approach treats themes as developed through active researcher engagement rather than discovered as fixed objects hidden in the data.

Qualitative content analysis

Qualitative content analysis systematically organizes material into categories. It may be more descriptive or more interpretive, depending on the research aim and approach.

Counts may be included, but the analysis should explain what categories mean in context rather than treating frequency as the only evidence.

Grounded theory

Grounded theory aims to generate theory through iterative collection, coding, constant comparison, memo writing, and theoretical sampling.

Researchers should avoid calling any inductive coding exercise “grounded theory.” Grounded-theory traditions have specific procedures and philosophical differences.

Framework analysis

Framework analysis is especially useful in applied research with predetermined objectives, multiple stakeholder questions, or team-based analysis.

Data is commonly summarized in a matrix so that researchers can compare themes across cases while retaining links to the source material.

Narrative analysis

Narrative analysis examines how people organize events into stories. It may consider sequence, turning points, characters, audience, cultural storylines, identity, and how a particular version of events is produced.

Fragmenting every transcript into small thematic codes may be inappropriate when the integrity of each story is analytically important.

Discourse and conversation analysis

Discourse analysis investigates how language constructs social reality, categories, identities, responsibility, normality, or authority.

Conversation analysis studies naturally occurring interaction in fine detail. It generally requires recordings and specialized transcription rather than relying only on researcher summaries.

Worked Coding Example

Imagine a study asking:

“How do postgraduate students experience academic feedback?”

A participant says:

“The comments were detailed, but I received them after the next assignment was due. I understood what I had done wrong, but I could not use the advice when it mattered.”

Possible initial codes

• Detailed feedback
• Feedback received too late
• Understanding mistakes
• Unable to apply advice
• Timing reduces usefulness

Possible category

Barriers to using feedback

Possible theme

Feedback was informative but disconnected from the moment of decision

Interpretation

The issue is not simply that feedback quality was low. The participant regarded the comments as useful in content but ineffective in practice because institutional timing prevented application.

This example illustrates the difference between:

• Quoting the participant
• Summarizing the statement
• Coding relevant features
• Grouping related codes
• Developing an analytical claim

How to Establish Trustworthiness

Trustworthiness is strengthened when researchers show clearly how data was produced, how interpretations were developed, how alternatives were considered, and how conclusions are supported by evidence. No single technique guarantees quality.

Common quality concepts include credibility, transferability, dependability, and confirmability, although terminology varies across qualitative traditions.

Reflexivity

Reflexivity involves examining how the researcher’s background, position, relationships, assumptions, theoretical commitments, and decisions influence the study.

A reflexive account may consider:

• Why the topic was selected
• The researcher’s relationship to participants
• Insider or outsider status
• Power differences
• Expectations before collection
• Emotional responses
• How interpretations changed
• How the research setting influenced what could be said

Triangulation

Triangulation may compare:

• Multiple data sources
• Different participant groups
• Interviews and observations
• Documents and spoken accounts
• Multiple researchers
• Different theoretical perspectives

Triangulation does not necessarily identify one final objective truth. It may reveal complementary, contradictory, or context-dependent perspectives.

Negative-case analysis

Researchers actively examine material that does not fit the developing explanation. A strong analysis explains variation rather than hiding inconvenient examples.

Audit trail

An audit trail documents how the project developed. It may include versions of the codebook, memos, theme maps, meeting notes, sampling decisions, and records of analytical changes.

Participant engagement or member reflection

Researchers may invite participants to comment on transcripts, summaries, interpretations, or practical implications.

This process should not be treated as a simple accuracy test. Participants may change their views, disagree with one another, or interpret the research purpose differently. Their responses become additional evidence to consider.

Thick contextual description

Detailed description helps readers evaluate whether findings may be meaningful in another setting. Researchers should explain the participants, institutional context, relevant social conditions, collection process, and limits of transfer.

Team-based analysis

A research team may compare interpretations, refine code definitions, challenge assumptions, and discuss disagreements.

Not every methodology requires numerical inter-coder reliability. In reflexive approaches, differences between researchers can support reflection rather than being treated only as errors requiring elimination.

Advantages of Qualitative Data

Rich contextual understanding

Qualitative data can show how an outcome developed, why it matters, and how it is experienced in a particular setting.

Participant language and perspective

Open-ended methods allow participants to express priorities in their own terms rather than responding only to researcher-defined categories.

Flexibility

Researchers can refine questions, follow unexpected issues, and investigate emerging patterns.

