The most common reason a mixed methods dissertation is challenged in the viva is not a weak survey or a thin set of interviews. It is that the two never meet. A candidate collects quantitative data, analyzes it, writes it up; collects qualitative data, analyzes it, writes it up; and places the two accounts side by side in the belief that having both makes the study mixed methods. It does not. What makes a study mixed methods is integration: the deliberate bringing together of the quantitative and qualitative strands so that the combination yields an insight neither could produce alone. A study that keeps its strands parallel is not one mixed methods study; it is two smaller studies sharing a cover page, and examiners recognize the difference at once. The primary tool for demonstrating genuine integration is the joint display, and building one that actually integrates, rather than merely arranging data in a table, is the skill this article teaches. Our dissertation chapter support is built to help candidates get this stage right, because it is where mixed methods work most often fail.
This guide explains what integration means, the three levels at which it happens, the three core designs, and how to build a joint display that generates real meta-inferences rather than sitting inert on the page.
Quick Answer:
Integration is the defining feature of mixed methods research: not the mere collection of both quantitative and qualitative data, but their deliberate combination. Hence, the whole exceeds the sum of the parts. Integration happens at three levels: design (connecting, building, merging, or embedding the strands), methods, and interpretation and reporting (through narrative, data transformation, and joint displays). A joint display is a table or figure that arrays quantitative and qualitative results together to draw meta-inferences, new conclusions produced by the integration itself. The most common failure examiners flag is the "two studies in one" report, where the strands never integrate, or a joint display that places data side by side without generating any meta-inference or stating the fit (confirmation, expansion, or discordance) between the strands.
Integration Is the Defining Feature, Not the Data
The single most important idea in mixed methods research is that collecting both types of data is not the point; integrating them is. This is easy to say and surprisingly hard to do, which is why so many studies fall short. The purpose of mixing methods is to produce understanding that a single method could not, and that added understanding only materializes when the strands are brought into genuine contact with one another. Methodologists express this as the principle that, in a well-integrated mixed-methods study, one plus one equals three: the integrated whole yields more than the quantitative and qualitative parts counted separately.
The failure mode has a name that candidates should internalize: the "two studies in one" problem, sometimes called the parallel-strands problem. A study exhibits it when the quantitative and qualitative components run alongside each other from start to finish and never actually connect, meeting, if at all, only in a discussion section that gestures vaguely at both. When that happens, the study forfeits the entire rationale for using mixed methods, because it never realizes the combination that justified the extra work. Reviewers describe such studies as being less than the sum of their parts. Avoiding this outcome is not a matter of writing a better discussion; it requires planning integration deliberately, at each of the three levels.
The Three Levels of Integration
Integration is not a single act performed once at the end. It happens, or fails to happen, at three distinct levels of a study, and a rigorous mixed methods design attends to all three. This framework comes from Fetters, Curry, and Creswell (Fetters, MD, Curry LA, Creswell JW. "Achieving integration in mixed methods designs: principles and practices." Health Services Research 2013;48(6 Pt 2):2134-2156), and it is the structure examiners expect a candidate to be able to articulate.
Integration at the design level concerns how the strands relate structurally across the study. There are four basic ways they can connect. Connecting joins the strands through sampling, when one database links to the other by, for example, selecting interview participants based on survey results. Building occurs when the results of one strand inform the data collection of the other, as when qualitative findings shape the items on a subsequent questionnaire. Merging combines the two databases for analysis. Embedding links collection and analysis at multiple points, with one strand nested inside a design led by the other.
Integration at the methods level concerns the practical linking of the data during collection and analysis, the actual mechanics by which the strands are joined rather than kept apart. Integration at the interpretation and reporting level concerns how integrated findings are presented to the reader and is achieved through three principal techniques: narrative approaches, data transformation, and joint displays. Narrative integration can weave the strands together theme by theme, present them in contiguous sections within a single report, or stage them across the phases of a larger study. Data transformation converts one data type into another, for instance, quantizing qualitative codes so that they can be analyzed together. Joint displays, the focus of this article, visually arrange the strands to provoke and demonstrate integration.
