03 · What You Need to Know
Understanding the Different Points of Mixed Methods Integration
What Counts as Integration in Mixed-Methods Research?
Integration occurs when quantitative and qualitative components are related in a way that contributes to the study's overarching research purpose.
The relationship may be procedural, such as using survey findings to select interview participants. It may be developmental, such as using qualitative findings to construct a questionnaire. It may also be analytical or interpretive, such as comparing quantitative results with qualitative themes.
Fetters, Curry, and Creswell (2013) distinguish integration at three broad levels: study design, methods, and interpretation and reporting. Their framework provides a useful basis for identifying where integration can occur without assuming that every study must use the same procedures.
The important question is not how many integration activities the researcher can identify. It is whether those activities address the specific problem that requires quantitative and qualitative evidence to be brought together.
Integration at the Study Design Stage
Integration begins conceptually when researchers determine why both approaches are needed and how their contributions will relate to one another.
At this stage, the researcher should identify the overarching question, the contribution expected from each component, the intended sequence or timing, and the general integration strategy.
For example, a researcher who intends to explain unexpected survey results through interviews must anticipate that the qualitative phase will depend on quantitative findings.
That dependency affects the research schedule, participant recruitment arrangements, and the information that must be retained for selecting interviewees.
By contrast, a researcher planning to compare survey responses with interview accounts collected during the same semester must ensure that both components investigate sufficiently related constructs and contexts.
Integration at the design stage does not necessarily involve combining datasets. It involves establishing the logic that makes their eventual relationship methodologically coherent.
Integration Through Connecting Samples
Connecting occurs when one component informs the sampling procedures of another.
For example, quantitative findings may identify participants representing high, low, typical, or unexpected outcome profiles. The researcher then selects interview participants because their experiences may help explain those profiles.
This approach is common in explanatory sequential mixed-methods research.
Connecting may also involve using qualitative findings to identify groups or cases relevant to subsequent quantitative investigation.
Researchers should distinguish sampling connections from analytical integration. Selecting interviewees based on survey results creates a methodological link, but the subsequent analysis should still explain how their accounts relate to the quantitative findings.
Detailed decisions about connecting samples across phases include participant overlap, selection criteria, consent, and the level of inference supported by the linkage.
Integration Through Building One Component From Another
Building occurs when findings from one component inform the data collection approach of another.
A qualitative study may identify constructs, language, or contextual conditions that guide questionnaire development. Alternatively, quantitative findings may inform the content of subsequent interview questions.
For example, interviews with teachers about responsible generative AI use might reveal distinctions among disclosure practices, verification procedures, and assessment responsibilities.
These findings could inform candidate questionnaire dimensions, which would then require appropriate evaluation before the resulting scores are interpreted.
Building is especially relevant when one component must produce information before the next can be designed adequately.
The relationship should be documented through a transparent account of which findings informed which decisions.
Researchers considering qualitative-informed instrument or intervention development should distinguish the initial developmental contribution from subsequent evaluation of the resulting product.
Integration Through Merging Findings
Merging involves bringing quantitative and qualitative evidence together for analysis or comparison.
This is particularly relevant when both components investigate related aspects of a phenomenon and the overarching question concerns how their findings correspond.
For example, a researcher might compare measured online learning engagement with students' accounts of meaningful participation.
The quantitative component may identify patterns of recorded activity. The qualitative component may reveal how students interpret and use the learning platform.
Merging allows researchers to examine whether the findings converge, complement one another, or suggest different interpretations.
However, meaningful merging requires attention to the relationship between constructs, samples, time periods, and units of analysis.
Two datasets cannot necessarily be compared directly simply because they concern the same broad topic.
Integration Through Embedding
Embedding involves incorporating one component within a broader research design, with connections potentially occurring at several stages.
For instance, an experimental evaluation may include qualitative interviews before implementation, observations during the intervention, and interviews after outcome measurement.
These activities could inform intervention refinement, investigate implementation conditions, and contribute to interpretation of the measured outcomes.
Embedding is not restricted to qualitative inquiry within quantitative research. A substantive quantitative component may also be incorporated within a predominantly qualitative investigation.
The defining issue is the relationship between the embedded component and the primary research purpose.
The design implications of embedding one component within another research design should be considered before deciding which integration points are necessary.
How Do Integration Approaches Relate to Research Purposes?
Different integration purposes generally favor different methodological connections.
| Research purpose |
Possible integration point |
What integration accomplishes |
| Explain particular quantitative findings |
Sampling and qualitative data collection |
Quantitative results guide the selection of cases and follow-up questions. |
| Develop a quantitative instrument |
Transition between qualitative analysis and instrument development |
Qualitative findings inform constructs, items, and subsequent measurement procedures. |
| Compare different representations of a phenomenon |
Analysis and interpretation |
Quantitative and qualitative findings are examined together. |
| Understand intervention implementation |
Before, during, or after intervention delivery |
Implementation evidence informs development or interpretation of outcomes. |
| Investigate disagreement between findings |
Integrated analysis and interpretation |
Researchers examine differences in constructs, contexts, samples, or possible explanations. |
These relationships are illustrative rather than exclusive. A study may use connecting and merging, building and merging, or several forms of integration across a multistage design.
