ACADEMIC WRITING GUIDE
How to Create a Mixed-Methods Joint Display for a Dissertation (With Worked Examples)
A dissertation can include a survey, interviews, and two carefully written findings sections but still leave readers wondering: What did the two sets of evidence tell us together? A mixed-methods joint display helps answer that question. It puts related quantitative and qualitative findings into a structured table or figure, then shows what their combination means for the research question.
The goal is not to decorate your chapter or to force interviews to “prove” a survey result. A good joint display makes the logic of integration visible. It shows where the two strands agree, where they differ, and where one adds information the other could not provide alone.
This guide walks through constructing a defensible joint display and provides three complete hypothetical examples: an explanatory sequential student-feedback study, a convergent study of campus support services, and an exploratory sequential study used to develop a questionnaire. All numerical results, quotations, and study scenarios below are invented solely for teaching.
What is a joint display in mixed-methods research?
A joint display is a structured presentation that deliberately brings qualitative and quantitative evidence together for analysis or interpretation. A common format has a column for numerical findings, another for qualitative themes or accounts, and a final column explaining the integrated inference. Other designs use diagrams, matrices, or case-by-case comparisons.
Guetterman, Fetters, and Creswell (2015) examined published examples and describe how joint displays can support meaningful integration rather than simple co-presentation of findings. Fetters, Curry, and Creswell (2013) explain that mixed-methods integration also happens through the study design and methods—for example, connecting survey respondents to interview sampling—not only at the reporting stage.
That distinction matters. A joint display cannot fix incompatible questions, weak measurements, an untrustworthy interview analysis, or an unexplained connection between samples. It makes existing integration transparent; it cannot manufacture integration after the fact.
A useful joint display answers three questions:
- What does each evidence strand show? Present the quantitative result and qualitative finding accurately, with denominators and contextual information where needed.
- How do they relate? Identify convergence, divergence, elaboration, or uncertainty without assuming the same constructs were measured.
- What new, bounded interpretation follows? Explain the integrated insight and its limits rather than restating the two cells.
If you are still planning a project, start with our dissertation methodology chapter guide to establish why both forms of evidence are needed.
When should you use a joint display?
A joint display is particularly useful when one dissertation research question needs both numbers and accounts of meaning or context. Consider the study's design before choosing its layout.
| Mixed-methods design | Typical sequence | What the display needs to communicate |
|---|---|---|
| Explanatory sequential | Quantitative results, then qualitative follow-up | How interviews explain, complicate, or fail to explain particular numerical patterns |
| Convergent | Quantitative and qualitative strands collected during a comparable period | Where findings converge, diverge, or address different facets of a shared question |
| Exploratory sequential | Qualitative exploration, then instrument or quantitative development | How qualitative insights shaped measures and how the later quantitative results relate back to them |
A joint display can be useful in intervention or case-study mixed methods too, but the layout must reflect the actual relationship between strands.
A single-method qualitative dissertation usually does not need a mixed-methods joint display. A thematic summary table may be helpful, but calling every table a joint display obscures what is being integrated.
Step 1: Start with the integration question
Write the question that requires both strands before building the table. For example:
How do first-year students' reported satisfaction levels relate to the experiences they describe when trying to use lecturer feedback?
The quantitative strand may measure satisfaction scores, while interviews explore why students interpret feedback as useful or confusing. Together they may reveal that relatively high satisfaction coexists with practical barriers to using feedback.
A vague integration aim—“compare the results”—is not enough. Decide what the second strand is intended to explain, test for consistency, extend, or challenge. The integration question should be aligned with the approved research questions and design, not retrofitted merely to make a table look symmetrical.
Our data analysis plan guide explains how to map research questions to appropriate evidence and outputs.
Step 2: Identify comparable findings without forcing false equivalence
Review the completed or planned analyses and choose a manageable number of findings relevant to the same conceptual issue. One row might pair survey respondents' ratings of access with interview accounts of trying to obtain access. Another could pair a difference between groups with the explanations some interviewees offered.
Document what the cells actually represent. A survey percentage concerns a defined denominator and measure. An interview theme is an analytic interpretation of selected accounts, not automatically a population percentage. A theme should not be described as “confirmed” simply because a quantitative result looks similar.
Avoid matching findings by superficial keywords. “Support” on a questionnaire may mean prompt academic advice, while interviewees may be discussing emotional support from peers. Those are not interchangeable measures. You may need separate rows or a clear note describing their limited comparability.
