Mixed Methods and Triangulation
The source devotes two slides to mixed methods and defines triangulation instead - twice, differently, in consecutive sentences. This page sets out both definitions, the three named kinds that get no explanation, the benefit that its own listed constraint cancels, and the dilemma slide that is unrecoverable.
What the two mixed-methods slides actually contain
The material sits on two consecutive slides, "Mixed Methods Approach (1)" and "Mixed Methods Approach (2)". There is no supporting prose: the week's study notes carry no research design content, so the slides are the whole treatment.
That is the first slide in its entirety. Read it once more and the difficulty is visible on the face of it: a slide titled with one term, carrying a boxed definition of a different term, which the framing sentence has already described as merely a rationale for the first.
Three claims about what triangulation is, in three consecutive sentences
The panel does not say one thing twice. It says two different things, and the sentence above it says a third. Each has different consequences for how you would design and defend a study.
THE THREE FORMULATIONS AND WHAT EACH COMMITS YOU TO
| Formulation | Source wording | What it would commit a design to |
|---|---|---|
| Procedural | "the use of more than one approach to answering the research question or responding to the research issue" | Two or more approaches each producing an answer, which then have to be compared, weighted or combined. |
| Perspectival | "looking at the problem or question under investigation from more than one perspective" | One object examined through several lenses, where difference between the views is informative rather than a discrepancy to resolve. |
| Verificationist | different approaches "are used to test findings" | A primary finding and a secondary check on it, where agreement supports the finding and disagreement counts against it. |
The three-way split is this page's reading of the source's own wording. The source presents all three as one idea and does not distinguish them.
Three kinds named, none of them explained
The panel closes with three bullets. Each is two words long. Nothing follows any of them - no explanation of what is varied in each, no procedure, no condition under which one is preferable, and no example.
THE THREE KINDS, AND WHAT THE SOURCE ATTACHES TO EACH
| Kind, as named | What the source says about it | What you would need to use it |
|---|---|---|
| Researcher triangulation | Nothing beyond the label. | Who the additional researchers are, how their work is kept independent, and how disagreement between them is resolved. |
| Theoretical triangulation | Nothing beyond the label. | Which theories, why those, and what it would mean if they yielded incompatible readings of the same data. |
| Method triangulation | Nothing beyond the label. | Which methods, in what order or in parallel, and how their outputs are brought together. |
Three is the count of printed bullets. The source states no count. The third column is this page's statement of what is missing, not a source list.
The stated benefits and the stated issues
The second slide is more substantial. It offers benefits under the heading "Triangulation has several benefits for your work" and then a list headed "Issues to consider".
Benefits, quoted from the slide
- "Check that the results obtained from the use of different logic and processes will validate your results"
- "Multiple methods provide a wider perspective of the issue you are researching"
- "You develop a more comprehensive understanding of the issues, which may in turn benefit the design of your data collection processes and tools"
Issues to consider, quoted from the slide
- "Different philosophical assumptions"
- "Practicalities of data collection could make simultaneous analysis difficult"
- "Statistical analysis relies on large samples: unusual in qualitative research"
The benefit and the constraint that cancel each other
Benefit 1 claims that results from different logic and processes will "validate your results". Issue 3 states that statistical analysis relies on large samples, which is "unusual in qualitative research". Both are on the same slide, four lines apart.
If method triangulation pairs a small qualitative component with a quantitative one, the source gives no account of how the small component validates anything statistically, or of what validation without statistics would consist of. It asserts the benefit and states the disabling constraint without connecting them.
The comparison table the argument rests on
The slide immediately before the mixed-methods pair carries a four-row comparison of the two approaches, attributed in the source to a 2010 research-methods guide at page 105. It is the table a reader would use to decide what combining approaches would actually combine.
THE FOUR-ROW COMPARISON, REPRODUCED IN FULL
| Quantitative Method | Qualitative Method |
|---|---|
| “Methodology: scientific method, hypothesis drive, deductive, reliable, valid, reproducible, objective, generalizable” | Methodology: ethnomethodology, ethnography, action research, inductive, subjective, intuitive |
| Methods: large scale, surveys, random control trials | Methods: Small scale, interviewing, observation, document analysis |
| Data type: generally quantitative | Data type: Generally qualitative |
| Analysis: statistics | Analysis: thematic exploration |
Quoted from elsewhere: the source attributes the table to a 2010 guide, p. 105. Spelling and capitalisation are the source's.
Three things about this table matter if you are going to build a combined design on it. First, the methodology row mixes categories: on the qualitative side, ethnomethodology, ethnography and action research are named methodologies while inductive, subjective and intuitive are descriptors of stance; on the quantitative side, reliable, valid, reproducible, objective and generalizable are claimed properties of the output rather than methodologies at all.
Second, that makes the table assert reliability and validity as intrinsic features of quantitative work - which contradicts the same week's own treatment of them as things a project must achieve, and its selection question asking whether any project "will be considered valid, reliable or trustworthy when reviewed". Both positions are in the week and neither is withdrawn; see Validity and Reliability.
