How to Combine Qualitative and Quantitative Data
Combining qualitative and quantitative data is the step that separates actual mixed methods research from "we did a survey and some interviews." It's also the step most teams struggle with. The collection part is familiar, most researchers know how to run a survey and how to conduct interviews. But merging two fundamentally different types of data into a unified set of findings? That's where things get murky.
This guide covers three practical approaches to combining qual and quant data, with specific techniques you can apply to your next mixed methods study.
Why Combining Data Is Hard
Qualitative and quantitative data don't share a common language. Quantitative data speaks in numbers, frequencies, and statistical relationships. Qualitative data speaks in themes, narratives, and meaning. You can't simply average them together.
The challenge isn't theoretical, it's procedural. Most teams analyze their quantitative data in one tool, their qualitative data in another, and then try to stitch findings together in a slide deck. That's not integration. That's two separate reports with a shared cover page.
Real combination requires a structured process for making the two datasets speak to each other.
Three Approaches to Combining Data
Approach 1: Merging (Side-by-Side Comparison)
Merging works best with concurrent triangulation and convergent designs, where both datasets address the same questions and are collected in parallel.
How it works: Analyze each dataset independently, then create a joint display that puts quantitative findings and qualitative findings side by side, organized by research question or topic.
Joint display example:
| Research Question | Quantitative Finding | Qualitative Finding | Meta-Inference |
|---|---|---|---|
| How satisfied are users with onboarding? | Mean satisfaction: 3.2/5; drops to 2.8 for enterprise users | Enterprise users describe "being left alone after signup" and wanting a dedicated contact | Enterprise onboarding needs a human touchpoint, satisfaction gap isn't about the product, it's about support |
| Which features drive retention? | Feature A usage correlates with 30% higher retention | Users describe Feature A as "the thing that makes this worth keeping" | Feature A is both behaviorally and perceptually sticky, protect it in the roadmap |
The meta-inference column is where the real value lives. It's your interpretation of what both findings together mean, something neither dataset could tell you alone.
Approach 2: Connecting (One Dataset Informs the Other)
Connecting is the natural approach for sequential explanatory design. The quantitative findings directly shape the qualitative phase.
How it works: Use quantitative results to determine who to interview, what to ask them, and what patterns to probe. Then use qualitative findings to explain, expand, or reframe the quantitative results.
Practical steps:
- Complete quantitative analysis and identify findings that need explanation.
- Select qualitative participants based on quantitative results (e.g., people who scored unusually high or low, members of an unexpected segment).
- Design interview questions that directly address the quantitative gaps.
- Analyze qualitative data through the lens of the quantitative findings.
- Write an integrated narrative that uses qualitative themes to explain quantitative patterns.
The connection here is sequential and causal: quant findings cause you to ask specific qualitative questions, and qual findings explain the quant results.
Approach 3: Data Transformation
Data transformation converts one type of data into the other's format so they can be analyzed together.
Quantitizing (qual → quant): Convert qualitative themes into numerical codes. For example, if you've coded interview transcripts and found five major themes, you can count how many participants mentioned each theme, calculate the percentage of text devoted to each, or create binary variables (mentioned theme / didn't mention theme) that you can cross-tabulate with quantitative variables.
Qualitizing (quant → qual): Convert quantitative data into narrative descriptions. This is less common but useful when you want to create participant profiles that combine survey scores with interview themes.
Data transformation is particularly useful for embedded designs where you need to integrate open-ended survey responses with closed-ended data. Code the open-ended responses into categories, then treat those categories as another variable in your quantitative analysis.
Building a Joint Display
Joint displays are the workhorse tool for data combination. They're structured tables or matrices that put both types of data in conversation with each other.
Types of Joint Displays
Statistics-by-theme display: Rows are qualitative themes; columns include the quantitative statistics that relate to each theme.
Case-by-case display: Rows are individual participants; columns show their quantitative scores and qualitative responses side by side. Works well when you've used the same participants for both strands.
Timeline display: Rows are time points; columns show quantitative measures and qualitative observations at each point. Useful for longitudinal mixed methods studies.
Process display: Rows are stages of a process (e.g., customer journey stages); columns show quantitative metrics and qualitative experiences at each stage.
Joint Display Best Practices
- Include a meta-inference column. Don't just put data side by side, interpret what the combination means.
- Note convergence and divergence. Flag where findings agree, partially agree, or conflict. Divergent findings are often the most valuable.
- Connect back to research questions. Every row or section of your joint display should map to a specific question you set out to answer.
- Keep it readable. A joint display that requires 20 minutes to parse isn't useful. Summarize rather than reproducing raw data.
Handling Conflicts Between Datasets
When qualitative and quantitative findings disagree, resist the urge to pick the one you like better. Instead:
1. Check for methodological explanations. Did the survey question and the interview question actually ask the same thing? Different wording can produce different results that look like a conflict but are actually measuring different constructs.
2. Consider scope differences. Your survey sampled 500 people; your interviews included 15. A finding that's dominant in interviews but doesn't show up in survey data might reflect a real but uncommon experience, or it might reflect selection bias in your qualitative sample. Review your sampling approach for both strands.
3. Look for nuance. Surface-level agreement can mask deeper disagreement, and vice versa. A survey showing high satisfaction and interviews revealing frustration might not conflict, users might be satisfied overall but frustrated with specific touchpoints.
4. Report the conflict transparently. Divergent findings are a feature, not a bug. They reveal complexity in your research topic and often point to areas that need further investigation.
For detailed integration frameworks, see our guide on qual-quant integration. For analysis techniques specific to each data type, check our mixed methods data analysis guide.
Tools and Templates for Data Combination
Your integration process is only as good as the infrastructure supporting it. If your quantitative data lives in a survey platform and your qualitative data lives in a separate tool, the combination step starts with a painful data export and reformatting exercise.
Quali-Fi was designed for exactly this challenge. Because Quali-Fi handles surveys, conjoint, focus groups, and IDIs in one platform, your quantitative and qualitative data are already connected at the participant level. Building joint displays and running cross-strand analysis becomes a natural part of the workflow instead of a separate integration project.
See how Quali-Fi connects qual and quant data
FAQs
What does it mean to combine qualitative and quantitative data?
Combining qual and quant data means systematically bringing findings from both data types together to produce integrated insights. This goes beyond presenting them separately, it involves techniques like merging (side-by-side comparison), connecting (one informs the other), or data transformation (converting one type into the other's format).
What is a joint display in mixed methods research?
A joint display is a table or matrix that presents qualitative and quantitative findings side by side, organized by research question, theme, or case. It typically includes a meta-inference column where you interpret what the combined findings mean. Joint displays are the primary tool for making data combination systematic and transparent.
How do I handle it when my qual and quant findings contradict each other?
Check whether the contradiction stems from methodological differences (different questions, different samples), scope differences (broad survey vs. Narrow interviews), or genuine complexity in your topic. Report conflicts transparently rather than ignoring one dataset. Divergent findings often point to the most important insights.
Can I convert qualitative data into numbers?
Yes, through a process called quantitizing. You can code qualitative themes and count their frequency, calculate the percentage of participants who mentioned each theme, or create binary variables for cross-tabulation. This is especially useful for integrating open-ended survey responses with closed-ended data in embedded designs.
When should I plan the combination strategy, before or after data collection?
Before. Deciding how you'll combine data after collecting it leads to ad hoc integration and missed opportunities. Your combination strategy should inform your study design, including what questions you ask, how you sample participants, and what variables you collect in both strands.