Research Design: Qualitative, Quantitative, and Mixed Methods Approaches

Research Design: Qualitative, Quantitative, and Mixed Methods Approaches by John W. Creswell is a comprehensive guide for researchers to understand the mixed methods approach. This book is useful for researchers in diverse fields such as occupational therapy, interpersonal communication, AIDS prevention, dementia caregiving, mental health, and middle-school science.

THIRD EDITION
RESEARCH
DESIGN
Qualitative, Quantitative,
and Mixed Methods Approaches
JOHN W. CRESWELL
UNIVERSITY OF NEBRASKA-LlNCOLN
~SAGE
Los Angeles • London • New Delhi • Singapore

Copyright © 2009 by SAGEPublicatíons, Inc.
Al! rights reserved. No part of this book may be reproduced or utilized in any form or
by any means, electronic or mechanícal, including photocopying. recording. or by
any information storage and retrieval system, without permission in writing from
the publisher.
Por iniormation:
SAGEPublications. Inc.
2455 Tel!er Road
Thousand Oaks, California 91320
E-mail: [e-mail]
SAGEPublications India Pvt.

Brief Contents
Analytic Contents of Research Techniques
Preface
Acknowledgments
About the Author
~ Part 1:Preliminary Considerations
1. The Selection of a Research Desiqn
2. Review of the Literature
3. The Useof Theory
4. Writing Strategies and Ethical Considerations
xv
xix
xxvii
xxix
~ Part 11.Designing Research
5. The Introduction
6. The Purpose storernent
7. Research Questions and Hypotheses
8. Quantitative Methods
9. Qualitative Procedures
10. Mixed Methods Procedures
3
23
49
73
95
97
111
129
145
173
203
Glossary
References
Author Index
Subject Index
. 227
237
247
251

~ CHAPTER TEN
Mixed Methods
Procedures
W ith the development and perceived legitimacy of both qual-
itative and quantitative research in the social and human
sciences, mixed methods research, employing the combina-
tion of quantitative and qualitative approaches, has gained popularity.
This popularity is beca use resear

204 Designing Research
Mixed Methods in the Social & Behavior Sciences (Tashakk:ori & Teddlíe, 2003)
was publíshed, providing the fírst comprehensive overview of this strategy of
inquiry. Now several journals emphasize mixed methods research. such as
the Joumal 01 Mixed Methods Beseardi; Quality and Quantity. and Field Methods.
while numerous others activelyencourage this form of inquiry (e.g.. Iniemaiional
[ournal 01 Social Research Methodology. Qualitative Health Research. Annals 01
Family Medicine). Numerous published research studies have incorporated
mixed methods research in the social and human sciences in diverse fields
such as occupational therapy (Lysack & Kreftíng, 1994). interpersonal com-
munication (Boneva, Kraut, & Frohlích, 2001). AIDS prevention (Ianz et al ..
1996), dementia caregiving (Weitzman & Levkoff. 2000). mental health
(Rogers, Day. Randall. & Bentall, 2003). and in míddle-school science (Houtz,
1995). New books arrive each year solely devoted to mixed methods research
(Bryman, 2006; Creswell & Plano Clark, 2007; Greene, 2007; Plano Clark

Mixed Methods Procedures 205
;..i{::;r; : ~\-:/f-:'".:i~{='~ ~}!'!_..-1~"i;'" -::~.;f,\~..l'/." ' 'Jl~" ,',;' ~.:."1-•• <; r : .t:> ,; ..-:;,1 '". :-':.~, ": ~. -', .
.::'fáble; 10.1 - -

".!
. j
J,.
d'
i,
~f.
\,
I
1:,
206 Designing Research
TYPES OF MIXEO METHOOS STRATEGIESANO VISUAL MOOElS
There have been several typologies for classifying and identifying types of
rnixed methods strategies that proposal developers might use in their pro-
posed mixed methods study. Creswell and Plano Clark (2007) identify 12
classification

