An Overview of Data Analysis in

Mixed Methods Research Projects

Supporting Postgraduate Students in Designing

and Conducting Mixed Methods Studies

Ben Daniel, PhD, SMIEEE

University of Northern British Columbia
BC, Canada

Email: [Link]@[Link]
 Workshop Overview

• Why mixed methods analysis?

• Stages of analysis
• Integration strategies
• Joint displays
• Meta-inferences
• Discussion
 What is Mixed Methods
Data Analysis?
• Mixed methods analysis is more
than analysing two datasets

It involves:
• Quantitative analysis
• Qualitative analysis
• Integration
• Interpretation
• Meta-inferences
 Key decisions for deciding on the type of mixed methods design to use (Creswell & Inoue(2024)
What is your “intent” for collecting both What are your “procedures” for linking the Mixed methods design
quantitative and qualitative data? quantitative and qualitative databases? recommended
• To see if results are similar or validated • Merging or bringing the two database • Convergent design
when quantitative and qualitative data are results together
collected
• To explain quantitative results in more • Using qualitative data and analysis to • Explanatory sequential design
depth with qualitative data explain quantitative results
• To explore before administering quantitative • To gather qualitative data and analyze it, • Exploratory sequential design
instruments to make them adapted for a then use the results to design or modify a
sample or population quantitative assessment before
administering the assessment
• To gain personal perspectives by adding • To collect qualitative data at different • A mixed methods intervention
qualitative data into a quantitative stages of the experiment (before, during, trial or experiment (hybrid
intervention trial or experiment and after) design)
• To make a complete evaluation of a program • To collect both quantitative and • Mixed methods evaluation study
by examining personal qualitative qualitative data during the evaluation (hybrid design)
experiences with the program as well as process at different stages
quantitative outcomes
• To understand the uptake of an intervention • Add both quantitative and qualitative • A mixed method intervention
by examining personal experiences and data into an intervention trial study (hybrid design)
outcomes in an implementation science
study
• To conduct a community participatory • To add quantitative and qualitative data • A mixed methods participatory
action study involving key participants and into one or more of the stages of action study
gathering both qualitative personal conducting a participatory action research
 1. Analyse quantitative data (What
happened?)

The Four Stages of Mixed 2. Analyse qualitative data (Why did

it happen?)
Methods Data Analysis
3. Integrate findings (How do they fit
together?)

4. Develop meta-inferences (What

new understanding emerges?)
 Choosing an
Analytical Strategy
 • Purpose: Identify patterns, relationships and effects

Quantitative
Analysis (Example) • Examples:
– Descriptive statistics
– t-tests
– ANOVA
– Regression
– SEM
 • Purpose:
Qualitative – Explain experiences, meanings and
processes
Analysis
• Examples
– Thematic analysis
− General Inductive Method
− Content analysis
− Grounded theory
Five Common Integration Strategies
Merging

Connecting

Building

Embedding

Transforming qualitative data into quantitative variables

 Some Examples
Strategy What it means Example

A study surveys 500 students about AI readiness (quantitative) and

Analyze qualitative and quantitative data separately, interviews 20 students (qualitative). Survey results show 78% feel
1. Merging then bring the results together to compare, confirm, or confident using AI, while interviews explain that confidence comes
explain findings. from prior exposure and peer support. The findings are merged in
the discussion.

A survey identifies students with high, medium, and low AI

Use the results from one phase to inform the next phase confidence. Researchers then select participants from each group for
2. Connecting
of the study. interviews to explore why their confidence differs. The quantitative
results connect to the qualitative sampling.
Researchers conduct interviews with faculty to identify key AI
Use findings from one method to develop the readiness factors. These themes are then used to develop a
3. Building
instruments or procedures for the next method. questionnaire that is distributed to a large sample of academics. The
qualitative findings build the quantitative instrument.

An experimental study evaluates an AI training programme using

One form of data plays a supportive role within a study pre- and post-tests (quantitative). Short interviews are embedded to
4. Embedding
that is primarily qualitative or quantitative. understand participants' experiences and explain why some
improved more than others.

