3.4 Choosing a design: a decision framework
With three core designs to choose from, how do you decide which is right for your study? The key question is: what do you need to know, and in what order?
Consider the following decision points:
- Start with your research question: Does your question have both ‘how many/how much’ and ‘why/how’ dimensions? If so, mixed methods may be appropriate.
- Consider what you already know: Do you have existing data or established measures? If you already have quantitative data showing a pattern you want to understand, explanatory sequential makes sense. If you are starting from scratch with little prior research, exploratory sequential may be better. If you want to compare or validate findings from different angles, convergent is appropriate.
- Think about timing and resources: Convergent designs require capacity to collect and analyse both datasets in the same timeframe. Sequential designs spread the work but take longer overall. What is feasible given your timeline and resources?
- Consider your audience: Who will use your findings? Some audiences value quantitative evidence more highly; others value qualitative depth. Your design can reflect these priorities.
The following table summarises when each design is most appropriate.
| Design | Use when you want to … | Key question |
|---|---|---|
| Convergent | Validate or triangulate findings from different sources | ‘Do my survey and interview findings point in the same direction?’ |
| Explanatory sequential | Understand quantitative results in more depth | ‘The numbers show X. Why? What does this mean?’ |
| Exploratory sequential | Build theory or develop measures from qualitative insights | ‘What themes emerge that I could then measure at scale?’ |
Here is a short decision walkthrough:
- Does your question genuinely require both QUAN and QUAL? If no, stop and use a single approach.
- Do you already have one type of data (for example, an existing survey)? If yes, lean towards explanatory or exploratory sequential; if no, all three designs remain on the table.
- Can you collect both data types in the same time window? If no, sequential; if yes, convergent or sequential are both feasible.
- Who is your audience and what kind of evidence will they trust most? This shapes which strand carries the weight: QUAN-led, QUAL-led or balanced.
Activity 3.3: Selecting a design
Read the following research scenario and consider which design would be most appropriate.
You are investigating why academic staff in your faculty report lower wellbeing than staff in comparable faculties. You have access to annual staff survey data showing the pattern clearly, but the survey does not reveal the underlying reasons. You want to understand the causes so that interventions can be designed.
Consider the following questions:
- What do you already know? What do you still need to find out?
- What type of data would help you answer the ‘why’ question?
- Which design would be most appropriate and why?
Write your reasoning in the box below, then check the discussion.
Discussion
This scenario points toward an explanatory sequential design because:
- you already have quantitative data (the staff survey) showing a clear pattern (lower wellbeing in your faculty)
- what you lack is an understanding of why this pattern exists
- qualitative data (interviews, focus groups) would help you explore the underlying reasons
- the logical sequence is that you start with what you know (the quantitative pattern) and then investigate further (qualitative exploration of causes).
An explanatory sequential design would involve analysing the existing survey data to identify specific dimensions of wellbeing that are lower and then conducting interviews with staff to understand what factors contribute to those patterns. The qualitative findings would explain and contextualise the quantitative results and inform intervention design.
A convergent design would be less appropriate here because you are not trying to triangulate or validate. You already trust the survey findings and you want to explain them. An exploratory design would be less appropriate because you are not starting from scratch or developing new measures.
Exploratory sequential design


