Mixed Methods Use in Humanitarian Program Impact ... · Mixed Methods Project: An...
Transcript of Mixed Methods Use in Humanitarian Program Impact ... · Mixed Methods Project: An...
Keynote Address
New Directions in Advancing the Methodology of Mixed Methods Research
- John W. Creswell, Ph.D. -
Professor of family medicine and co-director of the Michigan Mixed
Methods Program at the University of Michigan
New Directions in Advancing the
Methodology of Mixed Methods
Research
John W. Creswell, PhD
Senior Research Scientist, Michigan Mixed Methods
Research Program
University of Michigan, Ann Arbor, Michigan
MQIC Conference, Berlin, Germany, February 26 - 28, 2020
Mixed Methods Program
University of Michigan
Institute for Southeast
Asian Qualitative and
Mixed Methods Research,
Osaka and Honolulu
Take home points:
By the end of this presentation,
you will have:
✓ A basic understanding of mixed methods research
✓ 9 procedures to include in your mixed methods study
The reasons for these procedures
Specific examples of these procedures
Creswell’s Books: Research Design, Qualitative
Research, and Mixed Methods Research
2015 lst Edition
2019 6th Edition 2016 lst Edition
2018 3rd Edition 2018 4th Edition
2018 5th Edition
General Research Design
Qualitative
Research
Mixed Methods Research
Honolulu, US Osaka, Japan
China
Germany
Workshops/Speeches in Last Year
A Basic
Understanding
Of Mixed
Methods
Research
1980 1990 2000 2010
Founders of
Mixed Methods
Arrive (Different
Disciplines/Countries-
Books (1985-90)
Mixed Methods
Developed as
Methodology
(1998)
Mixed Methods
Handbook (SAGE)
(2003)
Journal of
Mixed Methods
Research (2007)-
New books on
Mixed Methods
2011 2013 2014
“Best Practices”
Of Mixed Methods
In Health Sciences
(NIH)
Mixed Methods
International Research
Association (MMIRA) (2013)
World-Wide
Expansion of
Mixed Methods
(Regional MMIRA
Conferences)
2015
NIH Mixed
Methods
US training
In health
sciences
A Research Methodology Developed in the Last
30 Years
APA
Standards
on Mixed
Methods
2018
Books Journal US funding International
AssociationTraining
Program
Standards
The Three Key Ideas to Understand Mixed Methods Research
1. We will integrate or mix the data
2. We use mixed methods designs to integrate the data
3. From this integration we obtain insights or
meta-inferences beyond what we learn from
either our quantitative and qualitative data
9 Procedures
To Include in
Your Mixed
Methods Study
When conducting a mixed methods study, map
onto your content the latest mixed methods
procedures or advances
▪ You will build innovative features into your project
▪ You will signal to reviewers that you know the latest
innovations in the field
▪ You will have stronger projects for submission
Why?
1. Identify and label your mixed methods design
2. Include a diagram of your mixed methods design
3. Consider using an implementation matrix
4. Reinforce integration
5. Develop a data table
6. Use rigorous procedures
7. Create a joint display to show integration
8. Include a mixed methods study aim
9. Add a conceptual framework into your design
What are these Advances?
The content of this presentation was developed in collaboration
with the University of Michigan Mixed Methods Program
Executive Committee
Jane Forman
Arden Morris
Sara Warber Program
Coordinator
Satoko Motohara
Research Fellow
Melissa
DeJonckheere
Program Staff
Rae
Sakakibara
Senior Research
Scientist
John W. Creswell
MMP Visiting Scholar
Sergi FàbreguesDirector and Assoc
Director
Michael D. Fetters
Timothy C. Guetterman
Michigan Mixed Methods Program, University of Michigan, 2019
1. Identify and Label Your
Mixed Methods Design
Designs are procedures for
conducting your research. We
have them in quantitative and qualitative
research. Why not have designs in
mixed methods research?
