Numbers first, then the explanation — and the participant-selection decision that decides whether the design was worth running.
An explanatory sequential design collects and analyses quantitative data first, then uses qualitative work to explain what the numbers showed — particularly the results that were surprising, weak or contradictory. The whole value of the design sits in one decision: phase two participants are selected on the basis of phase one results. Interviewing a fresh random sample instead is the most common way doctoral versions of this design end up as two unrelated studies.
Use it when you expect to find patterns you will not be able to interpret from the numbers alone. That is a reasonable expectation in most research on human behaviour, which is why this is the most common mixed-methods design in doctoral work.
Typical situations where it earns its place:
It is the wrong choice when you already know what explanation you expect — then you are illustrating rather than explaining — and when time is tight, because the phases genuinely cannot overlap.
This is what distinguishes the design from running a survey and some interviews. Phase one results decide who you speak to in phase two, and the selection logic has to be stated explicitly in the methodology chapter.
Established ways of selecting, each answering a different kind of question:
| Selection logic | Who you approach | Answers |
|---|---|---|
| Extreme cases | Highest and lowest scorers on the key variable | What distinguishes those at each end |
| Outliers and poor fit | Cases the model predicted badly | What the model is missing |
| Typical cases | Participants close to the group average | What the ordinary case looks like from inside |
| Contrasting groups | Members of groups that differed significantly | Why the difference exists |
| Confirming and disconfirming | Cases consistent with, and contrary to, the pattern | Where the explanation holds and where it breaks |
Two practical requirements follow, and both need planning at the proposal stage. You must be able to link phase two participants back to their phase one data, which means the survey cannot be fully anonymous — usually an identifier plus consent to be contacted again, with the implications for confidentiality explained to participants. And your interview guide is written after phase one analysis, because the whole point is to ask about the specific results you found.
That second point catches people out. If your interview schedule was written before the survey was analysed, the design has collapsed into two parallel studies, and an examiner will see it immediately.
The interim analysis is a real phase. Between collection points sits a period of analysing phase one, deciding what needs explaining, selecting participants and writing the guide. It routinely takes longer than expected and is routinely absent from doctoral timelines. Build it in explicitly.
The phases cannot be parallelised. Any delay in phase one moves everything after it. Where recruitment for phase one is slow — which is common — the whole project slips, and there is no way to work around it.
Ethics approval usually needs handling in two stages. You cannot fully specify phase two before phase one is analysed, since you do not yet know who you will approach or what you will ask. Committees deal with this regularly: the usual route is to describe the selection logic and the general area of questioning in the original application, then submit an amendment with the final guide. Ask your committee what they prefer before submitting, because being sent back costs weeks you cannot recover in a sequential design.
Attrition between phases is normal. Not everyone who agrees to be contacted will respond months later. Recruit phase one with that in mind, and collect more consents to follow-up than you expect to use.
The commonest structural mistake is reporting the two phases as separate studies with a short discussion at the end pointing out that both happened. The integration has to be visible in the results, not deferred.
What works:
The test for the whole thesis: could you state your conclusion in one sentence that needs both phases? If the sentence survives deleting either one, the design has not delivered what it cost.
A mixed methods design where quantitative data is collected and analysed first, and qualitative work then explains those results — particularly findings that were surprising, weak or contradictory. Its defining feature is that phase two participants are selected on the basis of phase one results, which is what makes it one study rather than two.
From your quantitative results, using a logic you state explicitly. Common approaches are selecting extreme cases at each end of a key variable, cases the model predicted poorly, typical cases near the average, members of groups that differed significantly, or a mix of confirming and disconfirming cases. This requires that phase one is not fully anonymous, so plan an identifier and consent to re-contact from the start.
No — and doing so undermines the design. The guide exists to ask about the specific results you found, so it is written after phase one analysis. An interview schedule prepared in advance shows that phase two was never genuinely responsive to phase one, which turns the study into two parallel pieces of work and is easy for an examiner to spot.
Committees handle this routinely. The usual route is to describe your participant selection logic and the general area of questioning in the original application, then submit an amendment with the final interview guide once phase one is analysed. Ask your committee what they prefer before submitting, since being sent back costs weeks that a sequential design cannot absorb.
Longer than people plan for, because the phases cannot overlap and there is a third stage between them that rarely appears in timelines: analysing phase one, deciding what needs explaining, selecting participants and writing the guide. Any delay in phase one recruitment pushes everything after it, with no way to work around it.
Anonymous phase one data and pre-written interview guides are the two mistakes that cannot be fixed afterwards. Describe your design and a PhD in your field will check the phases will actually connect.
Leave an email or a WhatsApp number — whichever you prefer — and tick how we should reach you. We reach out within 30 minutes.