Third, the data-type row is circular: the quantitative method is characterised by producing quantitative data. The hedge "generally" is doing all the work in that row, and it should be preserved whenever the row is quoted.
One more term to note: ethnomethodology appears in this table and is defined nowhere in either week. So do random control trials, thematic exploration and document analysis. They are labels, in the same way the week's method grid is labels - see The Method Catalogue.
The methodological dilemma slide
The slide that follows the mixed-methods pair is titled "Methodological dilemma" and its complete textual content is that title and the words "What do you think?". The substance sits in an image that is not present in the supplied text.
Building a combined design from what is here
The material supports less than it appears to. What follows is a way of using it that stays inside what it actually says; the sequence is this page's, not the source's.
Working sequence for a combined design
Say which sense of triangulation you mean
Procedural, perspectival or verificationist, in one sentence, citing the definition you are using. The source's panel supports the first two and its framing sentence supports the third, so quoting the panel alone will not settle it.
Name the kind, then specify it
Researcher, theoretical or method triangulation. Then supply what the source does not: who or which theories or which methods, and how their outputs will be brought together.
Decide in advance what disagreement means
This is the question the two benefits and the third issue leave open. If your two components disagree, is that a failure of validation, or a finding about the phenomenon? Answer it before you collect, and record the answer.
Cost the second component honestly
The source's second issue - practicalities of data collection making simultaneous analysis difficult - is the one that most often ends mixed designs in practice. Two components mean two instruments, two ethics positions and two analysis passes.
State the philosophical position once
The source raises differing philosophical assumptions and stops. You cannot, because a reader will ask. Take a position, state it early, and keep the rest of the design consistent with it.
Do not promise statistical validation from a small component
The material states plainly that statistical analysis relies on large samples and that this is unusual in qualitative work. A design that claims a small qualitative component will statistically validate a quantitative one is claiming something the source's own constraint denies.
What to carry forward
- Two slides on mixed methods and no definition of mixed methods. What is defined is triangulation, which the same slide calls a rationale for mixed methods rather than the thing itself.
- Triangulation is defined procedurally and perspectivally in consecutive sentences, with verification asserted above them. Pick one sense, state it, and keep it.
- Researcher, theoretical and method triangulation are named and nothing is attached to any of them. Specify what varies yourself.
- The first stated benefit and the third stated issue cancel: validation is promised, and the sample constraint that would prevent it is stated four lines later.
- The dilemma slide is unrecoverable, and no numeric sample size appears anywhere in the week. Do not import either gap into your own work as though it were content.
Frequently asked questions
Does the source define mixed methods?
No. Both slides are titled "Mixed Methods Approach" and neither defines the term. The only definition offered is of triangulation, which the framing sentence describes as "one of the first rationales" for a mixed methods approach - that is, explicitly not the same thing. You will need a definition of mixed methods from a text you can cite.
Which definition of triangulation should I use?
The source gives two in consecutive sentences and implies a third, and it does not choose between them. Decide which you mean - more than one approach producing answers, one problem viewed from several perspectives, or a second approach testing the first - and say so in writing. The important thing for a reader is that you are consistent, not which you pick.
What is method triangulation, specifically?
The source names it and says nothing further. No procedure, no ordering, no worked example and no guidance on combining outputs is supplied for any of the three named kinds. Treat the label as a heading you must fill from a cited source and from your own design decisions.
Can a small qualitative component validate a quantitative finding?
The material claims triangulation will validate your results and, on the same slide, states that statistical analysis relies on large samples which are unusual in qualitative research. It never connects the two. Both statements stand in the source, so avoid resting a design on the validation claim without saying what form the validation takes.
How large is a "large" sample?
The material never says. Across the research design and qualitative weeks no numeric sample size appears at all - only "large scale", "Small scale" and "small sample sizes". Any figure you use is one you have brought in and must justify against your own population and analysis plan.
What was the methodological dilemma?
Unknown. The slide's entire text is its title and the words "What do you think?", with the substance carried by an image not present in the supplied material. It cannot be reconstructed, and this page does not attempt a reconstruction.
References and source attribution
- O'Leary, Z. 2010, The essential guide to doing your research project, Sage Publications - the work the research design week attributes its quantitative and qualitative comparison table to, at p. 105.
- O'Leary, Z. 2017, The Essential Guide to Doing Your Research Project, 3rd ed., Sage Publications, London.
- Bryman, A. 2016, Social Research Methods, 5th ed., Oxford University Press, Oxford.
- Quinlan, C. 2011, Business Research Methods, 1st ed., Cengage Publishing.
- The supplied teaching source: the research design week's slide deck, slides 14 to 17, used here as the basis for every statement attributed to "the source".
Suggested questions for Ask KEVOS
- Draft a triangulation statement for my methodology chapter using one consistent definition.
- What would method triangulation look like for a study of project handover failures?
- Help me decide what to do if my interview findings contradict my survey results.
- Give me a cited definition of mixed methods, since the teaching material has none.
- Compare the three quantitative and qualitative tables in this material and show me where they disagree.