A mixed methods results chapter reports the findings of your quantitative and qualitative strands, then shows how the two were combined into conclusions that neither strand could support alone. It reports evidence. It does not interpret what that evidence means, which belongs in the discussion chapter. For most candidates that single boundary is where marks are lost.
Knowing how to write a mixed methods results chapter comes down to structure rather than analysis, which is why writing one takes longer than a single-strand results chapter: the integration work happens at the writing stage. You are not summarizing two projects. You are showing a reader where the numbers, the code list and the interview quotes meet, and what the study can claim once they meet.
The seven-stage process below takes a blank document to a submitted chapter. Most people who follow it finish in two to three focused sittings, though the first sitting is usually the one that stalls. That stall is nearly always a planning problem, not a writing problem, which is why stage 1 gets the most attention here.
A note on conventions before you start: the reporting rules for the statistical strand follow APA style, because that is what most universities and journals in the social sciences and health fields expect. The qualitative conventions I describe match thematic analysis reporting practice. Where your department prescribes something different, its rules win, and step 7 is where you check that.
Table of Contents
- 1What You Need
- 2Your research questions and design notes
- 3Completed analysis outputs for both strands
- 4Finished tables and figures, numbered and titled
- 5A reporting guideline or your department’s chapter template
- 6Your supervisor’s last round of comments
- 7A place where joint displays are editable
- 8A traceability list linking claims to sources
- 9Step-by-Step: How to Write a Mixed Methods Results Chapter
- 101. Review Your Research Questions and Mixed Methods Design
- 112. Organize the Chapter Around the Research Questions
- 123. Report the Quantitative Results
- 134. Report the Qualitative Results
- 145. Integrate Quantitative and Qualitative Findings
- 156. Write the Results in Clear Academic Prose
- 167. Check the Chapter for Accuracy and Coherence
- 17Common Mistakes
- 18Reporting two studies side by side
- 19Mixing methods without saying why
- 20Reporting methods again
- 21Overinterpreting
- 22Burying contradictory findings
- 23Quotation overload
- 24Silent non-significant results
- 25Interpretation in the results chapter
- 26Frequently Asked Questions
- 27How long should a mixed methods results chapter be?
- 28Is it okay to combine results and discussion in a mixed methods study?
- 29What is the purpose of a joint display in mixed methods research?
- 30Which order should I present quantitative and qualitative results in?
- 31How should I report conflicting results from my survey and interviews?
- 32What tense should I use in a mixed methods results chapter?
- 33Conclusion
What You Need
Assemble these before you open a blank document. The order matters less than having all of them within reach, because you will move between strands constantly while writing.
Your research questions and design notes
The research questions set the order of your sections. The design notes tell you why the strands sit in that order, which sampling approach you used in each, and where you planned to integrate. If you cannot state the integration point in one sentence, your chapter structure will not hold together.
Completed analysis outputs for both strands
For the quantitative strand, that means cleaned datasets, syntax files, the full output, and any model or regression results. For the qualitative strand, it means the final code list, the theme and subtheme structure, a code frequency table, and your analytic memos. Frequencies matter, because you will need them in the joint displays later.
Finished tables and figures, numbered and titled
Build these while you still have the data open. A table that is finished but unnumbered forces you back into the software every time you want to check a figure reference mid-paragraph.
A reporting guideline or your department’s chapter template
Take the required subheadings, the word allocation per section, and the formatting rules for tables and figures from the handbook rather than guessing. If your university has a thesis template for mixed methods chapters, that template resolves most of the formatting questions you would otherwise spend hours on.
Your supervisor’s last round of comments
Read them for movement requests specifically. A comment like make the quant section stronger usually means a missing effect size, not vague prose. Pulling the actual change list forward now saves a second rewrite cycle later.
A place where joint displays are editable
A spreadsheet or a table in your word processor, because integration is where the real thinking happens and you will rebuild these tables more than once.
A traceability list linking claims to sources
A simple table with three columns: the claim, the table or figure it comes from, and the interview transcript line behind any quotation. Building it now turns step 7 from a memory test into a check.
Step-by-Step: How to Write a Mixed Methods Results Chapter
The process runs in seven stages, from confirming the design through a final accuracy pass. The sequence follows your study design and your research questions rather than forcing both strands into one combined block, which is the single most common structural error in these chapters.
Two strands reported back to back with a short connecting paragraph are not integration. Integration is a designed point in the study where the two datasets were combined, and your chapter structure has to show where that was.
