Die Druckversion wird nicht mehr unterstützt und kann Darstellungsfehler aufweisen. Bitte aktualisiere deine Browser-Lesezeichen und verwende stattdessen die Standard-Druckfunktion des Browsers.

Educational Research

Introduction

Educational research is the systematic and critical study of learning, teaching, educational institutions, policies, technologies, curricula, and the social conditions in which education takes place. It uses evidence to describe what is happening, interpret meanings and experiences, test explanations, evaluate interventions, and improve theory or practice. Educational research is inherently interdisciplinary: researchers may draw on Education, Psychology, Sociology, Economics, Anthropology, History, Philosophy of education, Statistics, and Data science.

A strong study does more than collect data. It aligns a worthwhile problem with a clear research question, an appropriate design, defensible sampling and measurement, ethical procedures, a transparent analysis, and claims that are no stronger than the evidence permits. Because educational settings are complex, context matters. A result found in one institution, age group, curriculum, country, or historical period may not transfer unchanged to another.

The research cycle is iterative rather than strictly linear. Reading may change your question; piloting may reveal a weak measure; early analysis may expose missing perspectives; and peer review may lead you to refine a theory. Good research therefore combines planning with disciplined responsiveness.

This aiMOOC is designed for university students who need to read, critique, design, conduct, and communicate educational research. You will work with quantitative, qualitative, mixed-methods, evaluation, action-research, and evidence-synthesis approaches while keeping validity, ethics, equity, and practical significance in view.

Research Problems, Questions, and Claims

A research project usually begins with a problem that is educationally important and researchable. A broad concern such as "students are disengaged" is not yet a research question. You need to define the phenomenon, population, setting, and purpose precisely enough to guide evidence collection.

Useful research questions often fall into several families. Descriptive questions ask what is happening or how frequently it occurs. Comparative questions examine differences between groups, conditions, or time points. Relational questions investigate associations. Causal questions ask whether changing one factor produces a change in another. Interpretive questions explore meaning, experience, identity, or process. Design-oriented questions investigate how an intervention can be developed, implemented, and improved.

A good question is not judged by one formula. It should be clear, significant, feasible, ethically defensible, and aligned with the methods you can justify. Quantitative studies may specify hypotheses in advance. Qualitative studies may begin with open questions that are refined as understanding develops. Mixed-methods studies need an explicit reason for combining forms of evidence rather than merely collecting two types of data.

The central discipline is alignment: your question determines what evidence is relevant; your design determines what comparisons or interpretations are possible; your analysis determines what patterns you can estimate; and your conclusions must stay within those boundaries.

Concepts, Theory, and Operationalization

A construct is an abstract concept such as motivation, belonging, self-efficacy, cognitive load, instructional quality, or school climate. Constructs are not directly observed in the same way as height or elapsed time. You need a defensible connection between the construct and the indicators used to represent it.

A theoretical framework explains relationships among concepts using an established theory or body of theory. A conceptual framework is the organized set of ideas and relationships you use to structure a study; it may combine theory, prior evidence, and context-specific reasoning. These frameworks help you decide what to observe, what to compare, and how to interpret findings.

Operationalization translates a construct into observable indicators or procedures. For example, "engagement" might be represented by classroom observation codes, survey responses, attendance records, platform activity, interviews, or a combination. Different operationalizations can capture different aspects of the same concept, so you should justify why your indicators match the intended interpretation.

Literature Review and Evidence Mapping

A literature review is not a list of summaries. Its purpose is to establish what is known, how it is known, where findings agree or conflict, what limitations remain, and how your project contributes. A rigorous review makes the path from search to synthesis visible.

Begin by defining the scope and key concepts. Search more than one appropriate database when possible, document search terms and date limits, and use backward and forward citation searching. Record inclusion and exclusion decisions. Critically appraise study quality rather than treating all published findings as equally trustworthy. Then synthesize patterns across studies by concepts, methods, contexts, mechanisms, or outcomes.

A systematic review uses prespecified and transparent methods to identify, select, appraise, and synthesize evidence for a focused question. A scoping review is often used to map a broad field, clarify concepts, or identify gaps. A narrative review can be scholarly and rigorous but usually allows more interpretive flexibility. The label should match the actual method.