Attention to process

Qualitative research can examine how practices, identities, decisions, relationships, and institutions change over time.

Suitability for complex or sensitive topics

Detailed methods can reveal uncertainty, contradiction, ambivalence, and experiences that are difficult to represent through fixed-choice questions.

Theory and instrument development

Qualitative findings can generate concepts, improve survey questions, identify mechanisms, and inform later quantitative research.

Explanation of numerical patterns

Interviews, observations, and documents can help explain why a program produced different outcomes in different settings or why a statistical pattern occurred.

Limitations and Practical Challenges

Collection and analysis can require substantial time

Recruitment, interviewing, transcription, checking, coding, memo writing, comparison, and interpretation can be labor-intensive.

Findings are not automatically statistically generalizable

A purposive interview sample is usually not designed to estimate how common a view is in an entire population.

Researchers may instead pursue theoretical, analytical, contextual, or transfer-oriented forms of inference.

Researcher influence cannot be eliminated completely

Interview questions, relationships, theoretical assumptions, and analytical choices shape the study. The appropriate response is reflexivity and transparency rather than pretending the researcher is absent.

Large volumes of data can become difficult to manage

A relatively small number of interviews can produce hundreds of transcript pages. Clear file organization, code definitions, memos, and data-management procedures are essential.

Translation and transcription may alter meaning

Translation can change idioms, emotion, ambiguity, humor, and culturally specific concepts. Researchers should document who translated the data, how choices were reviewed, and which language was used for analysis.

Confidentiality may be difficult to protect

A quotation can reveal a participant through occupation, location, life history, or distinctive wording even after a name is removed.

Qualitative findings may be misused

Organizations may select dramatic quotations that support a preferred conclusion while ignoring contradictory evidence. Findings should be presented with context and variation.

Ethics, Privacy and Data Management

Qualitative datasets frequently contain sensitive and indirectly identifying information. Data management should therefore be built into the study design.

Obtain meaningful consent

Participants should understand:

• What will be recorded
• How the data will be used
• Who can access it
• Whether quotations may be published
• Whether data may be archived or reused
• How long files will be retained
• Limits to confidentiality
• How withdrawal will be handled

Consent for participation, analysis, publication, and future sharing may require separate explanation.

Minimize unnecessary collection

Do not collect names, exact addresses, employer details, photographs, or other identifiers unless they are necessary for the research aim.

Separate identifiers from research files

Store contact details and consent records separately from transcripts and analytic files. Use participant codes or pseudonyms in working documents.

Anonymize with context in mind

Deleting a name may not be enough. A combination such as profession, small town, rare diagnosis, and specific event may identify a participant.

Researchers may need to:

• Generalize locations
• Replace organizations
• Remove third-party details
• Alter non-essential characteristics
• Paraphrase highly searchable online quotations
• Blur faces or backgrounds
• Distort voices when ethically appropriate

Any alteration should preserve the analytical meaning as far as possible and be documented.

Protect original recordings

Audio and video may be more identifying than transcripts. Use encryption, access controls, approved institutional storage, secure transfer, and a documented deletion or retention schedule.

Document the dataset

Record:

• File names and formats
• Data sources
• Collection dates
• Transcription conventions
• Anonymization decisions
• Versions
• Exclusions
• Software used
• Data transformations
• Access restrictions

Good documentation supports analysis, collaboration, archiving, and responsible reuse.

Qualitative Data Analysis Software

Computer-assisted qualitative data analysis software, often called CAQDAS, helps researchers organize and retrieve material. It does not choose a defensible methodology or produce trustworthy interpretation automatically.

Common tools include:

• NVivo
• ATLAS.ti
• MAXQDA
• Dedoose
• Quirkos
• Taguette

Depending on the program, features may include:

• Importing text, audio, video, images and PDFs
• Coding and retrieving extracts
• Managing code hierarchies
• Writing memos
• Linking cases and attributes
• Comparing groups
• Searching text
• Creating matrices
• Visualizing relationships
• Supporting team collaboration
• Exporting reports

Do researchers need specialized software?

Not always.

A small project may be analyzed using:

• A word processor
• A spreadsheet
• Printed transcripts
• Index cards
• A reference manager
• A secure note-taking system

Specialized software becomes more useful when the project includes many files, multiple data types, repeated comparison, team coding, extensive metadata, or complex retrieval needs.

The software organizes analytical work; it does not replace the researcher’s responsibility to interpret the data.

Artificial Intelligence and Qualitative Data

AI can assist with selected qualitative-research tasks, but it should not be treated as an independent analyst. Researchers remain responsible for confidentiality, methodological alignment, contextual interpretation, verification, and disclosure.

