Research techniques are the specific procedures, tools, and analytical practices researchers use to find evidence, select cases, collect or generate data, analyze results, and evaluate their quality. Examples include literature searching, sampling, questionnaires, interviews, experiments, observation, thematic coding, statistical testing, triangulation, and sensitivity analysis.
Research techniques turn an abstract research plan into practical actions. A researcher may have a suitable topic and a well-written question, but the study will not produce useful evidence unless its sampling, measurement, collection, analysis, and quality-control procedures fit that question.
This guide explains the major types of research techniques, how they differ from methods and methodology, when each technique is appropriate, and how modern tools—including artificial intelligence—can be used responsibly.
Key takeaways
- A research technique is a specific procedure used during one or more stages of a study.
- Techniques should be selected from the research question and intended claim, not from personal preference alone.
- Surveys, interviews, observations, experiments, sampling procedures and analytical tests answer different kinds of questions.
- No technique is automatically valid or reliable; quality depends on design and execution.
- Mixed-methods research requires meaningful integration of qualitative and quantitative evidence.
- Ethical review, transparent reporting and responsible data management are part of rigorous technique use.
What are research techniques?
Research techniques are systematic procedures used to obtain, organize, analyze, verify, or communicate evidence. They are the practical operations through which a research design is implemented.
For example, a study examining university students’ experiences of online learning might use:
- Purposive sampling to recruit students with relevant experience
- Semi-structured interviews to collect detailed accounts
- Audio recording and transcription to preserve the data
- Reflexive thematic analysis to identify patterns of meaning
- Member reflection or peer discussion to examine interpretations
- A reporting checklist to describe the study transparently
The term is sometimes used broadly as a synonym for “research methods.” Other authors use method for the larger procedure—such as survey research—and technique for a specific operation within it, such as cognitive interviewing, random-digit sampling, questionnaire pretesting, weighting, or regression analysis.
Because usage differs among disciplines, the most important requirement is consistency: define what each term means in the study and explain exactly what was done. University and educational guides commonly describe research methods as the practical tools or procedures through which data are collected and analyzed.
Research technique, method, design, instrument and methodology
These terms are related but not identical.
| Term | Meaning | Example |
|---|---|---|
| Research methodology | The reasoning, assumptions and justification underlying how the study will produce knowledge | A pragmatic methodology supporting the use of qualitative and quantitative evidence |
| Research approach | The general logic connecting theory and evidence | Deductive, inductive or abductive reasoning |
| Research design | The overall structure of the study | Cross-sectional survey, randomized trial, ethnography or explanatory sequential mixed-methods design |
| Research method | A relatively broad way of obtaining or analyzing evidence | Survey research, interviewing, experimentation or case-study research |
| Research technique | A specific procedure used within a method or stage of research | Stratified sampling, cognitive interviewing, random assignment, thematic coding or regression |
| Research instrument | A device or structured tool used to record or measure data | Questionnaire, interview guide, rating scale, test, sensor or observation checklist |
| Research protocol | The documented sequence of procedures used to conduct the study consistently | Recruitment, consent, measurement, data storage and analysis instructions |
A questionnaire illustrates the distinction. Survey research may be the method, stratified random sampling a sampling technique, the questionnaire an instrument, Likert-type items a measurement format, and multiple regression an analysis technique.
Main types of research techniques
Research techniques can be classified by the stage of research in which they are used.
| Research stage | Main technique categories | Examples |
|---|---|---|
| Evidence discovery | Search and review techniques | Database searching, citation chaining, systematic searching |
| Case or participant selection | Sampling techniques | Simple random, stratified, purposive, snowball sampling |
| Concept measurement | Operationalization and instrument techniques | Scale development, rubric construction, sensor calibration |
| Data collection | Quantitative, qualitative and mixed techniques | Questionnaires, interviews, experiments, observation |
| Data preparation | Cleaning and management techniques | Validation rules, transcription, coding, de-identification |
| Data analysis | Statistical and interpretive techniques | Regression, thematic analysis, content analysis |
| Quality assessment | Validation and robustness techniques | Reliability testing, triangulation, sensitivity analysis |
| Reporting and sharing | Transparency techniques | Reporting checklists, preregistration, data and code sharing |
This lifecycle classification is more useful than treating research techniques as a short list of data-collection tools.
Literature-search and documentary research techniques
Database searching
Database searching uses planned combinations of keywords, subject headings, Boolean operators, date limits, and inclusion criteria to find relevant literature.
A transparent search normally records:
- Databases searched
- Search dates
- Complete search strings
- Language or publication limits
- Screening criteria
- Deduplication procedures
A broad narrative review may use a flexible search, while a systematic review requires a reproducible and explicitly reported strategy. PRISMA 2020 provides checklists and flow-diagram templates for reporting systematic reviews.
