Week 6

Question 1

What is Mixed Methods research

Answer 1

• Involved integrating both quantitative and qualitative research methods within a single study or series of studies
  
  • In a mixed methods study, rigorous quantitative data (numbers, measurements) and qualitative data (words, experiences) are collected and analysed and the findings are combined (integrated) to provide a richer understanding than either approach alone
  • The goal is to draw on the strengths of each approach e.g. combining statistical trends with personal experiences to gain a comprehensive perspective on a research problem
  • An important point: The two types of data are intentionally integrated (not just collected in parallel) during the research process
    This integration can occur at various stages (during data collection, at analysis or when drawing conclusions) to ensure that quantitative and qualitative components inform one another

Question 2

Why Mixed Methods

Answer 2

• In health research, many complex questions benefit from multiple perspectives
  
  • Using MM approach allows researchers to offset the limitations of using only quantitative or only qualitative methods
  • Quantitative data can provide breadth, generalisability and objective measurement, while qualitative data provide depth, context and insight into meanings or experiences
  • By combining them, MM research offers the best of both words, yielding more robust and contextualised findings
    e.g. quantitative results might tell what happened (e.g. a treatments effect size) and qualitative results can tell why or how it happened (e.g. patient perceptions of the treatment).

Question 3

Key strengths of MM

Answer 3

• Comprehensive understanding: Validate and enrich findings by seeing whether both data types agree (converge) or offer complementary information. This builds a more complete answer to the research question than a single method would
  
  • Offsets weaknesses: The weaknesses of one method can be mitigated by the strengths for the other. e.g. qualitative insights can explain unexpected statistical results, and qualitative data can test whether insights from interviews hold for a large group
  • Innovation and flexibility: MM are adaptable and not tied to one paradigm. They allow researchers to be creative in design - generating hypotheses from qualitative data and then testing them quantitatively in the same study

Question 4

Challenges in combining qualitative and quantitative approaches

Answer 4

• More time, effort and expertise as you are essentially doing two studies in one
  
  • Heavier workload - collecting and analysing two types of data is labour-intensive and may require a research team with diverse skills
  • Challenges with integrating data (what if the two results conflict or don’t easily align)
  • Interpreting differing or conflicting results can be difficult
  • Must address methodological rigor of both components: a weak qualitative or quantitative part can undermine the overall study quality

Question 5

Common MM designs

Answer 5

Convergent parallel, Explanatory sequential, Exploratory sequential, embedded

Question 6

What is Convergent parallel

Answer 6

• Quantitative and qualitative data are collected simultaneously (in parallel) and are analysed separately, then merged to compare results and draw overall conclusions
  This design (also called concurrent triangulation design) is useful when you want to directly compare or triangulate findings from both data types to see if they agree or provide complementary insights

Question 7

What is Explanatory

Answer 7

• The research is done in two phases
  
  • A quantitative phase is conducted and analysed and
  • Then a qualitative phase follows to help explain or elaborate on the quantitative results.
  • Or you collect quantitative data first (e.g. survey or experiment), look at the results then use results to design a qualitative question that probe the reasons behind the numbers
    Useful when you have quantitative findings that need further explanation or context

Question 8

What is Exploratory

Answer 8

• Reverse sequence of explanatory
  
  • You start with qualitative phase first to explore a topic and gather insights, then use those findings to inform a follow-up quantitative phase
  • e.g. you might begin with focus groups to identify themes or develop a theory, and then design a survey or measurement based on that to see how prevalent those themes are in a larger population
    This design is helpful for instrument development or when little is known about a topic

Question 9

What is Embedded design

Answer 9

• Oen type of data is embedded within a larger study that is primarily guided by the other type
  
  • e.g. you might run a primarily quantitative trial or survey, but embed some qualitative interviews within it (or vis versa)
  • The embedded approach collects both types roughly at the same time, but one is secondary
  • It is often used when one set of data alone is not enough
  • The secondary data provide additional context or insight to support primary data
    This approach is useful if you have limited resources or need to address a specific qualitative question within a quantitative study ( or vis versa)

Question 10

What is convergent parallel

Answer 10

• Quantitative data: Blue
  
  • Qualitative: Green
  • Are collected independently but concurrently, each analysed on its own and then compared and merged for interpretation
    Allows researchers to see if both types of data yield similar results or if they diverge leading to a more nuanced overall finding

Question 11

Steps 1,2, in planning a MM study

Answer 11

1. Define the research problem and rationale
   
   • Clearly state what you want to know and why a MM is warranted
   • Ask why do I need both Qual and Quant
     2. Develop specific research question
   • Often MM have sub-questions or objectives for each component
     Oen or more Quant questions and one or more qual questions as well as an over-arching MM question

Question 12

Step 3 in planning a MM study

Answer 12

3. Choose a MM design:
   
   • Based on your question and what you need from each type of data, select the design that fits best
   • The designs introduced in Module 1 each serve different purposes
   • e.g. if your primary aim is outcome-focused but you want explanations for those outcomes, an explanatory sequential design might be appropriate
   • Take into account timing (both methods be done at the same time or one after the other)
     And priority (will one type of data have more weight in addressing the main question or are they equal?)