Table 1: The Three Levels of Integration in Mixed Methods Research
Level | What It Concerns | How It Is Achieved |
|---|---|---|
Design | How the strands relate structurally across the study | Connecting, building, merging, or embedding |
Methods | The practical linking of data during collection and analysis | Linking sampling, instruments, and analysis across strands |
Interpretation and reporting | How the integrated findings are presented to the reader | Narrative (weaving, contiguous, staged), data transformation, joint displays |
The Three Core Designs, and How Integration Works in Each
Mixed-methods research is built on three core designs, and the way integration is achieved differs across them, so naming your design correctly is the first step toward integrating well. Choosing between qualitative and quantitative foundations in the first place is the upstream decision this all rests on, which our guide on whether your dissertation should be qualitative or quantitative works through in detail.
In a convergent design, the quantitative and qualitative strands are collected and analyzed separately during roughly the same phase, then brought together and merged for comparison and interpretation. Integration here is primarily by merging, and the joint display is the natural place for it, because the whole point of the design is to set the two sets of results against each other.
In an explanatory sequential design, quantitative data are collected and analyzed first, and qualitative data are then gathered to explain or elaborate on the quantitative results. Integration is achieved chiefly by connecting and building: the qualitative phase is designed on the basis of the quantitative findings, often by sampling qualitative participants according to their quantitative results, and its job is to explain what the numbers showed.
In an exploratory sequential design, the order reverses. Qualitative data come first and are used to build something, typically an instrument, a survey, or an intervention, that is then tested quantitatively. Integration is dominated by building, because the qualitative findings directly shape the quantitative phase that follows.
What a Joint Display Is, and What It Is Not
A joint display is a table or figure that arranges quantitative and qualitative results together so that their integration becomes visible and analyzable, and from which the researcher draws meta-inferences (Guetterman TC, Fetters MD, Creswell JW. "Integrating quantitative and qualitative results in health science mixed methods research through joint displays." Annals of Family Medicine 2015;13(6):554-561). The keyword is together. A joint display is not simply a table with quantitative results in one column and qualitative results in another. That arrangement is necessary but not sufficient, and stopping there is the most common way joint displays fail. What distinguishes a genuine joint display is that it does analytic work. By placing the strands adjacent to one another, it forces a comparison, and from that comparison, the researcher produces a new conclusion that neither strand would state on its own.
Several types of joint displays exist for different purposes. A side-by-side display arrays quantitative and qualitative results for the same constructs, allowing comparison. A statistics-by-themes display organizes qualitative themes against quantitative statistical results. A themes-by-statistics display does the reverse. There are displays built for instrument development, for embedding qualitative data into an experiment, and for case study designs, a cross-case comparison display. The most prevalent types in the published literature are the statistics-by-themes and side-by-side comparison formats. The type you choose should follow from your design and your integration purpose, not from convenience.
Building a Joint Display That Generates Meta-Inferences
The difference between a joint display that integrates and one that merely tabulates is the meta-inference. A meta-inference is a conclusion drawn from the integration itself, a statement about what the combined evidence means that goes beyond what either the quantitative or the qualitative result said alone. Producing meta-inferences is the entire purpose of the display, and a display that ends without them has not integrated anything. Here is how to build one that works.
Begin by completing each strand's analysis separately, so you have clear quantitative results and clear qualitative findings to bring together. Then build the table. Let the rows be the shared constructs, domains, or research questions that both strands address, and give the table columns for the quantitative and qualitative results for each. Populate each cell with the relevant finding, including a representative quotation for the qualitative side rather than a summary alone.