Citation searching
Citation searching follows relationships among publications.
- Backward citation searching examines the reference list of a useful source.
- Forward citation searching identifies later publications that cited it.
- Related-record searching uses database algorithms to find conceptually similar papers.
These techniques are valuable when terminology changes across time or disciplines.
Document and archival analysis
Documentary analysis examines existing materials such as:
- Government reports
- Institutional records
- Historical archives
- Policy documents
- Court decisions
- Newspapers
- Meeting minutes
- Websites
- Social-media posts
- Images and audiovisual materials
Researchers must evaluate provenance, purpose, completeness, authenticity, context and potential bias. A document created for administration or publicity should not automatically be treated as a neutral record.
Systematic review and evidence-synthesis techniques
Evidence synthesis may involve:
- Narrative synthesis
- Scoping review
- Systematic review
- Qualitative evidence synthesis
- Meta-analysis
- Bibliometric analysis
- Evidence mapping
These are not interchangeable. A meta-analysis statistically combines compatible quantitative estimates, whereas a qualitative synthesis interprets concepts or findings across qualitative studies.
Sampling techniques
Sampling determines which people, cases, documents, locations, events or observations enter the study.
Probability sampling techniques
Probability sampling uses a defined selection mechanism in which units have a known or calculable chance of selection.
Simple random sampling
Each eligible unit is selected through a random process.
Useful when: A reasonably complete sampling frame exists.
Advantage: Supports probability-based statistical inference when correctly implemented.
Limitation: A complete and accurate list of the population may be unavailable.
Systematic sampling
The researcher selects every kth unit after a random starting point.
Useful when: A list is available and does not contain a pattern that aligns with the interval.
Risk: Periodicity in the list can create bias.
Stratified sampling
The population is divided into meaningful subgroups, and samples are drawn within each subgroup.
Useful when: Important subgroups must be represented or compared.
Example: Sampling undergraduate, master’s and doctoral students separately.
Cluster and multistage sampling
Groups such as schools, hospitals, neighborhoods or classes are sampled before individuals within them.
Useful when: The population is geographically dispersed.
Limitation: Observations within clusters may be similar, so analysis must account for the clustered design.
Non-probability sampling techniques
Convenience sampling
Participants are selected because they are readily available.
Advantage: Fast and inexpensive.
Limitation: The sample may differ systematically from the target population, restricting generalization.
Purposive sampling
Cases are selected because they have characteristics or experiences relevant to the research question.
Useful for: Qualitative research, expert studies, case studies and studies of information-rich cases.
Quota sampling
The researcher fills predetermined categories without random selection inside each category.
Snowball or chain-referral sampling
Existing participants help identify others.
Useful for: Hard-to-reach or networked populations.
Risks: Network dependence, overrepresentation of closely connected groups and privacy concerns.
Theoretical sampling
Cases are selected iteratively to develop or refine an emerging theory, especially in grounded-theory research.
Sampling quality
A large sample does not correct a biased recruitment process. Researchers should distinguish:
- Target population
- Accessible population
- Sampling frame
- Invited sample
- Participating sample
- Analytic sample
For survey research, professional guidance recommends transparent reporting of the population, sampling construction, recruitment, response process, weighting, questionnaire wording and analytical adjustments.
Quantitative data-collection techniques
Quantitative techniques generate numerical data suitable for mathematical or statistical analysis.
Surveys and structured questionnaires
A survey collects standardized responses from a sample. Questionnaires may include:
- Multiple-choice questions
- Rating scales
- Frequency questions
- Rankings
- Numerical entries
- Short open-text items
Surveys are useful for estimating prevalence, describing attitudes, comparing groups and examining relationships.
A rigorous survey process includes:
- Defining the population and constructs
- Selecting an appropriate sampling frame
- Writing one clear concept per question
- Choosing response categories
- Conducting cognitive interviews
- Pretesting and piloting the instrument
- Monitoring fieldwork and missing data
- Applying suitable weighting or adjustment
- Reporting the full methodology transparently
AAPOR recommends simple, specific questions, questionnaire pretesting with people resembling the intended respondents, pilot testing of the complete procedure, and quality checks throughout collection and analysis.
Experiments
Experiments examine the effect of manipulating an independent variable on an outcome.
Core techniques may include:
- Random assignment
- Control or comparison groups
- Standardized treatment delivery
- Blinding where feasible
- Pretest and post-test measurement
- Manipulation checks
- Allocation concealment
- Protocol adherence monitoring
A well-designed randomized experiment can strengthen causal inference because random assignment helps balance competing explanations. However, causal interpretation still depends on implementation, attrition, measurement, interference, missing data and analytical choices.