For systematic reviews, a flow diagram can document how records moved from identification through screening to inclusion. PRISMA is a reporting guideline for systematic reviews; using a PRISMA-style diagram does not by itself make a review systematic.

Research Paradigms and Methodological Coherence

Research methods are connected to assumptions about what can be known and how knowledge is justified. You do not need to turn every project into a philosophy essay, but you should understand the assumptions built into your design.

Post-positivist research often seeks fallible explanations of patterns and causal relationships through systematic observation, measurement, comparison, and attempts to rule out alternative explanations. Interpretivist or constructivist research often examines how people make meaning within social and cultural contexts. Critical traditions examine power, inequality, ideology, and possibilities for transformation. Pragmatist approaches often emphasize the research problem and the consequences of inquiry, and they are frequently associated with mixed methods.

These traditions are not simple boxes. Qualitative methods do not belong to only one paradigm, quantitative methods do not require one philosophical position, and mixed methods is not philosophically neutral. Methodological coherence means that your questions, assumptions, procedures, analysis, and claims fit together.

Quantitative Research Designs

Quantitative research represents phenomena numerically and uses statistical reasoning to describe variation, estimate relationships, compare groups, or support causal inference. The strength of the claim depends on the design, not on the presence of a p value.

Experiments and Randomized Designs

In a randomized controlled trial, participants or clusters such as classes or schools are randomly assigned to conditions. Random assignment helps balance both observed and unobserved confounders on average, strengthening causal inference when implementation, attrition, measurement, and analysis are also handled well.

Educational trials are often cluster randomized because an intervention is delivered to whole classrooms or schools. Clustered assignment creates dependence among observations, so the analysis must account for that structure. Researchers should also document recruitment, allocation, follow-up, exclusions, and analysis transparently.

The diagram above comes from a general randomized-trial reporting tradition. It is useful for visualizing participant flow, but no single reporting template replaces a design-specific justification.

Quasi-Experimental and Observational Designs

Random assignment is often impossible or inappropriate in education. Quasi-experimental designs seek credible comparison conditions without full randomization. Examples include regression discontinuity, interrupted time series, difference-in-differences, matched comparison designs, and natural experiments. Each design depends on assumptions that must be argued and tested where possible.

Observational studies describe or model naturally occurring variation. Cross-sectional designs measure a population at one time; longitudinal designs follow units across time. Correlation or regression can estimate associations, but statistical adjustment does not automatically remove confounding. A causal claim requires a design and assumptions that justify interpreting the association causally.

Surveys and Secondary Data

Surveys can estimate attitudes, experiences, behaviors, or characteristics when the sample, questions, administration mode, and nonresponse are handled carefully. Administrative data, assessment records, learning-management-system logs, and national datasets can support large-scale analysis, but data created for operational purposes may contain measurement limitations, missingness, selection effects, or changing definitions.

Always distinguish random sampling from random assignment. Random sampling concerns how units enter the study and supports population generalization. Random assignment concerns how units enter conditions and supports causal comparison.

Qualitative Research Designs

Qualitative research investigates meaning, experience, interaction, culture, identity, practice, and process through rich forms of evidence such as interviews, observations, field notes, documents, images, artifacts, and recordings.

A case study investigates a bounded case in depth and in context. Ethnography examines cultural practices and meanings, often through sustained field engagement. Phenomenological research explores lived experience. Grounded theory develops theory through iterative data collection and analysis. Narrative inquiry examines how experience is organized and communicated through stories. Qualitative descriptive approaches may aim for a detailed account of events or perspectives without committing to a more specialized tradition.

Qualitative research is not simply "research without numbers." It requires systematic choices about sampling, relationships with participants, data generation, analytic procedures, reflexivity, and the evidential basis for interpretations.

Mixed-Methods Research

Mixed-methods research intentionally integrates quantitative and qualitative evidence to answer a question more fully than either strand could alone. The key word is integration. Two parallel datasets do not constitute a strong mixed-methods design unless the study explains how they connect.

A convergent design collects different forms of data during a similar phase and compares or combines the findings. An explanatory sequential design begins with quantitative results and follows with qualitative inquiry to explain patterns, mechanisms, or unexpected findings. An exploratory sequential design begins qualitatively and then develops or tests measures, categories, or hypotheses quantitatively.