Question 13

Step 4 in planning an MM study

Answer 13

4. Plan data collection for each component
   
   • Decide what data you will collect and how
   • This could be clinical measures, surveys, observations etc.
   • Could be interviews, focus groups, surveys
   • Think about sampling: you might use some people for both parts, or different
   • In sequential, the second phase may depend on the first
     Also consider logistics: If you are doing an intervention study, will you embed some interviews during or after the intervention. If doing a survey, will you include some open-ended questions along close-ended ones

Question 14

Step 5 and 6 in planning an MM study

Answer 14

5. Decide on integration strategy
   
   • Plan when and how you will bring results together
     6. Maintain rigout in both components
   • Be mindful of the quality criteria for both quantitative and qualitative methods
   • MM is only as strong as its parts
   • Consider issues like sample size and power for the quant piece and trustworthiness (credibility and dependability) for qualitative
     Ensure ethical approvals cover both data type

Question 15

Defining an MM research question

Answer 15

• Not every questions requires MM
  
  • MM often links how many/how effective aspect with a how or why aspect

Question 16

Selecting the right MM design

Answer 16

• Comes down to intent
  
  • Are you seeking confirmation (do the qual or quant findings validate each other?)
  • Are you explaining results
  • Are you exploring then measuring
    Do you have a primary approach but want a supplemental insight

Question 17

Techniques for integrating Q and Q data

Answer 17

Triangulation, Connecting builds, Embedding, Narrative weaving

Question 18

What is triangulation

Answer 18

• Side-by-side comparison of findings on the same phenomenon to see how they relate
  
  • e.g. present a statistic and see if interviews corroborate
  • Can confirm findings by showing agreement between data types of it can reveal complementary aspects of a finding
    Do this through using a joint display: A table or matrix where Qual and Quant results are aligned for each research question our outcome

Question 19

What is connecting builds (sequential integration)

Answer 19

• Integration often occurs by connecting the phases - the results of the first phase inform the data collection of the second
  
  • e.g. in an exploratory sequential design you might use the quantitative results to decide whom to interview or what questions to focus on in the qualitative follow-up
    The integration here is in the design itself: the link is that the qualitative data explains or builds upon specific quantitative results. When writing up the analysis, you would explicitly connect findings (e.g., “Survey results showed outcome X was low; interviews with participants who had low X revealed the following reasons…”).

Question 20

What is Embedding (integrative analysis in real time)

Answer 20

• Integration can occur by transforming data from one form to another to mix within the others analysis
  
  • For instance, you might quantitise qualitative data – converting themes or responses into numeric codes or counts that can be compared or statistically analyzed alongside quantitative data.
  • Conversely, you might qualitise quantitative data – using some quantitative results as narratives or cases within qualitative analysis (e.g., treating an extreme case’s scores as a “story” to explore). Embedded approaches sometimes use real-time integration:
    e.g., in the Cronin physio example, they embedded open-ended questions in a survey and then mixed the results in the interpretation​.

Question 21

What is narrative weaving

Answer 21

• Weave qual and quant findings together in a narrative discussion
  
  • Instead of presenting separate result sections, the researcher integrates by topic
  • e.g. one might write a result section organized by key themes or topics and for each theme, present the supporting quantitative result and qualitative quotes or explanations together
  • Requires careful planning

Question 22

Using software tools in MM

Answer 22

• Managing two datasets can be challenging, but software can help. For quantitative data, programs like SPSS, R, or Excel are common; for qualitative data, software like NVivo, Atlas.ti, or MAXQDA can assist in coding and organizing text.
  
  • Some tools are designed with mixed methods in mind. For instance, NVivo (a qualitative data analysis software) allows you to import quantitative data or demographic info and link it to qualitative responses, and even to run simple quantifications of coded qualitative data.
  • Dedoose is another tool explicitly built for mixed methods, where you can analyze text and also generate charts that correlate qualitative codes with quantitative characteristics.
  • Using these tools, you could, say, code interview transcripts in NVivo and at the same time compare code frequencies or sentiment with survey scores imported for each participant.
  • This makes integration more systematic. To see an example of how software facilitates mixed methods analysis, check out the video below

Question 23

Synthesizing findings for a comprehensive understanding

Answer 23

• Interpreting how the qual and quant results inform each other
  
  • Do the two sets of results agree (converge) on the conclusions, do they highlight different aspects
  • The integrated conclusion might be that the intervention, while not changing the standard functional measure, provided subjective improvements in quality of life
  • Consider relative priority of findings if one dataset was dominan
  • If your study was QUAN→qual (capitalisation indicating priority), you might interpret primarily in terms of whether the qual explains the quant.
    If equal, you integrate on more even footing. In any case, explicitly discuss how the two types of data relate to each other. This “mixing” in the conclusion is what truly makes the study mixed methods, rather than just two studies reported together.