Researchers should select integration points because they serve the question, not because a greater number of connections appears more sophisticated.
Must Integration Occur During Data Analysis?
Not every mixed-methods study requires the same form of integrated analysis, but the overall interpretation should demonstrate what is learned from the relationship between components.
For example, a qualitative phase may inform the development of a quantitative instrument. The developmental connection itself constitutes meaningful methods-level integration.
The final report should nevertheless explain how the qualitative findings shaped the instrument and what the quantitative evaluation subsequently established.
In a convergent study, integrated analysis may be more central because the research question requires direct comparison of the findings.
Thus, the analytical procedures should follow the integration purpose rather than a universal requirement to combine raw datasets.
What Is a Joint Display, and When Is It Useful?
A joint display is a table, figure, or other structured presentation that brings quantitative and qualitative evidence together to support integrated analysis and interpretation.
Guetterman, Fetters, and Creswell (2015) examined published joint displays and identified their usefulness in communicating relationships between different forms of evidence.
For example, a joint display might organize quantitative engagement profiles alongside corresponding qualitative themes and an integrated interpretation.
It can help researchers identify patterns that are difficult to communicate through separate results sections.
However, a joint display should do more than place numerical findings and interview quotations next to each other. Its analytical value depends on the relationships examined and the conclusions developed.
A joint display is therefore a means of integration, not a substitute for methodological reasoning.
How Do You Integrate Findings Collected From Different Samples?
Different samples do not automatically prevent integration, but they affect the kinds of comparisons and conclusions that are defensible.
Suppose a survey examines students' online learning experiences across several universities, while interviews are conducted with teachers at a subset of those institutions.
The findings may be integrated at an institutional or contextual level if the research questions and available evidence support that relationship.
However, the researcher should not interpret teachers' accounts as explanations of particular students' survey responses without appropriate linkage or supporting evidence.
Similarly, individual-level findings should not be treated as institutional-level relationships without considering aggregation and the possibility of ecological or individualistic inference errors.
Integration requires clarity about what is being compared and at which level the resulting inference applies.
Does Integration Require Converting Qualitative Data Into Numbers?
No. Some studies transform qualitative information into quantitative variables or organize quantitative findings into qualitative categories, but data transformation is not a universal requirement.
Researchers may integrate evidence through sampling connections, instrument development, joint displays, or interpretive narratives without converting the original data.
Quantitizing qualitative themes can be useful for particular questions, but it may also obscure contextual meaning when applied without an appropriate analytical rationale.
Likewise, converting quantitative results into broad categories may discard information and introduce arbitrary distinctions.
The decision to transform data should be justified by the intended inference and the characteristics of the evidence.
How Should Integration Be Reported?
Reporting should make clear where integration occurred, how it occurred, and what additional understanding resulted.
The Good Reporting of A Mixed Methods Study recommendations, commonly known as GRAMMS, emphasize describing the justification for mixed methods, the design's purpose and priority, the procedures for each component, the points of integration, and the insights gained from combining methods.
Researchers should distinguish integration planned before data collection from integration decisions made during the investigation.
For sequential studies, the report should explain how first-phase findings informed subsequent sampling or data collection.
For convergent studies, it should explain how findings were compared and how similarities or differences affected interpretation.
For embedded studies, it should describe how the supporting component contributed to the broader design.
The results should demonstrate an integrated contribution rather than simply presenting two separate accounts.
What If Integration Reveals Contradictory Findings?
Integration does not require quantitative and qualitative findings to agree.
For example, quantitative indicators may suggest high participation in an online course while qualitative interviews reveal that students perceive much of their activity as administratively required rather than educationally meaningful.
The findings may reflect different aspects of engagement rather than a direct contradiction.
Researchers should examine whether the components measure or investigate the same constructs, whether their samples and time periods are comparable, and whether the evidence supports alternative interpretations.
Sometimes integration produces a coherent explanation. In other cases, it reveals uncertainty or a need to reconsider the conceptual framework.
The appropriate response to conflicting quantitative and qualitative findings is investigation rather than automatic reconciliation.
Can Integration Be Planned Too Rigidly?
Yes. Integration planning should be sufficiently explicit to support methodological coherence while allowing appropriate responses to findings that emerge during the study.
For example, an explanatory sequential design may specify that interviews will investigate unexpected quantitative patterns without knowing in advance which particular patterns will appear.
The researcher can establish decision criteria and a sampling framework while retaining flexibility to address the actual findings.
However, flexibility should not become a justification for selectively integrating only results that support a preferred interpretation.
Watch Out
Do not leave integration entirely to the final discussion, but do not predetermine what the integrated findings must show. Plan the procedures and decision rules while allowing the evidence to challenge the original expectations.