Integration can occur through sampling connections, instrument development, joint displays, merged datasets, sequential decisions, or interpretation. Mixed-methods quality depends on the quality of each strand and the quality of the integration.

Sampling, Access, and Representation

Sampling determines whose experiences or records can contribute to the evidence. Start by distinguishing the target population, the accessible population, the sampling frame, and the final analytic sample.

Probability sampling methods include simple random, systematic, stratified, cluster, and multistage sampling. They support design-based estimates of population quantities when selection probabilities and nonresponse are properly handled. Nonprobability methods include purposive, convenience, quota, snowball, and theoretical sampling. These can be appropriate for many qualitative, exploratory, and hard-to-reach-population studies, but they support different kinds of generalization.

Sample size should be justified rather than copied from a rule of thumb. Quantitative studies may use precision targets, expected effect sizes, statistical power, clustering, repeated measures, and expected attrition. Qualitative studies often justify adequacy through the richness and relevance of information, the specificity of the sample, the scope of the question, and the analytic approach. No universal participant number guarantees saturation.

Access matters as much as sample size. School leaders, instructors, platform administrators, parents, or community organizations may act as gatekeepers. Their involvement can affect who participates and how voluntary participation feels. Researchers should consider underrepresentation, language access, disability access, digital access, and the possibility that recruitment procedures reproduce existing inequalities.

Measurement and Data Collection

Data quality depends on the relationship between your question, construct, instrument, context, and procedure. A familiar instrument is not automatically valid for every population or use.

Reliability and Validity

Reliability concerns the consistency or precision of scores or classifications under specified conditions. Depending on the instrument, relevant evidence may include internal consistency, inter-rater agreement, test-retest stability, or measurement precision.

Validity concerns the justification for interpreting and using scores or evidence in a particular way. Validity is not a permanent property stamped onto a test. Evidence may address content, response processes, internal structure, relationships with other variables, and the consequences of use. Fairness and measurement invariance can be especially important when comparing groups.

A measure can be reliable yet still support a poor interpretation. For example, a questionnaire may produce highly consistent scores while failing to represent the intended construct.

Likert-type items are common in education, but good survey design requires more than choosing response labels. Items should use clear language, avoid double-barreled wording and unnecessary jargon, minimize leading phrasing, and be tested with members of the intended population. Multi-item scales should be evaluated as scales rather than assumed to work because the items appear related.

Interviews, Focus Groups, and Observation

Interviews can range from structured to semi-structured to open-ended. A strong interview protocol uses questions that invite evidence relevant to the research question, avoids presupposing the desired answer, and includes probes that support depth and clarification.

Focus groups produce interactional data as well as individual responses. They can reveal shared norms, disagreements, and how ideas are negotiated, but confidentiality is harder to guarantee because participants hear one another.

Observation can be structured with predefined categories or open and ethnographic. In educational settings, observation changes the social situation to some degree, and researchers should consider reactivity, observer position, timing, and whether a single lesson represents ordinary practice.

Digital Trace and Administrative Data

Digital learning environments generate logs of clicks, submissions, messages, timestamps, and assessment events. These traces can be useful, but a click is not a direct measure of attention, learning, or motivation. Operational data must be interpreted through a theory of what the trace means.

Data linkage can create powerful longitudinal datasets while increasing privacy risk. De-identification lowers risk but does not guarantee anonymity, especially when multiple variables can be combined. Use data minimization, access controls, secure storage, and governance procedures appropriate to the sensitivity of the records.

Quantitative Data Analysis

Begin with data quality. Check coding, impossible values, duplicates, missingness, distributions, and whether the analytic dataset corresponds to the design you planned. Document exclusions and transformations.

Descriptive statistics summarize the observed data through counts, proportions, means, medians, variability, distributions, and visualizations. Inferential statistics estimate population quantities or test model-based hypotheses under assumptions. Confidence intervals communicate uncertainty around estimates. Effect sizes help you discuss magnitude rather than relying only on statistical significance.

The normal distribution is important in many statistical models, but real educational data do not become valid merely because they resemble a bell curve. Choose models that fit the scale, structure, and distribution of your outcome. Binary, ordinal, count, time-to-event, repeated-measures, and nested data may require specialized models.