Quasi-experimental techniques
Quasi-experiments estimate intervention effects without full random assignment.
Examples include:
- Interrupted time-series analysis
- Regression discontinuity
- Difference-in-differences
- Matched comparison groups
- Natural experiments
- Instrumental-variable approaches
These techniques can be valuable when randomization is impractical or unethical, but they rely on design-specific assumptions that must be examined and reported.
Structured observation
Structured observation records predefined behaviors or events using a schedule or checklist.
Examples:
- Counting classroom participation
- Timing task completion
- Recording clinical hand-hygiene behavior
- Coding types of interaction in a meeting
Observers need training, clear operational definitions and checks of consistency where appropriate.
Tests, scales and standardized measurements
Researchers use tests and scales to measure constructs such as knowledge, ability, attitudes, symptoms, motivation or satisfaction.
Important considerations include:
- Content validity
- Construct validity
- Criterion-related validity
- Internal consistency
- Test–retest stability
- Inter-rater agreement
- Measurement equivalence across groups
- Floor and ceiling effects
A high reliability coefficient does not prove that an instrument measures the intended construct.
Sensors, laboratory instruments and digital traces
Modern studies may collect data through:
- Wearable devices
- Physiological sensors
- Imaging equipment
- Environmental monitors
- Learning-management systems
- Server logs
- Mobile applications
- Geographic information systems
- Administrative databases
Researchers must document calibration, accuracy, missingness, software processing, device versions and privacy implications.
Qualitative data-collection techniques
Qualitative techniques are used to explore meanings, experiences, processes, practices and contexts.
Semi-structured interviews
Semi-structured interviews use a flexible guide of topics and open questions.
Best for: Individual experiences, decisions, interpretations and sensitive topics.
Strengths:
- Rich and detailed accounts
- Opportunity to ask follow-up questions
- Flexibility to explore unexpected issues
Limitations:
- Time-intensive collection and transcription
- Interviewer influence
- Dependence on memory and self-report
- Potential social-desirability effects
Researchers should record how the guide was developed, who conducted the interviews, the interview setting, recording procedures, transcription choices and the interviewer’s relevant relationship to participants.
Structured and unstructured interviews
A structured interview asks standardized questions in a fixed order, improving comparability.
An unstructured interview is more conversational and participant-led, supporting discovery but requiring considerable interviewer skill.
Focus groups
Focus groups use guided group interaction to generate data. Their value lies not merely in interviewing several people at once, but in examining agreement, disagreement, shared language and interaction among participants.
Useful for:
- Exploring social norms
- Developing questionnaires
- Testing services or messages
- Comparing group perspectives
Limitations:
- Dominant participants may shape discussion
- Confidentiality cannot be guaranteed fully
- Sensitive experiences may be withheld
- Group data should not be interpreted as independent individual responses
Observation
Observation examines behavior, interactions, settings or events as they occur.
Types include:
- Participant observation
- Non-participant observation
- Overt observation
- Covert observation, where ethically and legally permissible
- Naturalistic observation
- Structured observation
Observation can reveal differences between what people say and what they do, but researchers must consider reactivity, access, note-taking, positionality and interpretation.
Ethnographic techniques
Ethnography studies social and cultural practices in context, usually through sustained engagement.
Common techniques include:
- Participant observation
- Fieldnotes
- Informal conversations
- Interviews
- Artifact collection
- Spatial or social mapping
- Reflexive journaling
Ethnography is more than a brief site visit. It normally requires contextual immersion and interpretation of practices, meanings and relationships.
Case-study techniques
A case study investigates a bounded case such as an organization, program, event, community or individual.
Evidence may come from:
- Interviews
- Documents
- Observation
- Administrative data
- Physical or digital artifacts
- Quantitative indicators
Case studies are valuable for complex “how” and “why” questions, especially when context cannot be separated easily from the phenomenon.
Diaries and experience-sampling techniques
Participants record experiences over time through:
- Written diaries
- Audio or video diaries
- Mobile prompts
- Daily logs
- Ecological momentary assessment
These techniques reduce dependence on distant recall and capture change, but participant burden and incomplete entries can affect quality.
Visual and participatory techniques
Examples include:
- Photo elicitation
- Participatory mapping
- Drawings
- Timelines
- Storyboards
- Community workshops
- Photovoice
Such techniques can support expression that is difficult to capture in conventional interviews. Researchers must carefully address ownership, consent and the identifiability of people or places shown in visual materials.
Mixed-methods research techniques
Mixed-methods research deliberately combines qualitative and quantitative evidence to produce an integrated understanding.