Now add the two columns that convert a table into a joint display. The first is a fit column. For each row, state the relationship between the strands. Confirmation is when the two agree and reinforce each other. Expansion, sometimes called complementarity, is when they address different aspects and together broaden the picture. Discordance is when they disagree or contradict each other. Naming the fit is a required analytic judgment, not an optional label. The second is a meta-inference column. For each row, write the integrated conclusion, the insight that the combination of the two results produces. This column is where one plus one becomes three, and it is the part reviewers look for first.
Finally, iterate. Your first display will rarely be your best; the detail that matters usually exceeds an initial impression, so expect to build several versions as your understanding of the integration deepens. Where a meta-inference is not clearly supported once you re-examine the underlying results, mark it as tentative rather than asserting it, because an unsupported meta-inference is worse than none.
Table 2: A Worked Joint Display Template (Convergent Design)
Shared Construct | Quantitative Result | Qualitative Finding (with quote) | Fit | Meta-Inference |
|---|---|---|---|---|
Construct A | Survey score/effect size/test result | Theme with representative quotation | Confirmation | The integrated conclusion of the pairing produces |
Construct B | Quantitative result | Theme with representative quotation | Expansion | How the strands broaden the picture together |
Construct C | Quantitative result | Theme with representative quotation | Discordance | What the disagreement reveals about the phenomenon |
Struggling to make your mixed methods strands actually integrate? |
|---|
Send us your quantitative and qualitative results. A mixed methods methodologist will build the joint display for your design, name the fit for each row, and draft the meta-inferences that turn two parallel strands into one integrated finding that your committee will accept. Request mixed methods integration support and receive an itemized quote within 2 to 4 business hours, no obligation. |
Judging the Fit: Confirmation, Expansion, and Discordance
The fit between the strands deserves closer attention because handling it well, especially when it is uncomfortable, is a mark of a sophisticated mixed methods study. Confirmation, where the quantitative and qualitative results agree, is the easiest case and the one candidates hope for, because it strengthens the conclusion through convergence. Expansion, where the strands illuminate different facets of the phenomenon, is common and valuable because it delivers the broader understanding that motivated mixing methods in the first place.
Discordance, where the strands disagree, is the case candidates fear, and it is precisely the one that most rewards careful treatment. A discordant result is not a failure of the study; it is a finding. When the survey says one thing and the interviews say another, the tension is information, and the researcher's task is to examine why the two diverge and what the divergence reveals, rather than to bury it or explain it away. Some frameworks add a further category, a unique contribution, for a finding that appears in only one strand. Reporting discordance openly and analyzing it is one of the clearest signals to an examiner that the integration is real rather than cosmetic. It is the kind of interpretive nerve our overview of the most common dissertation mistakes shows candidates too often avoid.
Reporting Integration So Reviewers Can See It
Integration that happened but is not visible to the reader earns no credit, so reporting it explicitly is part of the task. The established reporting guidance for mixed methods studies is known as GRAMMS. It asks authors to justify why a mixed methods approach was used, to describe the design in terms of its purpose, priority, and sequence, and to describe each method. Most importantly, it asks authors to state where, how, and by whom integration occurred, along with any limitations of one method arising from its combination with the other and any insights the integration produced. That fourth point is the one most often neglected: a study must say, in plain terms, where integration took place and what it produced. A joint display answers much of this directly, which is part of why it is such a valuable device. However, the surrounding text must still make the integration and its payoff explicit rather than leaving the reader to infer them. Being able to articulate all of this clearly is also what carries a candidate through the viva, which our dissertation defense preparation guide prepares them for. Our guide on writing a strong research proposal helps set the integration plan up correctly from the start.