1. Review Your Research Questions and Mixed Methods Design
Start by sorting your research questions into quantitative, qualitative, and mixed. A question is quantitative if the answer is a number or a statistical relationship, qualitative if the answer is a description or an explanation of meaning, and mixed if answering it properly requires both.
Then write one sentence naming your integration point. Something along the lines of qualitative interviews were conducted with a subsample of survey respondents in order to explain the survey result that showed no significant relationship between workload and completion time. That sentence becomes the spine of the chapter and eventually the opening claim of the integration section.
Next, pull your sampling logic. Whether the strands used the same participants, a nested subsample, or fully separate groups changes what you can claim and it changes which references you make between sections. In an identical-sample design you can refer to an interview and a survey score for the same person. In a parallel-sample design you cannot, and writing as if you can is a methodological error an examiner will catch.
Finally, write down the analysis plan as you actually ran it, and flag anything you did not run. Analyses that were planned but abandoned belong in the methods chapter as a note, not in the results.
2. Organize the Chapter Around the Research Questions
Build the outline from the research questions in order, not from the order in which you analysed things. Each subsection answers exactly one question and is named after that question, so a reader scanning your subheadings can follow your logic without reading the prose.
A typical working outline for a convergent parallel design looks like this: participant and sample overview, quantitative results for question one, qualitative results for question two, integrated findings for question three, and a summary of meta-inferences. Adapt the labels to your own questions rather than adopting this one.
Now check for repetition. If a number appears in a table, it may still need mentioning once in the text, but the text version should be the finding rather than the table contents. If a quotation appears in the qualitative section, do not also paste it into the joint display. Decide which artefact carries each piece of evidence and let the other refer to it.
Write your word allocation into the outline itself. A rough split for a 6,000-word results chapter in a mixed design is 800 words for the sample overview, 1,600 for the quantitative strand, 1,400 for the qualitative strand, 1,600 for integration, and 600 for the meta-inference summary. Adjust for your design, since sequential designs put more weight on integration.
3. Report the Quantitative Results
Open with a short participant and sample overview, then work from descriptive statistics to inferential ones. Give the n for every analysis, and where a sample differs between analyses, say why. A test run on 312 cases after three were excluded for missing data needs that sentence.
Report the full statistical picture rather than the significance verdict. A t-test line should carry the test statistic, the degrees of freedom, the p-value, the effect size and the confidence interval for the mean difference. Reporting only a p-value tells the reader nothing about magnitude, which is a frequent examiner complaint.
For example: an independent-samples t-test showed that mean completion time did not differ significantly between students who used the two revision apps, M = 34.2 minutes versus 35.1 minutes, t(298) = -0.84, p = .40, Cohen’s d = 0.10, 95% CI [-0.02, 0.22].
Put the full table in an appendix or a table in the chapter, not in the paragraph. The text carries the finding; the table carries the complete set of values, including any you report but do not discuss.
For non-parametric tests, give the test statistic and its distribution explicitly, such as a Kruskal-Wallis H with the chi-square approximation. For models, report coefficients with confidence intervals plus the model fit statistics your field expects, and state which reference category you recoded against.
Report what you planned even when it is unremarkable. A null result reported clearly is a finding, and hiding it is both a methods problem and, in most cases, a publication problem later.
4. Report the Qualitative Results
Describe the sample in the terms that matter for your analytic approach: how many participants, how they were recruited, and what the analytic strategy was. If you used thematic analysis, name the variant you followed, whether codes were inductive, deductive, or a mix, and who coded the material.
Then present the theme structure. A table of themes, subthemes and the number of participants contributing to each one gives the reader your whole analytic map in one view, and it earns its space because you will reference it in the integration section.
Quote sparingly and with purpose. Two or three short quotations per major theme is enough, and each one should illustrate something the paraphrase cannot carry. Note the participant identifier and, where relevant, a demographic detail that does not identify them.
Keep the analytic process visible without turning the section into a methods reprise. A short statement of how many coding rounds you ran, whether a second coder checked a share of the material, and how disagreements were resolved belongs here, in two or three sentences.
Two restraint rules matter most. Do not promote a theme into a cause, since a theme describing how participants talked about something is not evidence that the thing happened. And do not stack quotations to imply weight, because five consecutive quotes from two people is not saturation and reads as padding.
5. Integrate Quantitative and Qualitative Findings
How to write a mixed methods results chapter comes down to this section. The integration is what makes the study mixed methods rather than two studies placed in one document, and it is where the meta-inferences live: the conclusions that draw on both strands together.