Mixed Methods Procedures 207
Timing Weighting Mixing Theorizing
No Sequence Equal Integrating
concurrent Explicit
Sequential-
Oualitative first Oualitative Connecting
Sequential- Implicit
Ouantitative first Ouantitative Embedding
Figure 10.1 Aspects to Cansider in Planning a Mixed Methads Design
SOURCE: Adapted from Creswell et al. (2003).
one or the other. A priority for one type depends on the interests of the
researcher, the audience for the study (e.g., faculty commíttee, professional
assocíatíon), and what the investigator seeks to emphasize in the study.
In practical terms, weight occurs in a rnixed methods study through such
strategies as whether quantitative or qualitative information ís emphasized
fírst. the extent of treatment of one type of data or the other in the project,
or the use of primarily an inductive approach (i.e., generating themes in
qualitative) or a deductive approach (i.e., testing a theory). Sometímes the
researcher intentionally uses one form of data in a supportive role to a larger
study, as ís found in some experimental trials (see Rogers et al., 2003).
Mixing
Mixing the data (and in a larger sense, mixing the research questíons,
phílosophy, the interpretation) is difficult at best when one considers that
qualitative data consists of text and images and quantitative data, num-
bers. There are two different questions here: When does a researcher rnix
in a mixed methods study? And how does mixing occur? The first question
" .
is much easier to answer than the second. Mixing of the two types of data
míght occur at several stages: the data collectíon, the data analysís, in ter-
pretatíon, or at all three phases. Por proposal developers using mixed
methods, it is important to discuss and present in a proposal when the mix-
íng will occur.
How the data are mixed has received considerable recent attention
(Creswell & Plano Clark, 2007). Mixing means eíther that the qualitative

208 Designing Research
and quantitative data are actually merged on one end of the contínuum,
kept separate on the other end of the continuum, or combined in some way
between these two extremes. The two data bases might be kept separate but
connected; for example, in a two-phase project that begins with a quantita-
tive phase, the analysis of the data and its results can be used to identify
participants for qualitative data collection in a follow-up phase. In this sítu-
atíon, the quantitative and qualitative data areconnectedjluring the phases
of research.Connectedl in mixed methods research means a mixing of the
quantitative and qualitative research are.connected between a data .analy-
gs of the fírst phase of research and the data c<?l!ec~~nof the second phase
of research. In another study, the researcher might collect both quantíta-
tive and qualitative dataconcurrently' and 4:!t~g!:i!t~o!;'!ile.xgt::\thetwo data-
bases by transforming the qualitative themes into counts and comparíng
these counts with descriptive quantitative data. In this case, the mixing con-
sists of íntegratíng the two databases by actually merging the quantitative
data with th~ qiI~tative data. In a final scenarío, the researcher might have
a primary aim to coIlect one form of data (say quantitative) and have the
other form of data (say quaIitative) provide supportiveinformation. Neither
integrating the data nor connecting across phases is being utilized. Instead,
the researcher ísembedding a secondary form of data within a larger
study having a different form of data as the primary database. The sec-
ondary database provídes a supporting role in the study.
Theorizing or Transforming Perspectives
A final factor to consider is whether a larger, theoreticaI perspective
guides the entire designo It may be a theory from t~e social sciences (e.g.,
adoption theory,leadership theory, attribution theory) or a broad theoret-
ícal lens, such as an advocacy/participatory lens (e.g., gender, race, class;
see Chapter 3 ).i\.Il researchers bríngtheoríes, frªmewQrks_éip.dhunches to ..
theirJp._<iuiries, and these theories may be made explicit in a mixed meth-
ods study or be implicit and not mentioned. We wiIl focus here on the use
of explicit theories. In mixed methods studíes. the theories are found typí-
cally in the beginning sections as an orienting lens that shapes the types of
questions asked, who participates in the study, how data are collected, and
the implications made from the study (typicaIly for change and advocacy).
They present an overarching Q¡erspectiveused with aIl of the mixed meth-
ods strategies of inquiry (to be presently discussed). Mertens (2003) pro-
vides a good discussion as to how a transforming lens shapes all phases of
the research process in mixed methods research.
Alternative Strategies and Visual Models
These four factors-timing. weíght, míxíng, and theorizing-help to
shape the procedures of a mixed methods study. Although these do not