Common designs include:
Convergent design
Qualitative and quantitative data are collected during a similar period, analyzed separately and then compared or integrated.
Example: A student-satisfaction survey is compared with interview themes.
Explanatory sequential design
Quantitative data are collected first. Qualitative research then explains important, unexpected or unclear numerical findings.
Example: A survey finds low engagement among a subgroup, followed by interviews exploring why.
Exploratory sequential design
Qualitative work is conducted first to understand a poorly defined issue or develop concepts. Quantitative work then measures or tests the emerging ideas.
Example: Interviews identify dimensions of research anxiety, which are used to develop and validate a questionnaire.
Embedded design
One form of evidence supports a larger design.
Example: Interviews embedded within a randomized trial explore how participants experienced the intervention.
Integration techniques
Mixed methods should specify how evidence will be connected. Techniques include:
- Using qualitative findings to design a questionnaire
- Selecting interview participants from survey results
- Comparing datasets in a joint display
- Transforming qualitative categories into numerical variables
- Following a theme across multiple datasets
- Developing a combined meta-inference
- Examining convergence, complementarity and disagreement
Simply placing an interview study beside a survey does not create meaningful mixed methods. Integration is a defining feature.
Measurement and instrument-development techniques
Before collecting data, researchers must translate abstract concepts into observable indicators.
Operationalization
Operationalization specifies how a concept will be represented or measured.
For example, “academic engagement” might be operationalized through:
- Attendance records
- Time spent on learning activities
- A validated engagement scale
- Classroom observation
- Interview accounts
Each indicator represents only part of the concept. Researchers should explain why the selected indicator is appropriate.
Item generation
Questionnaire items may be developed from:
- Theory
- Previous instruments
- Literature reviews
- Interviews or focus groups
- Expert consultation
- Stakeholder participation
Expert review
Experts assess relevance, clarity, coverage and potential ambiguity. Expert agreement is useful evidence, but it should not be treated as complete proof of validity.
Cognitive interviewing
Participants explain how they understand a question and arrive at an answer. This can reveal:
- Ambiguous terminology
- Recall difficulties
- Unclear reference periods
- Inadequate response options
- Sensitive or burdensome wording
Pilot testing
A pilot test evaluates the feasibility of the full procedure, including recruitment, consent, timing, instrument administration, data storage and analysis preparation.
A pilot is not merely a smaller final study. Its purpose should be defined in advance, and substantive conclusions should not be overclaimed from an underpowered feasibility sample.
Quantitative data-analysis techniques
The analytical technique should match the research question, study design, variables and data-generating process.
Descriptive analysis
Descriptive techniques summarize the observed data.
Examples include:
- Frequencies and percentages
- Mean, median and mode
- Range and interquartile range
- Variance and standard deviation
- Tables and graphs
Descriptive statistics do not by themselves test hypotheses or establish causation.
Estimation and confidence intervals
Estimation techniques quantify an effect or population parameter and its uncertainty.
Examples:
- Mean difference
- Risk difference
- Odds ratio
- Correlation coefficient
- Regression coefficient
- Confidence interval
A confidence interval should be interpreted alongside design quality, model assumptions and practical importance.
Hypothesis testing
Common tests include:
- t tests
- Analysis of variance
- Chi-square tests
- Nonparametric tests
- Tests of regression coefficients
A small p value does not measure effect size, practical importance, study quality or the probability that the research hypothesis is true.
Correlation and regression
Correlation summarizes association between variables. Regression estimates relationships while representing one or more predictors.
Common forms include:
- Linear regression
- Logistic regression
- Poisson or count regression
- Survival analysis
- Multilevel modeling
- Time-series analysis
Model selection should follow the outcome type, design, distribution, clustering and theoretical question—not whichever test produces statistical significance.
Missing-data techniques
Researchers should report:
- The amount and pattern of missingness
- Reasons for missing data where known
- Whether cases or variables were excluded
- Any imputation procedure
- Sensitivity of conclusions to missing-data assumptions
Robustness and sensitivity analysis
Sensitivity analysis examines whether conclusions change under reasonable alternative decisions.
Examples include:
- Alternative variable definitions
- Different model specifications
- Inclusion or exclusion of influential observations
- Alternative missing-data assumptions
- Adjustment for clustering
- Placebo or falsification tests
Qualitative data-analysis techniques
Thematic analysis
Thematic analysis identifies and interprets patterned meaning across a dataset.
A commonly used process involves:
- Becoming familiar with the data
- Coding relevant features
- Developing candidate themes
- Reviewing themes
- Defining and naming themes
- Producing the analytical account
Different forms of thematic analysis make different assumptions. Researchers should state whether the analysis is primarily inductive or deductive, semantic or latent, and reflexive or codebook-oriented rather than citing “thematic analysis” as though it were a single uniform procedure.