Four integration types cover most designs, and they are named in the Fetters, Curry and Creswell tradition. Merging brings both datasets together in one table for comparison. Connecting links the output of one analysis to the next, for example taking a significant survey result and selecting interview participants who scored high or low on that variable. Embedding nests one strand inside the other. Transforming converts one type of data into the other, most often by turning a quantitative variable into a categorical grouping that the qualitative analysis can be organised around.
You can report which of these you used, and where in the chapter each one appears. Examiners want to see the vocabulary made concrete.
| Integration type | What it looks like in the chapter | What you report |
|---|---|---|
| Merging | A joint display with one column per strand | Which findings agree, differ, or add information neither strand holds alone |
| Connecting | A sentence or table linking one result to the sampling or analysis of the next | The link itself: which result drove which follow-up analysis |
| Embedding | A subsection of one strand nested inside the other strand’s reporting | Where the nested strand sits and why |
| Transforming | A described recoding step, then the second strand’s structure | How the data were converted and what the resulting categories are |
At least one joint display belongs in most mixed methods chapters. The minimum useful version is a three-column table: the quantitative finding, the qualitative finding, and what the two together support.
| Quantitative finding | Qualitative finding | Integrated conclusion |
|---|---|---|
| No significant relationship between weekly study hours and exam score, p = .18 | Participants with high study hours described scheduling study around work and family commitments | Study hours alone did not predict performance; the constraints participants described explain why the measure was insufficient |
| Self-efficacy scale showed a moderate positive association with retention, r = .42, p < .001 | Participants who described early small wins attributed their persistence to them | Confidence appears to operate through early success experiences rather than through study volume |
When the strands disagree, that belongs in the chapter, not buried in limitations. Report both results, then state plainly what accounts for the difference, and say what the combination changes about your conclusion. A survey showing no effect alongside interviews explaining why is often the most interesting result in the whole study.
To see the difference, here is a weak integration paragraph and its rewrite. The weak version: the interviews showed that students were stressed, and the survey showed that stress was related to performance. They were both related. The rewrite:
Quantitative results showed a significant negative association between perceived stress and mean semester grade, r = -0.47, p < .001, n = 298, with 95% CI [-0.55, -0.38]. Interviews with 18 students drawn from the high and low stress quartiles indicated that exam preparation was the mechanism through which stress operated. Students in the high-stress quartile described revising as an activity they delayed until they felt able to focus, creating a cycle in which low readiness produced more delay, while students in the low quartile described scheduling revision in fixed blocks irrespective of motivation.
Both strands therefore point to preparation routines rather than stress itself as the actionable lever. The survey result identifies the size of the relationship and the interview data specify the behaviour inside it.
The rewrite does three things the original does not: it names the integration method, it states what each strand contributes, and it draws a conclusion that needed both.
6. Write the Results in Clear Academic Prose
Move from broad to specific within each subsection. Open with the answer to the research question in one sentence, then give the supporting detail, then point the reader to the table or figure.
Keep tense consistent. Use past tense for what you did and what you found in both strands, and present tense when describing what a theme contains or what participants described, as in the theme of delayed revision, which captures how participants talked about postponing study. Mixing tenses randomly reads as carelessness, so pick the pattern and apply it.
Write in third person or first person plural, following your discipline’s convention, and keep it the same throughout. Examiners read for what you claim, not for who claims it.
Watch the verbs. The quantitative strand wants tested, showed, differed, predicted. The qualitative strand wants described, framed, articulated. Both strands can be said to indicate, and both should avoid proving, caused and led to unless a design genuinely supports causation.
Signpost transitions between strands explicitly. A sentence such as the survey result above identified the group most affected; the interviews that follow examine how that group described the problem does more work than any transition word, because it tells the reader why the next subsection exists.
7. Check the Chapter for Accuracy and Coherence
Run a verification pass with your traceability list open. Every statistic matches the output file, every table and figure is cited in the text at least once, every quotation matches the transcript, and every participant identifier in a quote still protects their identity after you added demographic detail.
Check that the numbers reconcile. If a table reports n = 312 and the text reports n = 298 for a later analysis, the reader needs a sentence explaining the difference. Silent mismatches are the fastest way to lose an examiner’s confidence.
Then read for the results and discussion boundary. Cut any sentence that explains why a result might have happened, any comparison to other studies, and any recommendation for practice or policy. Move them to the discussion chapter, keeping your chapter pointing forward to it rather than drifting into it.