Educational data are often hierarchical: students are nested in classes, classes in schools, and observations within students. Ignoring this dependence can produce misleading standard errors. Multilevel models, cluster-robust procedures, or design-based methods may be appropriate.

Missing data deserve explicit attention. Complete-case analysis can be biased when missingness is related to the outcome or predictors. Explain the missing-data pattern, justify the method used, and consider sensitivity analyses when assumptions are uncertain.

Qualitative Data Analysis

Qualitative analysis turns records into an argued interpretation through systematic engagement with the data. The exact procedure depends on the methodology.

Coding labels segments of data for analytic purposes. Codes may be developed inductively from the material, deductively from theory, or through a combination. Themes are not merely the most frequent words; they are patterns of meaning that help answer the research question.

In reflexive thematic analysis, the researcher is an active interpreter rather than a neutral coding machine. Analytic memos, repeated reading, theme development, comparison with negative or divergent cases, and explicit reflexivity can strengthen the auditability of the interpretation.

Other approaches include qualitative content analysis, discourse analysis, conversation analysis, narrative analysis, and grounded-theory procedures. Do not mix labels casually: each approach carries assumptions about what counts as data, what the analytic object is, and what kind of claim is being made.

Quality, Validity, and Trustworthiness

Quality criteria should fit the research tradition. In many quantitative designs, you may discuss internal validity, external validity, construct validity, and statistical conclusion validity. Internal validity concerns whether a causal interpretation is credible. External validity concerns how findings may generalize across populations, settings, treatments, or times. Construct validity concerns whether operationalizations support the intended conceptual interpretation.

Qualitative researchers may use concepts such as credibility, dependability, confirmability, and transferability, although terminology varies across traditions. Useful practices can include prolonged engagement, thick description, reflexive memoing, transparent decision trails, peer debriefing, member reflection when appropriate, and attention to disconfirming evidence.

Triangulation means examining a phenomenon through multiple sources, methods, investigators, or theoretical perspectives. Agreement can increase confidence, but disagreement is also informative. Multiple methods do not automatically reveal one final truth; they may illuminate different dimensions of the educational situation.

Replication, robustness checks, sensitivity analyses, and transparent reporting strengthen the cumulative evidence base. A well-designed study can still produce an uncertain or null result, and such results remain informative when reported honestly.

Ethics, Power, and Research Integrity

Educational research frequently involves people in relationships where power is unequal: students and instructors, children and adults, trainees and supervisors, employees and managers, or communities and institutions. Ethical research therefore requires more than a signed form.

Core responsibilities include informed and voluntary consent, appropriate assent for children where relevant, clear information, proportionate risk, privacy, confidentiality, secure data handling, fair recruitment, respect for withdrawal rights, and plans for unexpected disclosures or distress. The exact requirements depend on local law, institutional rules, the population, and the study design.

The British Educational Research Association's 2024 ethical guidelines emphasize ongoing, reflexive ethical decision-making as well as responsibilities to participants, stakeholders, the research community, publication, the environment, and researchers themselves. The American Educational Research Association also provides a professional code of ethics and guidance on human research protections.

Ethics review is a minimum governance process, not a one-time guarantee that every later decision is ethical. New issues can arise during recruitment, observation, interviewing, data linkage, analysis, or dissemination. Researchers should reassess risks when the project changes.

Digital research creates additional questions. Public visibility does not always mean that people reasonably expected their posts to become research data. Artificial intelligence can introduce privacy, bias, provenance, authorship, and reproducibility concerns. If you use AI tools, follow institutional requirements, protect confidential data, verify outputs, and disclose uses that materially affect the research process.

Research integrity also includes accurate records, honest analysis, appropriate authorship, disclosure of conflicts of interest, correction of errors, and avoidance of fabrication, falsification, and plagiarism.

Reflexivity, Positionality, and Equity

Reflexivity is the disciplined examination of how your position, assumptions, relationships, incentives, and methodological choices shape the research. It is important in qualitative research and also matters in quantitative work through decisions about constructs, datasets, model specifications, exclusions, comparison groups, and interpretation.

A positionality statement can describe relevant aspects of the researcher's relationship to the setting or topic, but it should not become a biography disconnected from methodological consequences. Ask what your position enabled you to see, what it may have obscured, how participants experienced your role, and what safeguards or collaborative practices you used.