Step 3: Choose columns that demonstrate the reasoning
For many dissertations, four columns work well:
| Display element | What to record | Common pitfall |
|---|---|---|
| Issue or research-question component | The concept each row addresses | Mixing unrelated concepts in a row |
| Quantitative result | Statistic, denominator, time point, and relevant uncertainty | Presenting a percentage without saying who was counted |
| Qualitative finding | Theme, short definition, and carefully selected illustrative evidence | Treating one invented or unusual quote as population-wide proof |
| Integrated inference | Fit, difference, added insight, and remaining uncertainty | Simply writing “results agree” without saying what was learned |
A fifth column for implications or limitations may be appropriate, but avoid squeezing too much prose into a single cell. Use clear headers and concise entries; explain substantial reasoning in the paragraph below the display.
If you need guidance on table numbering and clear labels, see how to present research data in tables and figures.
Step 4: Check the provenance and quality of each cell
Before integrating findings, ask where every claim came from.
- Were the survey denominator, response options, and analytic sample correctly reported?
- Was the interview theme developed using the named analytic approach, with enough evidence to justify it?
- Were the same people involved in both phases, or were separate samples used?
- Is the time period comparable?
- Are groups and measures genuinely comparable?
- Has the researcher considered contrary cases and potential sampling or measurement limitations?
Where interviews followed a survey, state whether interview participants were selected based on their responses and whether consent permitted linkage. Do not suggest person-level correspondence if records were not linked.
When the strands disagree, verify differences in wording, participants, collection periods, and definitions before calling the result a substantive contradiction. Integration requires attention to the quality of both strands—not treating one as automatically the truth.
Step 5: Write an integrated inference, not a summary of two cells
Compare these two interpretations:
Weak: “Survey participants liked feedback, and the interviews showed that feedback was sometimes confusing.”
Stronger: “Although the survey indicated generally favourable ratings, interview accounts suggest that a positive overall rating did not always mean students knew how to apply the comments. The evidence therefore distinguishes perceived helpfulness from practical use; the study cannot determine whether feedback improved subsequent grades.”
The second statement contributes something neither strand could establish alone. It also makes its causal limitation explicit.
Terms such as convergence, divergence, complementarity, and expansion can help, but explain what the relationship actually is. Differences may be analytically informative rather than errors to conceal.
Worked example 1: Explanatory sequential joint display
Hypothetical study and connection between phases
A university student-feedback study surveys 120 eligible first-year undergraduates about written feedback. The survey asks whether feedback was helpful and whether students felt able to act on it. After analysis, the researcher purposively invites consenting respondents with varied survey experiences to interviews. Twelve students are interviewed. These numbers are invented for demonstration, not empirical findings or a suggested universal sample size.
The survey provides breadth within its responding sample. Interviews offer detailed accounts of how students understood or applied comments. The study does not randomly assign feedback, so it cannot support causal effectiveness claims.
Example display
Table 1. Illustrative explanatory sequential joint display: academic feedback
| Integrated issue | Quantitative finding (hypothetical) | Qualitative finding (hypothetical) | Integrated inference |
|---|---|---|---|
| Overall helpfulness | 90/120 respondents (75%) rated feedback helpful | Interviewees often appreciated detailed comments but described uncertainty about priorities | Favourable ratings and difficulty acting on feedback can coexist |
| Ability to act | 54/120 (45%) reported confidence applying feedback to the next task | Interviewees described wanting clearer examples of how to revise an argument | Perceived helpfulness does not necessarily imply actionable understanding |
| Timeliness | 48/120 (40%) reported receiving feedback early enough to use for their next assignment | Some participants said feedback arrived after they had begun the next task | Timing may help explain barriers to use; the data do not show that earlier feedback caused better marks |
How the interpretation is developed
First, verify that each percentage uses the stated denominator of 120. The calculations are 90 ÷ 120 = 75%, 54 ÷ 120 = 45%, and 48 ÷ 120 = 40%.
Second, distinguish the types of claims. The survey reports responses among the observed sample. The interview accounts explain possibilities and meanings; twelve selected participants cannot establish how common those experiences are in the full university population.
Third, identify the new insight: a student may rate feedback positively yet struggle to apply it. That insight supports a question for future teaching practice, not an experimental conclusion that any intervention improves performance.
Sample dissertation paragraph
The hypothetical joint display indicates that positive ratings of written feedback did not consistently coincide with confidence in applying it. Although 90 of 120 survey respondents rated feedback helpful, only 54 reported confidence in using it for subsequent work. Follow-up interview accounts suggested that vague priorities and insufficient worked examples contributed to some students' uncertainty. Together, the strands distinguish general approval from practical usability. Because the data were observational and interview participants were purposively selected, these findings do not establish causal effects or population-wide prevalence of the interview themes.
This paragraph interprets a relationship across the two forms of evidence instead of repeating every table entry.
Worked example 2: Convergent joint display with agreement and disagreement
Hypothetical study
A campus support-service study collects a survey of 80 respondents and conducts 10 independently recruited interviews during the same term. The strands address overlapping questions about service access, clarity, and privacy, but the samples are not linked at individual level. All numbers and accounts are invented.