Finally, check that each subsection ends by answering its research question, and that the closing section states your meta-inferences as conclusions from the evidence rather than as opinions about it.
Common Mistakes
These are the criticisms that come up most often in supervision and examination, with the fix for each.
Reporting two studies side by side
The strands appear in sequence with a thin paragraph between them and no visible connection. Fix it by adding an integration subsection and at least one joint display, then rewriting the two strand sections so each references the point at which it connects.
Mixing methods without saying why
The design is never named, so the reader cannot judge whether mixing was appropriate. Fix it with one sentence in the opening of the chapter stating the design type and the reason for choosing it.
Reporting methods again
Sample details, instruments and procedures get repeated from the methods chapter. Fix it by keeping the methods chapter as the single source and retaining only the sample overview numbers a reader needs to follow the results.
Overinterpreting
Claims like the interviews demonstrate that students are overworked go beyond what the data support. Fix it by describing what participants said and reserving causal or explanatory language for the discussion.
Burying contradictory findings
The two strands disagree and the chapter mentions the conflict only in a limitations subsection. Fix it by reporting the disagreement in the integration section, where it belongs, and stating what the study can now claim.
Quotation overload
Sections run on with block quotes and few theme labels. Fix it by naming the theme, summarising the pattern in prose, and using one or two quotations to bring it to life.
Silent non-significant results
Tests that came back null are omitted from the prose. Fix it by reporting them with their statistics and effect sizes, which lets you discuss them honestly later rather than leaving a gap a reader will notice.
Interpretation in the results chapter
Sentences about what the findings mean for practice appear throughout. Fix it by moving them to the discussion and checking the chapter reads as evidence only.
One more worth naming: writing the chapter before your tables are final. Numbers change during review, and prose built around old versions creates silent inconsistencies. Finalise tables and figures first, then write the paragraphs that reference them.
Frequently Asked Questions
How long should a mixed methods results chapter be?
Length depends on your design and your institution’s limits rather than a fixed number. A common allocation puts roughly 25 to 30 per cent of the chapter on the quantitative strand, 20 to 25 per cent on the qualitative strand, and 25 to 30 per cent on integration. Sequential designs need a larger integration share because the connection between phases is the finding. Check your handbook before allocating, since a stated word limit overrides any general rule.
Is it okay to combine results and discussion in a mixed methods study?
Yes, in some disciplines and with most journals, but separate chapters are safer for a thesis. Combining them is common in applied health and education journals that use an integrated results and discussion style. If you combine, you still need a clear sentence-level boundary: report first, interpret after, within each paragraph. Whichever you choose, tell your supervisor before you draft rather than restructuring later.
What is the purpose of a joint display in mixed methods research?
A joint display places quantitative and qualitative findings side by side in one table so a reader can see the relationship between them directly rather than inferring it across pages. The usual columns are the quantitative finding, the qualitative finding, and the integrated conclusion. It is the clearest evidence that integration happened, which is why examiners look for it. Most mixed methods chapters benefit from at least one, even a small one.
Which order should I present quantitative and qualitative results in?
The order should follow your design and your research questions, not a house preference. In a convergent design either order works, though putting the quantitative strand first often gives the reader an anchor. In a sequential explanatory design the quantitative results come first because they drove the interviews. In a sequential exploratory design the reverse holds. State the reason for your order in one sentence at the start of the chapter.
How should I report conflicting results from my survey and interviews?
Report both results with their full statistics, then state what accounts for the difference, then say what the combination changes about your conclusion. Options include a methodological explanation, such as different measures of the same idea, or a substantive one, such as what the survey could not capture. Keep the disagreement in the integration section of the results chapter rather than a limitations subsection, where it is easy to miss.
What tense should I use in a mixed methods results chapter?
Use past tense for what you did and what you found in both strands, and present tense when describing what a theme contains or what participants described. So a tested relationship is past tense, while a theme of delayed revision, which captures how participants talked about postponing study, is present tense. Apply the pattern consistently across both strands, since switching between them randomly reads as carelessness.
Conclusion
Start with stage 1, because that is where how to write a mixed methods results chapter stops being a guess and becomes a plan. Write your integration point in one sentence, sort your research questions into strands, and check your sampling design, because the structure of everything after that follows from those three decisions.
Then finalise your tables before you draft a paragraph, and give the integration section the same care as the two strands. Most people find the quantitative and qualitative sections straightforward once the analysis is done. The meta-inferences, the sentences that exist only because you mixed methods, are the part worth extra time, and they are the part examiners read first.