Equity-sensitive research asks who benefits, who bears risk, whose knowledge is treated as legitimate, who is missing from the sample, whether instruments work comparably across groups, and how findings might be used. Inclusive design may require accessible materials, translation, community consultation, compensation practices, and dissemination in formats participants can actually use.

Practitioner Inquiry, Action Research, and Design-Based Research

Educational research is not limited to studies conducted from outside a classroom. Action research is often carried out by practitioners who investigate their own context in cycles of planning, action, observation, and reflection. Its strength can be close relevance to practice; its challenges include role conflict, voluntariness, confirmation bias, and limited transfer beyond the setting.

Design-based research develops and studies educational interventions through iterative cycles in real settings. It typically aims both to improve a design and to refine theoretical understanding about how and why learning occurs. Because the intervention changes during the project, transparent documentation of design decisions is essential.

Evaluation asks about the merit, worth, implementation, outcomes, or impact of a program or policy. Evaluation may use experimental, quasi-experimental, qualitative, mixed, economic, and implementation methods. A useful evaluation begins with the decisions stakeholders need to make and the theory of how the program is expected to work.

Evidence Synthesis, Meta-Analysis, and Research Use

A single study rarely settles an educational question. Evidence synthesis examines findings across studies while considering differences in design, population, intervention, outcome, and context.

In a meta-analysis, compatible quantitative effect estimates are combined statistically. A forest plot displays study estimates and uncertainty and may include a pooled estimate. Heterogeneity is not a nuisance to hide; it can indicate that effects vary meaningfully across contexts, populations, measures, or implementations.

A pooled effect should not be interpreted without examining study quality, risk of bias, publication bias, heterogeneity, and the substantive comparability of studies. Large datasets and many studies do not remove systematic error.

For research use, distinguish statistical significance, effect size, practical significance, and implementation feasibility. A modest average effect may matter if an intervention is inexpensive and scalable, while a larger effect may be less useful if it requires unrealistic resources. Educational decisions should combine the best available evidence with professional expertise, local context, learner needs, values, costs, and possible unintended consequences.

Open and Transparent Research Practices

Transparency makes it easier to evaluate how a result was produced. Depending on the study, useful practices may include preregistration, registered reports, version-controlled code, open materials, data dictionaries, analytic notebooks, and clearly labeled exploratory analyses.

Open data is not always ethical or legal. Student records, interview transcripts, small-group datasets, and linked administrative records may remain identifiable even after direct identifiers are removed. Responsible openness means sharing what can be shared safely and explaining restrictions on what cannot.

Preregistration records planned questions, outcomes, hypotheses, sampling rules, and analyses before results are known. It can reduce undisclosed analytical flexibility, but it does not forbid justified changes. Deviations should be documented and explained. Exploratory work is scientifically valuable when it is labeled as exploratory.

Reporting and Communicating Educational Research

A strong report makes the logic of the study traceable. Readers should be able to understand the problem, theoretical framing, context, participants, procedures, instruments, analysis, ethics, results, uncertainty, limitations, and the basis for each conclusion.

Avoid causal verbs when the design supports only association. Report effect sizes and uncertainty where appropriate. In qualitative work, connect interpretations to sufficient evidence while protecting confidentiality. In mixed-methods work, show what was learned through integration rather than writing two unrelated mini-studies.

Limitations should be specific and consequential. Instead of saying "the sample was small," explain what the sample size limits: precision, subgroup analysis, transferability, or detection of rare outcomes. Instead of saying "more research is needed," state what uncertainty remains and what design would reduce it.

Communicate beyond academic journals when appropriate. Participants, educators, families, administrators, policymakers, and community partners may need plain-language summaries, visualizations, workshops, briefs, or videos. Ethical dissemination includes accurate uncertainty and avoids turning tentative findings into slogans.

Research Standards and Further Reading

BERA Ethical Guidelines for Educational Research: Current guidance for ethical educational research, including consent, privacy, inclusion, publication, digital research, and researcher responsibilities.
AERA Professional Ethics: The American Educational Research Association's ethics resources and human research protections guidance.
What Works Clearinghouse Procedures and Standards Handbook: Standards used to review and synthesize evidence on educational programs, products, practices, and policies.
PRISMA 2020: Reporting guidance and flow diagrams for systematic reviews.
Standards for Educational and Psychological Testing: Open-access guidance developed by AERA, APA, and NCME for the development and use of educational and psychological tests.