Two Categories of Mixed Methods Designs: Core and
Complex (Creswell & Plano Clark, 2018, 3rdEdition)
Convergent Design
Explanatory
Sequential Design
Exploratory Sequential
Design
Intervention Trial
Multiple Case Study
Program Evaluation
CBPR Participatory Studies
Embedded
Within
Chapter 3: CORE MIXED
METHODS DESIGNS
CHAPTER 4:
COMPLEX
DESIGNS THAT
INCLUDE CORE
DESIGNS
Merge
Quantitative
Data Collection
and Analysis
Qualitative
Data Collection
and Analysis
A Convergent Design
Phase 1
Intent: To compare
results from both
databasesInterpretation
1
Merge
Quantitative
Data Collection
and Analysis
Qualitative
Data Collection
and Analysis
A Convergent Transformative Design
Phase 1
Intent: To develop
a more comprehensive
quantitative assessmentInterpretation
Quantitative Data
Transformation
2
Quantitative
Data Collection
and Analysis
Qualitative
Data Collection
and Analysis
Explained
byInterpretation
Explanatory Sequential Design
Phase 1 Phase 2
Intent: Use the qualitative data to help explain
the quantitative results
3
20
Qualitative
Data Collection
and Analysis
Quantitative
Phase –
Instrument design,
new variables or
intervention activities
Builds into
Interpretation
An Exploratory Sequential Design
Quantitative
Phase –
Instrument design
or intervention design
Phase 1 Phase 2
Phase 3 Intent: To explore first before
building into and testing a quantitative phase
4
How Do You Decide What Core Design to Use?
✓ How do you plan to “mix” the two databases?
✓ Merge or combine them?
✓ Connect them?
✓ If you plan on connecting them, what goes first, quantitative or
qualitative?
✓ Do your research experiences (and field) lean toward quantitative,
toward qualitative or both? (this might influence what you start with)
✓ When mixed methods is used in your field, what are the popular
core designs used?
How Do You Decide Whether You Have a Core
Design or a Complex Design?
✓ Am I simply integrating qualitative
and quantitative data in my project +Qual Quan
+Qual Quan
A Process of StepsOR
✓ Am I integrating qualitative and
quantitative data AND using this
integration within a larger
process of steps of research?
ASK YOURSELF:
Intervention Mixed Methods Design
Qualitative
Interviews
Before Experiment
(Exploratory
Sequential Design)
Experiment with an
Intervention and Pre- and
Post-Test Measures
Qualitative Interviews
During Experiment
(Convergent Design)
Qualitative
Interviews
After Experiment
(Explanatory
Sequential Design)
• Recruit participants
• Develop workable interventions
• Develop good pre-post-test
measures
• Examine participants’
experiences
• Modify treatment
• Explain outcomes
• Modify experiment
Complex Design:
A Mixed Methods Intervention (Experimental) Design Diagram
Integration Integration
Integration
2016
• N=20 Program Part.
• N=4 Experimental/
N=4 Wait List
Pre-Post-test Measures:
GHQ12/PANAS/FIQ/
Social validity
Phase4
Final InterventionTrial
QualitativeFollowup toTrial
• N=6Qual data:. Narrative Data
• TEA/TLMG
Phase3
2014-15
2014-
15
Phase2
PreliminaryIntervention Trial
Single group (quasi-
experimental design)
N=44Quan data:
Pre- Post-test Measures :
GHQ12/PANAS/FIQ/Social Validity
Qual data: Homework
Needs AssessmentN=58Qual data:Text DataKJ methods
Phase1
2013-
15
Exploratory Sequential Design
Explanatory Sequential Design
Exploratory Sequential Design
Integration Integration Integration
A Mixed Methods Intervention Design with Sequential Components to Develop and Test a Behavioral Program Intervention (Hirose
Creswell dissertation, Japan)
Overall Explanatory Sequential Design
= Intent to Develop and Test the Final Program Intervention
Interpretation
Needs
Assessment
Qualitative Stage
• Interviews
• Observations
• Documents
Theory Conceptualization
Specific to Setting
Quantitative stage
• Literature review stage
Instrument and Measures
Development
Quantitative stage
• Measures and
instruments
Program Implementation and
test
Quantitative stage
• Experimental intervention
based on quantitative
measures
Program Follow up and Refinement
Qualitative stage
• Interviews
• Observations
• Documents
Exploratory Sequential Design
Explanatory Sequential Design
A MIXED METHODS EVALUATION COMPLEX DESIGN
Names for the Complex Designs
▪ “Mixed methods intervention design with an exploratory
sequential core component” (or convergent, or explanatory,
or some combination)
▪ “Mixed methods comparative case study design with a
core convergent component”
▪ “Mixed methods evaluation design with both convergent and
sequential core components”
Complex Process
Core
Design
2. Include a diagram of
your mixed
methods design
How can you simplify
the complex procedures
of mixed methods so that
your design can be understood
by a reader?
