Choosing mixed methods research software means building a workflow in which a statistics package and a qualitative package talk to each other, not finding one package that does everything. In most theses the quantitative strand runs in SPSS or Stata and the qualitative strand in MAXQDA; integration, the part examiners probe, happens where the two meet. This guide maps each design to its software steps, walks through an explanatory sequential study and shows how to report the result. For the coding itself, see our thematic analysis guide.
What mixed methods research software must actually do
A mixed methods study is not two studies stapled together. The workflow must cover four jobs, or integration collapses into two parallel results chapters.
- Manage both strands: numeric data analysed from saved syntax; transcripts, memos and a versioned codebook.
- Link cases: each participant carries the same ID (e.g. T017) in the survey file and in the transcript name. Everything downstream depends on this.
- Integrate: compare codes across quantitative groups, turn codes into variables where justified and build joint displays.
- Export: send quantitized variables back to the statistics package and tables to the write-up without retyping.
MAXQDA's Mixed Methods menu covers the last three jobs: it imports survey scores from SPSS or Excel files as document variables, activates documents by variable values, cross-tabulates codes by group and exports variables. Its statistics module handles descriptive work, but inferential analysis usually stays in SPSS or Stata, with AMOS or Mplus for latent variables.
Mixed methods analysis software by design: which step runs where
The three core designs divide the work differently. A survey with open-ended items is a convergent variant: MAXQDA can import the answers as documents and the closed items as variables.
| Design | SPSS / Stata | MAXQDA | Integration output |
|---|---|---|---|
| Convergent | Analyse the survey independently: descriptives, group comparisons, regression | Code interviews independently on the same topics | Side-by-side display per topic: convergence, expansion or discordance |
| Explanatory sequential | Analyse the survey first; create a group variable; flag extreme, typical or deviant cases | Interview the flagged cases; import their scores as document variables; compare themes by group | Themes-by-group display explaining the quantitative pattern |
| Exploratory sequential | Test the instrument built from the themes (EFA in SPSS, CFA in AMOS or Mplus) | Code the interviews first; turn themes into candidate dimensions and items | Theme → item → factor table showing how the instrument was built |
A strong methods chapter names all three levels of integration: design (the sequence and priority of the strands), method (connecting through case selection, building an instrument, merging data) and interpretation (joint displays and meta-inferences). Software supports the last two; the first is a decision you document before opening any file.
Walkthrough: an explanatory sequential study from SPSS to MAXQDA
Take a hypothetical thesis: a teacher survey measures burnout, and interviews explain why some teachers score far higher.
- Analyse the survey in SPSS from a saved syntax file (see our SPSS thesis analysis guide) and create a group variable, e.g. burnout more than 1 SD above or below the mean.
- Select cases by rule: extreme (both tails), typical (near the mean) or deviant (large regression residuals). Write the rule into the syntax so it can be rerun.
- Name transcripts by case ID. T017 in MAXQDA must match T017 in the survey file exactly; mismatched IDs are the usual reason variables fail to attach.
- Code in MAXQDA and check agreement with a second coder on a subset; our MAXQDA intercoder reliability guide covers the kappa settings.
- Import survey scores as document variables: burnout score, group and key covariates.
- Compare themes by group. Activate high-burnout, then low-burnout documents and retrieve their segments; use Crosstabs for code frequencies per group and the joint display that arranges qualitative themes by quantitative groups.
- Draft the joint display: for each theme, state whether it explains, contradicts or extends the survey finding.
With one or two dozen interviews, group counts are descriptive: skip significance tests and let the quotations carry the explanation.
Quantitizing and qualitizing without overreach
Quantitizing turns codes into numbers. MAXQDA can transform a code into a document variable holding presence/absence (0/1) or the number of coded segments per document. In our hypothetical study, 'work spilling into evenings' might appear in 9 of 12 high-burnout and 3 of 12 low-burnout interviews. Presence/absence is usually safer: raw frequencies also track interview length and talkativeness. Exported to SPSS, these variables support descriptive cross-tabs; formal tests need large qualitative samples, such as hundreds of open-ended answers.
Qualitizing runs the other way: numbers become narrative categories. Respondents grouped into profiles (cluster analysis in SPSS, latent profile analysis in Mplus) are described and named from their members' interviews; MAXQDA's Typology Table shows variable summaries per document group side by side. Either way, report the transformation rule; reviewers see an undocumented conversion as a validity threat.
Joint displays: building and reporting them
A joint display places quantitative and qualitative results in one frame, with a column stating what the combination means. MAXQDA generates side-by-side and group-comparison displays; the thesis version is usually finished in Word or Excel.
| Theme | High-burnout group | Low-burnout group | Meta-inference |
|---|---|---|---|
| Work spilling into evenings | 9 of 12 interviews; 'the marking comes home with me every night' (T017) | 3 of 12 interviews | Explains the survey gap |
| Principal support | Present but merely formal | Frequent, concrete examples | Confirms and refines: the quality of support matters, not just its presence |
| Parental pressure | Raised in both groups | Raised in both groups | Expands: absent from the survey model |
| Career stage | Early-career strain in most interviews | Rarely raised | Discordant: no experience effect in the survey; revisit the measure |
- Place it in the results or discussion chapter and follow APA 7 table format.
- Give every row a meta-inference: confirmation, expansion or discordance.
- Report discordance openly; it is often the most publishable finding.
- With small samples, give denominators ('9 of 12', not '75%').
Reproducibility and the method-section sentence
Reproducible means someone else can rerun both strands and see how they were joined. Archive:
- the SPSS syntax or Stata do-file, including the case-selection rule;
- the MAXQDA project, the exported codebook with its version history, and your memos;
- a case-ID key stored apart from identifying data, plus a Python or Excel check that every interviewed ID is in the survey file;
- a dated decision log (MAXQDA's logbook works well) recording every recode, merge and transformation.
In the method section, name both packages with their versions and state the integration procedure: “Survey data were analysed in IBM SPSS Statistics (Version [xx]). Following an explanatory sequential design, participants with extreme burnout scores (±1 SD) were interviewed, and transcripts were coded in MAXQDA [year] (VERBI Software). Survey scores were imported as document variables, themes were compared across groups and integration was reported in a joint display.” For end-to-end support, see our qualitative and mixed methods analysis service.
Integration is not a chapter you write at the end; it is a case ID you define at the start.
Frequently Asked Questions
What is the best software for mixed methods research?
No single package handles both strands equally well. The most defensible thesis workflow pairs SPSS or Stata with MAXQDA, which codes the qualitative data and links the two strands. A shared case ID and a documented integration step matter more than the brand.
Can MAXQDA do statistical analysis, or do I still need SPSS?
MAXQDA's statistics module covers descriptive statistics and common basic analyses, which is often enough for a modest quantitative strand. For regression or latent variable models, use SPSS, Stata, AMOS or Mplus and import the results into MAXQDA as variables.
How do I report a joint display in a mixed methods thesis?
Present it as an APA-formatted table in the results or discussion chapter, with one row per theme. Include the quantitative result, qualitative evidence with a short quotation and a meta-inference column: confirmation, expansion or discordance.
What mixed methods analysis support does Celsus offer?
Celsus designs the integration plan, analyses the quantitative strand in SPSS or Stata, codes the qualitative strand in MAXQDA with intercoder checks and builds the joint displays. Delivery includes syntax files, the MAXQDA project and a method-section draft you can defend.