Interactive Tasks

Quiz: Test Your Knowledge

Which feature best characterizes educational research? (Systematic collection and analysis of evidence about education) (!Personal opinion about teaching without evidence) (!A method that always requires an experiment) (!A process limited to university laboratories)

What is the main purpose of random assignment in an experiment? (To create comparable conditions for causal inference) (!To guarantee a representative population sample) (!To eliminate every form of measurement error) (!To make qualitative evidence unnecessary)

What makes a mixed methods study genuinely integrated? (It connects qualitative and quantitative evidence to answer one study purpose) (!It collects two datasets that are never compared) (!It uses a survey with many questions) (!It reports statistics before an interview)

What does random sampling primarily support? (Generalization from a sample to a defined population) (!Causal attribution between treatment conditions) (!Perfect measurement of every construct) (!Automatic removal of participant attrition)

Which statement about validity is most accurate? (Validity concerns the justification for an interpretation or use of evidence) (!Validity is a permanent property of a questionnaire) (!Validity means that scores never change over time) (!Validity is identical to internal consistency)

What is the purpose of synthesizing a literature review? (To identify patterns disagreements gaps and methodological limits across studies) (!To list article titles in publication order) (!To count citations without reading study methods) (!To replace the need for a research question)

What does reflexivity ask a researcher to examine? (How researcher position and decisions shape the inquiry) (!How to hide every methodological limitation) (!How to avoid documenting analytic choices) (!How to make all participants agree)

How should informed consent be understood in ethical research? (As an ongoing process of information choice and respect) (!As a signature that removes all later ethical duties) (!As permission to publish identifiable data without limits) (!As a substitute for institutional review)

Why report an effect size with uncertainty? (To communicate the estimated magnitude and its precision) (!To prove that the result is important in every context) (!To guarantee that no confounding exists) (!To avoid describing the research design)

Which statement is safest when a study finds an association without a causal design? (The variables are related in the observed data) (!Changing one variable definitely causes the other to change) (!The association proves the mechanism) (!The result applies to every educational setting)

Memory Game

Randomization	Assignment by chance to study conditions
Triangulation	Examination of a phenomenon through multiple sources or perspectives
Reflexivity	Systematic consideration of the researcher's role and assumptions
Saturation	A debated marker of analytic adequacy in some qualitative traditions
Confounding	Mixing of an exposure effect with another related influence
Preregistration	Time-stamped documentation of planned research decisions before results are known

Drag and Drop

Match the correct terms.	Topic
Random assignment	Allocation of units to study conditions by chance
Purposive sampling	Selection of information rich cases for a defined qualitative purpose
Operationalization	Translation of an abstract construct into observable indicators
Member reflection	Participant engagement with interpretations or findings when methodologically appropriate
Confidence interval	Range that expresses uncertainty around an estimated quantity

Crossword Puzzle

Sampling	What process selects cases or units for a study?
Validity	What concept concerns whether an interpretation or use of evidence is justified?
Ethics	What field guides responsible treatment of participants and research conduct?
Interview	What data collection method uses researcher questions and participant responses?
Triangulation	What approach compares evidence from multiple sources methods or perspectives?
Reflexivity	What practice examines how researcher position and decisions shape inquiry?

LearningApps

Cloze Text

Educational research begins with a focused

that identifies what the inquiry seeks to understand. A defensible design aligns evidence collection with the kind of

the researcher wants to make. Random

strengthens causal inference by creating comparable study conditions on average. Random

primarily supports generalization from a sample to a defined population. A construct becomes observable through careful

. Quantitative estimates should be reported with measures of

. Qualitative researchers use

to examine how their position and decisions shape interpretation. Mixed methods requires explicit

of qualitative and quantitative evidence. Ethical participation depends on informed and voluntary

. Transparent research distinguishes planned analyses from later

.

