The missing explanation in a number
A survey might show that participation in a program fell after the first month. Interviews might help explain how people experienced the schedule, material, or registration process. The numbers describe a pattern; the interviews can investigate meanings and circumstances connected to that pattern.
That combination becomes mixed methods research when the study deliberately connects the approaches. NIH’s mixed methods guidance emphasizes integration rather than simply placing a quantitative and qualitative component beside each other. The purpose is to answer a question more fully through their relationship.
Explanatory designs start with a pattern
In a hypothetical explanatory sequence, researchers first analyze attendance data and identify different participation patterns. They then sample participants for interviews to understand those patterns. The quantitative result helps shape the qualitative sampling and questions.
The interviews do not automatically prove why attendance changed. Participants may offer different explanations, and the sample may not represent everyone. Their accounts can still identify mechanisms, constraints, or interpretations that the attendance records could not show.
A useful integrated conclusion might say that the decline was concentrated in certain time slots and that interviewed participants described scheduling conflicts. That connects evidence while preserving the difference between an observed pattern and a demonstrated causal explanation.
Exploratory designs can begin with experience
The sequence can also run the other way. Interviews may reveal how people describe an unfamiliar issue, which then informs a survey intended to examine how those experiences are distributed. The qualitative component helps define concepts rather than merely decorating the numerical results with quotations.
For example, a fictional study of service access might discover that “difficulty getting an appointment” includes several different problems: booking technology, unsuitable times, or uncertainty about eligibility. A later survey could distinguish those experiences rather than asking one ambiguous question.
The resulting questionnaire still needs appropriate development and evaluation. Deriving items from interviews does not automatically establish reliability, validity, or representativeness. Each stage creates its own evidence and its own limitations.
Convergent designs bring findings together
In a convergent design, quantitative and qualitative components may be conducted during a similar period and then compared or integrated. The methods can examine complementary aspects of the same issue. Agreement may strengthen an interpretation, but disagreement can also be informative.
Suppose a satisfaction survey is generally favorable while interviews describe substantial frustration. Possible explanations include who responded, how the questions were phrased, what the scale measured, or whether people valued the service while disliking one part of it. The discrepancy creates a research question rather than a reason to discard one component automatically.
What integration looks like on the page
| Quantitative finding | Qualitative finding | Integrated question |
|---|---|---|
| Attendance falls after month one. | Participants describe schedule changes. | Does timing help explain the observed pattern? |
| Average satisfaction is high. | Some interviews describe access problems. | Are satisfied users and people unable to access the service represented differently? |
| Outcomes vary across locations. | Staff describe different implementation practices. | Which contextual differences deserve closer examination? |
These are fictional examples. A joint display like this can make the reasoning visible, but a table alone does not establish that integration was rigorous. The study still needs a defensible connection between samples, measures, and interpretations.
Why adding a second method is not automatically better
Each component takes time and expertise. A small qualitative component added without a clear purpose can produce superficial interpretation. A survey added to a deeply contextual question may measure a construct poorly. More methods create more opportunities for insight, but also more ways to introduce weak evidence.
The reason for combining approaches should follow from the research question. If one well-designed method can answer it adequately, a second method is not required as a badge of sophistication. If the question concerns both distribution and experience, a deliberate combination may be especially informative.
The limits remain visible
Mixed methods does not automatically make a convenience sample representative, an association causal, or an interview account exhaustive. The components can address different uncertainties while leaving others unresolved. The final interpretation needs to state which conclusions come from which evidence and how the connection was made.
That is what makes the approach useful: it can preserve the specificity of people’s accounts while also examining broader patterns, without pretending either form of evidence answers every part of the question alone.