Qualitative content analysis
Content analysis organizes text, images or media into categories. It may be:
- Conventional or inductive
- Directed by an existing framework
- Summative, including counts and interpretation
Counting references to a topic can support analysis, but frequency alone does not necessarily indicate importance.
Grounded-theory techniques
Grounded-theory studies may use:
- Initial and focused coding
- Constant comparison
- Memo writing
- Theoretical sampling
- Category development
- Examination of relationships and processes
Researchers should avoid claiming grounded theory when they have only coded interview data thematically without its iterative theory-development procedures.
Narrative analysis
Narrative analysis examines how experiences are organized and communicated as stories. It may focus on sequence, plot, identity, turning points, audience and cultural narratives.
Discourse analysis
Discourse analysis studies how language constructs meanings, identities, relationships or social realities. It requires attention to language use and context rather than treating text as a transparent report of internal attitudes.
Framework analysis
Framework analysis uses a structured matrix to compare themes across cases while preserving links to the original data. It is useful in applied and policy research with predefined questions and multiple stakeholder groups.
How to choose the right research technique
Choose a research technique by identifying the claim you need to make, the evidence required to support it, the accessible units of study, ethical constraints, and the analysis that will answer the research question.
Step 1: Clarify the research question
Different questions require different evidence.
- How many or how often? Survey, administrative data or structured observation
- Is there an association? Correlational design and regression
- Did an intervention cause change? Experiment or strong quasi-experiment
- How do people experience something? Interviews, diaries or observation
- How does a process unfold? Longitudinal, case-study or ethnographic techniques
- What does existing research show? Systematic review or evidence synthesis
Step 2: Define the intended claim
Decide whether the goal is to:
- Describe
- Compare
- Explain
- Predict
- Estimate an effect
- Interpret meaning
- Develop theory
- Evaluate a program
- Synthesize existing evidence
Do not select a descriptive technique and later make an unsupported causal claim.
Step 3: Identify the unit of analysis
The unit may be:
- Individual
- Household
- School
- Organization
- Country
- Document
- Social-media post
- Interaction
- Event
- Biological specimen
- Repeated measurement
Sampling, measurement and analysis must correspond to that unit.
Step 4: Determine the required data
Ask whether the question needs:
- Numerical measurements
- Textual or visual accounts
- Behavioral observations
- Existing records
- Experimental outcomes
- Multiple complementary forms of evidence
Step 5: Assess feasibility and ethics
Consider:
- Participant access
- Time
- Budget
- Equipment
- Researcher skills
- Data sensitivity
- Risk to participants
- Required ethical approval
- Data-storage capacity
- Language and accessibility
The theoretically strongest design may be unsuitable if it cannot be implemented ethically or competently.
Step 6: Plan sampling and recruitment
Define:
- Population or case boundaries
- Inclusion and exclusion criteria
- Sampling frame
- Sampling technique
- Recruitment procedures
- Anticipated nonresponse or attrition
- Rationale for sample size or information adequacy
Step 7: Match analysis to collection
The analysis should be planned before data collection where possible.
For example:
- A paired design requires paired analysis.
- Cluster sampling may require cluster-adjusted analysis.
- Open interview questions require a qualitative analysis plan.
- A repeated-measures study requires techniques that represent within-person dependence.
Step 8: Build in quality checks
Include appropriate procedures such as:
- Instrument pretesting
- Pilot testing
- Calibration
- Observer training
- Inter-rater assessment
- Reflexive notes
- Audit trails
- Manipulation checks
- Assumption checks
- Sensitivity analyses
- Triangulation
- Preregistration
Research-question-to-technique table
| Research objective | Suitable techniques | Example | Important limitation |
|---|---|---|---|
| Estimate prevalence | Probability survey, administrative records | Estimate the proportion of students using AI tools | Nonresponse or incomplete coverage can bias estimates |
| Explore experiences | Semi-structured interviews, diaries | Understand doctoral students’ supervision experiences | Findings are context-dependent and interpretive |
| Test an intervention | Randomized experiment | Compare two teaching strategies | Attrition or contamination may weaken inference |
| Evaluate a policy without randomization | Interrupted time series, difference-in-differences | Examine outcomes before and after a policy change | Causal inference depends on strong assumptions |
| Observe behavior | Structured or naturalistic observation | Record classroom participation | People may change behavior when observed |
| Study culture or practice | Ethnography, participant observation | Examine collaborative practice in a laboratory | Requires sustained access and reflexive interpretation |
| Analyze documents or media | Content, discourse or documentary analysis | Study policy framing in government reports | Available documents may be incomplete or strategic |
| Develop a theory | Grounded-theory techniques | Model how early-career researchers adopt open science | Requires iterative sampling and analysis |
| Combine breadth and explanation | Mixed-methods design | Survey researchers, then interview selected groups | Integration must be planned and justified |
| Synthesize published evidence | Systematic review, meta-analysis | Evaluate evidence on a teaching intervention | Conclusions depend on included-study quality and reporting |
Worked example: selecting techniques for an education study
Research topic
The relationship between generative-AI use and undergraduate academic writing.