Potential uses

AI-supported tools may help with:

• Draft transcription
• Speaker separation
• Translation drafts
• Removing formatting inconsistencies
• Locating passages related to a concept
• Suggesting provisional descriptive codes
• Comparing code definitions
• Clustering similar extracts
• Summarizing long documents
• Checking whether a report explains its procedures
• Creating tables from an approved analysis

Important risks

Confidentiality

Uploading raw interviews to a public or unapproved AI service may expose personal, confidential, copyrighted, commercially sensitive, or legally protected material.

Researchers should follow institutional ethics approval, consent terms, contracts, data-protection law, funder requirements, and approved data-processing arrangements.

Fabricated or distorted output

Generative systems may invent quotations, omit qualifiers, merge participants, alter wording, or provide an interpretation unsupported by the source.

Every output must be checked against the original data.

Loss of context

A model may ignore interaction, chronology, silence, irony, cultural meaning, body language, or the relationship between speaker and setting.

Bias

AI outputs may reproduce linguistic, cultural, social, or demographic biases in training data and system design.

Premature closure

Rapidly generated categories can encourage researchers to accept familiar patterns before becoming deeply familiar with the material.

Unclear methodological role

An AI-generated list of topics is not automatically thematic analysis, grounded theory, discourse analysis, or another established method.

A responsible AI-assisted workflow

1. Confirm that the tool is permitted for the data.
2. Remove or protect identifying and confidential information.
3. Define the specific task assigned to the tool.
4. Retain the original data and outputs.
5. Check every summary, code and quotation.
6. Compare AI suggestions with human reading.
7. Search for missing, contradictory and minority cases.
8. Keep a record of prompts, model or software version, dates and researcher changes.
9. Explain material AI use in the methods or acknowledgments section.
10. Ensure that researchers—not the tool—make and defend the final interpretation.

Emerging studies suggest that AI may accelerate parts of coding and theme development, but also show problems involving fragmentation, shallow latent interpretation, and unclear theme boundaries. These findings should be treated as developing evidence rather than universal validation.

Qualitative Data in Modern Research

Qualitative data now extends well beyond typed interview transcripts.

Modern projects may combine:

• Online interviews
• Digital ethnography
• Social-media content
• Mobile diaries
• Screen recordings
• Geotagged photographs
• Sensor information accompanied by participant explanations
• Collaborative digital whiteboards
• Virtual-reality interactions
• Large document collections
• Multilingual transcripts
• Machine-assisted retrieval
• Mixed-methods dashboards

These developments create opportunities to study interaction and context at greater scale. They also increase risks involving platform consent, re-identification, intellectual property, algorithmic bias, data jurisdiction, and the loss of meaning through automated processing.

Researchers should not assume that “more data” creates a stronger qualitative study. Data relevance, depth, quality, ethical integrity, and analytical coherence remain more important than volume alone.

Common Mistakes

Treating every non-numerical value as rich qualitative data

A category label and a detailed interview account are not analytically equivalent.

Calling nominal and ordinal variables the only types of qualitative data

That classification is useful in statistics but does not represent the full range of qualitative research materials.

Using an interview without identifying an analytical method

Researchers must explain how they moved from recordings or transcripts to findings.

Presenting codes as final themes

Codes identify relevant features. Themes should communicate a broader pattern or organizing concept.

Creating themes that merely repeat interview questions

A heading such as “Challenges” may organize material but provides little analysis. Strong themes explain what kind of challenge exists, how it operates, or why it matters.

Assuming the most frequent response is the most important

Importance depends on the research question, context, consequences, variation, and explanatory value—not frequency alone.

Using quotations without interpretation

A quotation is evidence, not a substitute for analysis.

Claiming objectivity by removing reflexive discussion

Qualitative quality is often strengthened when researchers explain their role and decisions.

Selecting a sample-size number without methodological justification

A number borrowed from another study may not fit a different aim, participant group, data-collection method, or analytical approach.

Uploading confidential transcripts to unapproved AI tools

Convenience does not override consent, ethics, privacy, or data-protection requirements.

Assuming software performed the analysis

Software stores, retrieves, connects, and displays data. Researchers still make the conceptual and interpretive decisions.

How to Report Qualitative Data

A qualitative report should provide enough information for readers to understand what was done and evaluate the resulting claims.