What Good Mixed Methods Integration Looks Like
Set against the parallel-strands failure, a well-integrated mixed methods study has a clear profile. It names its design and states, using the language of connecting, building, merging, or embedding, how the strands relate. It plans integration from the outset rather than improvising it at the end. It presents at least one genuine joint display in which the strands are set against each other by shared construct, the fit of each pairing is named, and a meta-inference is drawn for each. It treats discordance as a finding to be analyzed rather than a problem to be hidden. And it reports, in the text and in line with GRAMMS, exactly where integration occurred and what new understanding it produced. A study with these features demonstrates the one-plus-one-equals-three payoff that justifies mixed methods, and it holds together under examination because the integration is built into the design rather than bolted on afterward. That integration is one of the central things our broader dissertation support is built to strengthen.
Frequently Asked Questions
What does integration mean in mixed methods research?
Integration is the deliberate combining of the quantitative and qualitative strands of a study so that the result yields understanding that neither could produce alone. It is the defining feature of mixed methods research: collecting both data types is not enough. Methodologists describe well-integrated work as achieving "one plus one equals three," meaning the integrated whole exceeds the sum of the separate parts. Integration happens at the design, methods, and interpretation levels.
What is a joint display in mixed methods?
A joint display is a table or figure that arranges quantitative and qualitative results together so their integration becomes visible and analyzable, and from which the researcher draws meta-inferences. It is not simply a table with numbers in one column and quotes in another; it does analytic work by forcing a comparison between the strands and producing a new conclusion from that comparison. Common types include side-by-side, statistics-by-themes, and themes-by-statistics displays.
What is a meta-inference?
A meta-inference is a conclusion drawn from the integration of the two strands, a statement about what the combined evidence means that goes beyond what either the quantitative or qualitative result said on its own. Producing meta-inferences is the purpose of a joint display; a display that arranges data without drawing meta-inferences has not integrated anything. In a well-built display, each row carries its own meta-inference in a dedicated column.
What is the "two studies in one" problem?
It is the most common failure in mixed methods research: the quantitative and qualitative strands run in parallel from beginning to end and never actually connect, meeting only, if at all, in a vague discussion section. A study that does this forfeits the rationale for mixing methods, because it never produces the integrated understanding that justified the extra work. Reviewers describe such studies as less than the sum of their parts. The fix is to plan integration deliberately at each level.
What are the three core mixed methods designs?
The three core designs are convergent, explanatory sequential, and exploratory sequential. In a convergent design, both strands are collected in the same phase and merged for comparison. In an explanatory sequential design, quantitative data come first, and qualitative data are gathered to explain them. In an exploratory sequential design, qualitative data come first and are used to build something, such as an instrument, that is then tested quantitatively. Integration works differently in each.
How do I handle it when my quantitative and qualitative results disagree?
Treat the disagreement, called discordance, as a finding rather than a failure. When the strands diverge, the tension is information: your task is to examine why they differ and what the divergence reveals about the phenomenon, not to hide it or explain it away. Analyzing discordance openly is one of the clearest signals to an examiner that your integration is genuine. Name the fit as discordant in your joint display and draw a meta-inference from it.
What is GRAMMS?
GRAMMS (Good Reporting of a Mixed Methods Study) is the established reporting guidance for mixed methods research. It asks authors to justify the mixed methods approach, describe the design by purpose, priority, and sequence, and describe each method. It also asks them to state where, how, and by whom integration occurred, along with any limitations arising from the combination and any insights gained from it. The requirement to report where integration took place and what it produced is the item most often neglected.
You May Also Find Useful
Making the Strands Meet
A mixed methods study earns its name at the point where the two strands come together and produce something neither could alone. Plan that integration from the start, and name how your strands connect. Then build a joint display that does more than tabulate: set the results against each other by shared construct, name the fit of each pairing, and draw a meta-inference for every one. Treat discordance as a finding, and report plainly where the integration happened and what it yielded. Do that, and your study stops being two smaller studies sharing a cover page and becomes the integrated whole that mixed methods promises.
If you want a mixed methods methodologist to build your joint display and draft the meta-inferences that make your integration visible, send us your results. You will have an itemized quote within 2 to 4 business hours, with no obligation.