Weak starting question
“Is AI good or bad for students?”
This question is too broad, uses undefined concepts and does not identify an outcome or population.
Improved questions
- How frequently do undergraduates use generative AI during academic writing?
- Is reported use associated with writing self-efficacy?
- How do students explain the benefits, risks and decision-making involved?
Suitable design
An explanatory sequential mixed-methods design.
Quantitative phase
- Sampling: Stratified sampling by year of study and discipline
- Instrument: Questionnaire measuring usage patterns and writing self-efficacy
- Quality techniques: Cognitive interviews and pilot testing
- Analysis: Descriptive statistics and regression with prespecified variables
Qualitative phase
- Sampling: Purposive selection from contrasting survey-response profiles
- Collection: Semi-structured interviews
- Analysis: Thematic analysis
- Quality techniques: Reflexive memoing, transparent coding decisions and negative-case examination
Integration
Survey findings would identify broad patterns. Interviews would explore why those patterns occur and how students interpret responsible and inappropriate uses. A joint display could compare statistical results with qualitative themes.
Ethical issues
The protocol should avoid collecting unnecessary identifiable information, clarify that participation will not affect grades, explain data storage, and address the possibility that students disclose conduct that violates institutional rules.
Quality standards for research techniques
Quantitative quality
Common concepts include:
- Reliability: Consistency of measurement under specified conditions
- Internal validity: Credibility of the causal or relational inference within the study
- Construct validity: Whether measures and procedures represent the intended concept
- External validity: Applicability beyond the studied setting or sample
- Statistical-conclusion validity: Appropriateness of the analysis and inference
Qualitative quality
Depending on the qualitative tradition, researchers may discuss:
- Credibility
- Dependability
- Confirmability
- Transferability
- Reflexivity
- Contextual adequacy
- Interpretive depth
- Transparency
These criteria should not be converted into a mechanical checklist detached from the study’s epistemological position.
Triangulation
Triangulation compares evidence across:
- Methods
- Data sources
- Researchers
- Theoretical perspectives
- Time points
- Settings
Agreement may increase confidence, but disagreement can also be analytically valuable. Triangulation should investigate why evidence converges or diverges rather than treating majority agreement as automatic truth.
Reflexivity
Reflexivity requires researchers to examine how their assumptions, position, relationships and decisions shape the research process.
A reflexive account may discuss:
- The researcher’s relationship to the topic
- Access and power differences
- Interview interactions
- Changes to the analytical framework
- Interpretive uncertainties
- Alternative explanations
Research ethics and responsible technique use
Human-participant research should address respect, risk, benefit, fairness and voluntary participation. The Belmont Report’s foundational principles are respect for persons, beneficence and justice, with applications including informed consent, risk–benefit assessment and participant selection.
Key considerations include:
- Institutional ethical approval where required
- Meaningful informed consent
- Voluntary participation and withdrawal
- Proportionate risk
- Fair recruitment
- Privacy and confidentiality
- Data minimization
- Secure storage
- Appropriate anonymization or pseudonymization
- Special protections for vulnerable participants
- Disclosure of conflicts of interest
- Responsible compensation
- Community or stakeholder consultation where appropriate
Publicly accessible data are not automatically ethically unrestricted. Researchers should consider platform expectations, identifiability, sensitive content, terms of service, quotation traceability and potential harm.
Digital research tools
Literature and reference management
Common functions include:
- Searching academic databases
- Managing references
- Removing duplicates
- Annotating papers
- Screening records
- Mapping citation relationships
Tools improve organization but do not determine relevance or study quality automatically.
Survey and data-collection platforms
Digital platforms can provide:
- Skip logic
- Randomization
- Validation rules
- Multilingual forms
- Timestamping
- Audit logs
- Export formats
Researchers should examine data location, encryption, access controls, accessibility, platform retention and regulatory requirements.
Statistical and programming tools
Examples include:
- R
- Python
- Stata
- SPSS
- SAS
- MATLAB
- Julia
Researchers should report software, relevant packages, versions and major analytical settings. Code-based workflows can improve auditability when scripts and data-processing decisions are documented.
