Begin with the question
State the population, phenomenon and context. Distinguish description, comparison, association, prediction and causal explanation. Each purpose requires different evidence; adding a sophisticated technique does not repair an unclear question.
Create an alignment map
- Question: what exactly do you want to learn?
- Objective: what observable task will address that question?
- Evidence: what variables, records, texts or observations are needed?
- Design: how will you obtain and compare that evidence?
- Analysis: which method fits the design and measurement?
- Claim: what can and cannot be concluded?
Worked example
For a question about an association between financial literacy and saving behaviour, an objective could be to estimate that association in a defined population. Specify how both concepts are measured, how the sample is recruited and which confounders matter. An observational association alone does not establish that literacy causes saving.
Check before collecting data
- Verify ethics approval, consent and access requirements.
- Justify sample size and sampling choices.
- Plan missing-data handling, exclusions and primary analysis before inspecting results.
- Ask your supervisor to review the alignment map against institutional requirements.
Need personalised guidance?
Talk it through with a mentor.
Share your programme, project area, current stage and deadline.