Table 2. Illustrative convergent joint display: campus support services
| Question component | Quantitative strand (hypothetical) | Qualitative strand (hypothetical) | Relationship and cautious inference |
|---|---|---|---|
| Appointment access | 60/80 (75%) rated booking easy | Interviewees commonly described booking as straightforward | Convergence: both strands suggest ease of initial booking among people studied, without establishing campus-wide access |
| Information clarity | 52/80 (65%) said instructions were clear | Several interviewees described confusing eligibility rules | Divergence: clarity ratings and specific reported confusion differ; different samples and question wording may contribute |
| Privacy | 64/80 (80%) felt comfortable with confidentiality arrangements | Interviewees described avoiding visible waiting areas | Expansion: a general privacy rating does not capture every concern about public visibility |
How to treat the disagreement
It would be wrong to declare the qualitative strand “disproves” the survey finding. The survey asks about instructions as measured by a specific item; interviews permit people to discuss particular rules that caused difficulty. The samples also differ. Researchers should inspect wording and recruitment before attributing disagreement to a substantive difference.
Divergence may justify a limited interpretation: broad satisfaction ratings can obscure particular barriers. It does not justify claiming that most users were confused or that the service violated privacy obligations.
Sample dissertation paragraph
The hypothetical convergent display shows partial agreement between the survey and interview strands. Most respondents described booking as easy, consistent with interview accounts of straightforward initial access. However, a majority rating of clarity coexisted with interview descriptions of confusing eligibility conditions. These findings may concern different aspects of the service, and the independently recruited samples cannot be matched participant by participant. The integrated account therefore identifies specific issues for further investigation without treating the interview accounts as estimates of prevalence.
Worked example 3: Exploratory sequential joint display for questionnaire development
Hypothetical study
A researcher first interviews eight students about obstacles to participating in online seminars. Interview analysis identifies several concerns. The researcher then drafts survey items reflecting these ideas, checks the wording for relevance and clarity, and pilots an instrument with a separate group before fielding a hypothetical survey of 150 eligible respondents. These phases and numbers are fictional; they are not claims of instrument validation.
Here, the joint display tracks building: how qualitative insights informed questionnaire content, followed by the descriptive survey pattern. It does not imply that the initial interviews measured population frequency.
Table 3. Illustrative exploratory sequential joint display: online seminar participation
| Qualitative insight (invented) | Draft survey item developed from it | Quantitative response (hypothetical) | Integrated interpretation and limitation |
|---|---|---|---|
| Fear of being judged for mistakes | “I hesitate to speak because I worry about giving an incorrect answer.” | 96/150 (64%) selected agree or strongly agree | The concern informed a measurable item; endorsement in this sample suggests relevance, not proof of cause |
| Unstable connectivity | “Connection problems have interrupted my participation in live seminars.” | 69/150 (46%) agreed | Interviews identified a possible barrier that respondents also reported; self-report does not document actual connection failures |
| Unclear participation expectations | “I understand when and how I am expected to contribute during live seminars.” | 105/150 (70%) agreed | Positive clarity responses may coexist with other participation barriers; the item needs appropriate quality assessment |
Why this is integration
The interview strand did more than produce a quotation to place next to a percentage: it informed item construction. The display records the link from qualitative interpretation to a defined measure, and the later survey provides a different form of evidence about how respondents answered those items.
An exploratory study must still evaluate whether the items represent the intended constructs. An illustrative pilot or high endorsement rate alone does not establish validity, reliability, or generalisability. Describe cognitive interviews, piloting, item revisions, and psychometric assessment only if they actually occurred.
Sample dissertation paragraph
In this hypothetical exploratory sequential design, initial interview analysis identified fear of errors, connectivity difficulties, and uncertainty about participation expectations as concerns relevant to online seminar engagement. These insights informed the wording of candidate questionnaire items. Subsequent descriptive responses suggested that worry about incorrect answers was frequently endorsed among surveyed students, while clarity of participation expectations was endorsed by many respondents. The joint display documents how qualitative development and quantitative description were connected; it does not establish the instrument's psychometric properties or demonstrate that the identified concerns caused low participation.
How to explain a joint display in your dissertation chapter
Do not insert a large table without a lead-in or analysis. Most readers need three linked elements:
Before the display: Identify the integration question, study design, data sources, and how the strands were connected or compared. Explain whether denominators and participants are linked or separate.
Within the display: Give a descriptive title, clear headings, concise findings, and a column for the integrated inference. Cite or cross-reference supporting analysis and quotations elsewhere in the chapter rather than packing every source detail into the table.