Qualitative analysis software
Programs such as NVivo, ATLAS.ti, MAXQDA and open-source alternatives can help store, retrieve, code and compare qualitative material.
The software organizes analysis; it does not decide what a theme means or remove the need for methodological judgment.
Version control and reproducible notebooks
Version-control systems and computational notebooks can record:
- Changes to code
- Data-cleaning steps
- Analysis outputs
- Contributor activity
- Software environments
- Connections between data, code and figures
The National Academies distinguishes reproducibility and replicability and recommends greater methodological transparency, access to data and code where appropriate, and institutional support for rigorous research practice.
Artificial intelligence in modern research
AI tools may support:
- Keyword generation for literature searches
- Screening assistance
- Questionnaire brainstorming
- Transcription
- Translation
- Coding suggestions
- Programming assistance
- Data visualization
- Language editing
- Summarization of researcher-provided notes
AI output should be treated as unverified assistance, not as evidence.
Responsible AI practices
- Protect confidential information. Do not upload identifiable participant data, unpublished sensitive material or protected records to an unapproved service.
- Verify every citation. Generative systems can invent authors, titles, quotations, findings and identifiers.
- Retain human responsibility. Researchers remain accountable for design, analysis, interpretation and reporting.
- Document material use. Record the tool, version or access date, task, prompts where relevant, and how output was checked.
- Check for bias. AI-supported coding, classification and translation can reproduce linguistic or social biases.
- Follow institutional and journal rules. Permitted uses and disclosure requirements differ.
- Preserve an audit trail. Keep original data, human decisions and revisions separate from generated suggestions.
- Do not list an AI tool as an author. Authorship requires accountability that a tool cannot accept.
UNESCO’s guidance advocates human-centered, safe, equitable and privacy-aware use of generative AI in education and research. Its official page was last updated January 16, 2026.
Survey researchers are already using generative AI for questionnaire development, interviewing support, open-response coding, analysis and writing assistance, making validation and disclosure particularly important.
Open science and transparent research techniques
Open-science practices can include:
- Preregistration
- Registered Reports
- Open protocols
- Open materials
- Open data
- Open code
- Preprints
- Transparent peer review
- Contributor statements
Preregistration creates a time-stamped record of hypotheses, methods and analysis plans before the relevant analysis, helping distinguish confirmatory decisions from later exploration. It does not prevent justified changes; deviations should be explained.
Data sharing should follow consent, privacy, legal and ethical constraints. The NIH Data Management and Sharing Policy, effective since January 25, 2023, expects covered investigators and institutions to plan how scientific data will be managed and shared, subject to justified limitations.
Reporting guidelines
Reporting guidelines help authors describe what they did clearly enough for readers to evaluate the research.
Examples include:
- CONSORT 2025: Randomized trials
- STROBE: Observational studies
- PRISMA 2020: Systematic reviews
- COREQ: Interview and focus-group studies
- SRQR: Qualitative research
- APA JARS: Quantitative, qualitative and mixed-methods research
- CONSORT-AI: Trials involving AI interventions
EQUATOR maintains a searchable library of reporting guidelines. CONSORT was updated in 2025, while PRISMA 2020 remains the principal general reporting guideline for many systematic reviews.
A reporting checklist does not repair a weak design. It improves transparency about the design and execution that actually occurred.
Advantages of systematic research techniques
Appropriately selected and applied techniques can:
- Align evidence with the research question
- Make procedures more consistent
- Reduce avoidable error
- Support transparent evaluation
- Improve comparability
- Reveal assumptions and limitations
- Facilitate replication or secondary analysis
- Help researchers distinguish evidence from interpretation
Limitations of research techniques
Every technique has boundaries.
- Self-reports may be affected by recall and presentation.
- Observation can be influenced by researcher interpretation or participant reactivity.
- Experiments may create artificial conditions or face ethical limits.
- Convenience samples limit population inference.
- Statistical models depend on assumptions and measurement quality.
- Qualitative interpretations are shaped by context and researcher positioning.
- Existing datasets may omit important variables.
- Digital data may exclude people with limited access.
- AI-assisted procedures may introduce hidden errors or bias.
- Mixed-methods projects require more planning and integration expertise.
The goal is not to find a technique with no limitations. It is to select a defensible technique, reduce avoidable weaknesses and report remaining uncertainty honestly.
Common mistakes
Choosing a technique before defining the question
A researcher decides to “do a questionnaire” and then searches for a question that fits it.
Better practice: Define the intended contribution and required evidence first.
Treating design labels as proof of quality
Calling a study experimental, mixed methods or systematic does not guarantee rigor.
Better practice: Examine the actual sampling, procedures, measurements, analysis and reporting.