Aspects of a Diagram to Consider:
❑ A diagram is a visual tool to help understand your project
(much like a theoretical model, or consort sampling figure)
❑ A diagram has many applications
❑ A diagram does not need to be overly complex – keep it simple
❑ A diagram needs to relate to your study aims
❑ A diagram needs to showcase your major procedures
❑ A diagram builds on and simplifies complex
procedures
A diagram is a figure inserted into the “Approach” section or
appended to the project that conveys the major procedures
of your mixed methods design.
• Data Collection
• Data Analysis
QualitativeQuantitative
Interpretation
Merging Data
Aim 1. Explain determinants
of aging through a survey.
Aim 3. Develop a complete
understanding of aging
Aim 2. Exploring personal
experiences with aging
Figure 1. Explaining Aging:
A Convergent Mixed Methods Design
How to Draw a Diagram of Your Convergent Design
Timeline
Start
Oct ‘17
End
Oct. ‘18
2 mos.
4 mos.
6 mos.
8 mos.
10 mos.
12 mos.
• Data Collection
• Data Analysis
• Data Collection
• Data Analysis
Quantitative Qualitative
• Data Collection
• Data Analysis
Connecting Interpreting
How Qual
Explains
the Quan
Aim 1. Explain
determinants of aging
through a survey.
Aim 3. Interpret why
social isolation does not
impact aging through
interviews.
Aim 2. Explain surprising
results in survey (e.g.,
family, social isolation,
etc.
Figure 2. Explaining Determinants of Aging:
An Explanatory Sequential Mixed Methods Design
How to Draw a Diagram of Your Explanatory Sequential Design
Timeline
11 mos.7 mos. 9 mos.5 mos. End
Nov. 2018
Start
Oct. 2017
Source: Ivankova, N. V. (2015). Mixed Methods Applications in Action Research. Los
Angeles, CA: Sage, p. 74
Ivankova (2015) diagram: Glason, J. Change, Chenoweth, et al. (2006) Study of Nursing Care
for Older Patients in an Acute Medical Ward, Journal of Clinical Nursing
Collecting and
analyzing
quantitative
and qualitative
data
Collecting and
analyzing
quantitative
and qualitative
data
If You Have Complex Design: Create Diagram of
Complex Procedures and Embed Your Core
Designs in it
An Example of a Mixed Methods Community-
Based Participatory Research (CBPR) Study
Core
MM
Design
Core
MM
Design
3. Consider using an implementation matrix
What if you had in your project a figure that related the
major parts of your study to provide an overview?
Study
Aims
Data
Collection
Data
AnalysisOutcomes Integration
Points
A FIGURE OR MATRIX
Aspects of an Implementation Matrix to
Consider:
❑ This table is an alternative to a mixed methods diagram
(entire project versus methods)
❑ Existing implementation matrices often do not contain all of
essential elements of mixed methods research
❑ This table provides a helpful way for a reviewer to easily
see the major elements of a study.
