EDUF4044 Chap.5 Choosing a Research Design
Choosing a Research Design
This chapter selects research design from the evidential job rather than from a preferred method. Questions about experience, pattern, change and context point toward different combinations of interviews, surveys, observations, experiments and case study.
The chapter tests alignment across constructs, participants, time, data and analysis, and it explains why mixed methods requires planned integration rather than the mere presence of several data sources.
What this chapter covers
- 01
Claims and evidential jobs
- 02
Question-design alignment
- 03
Participants and sampling
- 04
Time and comparison
- 05
Mixed evidence and integration
- 06
Internal and external boundaries
Worked example · free
Designing a retrieval-practice inquiry
- +1Define the outcomes separately. These marks are not an official marking scheme; they organise a fresh design-choice exercise.
- +1Use repeated achievement measures cautiously.
- +1Gather focused student experience data.
- +1Document implementation fidelity.
- +1Integrate evidence for a local decision.
Key terms
- Mixed methods
- A design that intentionally integrates qualitative and quantitative evidence to answer related questions.
- Counterfactual
- A reasoned account of what would have happened without the intervention or exposure.
- Implementation fidelity
- Evidence about whether an intervention occurred as planned and with the intended dosage.
- Convenience sample
- Participants selected mainly because they are accessible to the researcher.
Choosing a Research Design FAQ
Which research design is strongest?
Strength depends on the question and claim. Experiments can strengthen causal inference, interviews can illuminate experience, and case studies can explain context and process. A design is strong when its logic fits the uncertainty and its limitations are managed transparently.
Does using several methods create triangulation?
Not automatically. Methods may examine different constructs or participants. Explain whether they test convergence, provide explanation or reveal distinct dimensions, and show how the findings will be compared to answer the same decision.
Can practitioner research use convenience sampling?
Yes, when access is part of the design and claims remain local. Describe who could participate, who did participate and which voices may be missing. Do not present convenience as population representativeness.
Assessment move
For each possible design, write what it can show, what it cannot show and which practical trade-off matters most. Draft one alignment sentence naming the question, participants, evidence, time and analysis. Reject any method that measures a different construct or demands access you do not have. Explain mixed-method integration before collecting either form of evidence. Create a design comparison table for your main question.
Across the top, place interview, survey, observation, experiment and case study only where each is plausible. Down the side, compare construct fit, participant access, time structure, analytic workload, ethical burden and claim strength. Do not award an overall score; write the trade-off that matters for your decision. Next, draw the study timeline.
Mark recruitment, baseline, intervention or event, each data point and interpretation. The drawing will reveal whether a proposed change question has enough time information and whether data sources refer to the same period. For sampling, describe the pathway from the wider group to invited participants and then to actual contributors. Note where convenience, self-selection or exclusion enters.
For mixed evidence, write expected relationships before collection: convergence, explanation, expansion or productive disagreement. Decide how a conflict will change the final judgement rather than treating it as a problem to hide. Review the design from the unit of inference: response, learner, lesson, class or school. Many rows of data from one class do not justify a school-wide conclusion. Finally, conduct a claim swap.
Replace the intended method with a weaker one and notice which verbs must change; then add the chosen design features back one at a time. This exercise makes the inferential contribution of comparison, repetition, sampling and context explicit. Practise defending one rejected design as well as the selected design.
Explain why the rejected option is attractive, then identify the decisive mismatch in construct, access, time or inference. This produces a stronger justification than a generic list of disadvantages. Ask what evidence would change the professional decision and whether the proposed measures can register that change. If not, redesign the outcome before expanding the sample.
End with a feasibility budget that assigns realistic time to recruitment, collection, cleaning, transcription, analysis and reporting, leaving space for school interruptions and incomplete responses. Review one published study by redesigning it for a different claim. Notice which participants, measures, comparisons and analyses must change.
This makes alignment concrete and shows why methods cannot be moved unchanged between questions. Finish by describing the ethical burden added or removed by the redesign and whether the stronger claim is worth that burden. Close by writing the selected design’s decisive strength, its decisive limitation and the claim that remains defensible between them.
Working through Choosing a Research Design in EDUF4044? Sia is AskSia’s AI Education tutor — ask any EDUF4044 Choosing a Research Design question and get a clear, step-by-step explanation grounded in how EDUF4044 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