Report the research approach

State:

• The research question
• Qualitative methodology or design
• Theoretical or epistemological position when relevant
• Why the approach was appropriate

Explain sampling and context

Describe:

• Eligibility criteria
• Recruitment
• Sampling strategy
• Sample characteristics
• Study setting
• Sample-adequacy rationale
• Relevant non-participation or attrition

Explain data collection

Report:

• Who collected the data
• Interview or observation procedures
• Location or digital platform
• Recording methods
• Duration and timing
• Topic guide development
• Field notes
• Changes made during collection
• Researcher-participant relationship where relevant

Explain analysis transparently

Include:

• The analytical method
• Whether coding was inductive, deductive or combined
• Semantic or latent emphasis where relevant
• How codes and findings were developed
• Team roles
• Software used
• Reflexive practices
• Treatment of disagreements
• Consideration of negative cases
• How quotations were selected
• Any AI-assisted procedures

Present findings with evidence and variation

Use:

• Clear analytical headings
• Explanatory text
• Relevant quotations or visual evidence
• Participant identifiers that protect privacy
• Examples of differences and exceptions
• Connections between themes
• Appropriate limitations

Use a relevant reporting guideline

Depending on discipline and design, useful resources include:

• JARS–Qual for qualitative research in psychology and related fields
• SRQR for broad qualitative-reporting guidance
• COREQ for studies using interviews and focus groups, particularly in health research

A checklist should support transparent reporting rather than force every qualitative methodology into an identical structure.

Conclusion

Qualitative data captures meanings, experiences, actions, language, representations, and contexts that cannot be understood adequately through measurement alone. It may consist of text, audio, video, images, observations, documents, or digital interactions.

Strong qualitative research does more than collect open-ended responses. It aligns the research question, sampling, collection, analysis, ethics, and reporting; preserves context; documents interpretation; considers contradictory evidence; and explains the limits of the findings. Software and AI can support this work, but methodological judgment and responsibility remain with the researcher.

References

• American Psychological Association. (n.d.). Reporting your research: How to use APA Style Journal Article Reporting Standards (JARS). The official resource includes JARS–Qual guidance.
• Bingham, A. J. (2023). From data management to actionable findings: A five-phase process of qualitative data analysis. International Journal of Qualitative Methods, 22. https://doi.org/10.1177/16094069231183620
• Braun, V., & Clarke, V. (2021). One size fits all? What counts as quality practice in (reflexive) thematic analysis? Qualitative Research in Psychology, 18(3), 328–352. https://doi.org/10.1080/14780887.2020.1769238
• Braun, V., & Clarke, V. (2022). Thematic analysis: A practical guide. SAGE.
• Guest, G., Namey, E., & Chen, M. (2020). A simple method to assess and report thematic saturation in qualitative research. PLOS ONE, 15(5), e0232076. https://doi.org/10.1371/journal.pone.0232076
• Hennink, M. M., Kaiser, B. N., & Marconi, V. C. (2017). Code saturation versus meaning saturation: How many interviews are enough? Qualitative Health Research, 27(4), 591–608. https://doi.org/10.1177/1049732316665344
• Levitt, H. M., Bamberg, M., Creswell, J. W., Frost, D. M., Josselson, R., & Suárez-Orozco, C. (2018). Journal article reporting standards for qualitative primary, qualitative meta-analytic, and mixed methods research in psychology: The APA Publications and Communications Board task force report. American Psychologist, 73(1), 26–46. https://doi.org/10.1037/amp0000151
• Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. SAGE.
• Malterud, K., Siersma, V. D., & Guassora, A. D. (2016). Sample size in qualitative interview studies: Guided by information power. Qualitative Health Research, 26(13), 1753–1760. https://doi.org/10.1177/1049732315617444
• Miles, M. B., Huberman, A. M., & Saldaña, J. Qualitative data analysis: A methods sourcebook (4th ed.). SAGE. Confirm the publication year against the exact edition used before publishing the final reference. The publisher identifies it as a principal sourcebook for qualitative data analysis and display.
• O’Brien, B. C., Harris, I. B., Beckman, T. J., Reed, D. A., & Cook, D. A. (2014). Standards for reporting qualitative research: A synthesis of recommendations. Academic Medicine, 89(9), 1245–1251. https://doi.org/10.1097/ACM.0000000000000388
• Tong, A., Sainsbury, P., & Craig, J. (2007). Consolidated criteria for reporting qualitative research (COREQ): A 32-item checklist for interviews and focus groups. International Journal for Quality in Health Care, 19(6), 349–357. https://doi.org/10.1093/intqhc/mzm042