Two studies in one document is not mixed methods
The most common criticism of mixed methods dissertations is that the qualitative and quantitative strands never actually meet. Both are competently done, they sit in separate chapters, and the conclusion mentions both. Integration is the thing being assessed, and it has to be designed in rather than added at the end.
Get your free quote
We reply within 1 business hour, no delays.
No card details required · 100% confidential · On time, or it's free
Why are you mixing, and what does each strand do for the other?
There are good answers. The qualitative strand explains a surprising quantitative result. The quantitative strand tests whether a pattern found in interviews generalizes. The qualitative strand builds an instrument the quantitative strand then validates. Each of those implies a different sequence and a different point of integration.
There is also a common bad answer, which is that more methods seemed more thorough. That produces exactly the parallel-strands problem, because without a reason for mixing there is no natural place for the strands to meet.
Deciding this at the design stage determines your sequence, your sampling, and where integration happens. Deciding it after both datasets are collected means integration has to be retrofitted, which is possible but visibly harder and usually less convincing.
- A stated rationale for mixing, not just a stated design name
- Sequence and priority decided deliberately and justified
- Integration planned at specific points rather than left to the conclusion
- Joint displays that show the strands actually meeting
What we work through
Best before collection, still worthwhile after
- Your rationale for mixing, written so a committee accepts it
- Design type, sequence and priority, with the trade-offs named
- Sampling across strands, including whether the same people appear in both
- Integration points and how each will be evidenced
- Joint displays and how to present them
Four shapes, and what each is for
Explanatory sequential
Quantitative first, then qualitative to explain it. The most common choice in dissertations, and it works well when your survey produces a result you did not expect and need to understand.
Exploratory sequential
Qualitative first, then quantitative. Suits building an instrument or a model from interviews and then testing whether it holds at scale.
Convergent parallel
Both strands at once, compared at analysis. Faster, and the hardest to integrate convincingly, because there is no built-in moment where one strand informs the other.
Embedded
One strand plays a supporting role inside a design led by the other, such as qualitative interviews inside a trial. The supporting strand still needs a clear job.
Where the strands are supposed to meet
Integration happens at specific points. Naming them is what turns two studies into one.
At design
One strand's questions are built from the other's findings or from a shared framework. This is the earliest and strongest form, and it has to be planned.
At sampling
Qualitative participants selected on the basis of quantitative results, such as the outliers or the cases that contradict the trend. Immediately makes the strands dependent on each other.
At analysis
Themes compared against statistical findings, quantitative results used to weight qualitative interpretation, or qualitative codes quantified where that is appropriate to your position.
In a joint display
A table or figure placing both strands side by side against the same question. This is the single most persuasive piece of evidence that integration happened, and most dissertations that lack one are graded down for it.
In interpretation
A discussion organized by research question rather than by method, so each answer draws on both strands. A discussion with a qualitative half and a quantitative half is the parallel-strands problem restated.
Your design, your analysis, your integration
- We do not run either strand of your analysis for you.
- We do not write your methodology, findings or discussion chapters.
- We do not fabricate, alter or supply data.
- We do not manufacture agreement between strands that disagree.
- We teach the design, work through your own data with you, and review what you write.
When strands genuinely disagree, that is a finding rather than a problem. Divergence handled openly is one of the more interesting things mixed methods work can produce, and hiding it is both dishonest and a wasted opportunity. Read the full policy.
Frequently Asked Questions
Integration. The strands have to inform each other at some identifiable point, whether in design, sampling, analysis or interpretation. Two analyses reported in sequence with a conclusion mentioning both is the definition of the problem, not the method.
It follows from why you are mixing. If you have a quantitative result you cannot explain, explanatory sequential. If you need to build something before you can measure it, exploratory sequential. If time is short and you accept harder integration, convergent. Pick the reason first and the design name follows.
No, and it depends on your design. In explanatory sequential work it is common and useful to select qualitative participants from the quantitative sample, often deliberately choosing unusual cases. In other designs entirely separate samples are appropriate. Either way, say which and why.
A table or figure putting both strands next to each other against the same research question, so a reader can see where they converge and diverge. It is the clearest available evidence that integration actually happened, and it is worth building even where your committee has not asked for one.
Report it and take it seriously. Divergence can mean the strands measured slightly different things, that context matters in a way the survey could not capture, or that one instrument was weak. Working through why is usually more interesting than agreement would have been.
Yes, meaningfully. Two datasets, two analyses, two sets of assumptions, plus the integration work on top. It is worth doing when the question genuinely needs it and a poor choice when it is being done for thoroughness.
Tell us why you are mixing
That answer decides your sequence, your sampling and where integration happens. If you can answer it clearly the rest of the design follows, and if you cannot that is the first thing to work on.
Talk About My Design