❑ An implementation matrix is popular in the health sciences.
An implementation matrix is a table that summarizes the major
elements in a project in an “Approach” section
Source: Creswell, J. W., Klassen, A. C., Plano Clark, V. L., & Smith, K. C. for the Office of
Behavioral and Social Sciences Research. (2011, August). Best practices for mixed
methods research in the health sciences. Washington, DC: National Institutes of Health.
Retrieved from: https://obssr.od.nih.gov/training/mixed-methods-research/
Source: Baez, J. C. Robert Wood Johnson Fellow
Project on Unidenitified Immigrants. 2018.
Example of an Implementation Matrix
Source: Justine Wu, MD, MPH, Assistant Professor, Family Medicine
and Obstetrics and Gynecology, University of Michigan, presented at
Designing a Mixed Methods Project: An Interactive-Participatory
Workshop, November 9-11, 2016.
Example of an Implementation Matrix
Source: Justine Wu, MD, MPH, Assistant Professor, Family Medicine and
Obstetrics and Gynecology, University of Michigan, presented at Designing a
Mixed Methods Project: An Interactive-Participatory Workshop, November 9-
11, 2016.
An Example of a Filled in Implementation Matrix
4. Reinforce integration throughout your
study
If integration is a centerpiece
for conducting mixed methods
research, why not convey its
importance several times in
your project?
How do we bake bread?
We IntegrateThe Ingredients
Source: Hirose, M.
We can integrate (or connect) the raisins into
the bread batter – the raisins stay intact
✓ Qual and quan stay separate but are
connected
We can integrate (or merge) the
flour into the batter, and the
flour dissolves
✓ Qual and quan merge into each other
We can include all of
the ingredients into the bread
✓ Embed both quan
and qual into the bread
Aspects of Integration to Consider:
❑ Integration is the centerpiece of a good mixed methods
study.
❑ Rigorous mixed methods research requires the identification
of where integration occurs in a study and the form it takes
❑ Consider stating and restating integration in several ways in
your project (e.g., title, study aim, justification, diagram)
❑ Integration should relate to your type of mixed methods
design
❑ Integration ”scripts” can help to convey your integration
Integration is the point in the research process in which the
quantitative data and the qualitative data meet. From this
integration additional insight (meta-inferences) results beyond the quantitative
and qualitative data.
Type of Design Where Integration Occurs
Convergent Design When quantitative and qualitative
data results are matched (compare)
Explanatory Sequential Design When quantitative results lead to the
qualitative data collection follow up
(explain)
Exploratory Sequential Design When the qualitative exploration
leads to the quantitative feature (e.g.,
instrument, etc.) (explore)
Intervention Trial Design When qualitative data is added into
an intervention trial (embedded
before, during, after)
Case Study Design When the quantitative and qualitative
results are compared to yield cases
(multiple convergence)
Social Justice/Participatory Design When the qualitative and quantitative
data are combined in different phases
of the research (multiple
convergence)
Evaluation Design When the qualitative and quantitative
data are combined in different stage
of the evaluation project (multiple
convergence)
Integration by Type of Design
Integration in an Explanatory Sequential Design
Collect/Analyze
Quantitative DataCollect/Analyze
Qualitative Data
• Collect N=250 surveys
• Measure behaviors leading
to lowered cholesterol
• Analyze data descriptively
and multivariately
• Collect N=30
interviews with
patients based
on surprising
quantitative results
• Code data and
identify themes
Explanatory Sequential Mixed Methods
Design
Interpret how
qualitative results help
explain quantitative results
Theory of
Behavioral Change
Integration
Phase 1
Phase 2
Conceptual Framework
or Theory
Scripts for Mixed Methods Integration Statements
▪ “Integration involved merging the results from the quantitative and qualitative
data so that a comparison could be made, and a more complete
understanding emerge than what was provided by the quantitative or the
qualitative results alone.” (Convergent Design)
▪ “Integration involved connecting the results from the initial quantitative phase
to help plan and explain the follow up qualitative data collection phase. This
plan would include what questions need to be further probed and what
individuals can help best explain the quantitative results.” (Explanatory
Sequential Design)
▪ “Integration involved gathering initial qualitative data, analyzing it, and then
using the qualitative results to build a new intervention (or measure or
instrument) that can be cultural specific that will be tested quantitatively with
a large sample.” (Exploratory Sequential Design)
5. Develop a data table
If you are going to integrate
your data, wouldn’t you want
to first determine what
constitutes your quantitative
and qualitative data before
you integrate the data?