After the display: Discuss the patterns that matter. Explain agreement, differences, expansion, and limits. Relate the interpretation to the research question without making unsupported claims.
For wider guidance on structuring your results, see our research findings chapter guide.
Common mistakes when creating joint displays
Putting numbers and quotes side by side without integration
A table with a percentage and a quote can be useful but may still be two parallel mini-results. Add a clear account of what the strands reveal together, including uncertainty and differences.
Calling qualitative themes statistically representative
An interview theme is not a survey proportion unless a defensible additional method and denominator support that quantification. Do not turn “several participants said” into a percentage of a larger population.
Forcing agreement
A mismatch may reflect distinct questions, samples, periods, or constructs. It may also reveal a substantive tension. Record and interpret it rather than hiding it.
Ignoring sample linkage
If different people completed surveys and interviews, do not imply that the same respondent both gave a rating and made an interview statement. If participants were intentionally linked, describe consent, confidentiality safeguards, and how linkage was handled.
Reporting percentages without denominators
“75% found it helpful” is less informative than “90/120 respondents (75%) found it helpful.” If missing values change the denominator for particular items, state that variation.
Treating the display as proof of causality
A joint display integrates forms of evidence; it does not by itself change an observational research design into an experiment. Interpret associations and explanations within the study's limitations.
Reproducing sensitive participant information
Even altered names may leave revealing combinations of roles, locations, or events. Use de-identified excerpts only as allowed under relevant approval, consent, and institutional procedures. All quotations in this article are invented instructional examples.
Practical joint-display checklist
Before submission, check that:
- Each display addresses an explicit integration question.
- The study design and the link between strands are accurately described.
- Quantitative values have verified denominators, time points, and appropriate uncertainty.
- Qualitative findings are supported by the actual analysis, not selected merely to agree with survey numbers.
- Similarities and differences are interpreted, not concealed.
- The integrated-inference column explains what is learned jointly.
- Independent samples are not presented as linked participants.
- The display has a number, title, accessible headings, and an accompanying interpretation.
- Participant privacy and approved data-sharing rules are respected.
- Findings are not presented as causal or more generalisable than the design permits.
Frequently asked questions
Is a joint display compulsory in a mixed-methods dissertation?
No single table format is compulsory across universities and designs. However, a joint display can make integration substantially clearer when the research questions require two strands of evidence. Check your university's handbook and explain any alternative integration strategy.
Can I use a joint display if the findings contradict each other?
Yes. Divergence is often important. Check whether measures, samples, and time frames are comparable, then describe the differences and possible explanations without assuming either strand is automatically correct.
Can I build the display in Word or Excel?
Yes. Either can work if headings, alignment, sources, and table accessibility are retained. Use the format your institution accepts and make sure wide tables remain readable in the submitted PDF or digital document.
How many joint displays should I include?
Use as many as needed to answer the integration questions without duplication. A focused dissertation may need only one well-explained display; a larger mixed-methods study may require several different displays.
Is it acceptable to combine unrelated surveys and interviews?
Only when there is a defensible shared question, compatible concepts, and an explanation of how the evidence relates. Merely using two data-collection methods is not enough to establish a coherent mixed-methods design.
What belongs in the discussion after a joint display?
Explain the integrated inference, meaningful disagreements, relevant context and methodological limitations. Connect these interpretations to the dissertation's question and appropriate literature. Do not repeat every table cell.
Final takeaway
A credible mixed-methods dissertation should show what the two strands contribute together. Begin with an integration question, select appropriately related findings, preserve the distinctive meaning of each method, and explain convergence, tension, and added insight. A well-constructed joint display is not a cosmetic table: it is a transparent record of your integrative reasoning.
If you need help reviewing your dissertation's structure, explaining mixed-methods integration, or improving the clarity of its tables and interpretation, Acadelyra's academic-support services can provide educational editing and feedback while leaving the research, analysis, and academic responsibility with you.
References and further reading
- Fetters, M. D., Curry, L. A., & Creswell, J. W. (2013). Achieving integration in mixed methods designs—principles and practices. Health Services Research, 48(6 Pt 2), 2134–2156. https://pmc.ncbi.nlm.nih.gov/articles/PMC4097839/ (DOI: 10.1111/1475-6773.12117)
- Guetterman, T. C., Fetters, M. D., & Creswell, J. W. (2015). Integrating quantitative and qualitative results in health science mixed methods research through joint displays. Annals of Family Medicine, 13(6), 554–561. https://pmc.ncbi.nlm.nih.gov/articles/PMC4639381/ (DOI: 10.1370/afm.1865)
Editorial note: Every survey value, interview account, research scenario, and reporting paragraph in this guide is a hypothetical worked example. Follow your institution's methodology, ethics, formatting, and disclosure requirements when reporting real research.