Open-Ended Tasks

Easy

Research Question Audit: Difficulty: Easy. Take a broad educational issue, write three possible research questions, classify each as descriptive, relational, causal, interpretive, or design-oriented, and explain what evidence each would require.
Source Map: Difficulty: Easy. Select six scholarly sources on one educational topic and create a one-page concept map showing agreements, disagreements, methods, populations, and one unanswered question.
Observation Protocol: Difficulty: Easy. Design a short classroom-observation protocol for a public or simulated learning event, identify what each observation category can and cannot show, and produce a simple visual version of the protocol.
Survey Item Critique: Difficulty: Easy. Find or draft eight survey items about a learning construct, identify wording problems, revise the items, and explain how you would pilot them with university students.

Standard

Mini Interview Study: Difficulty: Standard. With appropriate consent and course approval, interview two volunteers about a learning experience, transcribe a short segment, code it, and write a reflexive note about how your questions shaped the data.
Sampling Plan: Difficulty: Standard. Design a probability or purposive sampling plan for a realistic university study, identify the target population and sampling frame, anticipate nonresponse or access problems, and justify sample adequacy.
Open Data Analysis: Difficulty: Standard. Use an openly available educational dataset to create one visualization and one statistical estimate, document all cleaning decisions, and write a paragraph distinguishing association from causation.
Research Ethics Interview: Difficulty: Standard. Visit or contact a university research ethics office, library research-support unit, or data-protection team, interview a staff member about one recurring ethical challenge, and summarize the practical lessons for student researchers.

Advanced

Educational Experiment Design: Difficulty: Advanced. Design a small educational intervention study, specify assignment, outcomes, threats to validity, power considerations, attrition handling, and an ethical implementation plan; run only if formal approval is obtained.
Mixed Methods Proposal: Difficulty: Advanced. Develop a convergent or sequential mixed-methods proposal, include a joint display that demonstrates integration, and explain what the combined evidence can answer that either strand alone cannot.
Systematic Review Protocol: Difficulty: Advanced. Write a protocol for a focused educational review with databases, search terms, eligibility criteria, screening procedures, quality appraisal, synthesis plan, and a draft PRISMA-style flow structure.
Research Translation Video: Difficulty: Advanced. Produce a three-minute evidence-based video for educators or policymakers that translates a complex finding, shows uncertainty, names limitations, and explains what action the evidence does and does not justify.

Learning Assessment

Design Alignment Assessment: Given an educational problem, produce a research question, design, sampling plan, measurement strategy, analysis plan, and claim statement, then justify the alignment among all six components.
Causal Reasoning Assessment: Critique a study that reports an association as a causal effect, identify plausible confounders and design limitations, and propose a stronger design that would improve causal identification.
Qualitative Interpretation Assessment: Analyze a short interview or observation dataset, develop an evidence-based interpretation, identify a disconfirming case, and explain how reflexivity affects the conclusion.
Measurement Evaluation Assessment: Evaluate a proposed educational scale using evidence about reliability, validity, fairness, and context of use, then recommend whether and how the scale should be used.
Ethical Decision Assessment: Work through a scenario involving student participants, power imbalance, digital trace data, and unexpected disclosure; propose a response that addresses consent, privacy, harm, institutional rules, and participant rights.
Evidence Synthesis Assessment: Compare findings from several studies with different designs and contexts, judge the strength and transferability of the overall evidence, and write a recommendation that distinguishes certainty from practical judgment.

Evidence of Learning

Important evidence of learning includes your ability to explain how research questions, designs, data, analyses, and claims fit together; distinguish description, association, interpretation, and causation; and evaluate the quality of quantitative, qualitative, and mixed-methods evidence.

You should be able to produce research products such as a coherent proposal, literature synthesis, sampling plan, instrument or interview guide, reproducible analysis, coding framework, reflexive memo, ethics plan, visualization, and audience-appropriate research summary.

Transfer is demonstrated when you can apply these principles to an unfamiliar educational problem, recognize when a favored method is not appropriate, adapt a design to a new context without overstating generalizability, and make a reasoned decision under uncertainty.

A strong portfolio therefore shows not only what you know, but also how transparently you reason, how responsibly you work with participants and data, how accurately you communicate limitations, and how you revise your conclusions when the evidence changes.

OERs on the Topic

Additional open resources include the ethical, reporting, evidence-review, and testing standards linked earlier in this course. When using any OER, check its license, date, authorship, and suitability for your local context.