Aspects of data to consider:
❑ Clearly separate your quantitative and qualitative data
types (because you will need to be specific about integrating
them)
❑ Qualitative data consists of open-ended information;
quantitative data, close-ended information
❑ In preparation for integration and many parts of mixed
methods, labeling your quantitative and qualitative data is
helpful.
❑ Create a data table to clearly identify your types of data.
The data in a mixed methods study consists of collecting and
analyzing both quantitative (closed-ended) and qualitative (open-ended)
data, and interpreting their integration.
Instrument dataInstruments
Observational checklists
Document data
Qualitative Data (Open –Ended)
Interviews
Observations
Documents
Visual – Social Media Data
Quantitative Data
(Closed-Ended)
Types of Data
Data Source
Table
Based on:
Wittink MN, Barg FK,
Gallo JJ. (2006)
Unwritten rules of talking
to doctors about
depression: integrating
qualitative
andquantitative
methods. Ann Fam Med.
Jul – Aug; 4(4):302-9.
6. Use rigorous qualitative and
quantitative procedures
Just because you are
collecting and analyzing
both quantitative and
qualitative data, you still
need to be rigorous in both
approaches.
US Professional Association Developing Standards:APA Qualitative Article Reporting Standards (QARS)
New Edition:October, 2019
Task Force Developed Standards:
✓Qualitative Research✓Qualitative Meta-
Synthesis Research✓Mixed Methods Research
7. Create a Joint Display
To analyze data in mixed methods
research, you have (1) quantitative
data analysis (2) qualitative data
analysis and (3) mixed methods
integration data analysis. How do
you show integration in your
results?
Qualitative Results
Quantitative
Results
Interpretation
(Meta-Inferences)
A joint display is a table (or
graph) that arrays the
quantitative data with the
qualitative data so that an
interpretation can be formed
across the two databases
Scores
High
Medium
Low
A Joint Display Representing Integration Mixed
Methods Data Analysis in a Convergent Design
Theme 1 Theme 2 Theme 3 Theme 4
Quote Quote Quote Quote
Quote Quote Quote Quote
Quote Quote Quote Quote
An Example of a Joint Display for a Convergent Design
8. Include a Mixed Methods Study Aim
When you integrate, what question (or aim)
are you attempting to answer with
your integration procedures?
❑ It is a type of study aim that is little known.
❑ Follow the content then method approach
❑ For the quantitative aim, think in terms of variables and
their relationship
❑ For the qualitative aim, think about the central phenomenon
❑ For the mixed methods aim, think about your integration and what you will accomplish with the integration
❑ Resist attempts by others to reduce the number of your aims
❑ Order your aims to fit your design (could be qual first, or quan first, with mixed methods sliced in between)
Aspects of mixed methods aim to
consider:A mixed methods study aim is the aim in the project that captures
the integration of the quantitative and qualitative databases. As an
aim, it needs to be answered through data analysis in the study.
Type of Design Study Aim for the Mixed
Methods Design
Convergent Design The aim conveys a comparison of the
quantitative and qualitative results
Explanatory Sequential Design The aims describes how qualitative
results help to explain the quantitative
results
Exploratory Sequential Design The aim discusses how the qualitative
exploration leads to the quantitative
feature (e.g., instrument, etc.)