Question 24

Challenges in data integration and how to address them

Answer 24

• When results diverge and contradict each other
  
  • Consider it insight: Conflicting results can indicate subgroups or unseen factors
  • Re-examine whether methods truly addressed the same question or whether measurement issues exist
  • Be transparent about conflicts and, if possible, collect additional data or literature to help interpret them
  • Comparing data of different forms
  • Convert one form to another
  • Bias as integration can occur if a researcher subconsciously favours one dataset
    To avoid this: Use a structure approach for integration e.g. write separate integration results

Question 25

Criteria for quality in MM research (CCAAI)

Answer 25

Clear rationale for mixing, appropriate design and execution, integration and mixing, consistency and coherence of results, adherence to MM reporting standards

Question 26

Criteria for MM: Clear rationale for mixing and appropriate design and execution

Answer 26

Clear rationale for mixing: - Does the study state why it used MM Appropriate design and execution: - Why was the chosen MM suitable for the research question - Was each part carried out rigorously - Appropriate sample size, measures and analysis, sampling etc Both components should meet their respective quality standards

Question 27

Criteria for MM: Integration and mixing

Answer 27

- Critically check how and where the study integrated the two data sets - Did the researchers integrate during analysis or only in the discussion? Do they present merged findings or explicitly compare the qualitative and quantitative results - Good MM will describe the integration process - If the paper has separate results section with no connection = red flag Study should reflect on how the integrated findings provided insights that neither method alone would have achieved

Question 28

Criteria for MM: Consistency and coherence of results

Answer 28

- Do qual and quant results together make sense logistically - High quality studies will discuss whether the two sets of results confirm each other (consistency), or if not how they complement each other Coherence doesn’t mean the results have to be identical but the interpretation should address any differences in a reasoned way

Question 29

Criteria for MM: Adherence to MM standards

Answer 29

- There are emerging reporting checklists that expect authors to clearly report all aspects of their study - e.g. the study should identify itself as MM, name the design and detail the sequence/timing of qual and quant parts - It should report results for each part and the integrated results - When appraising, see if the study is transparent about these elements - Lack of clarity (e.g. you can't tell which data was collected first, or how the interview questions were derived) can indicate a methodological issue or simply poor reporting

Question 30

Common pitfalls in MM: Qual and quant ships passing in the night

Answer 30

- When a study claims to be MM but the two components are essentially conducted in isolation and never truly integrated - You might see two separate studies in one paper that don't inform each other - If in the discussion or results, the authors never bring the datasets together, that's a pitfall As a reader you should expect some combined analysis or at least a comparative discussion

Question 31

Common pitfalls in MM: Imbalance or unequal rigor

Answer 31

- When appraising, judge each component on its own merits, if one is weak, that’s a quality concern When quant or qual might be strong and the other is an afterthought

Question 32

Common pitfalls in MM: Inadequate integration

Answer 32

- Not actually combining data sets - Look at whether the conclusions of the paper actually draw on both quant and qual Tell-tale sign: if the conclusions or abstract only mention one set of results, as if the other didn't exist

Question 33

Common pitfalls in MM: Conflicting paradigms ignored

Answer 33

- MM often bring together different world views (positivist for quant, interpretivist for qual) - A subtle pitfall is when the study's framing doesn't reconcile this e.g. treating qual data in a purely positivist way without acknowledging its interpretive nature - A practical symptom might be if the study tries to quantify everything (even where depth is needed) or conversely if it dismisses numerical trends without reason Good practice Is adopting a pragmatic paradigm or similar that explicitly values both

Question 34

Common pitfalls in MM: Overgeneralization

Answer 34

- Risk of letting rich qual narrative cause overgeneralisation of quant findings or vice versa - Check conclusions stay grounded: quantitative inferences should stick to their population/metrics and qualitative insights should not be stretched beyond their context - Ideally the MM conclusions will delineate what part of the inference comes from generalisable evidence and what part comes from contextual evidence

Question 35

Using MM findings in EBP

Answer 35

- Healthcare decisions benefit from both evidence of effectiveness and understanding of context/implementation - In AH if an intervention shows moderate improvement in patient outcomes and the feedback showed they enjoyed it then a clinician may be more confident adopting this practice - MM can bridge the gap between research and practice by providing actionable insights - The numbers might tell whether something works and the narratives tell for whom, under what conditions and how to make it work better In EBP we consider patient values and context alongside the best research evidencen