Interview responses are coded into categories such as "High AI

readiness," "Moderate readiness," and "Low readiness." These codes
5. Transforming qualitative data into Convert qualitative data into numerical codes so that
are assigned values (3, 2, 1) and analysed using descriptive statistics
quantitative variables statistical analyses can be performed.
or regression. Another example is counting the frequency of themes
(e.g., "governance" mentioned by 85% of participants).
Joint Displays
Example of a template for a joint display for a convergent
design(Creswell & Inoue, 2024)
 Example: Convergent Design (Merging)
Research Question: How ready are universities to adopt AI?

Quantitative Findings Qualitative Findings Integrated Interpretation

Experts emphasised leadership Both datasets confirm that

88% agreed that leadership is
commitment as the key driver of strategic leadership is
critical (Mean = 4.6)
institutional change. fundamental to AI readiness.

Quantitative and qualitative

Experts noted inconsistent staff
AI literacy received lower findings converge, suggesting AI
capability and limited
agreement (Mean = 3.5) literacy remains an area for
professional development.
improvement.

Participants highlighted clear Governance emerged as a core

Governance scored highly (Mean =
policies and accountability institutional capability across both
4.4)
structures. datasets.
Example of a template for an explanatory sequential design joint
display(Creswell & Inoue, 2024)
Example of a template for a joint display for a convergent design
with meta-inferences added(Creswell & Inoue, 2024)
 Developing Meta-Inferences
• Meta-inferences combine evidence from both datasets.

Example:

– Quant: AI improved engagement.

– Qual: Students valued immediate feedback.

– Meta-inference: AI enhances engagement through adaptive feedback and self-

regulation.
 Worked Example: Healthcare

Research Question: Does Quantitative: HbA1c, attendance Qualitative: Patient interviews Meta-inference: Better outcomes
telehealth improve diabetes and medication adherence. and clinician focus groups. arise from continuity of care and
management? patient confidence.
Reflect on your student or your own work and discuss
how you can approach analysis of the data
Example: Education

RQ: Does AI feedback Quant: Grades, engagement,

improve learning? SRL survey.

Meta-inference: AI improves
Qual: Interviews, AI logs. learning through adaptive
feedback and reflection.
 Example: Medicine

RQ: Does telehealth Quant: HbA1c, Qual: Patient Integrated finding:

improve diabetes care? attendance. interviews. Continuity of care
explains improved
outcomes.
 RQ: How does AI affect decision making?

Example: Quant: Productivity survey.

Business Qual: Executive interviews.

Integration shows leadership support

drives adoption.
 Example: Psychology

• RQ: Does mindfulness reduce anxiety?

• Quant: Anxiety scale.
• Qual: Interviews.
• Integration explains coping strategies behind
score changes.
 Example:
Engineering
Open Discussion
 Common Pitfalls

▪ What to Avoid or Minimise:

– Reporting datasets separately
– Not integrating the dataset
– Using qualitative data only as quotes
– Ignoring conflicting findings
– Presenting weak or one-sided conclusions
 Key Message
• Integration is the hallmark of mixed
methods

• Joint displays strengthen reporting

• Meta-inferences create new

knowledge

• Mixed methods answers what, why

and how.
 Some Key Readings
• Creswell, J. W., & Inoue, M. (2024). A process for conducting mixed methods data
analysis. Journal of general and family medicine, 26(1), 4–11.
• Bryman A. Integrating quantitative and qualitative research: how is it done? Qual Res.
2006;6(1):97–113.
• Fetters MD, Curry LA, Creswell JW. Achieving integration in mixed methods designs—
principles and practices. Health Serv Res. 2013;48(6):2134–2156.
• Younas, A., Fàbregues, S., Munce, S., & Creswell, J. W. (2025). Framework for types of
metainferences in mixed methods research. BMC medical research methodology, 25(1), 18.