Intervention Trial Design The aim describes how the qualitative
data augments the intervention trial
(before, during, after)
Case Study Design The aims describes how the
comparative cases will be developed
based on quantitative and qualitative
results.
Social Justice/Participatory Design The aim indicates how the qualitative
and quantitative data will be integrated
into the social justice/participatory
approach.
Evaluation Design The aim identifies how the qualitative
and quantitative data will be integrated
into the evaluation design.
Mixed Methods Study Aim relates to your Design
Sample Mixed Methods Study Aims
▪ To compare the qualitative interviews about the elderly’s self-esteem with
their perceptions on a survey about their self-esteem. (Convergent Design)
▪ To explain the results from the survey about self-esteem with the elderly with
qualitative focus groups. (Explanatory Sequential Design)
▪ To explore the meaning of self-esteem for the elderly, to develop a survey
based on their qualitative interviews, and then to administer the contextually-
specific survey. (Exploratory Sequential Design)
Content Order for the Study Aims
▪ Template =
Content followed by methods (Best Practices, 2011)
Example (Explanatory Sequential Design Example):
1. Relate personal isolation to depression among older adults based on
survey data. (QUAN aim)
2. Explain how personal isolation affects depression among older adults by
obtaining personal experiences. (MIXED METHODS aim)
3. Explore personal isolation among older adults using semi-structured
interviews. (QUAL aim)
4. Develop an intervention (experiment) to compare different groups of
isolated older adults (FUTURE RESEARCH aim)
9. Add a Conceptual Framework into Your Design
What explains what you hope to
find? (based on the literature)
Aspects of a Conceptual Framework to
Consider:
▪ Preferably select one overarching framework
▪ Consider how it relates to your quan/qual/mixed procedures
▪ Point to where in your diagram (if used), the framework is being used
▪ Keep referring to the framework if possible in different sections of your project
▪ Many conceptual frameworks are social science frameworks – find a popular one
▪ Draw a visual of your framework and insert it into your project
▪ Realize that this could be a framework or a theory
▪ Realize that it could be inductively developed (as in qualitative research) or deductively used (as in quantitative research)
▪ Put an arrow from the label of your conceptual framework into your
diagram where it fits best
A conceptual framework in a mixed methods study provides
a priori explanation as to what you hope to find in your
quantitative, qualitative, and mixed methods analysis.
Place your Conceptual Framework into Your Design Diagram
Collect/analyze
Quantitative DataCollect/analyze
Qualitative Data
• Collect N=250 surveys
• Measure behaviors leading
to lowered cholesterol
• Analyze data descriptively
and multivariately
• Collect N=30
interviews with
patients based
on surprising
quantitative results
• Code data and
identify themes
Explanatory Sequential Mixed Methods Design
Interpret how
qualitative results help
explain quantitative results
Theory of
Behavioral ChangeIntegration
Phase 1
Phase 2
The Conceptual Framework can Help Explain Both the Quantitative
and Qualitative Sections of Your Mixed Methods Study
Source: Betancourt, T. S., Meyers-Ohki, S. E., Stevenson, A., et al. (2011). Using Mixed-Methods
Research to Adapt and Evaluate a Family Strengthening Intervention in Rwanda. African Journal of
Traumatic Stress, 2(1), p. 38.
1. Identify and label your mixed methods design
2. Include a diagram of your mixed methods design
3. Consider using an implementation matrix
4. Reinforce integration
5. Develop a data table
6. Use rigorous procedures
7. Create a joint display to show integration
8. Include a mixed methods study aim
9. Add a conceptual framework into your design
What are these Advances?
New Directions in Advancing the
Methodology of Mixed Methods
Research
John W. Creswell, PhD
Senior Research Scientist, Michigan Mixed Methods
Research Program
University of Michigan, Ann Arbor, Michigan
MQIC Conference, Berlin, Germany, February 26-28, 2020
Mixed Methods Program
University of Michigan
Institute for Southeast
Asian Qualitative and
Mixed Methods Research,
Osaka and Honolulu