Confusing an instrument with a method
A questionnaire is an instrument, while survey research is a broader method or design.
Using convenience data to make population-wide claims
Large online samples can still be systematically unrepresentative.
Equating correlation with causation
An association may arise from confounding, reverse direction, selection or measurement error.
Selecting an analysis after seeing the results
Undisclosed outcome switching and repeated testing can produce misleading findings.
Better practice: Prespecify confirmatory analyses and clearly label exploratory work.
Using software without checking assumptions
A program can execute an unsuitable test correctly.
Ignoring contradictory evidence
Negative cases, divergent themes and alternative model specifications can improve interpretation.
Treating AI-generated text as a source
AI output must be verified against original, authoritative materials.
Omitting procedural detail
Readers need enough information to understand who or what was studied, how data were generated, how they were analyzed and what limitations remain.
Research-technique selection worksheet
Use the following template when planning a study:
| Planning question | Researcher’s answer |
|---|---|
| What is the exact research question? | |
| What claim will the study support? | |
| What is the unit of analysis? | |
| What type of evidence is required? | |
| What population, cases or materials are relevant? | |
| Which sampling technique is suitable? | |
| Which data-collection technique fits the question? | |
| Which instrument or protocol is required? | |
| How will the instrument be pretested or calibrated? | |
| Which analysis will answer the question? | |
| What assumptions does the analysis make? | |
| What ethical approval or consent is required? | |
| How will privacy and data security be protected? | |
| Which quality checks will be used? | |
| How will deviations and limitations be reported? | |
| Can the protocol, materials, data or code be shared safely? | |
| Which reporting guideline applies? |
Research techniques across disciplines
Social sciences
Common techniques include surveys, interviews, focus groups, ethnography, observation, content analysis and regression.
Education
Researchers may use classroom observation, achievement tests, learning analytics, interviews, quasi-experiments, action research and mixed methods.
Healthcare
Techniques include clinical trials, cohort studies, diagnostic measurement, patient interviews, qualitative evidence synthesis and routinely collected health-data analysis. Ethical approval, privacy and discipline-specific reporting standards are especially important.
Business and management
Common techniques include customer surveys, organizational case studies, experiments, interviews, market-data analysis, forecasting and process evaluation.
Humanities
Researchers may use archival analysis, textual criticism, discourse analysis, historical comparison, close reading, digital humanities and visual analysis.
Natural sciences
Techniques include controlled experimentation, laboratory measurement, field sampling, microscopy, spectroscopy, simulation and computational modeling.
Computing and data science
Researchers may use benchmarking, simulation, formal proof, user studies, A/B tests, log analysis, machine-learning evaluation and software experiments. Data leakage, benchmark contamination, reproducible environments and transparent evaluation metrics require particular attention.
Conclusion
Research techniques are the practical procedures that connect a research question to defensible evidence. They include more than surveys, interviews and experiments: sampling, measurement, analysis, validation, reporting and data-management procedures are equally important.
The best technique is not the most advanced or popular one. It is the technique that fits the question, intended claim, evidence, context, ethical requirements and available expertise—and whose limitations can be explained transparently.
References
- American Association for Public Opinion Research. (n.d.). Best practices for survey research.
- Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77–101. https://doi.org/10.1191/1478088706qp063oa
- Center for Open Science. (n.d.). Open science.
- Creswell, J. W., & Plano Clark, V. L. (2018). Designing and conducting mixed methods research (3rd ed.). SAGE.
- EQUATOR Network. (2025). CONSORT 2025 statement: Updated guideline for reporting randomised trials.
- Miao, F., & Holmes, W. (2023). Guidance for generative AI in education and research. UNESCO.
- National Academies of Sciences, Engineering, and Medicine. (2019). Reproducibility and replicability in science. National Academies Press. https://doi.org/10.17226/25303
- National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research. (1979). The Belmont report. U.S. Department of Health and Human Services.
- National Institutes of Health. (2023). Data management and sharing policy.
- Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., et al. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ, 372, n71. https://doi.org/10.1136/bmj.n71
- Tong, A., Sainsbury, P., & Craig, J. (2007). Consolidated criteria for reporting qualitative research: 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
- von Elm, E., Altman, D. G., Egger, M., Pocock, S. J., Gøtzsche, P. C., & Vandenbroucke, J. P. (2007). The Strengthening the Reporting of Observational Studies in Epidemiology statement. PLoS Medicine, 4(10), e296. https://doi.org/10.1371/journal.pmed.0040296
- Wilkinson, M. D., Dumontier, M., Aalbersberg, I. J., et al. (2016). The FAIR Guiding Principles for scientific data management and stewardship. Scientific Data, 3, 160018. https://doi.org/10.1038/sdata.2016.18