1
Mixed Methods Procedures 209
exhaust all the possibilities, there are six major strategies for inquirers
to choose from in designing a research proposal; they are adapted from
Creswell et al. (2003). A proposal would contain a description of the strat-
egy and a visual model of it, as well as basic procedures that the investiga-
tor will use in implementing the strategy. Each strategy is briefly described
and illustrated in Figures 10.2 and 10.3. The words qualitative and quanii-
tative have been shortened in the figures to read "qual" and "quan,"
respectively (see the discussion following the figures).
These mixed methods strategies can be described using notation that
has developed in the mixed methods field. Mixed methods notation
provides shorthand labels and symbols that convey important aspects of
mixed methods research, and it provides a way that mixed methods
researchers can easily communicate their procedures. The following nota-
tion is adapted from Morse (1991), Tashakkorí and Teddlie (1998), and
CresweH and Plano Clark (2007) who suggest the foHowing:
4) A "+" indicates a simultaneous or concurrent form of data collec-
tion, with both quantitative and qualitative data. collected at same time.
• A" -7" indicates a sequentiaI form of data collectíon, with one form
(e.g., qualitative data) building on the other (e.g., quantitative data).
Sequential Explanatory Design (a)
1 QUAN 1--1 qua!
QUAN QUAN qua! qua! I . f
Data -- Data -- Data -- O t ntetptetetion o
aa -- E . Anelvsi
Collection Analysis Collection Al. ntire na ysts
naysls
Sequential Exploratory Design (b)
QUAL
Data --
Collection
1QUAL 1--1 quan
QUAL quan quan Interpretation of
Data -- Data -- Data --Analysis Collection Analysis Entire Analysis
Sequential Transformative Design (e)
QUAL -- quan
Social science theory, qualitative theory, advocacy worldview
,. .
QUAN --> qual
Social science theory, qualitative theory, advocacy worldview
Figure 10.2 Sequential Designs
SOURCE: Adapted from Creswell et al. (2003).
.~.

21O Designing Research
ConeurrentTriangulation Design (a)
QUAN I
QUAN
Data Col/ection
+ QUAL
QUAL
Data Col/ection
~
QUAN
Data Analysis
~
QUAL
Data Analysis
Data Results Compared
Coneurrent Embedded Design (b)
qual quan
QUAN QUAL
Analysis ot Findings Analysis ot Findings
Coneurrent Transformative Design (e)
QUAN + QUAL
Social science theory, qualitative
theory, advocacy worldview
QUAN
Social science theory, qualitative
theory, advocacy worldview
Figure 10.3 Concurrent Designs
SOURCE: Adapted from Creswell et al. (2003).
• Capítalization indicates a weight or priority on the quantitative or
qualitative data, analysis, and interpretation in the study. In a mixed meth-
ods study, the qualitative and quantitative data may be equally empha-
sized, or one may be more emphasized than the other. Capitalization
indicates that an approach or method is emphasized.
• "Quan" and "Qual" stand for quaniiiaiive and qualitative, respectively,
and they use the same number of letters to indicate equality between the
forms of data.
• A QUAN/qual notation indicates that the qualitative methods are
embedded within a quantitative designo
e Boxes highlight the quantitative and qualitative data collection and
analysis.
In addition, incorporated into each figure are specific data collection,
analysis, and interpretation procedures to help the reader understand the
more specific procedures used. In this way, a figure has at least two ele-
ments: the general procedure of mixed methods being used and the more
specific procedures of data collection, analysis, and interpretation.

!
I
J
Mixed Methods Procedures 211
Sequenfial Explanafory Sfrafegy
The sequential explanatory strategy is a popular strategy for mixed
methods design that often appeals to researchers with strong quantitative
leanings. It is characterized by the collection and analysis of quantitative
data in a first ph--ª.seof research followed by the collection and analysis of
qualítative data in a second phase that builds on the results of the initial
quantitative results. Weight typically is given to the quantitative data.
and the mixing of the data occurs when the initial quantitative results
informs the secondary qualitative data collectíon. 'I'hus, the two forms of
data are separate but connected. An explicit theory may or may not
inform the overall procedure. The steps of this strategy are pictured in
Figure 10.2a.
A sequential explanatory design is typically used to explain and inter-
pret quantitative results by collecting and analyzing follow-up qualitative
data. It can be especially useful when unexpected results arise from a
quantitative study (Morse, 1991). In thís case. the qualitative data collec-
tion that follows can be used to examine these surprising results in more
detail. This strategy may or may not have a specific theoretical perspective.
The straightforward nature of this design is one of its main strengths. It is
easy to implement beca use the steps fall into clear, separate stages. In addi-
tíon, thís design feature makes it easy to describe and to report. The main
weakness of this design is the length of time involved in data collection.
with the two separate phases. Thís is especially a drawback if the two
phases are given equal priority.
Sequenfial Exploratory StrQfegy
Thís next strategy is similar to the explanatory sequential approach
except that the phases are reversed. The sequential exploratory strategy
involves a fírst phase of qg~mªtüLe data collection and analysis. followed
by a second phase of quantitative data collection and analysis that builds
on the results of the fírst qualítatíve phase. "Ye,.ightis generally placed on
the Arst phase, and the data are mixed through being connected between
the qualitative data analysis and the quantitative data collection. The
design may or may not be implemented within an explicit theoretical per- .
spective (see Figure 10.2b).
At the most basic Ievel, the purpose of this strategy is to use quantita-
U'le_ data and results to assist in the interpretation Jf qualitativefindings.
Unlike .the sequential explanatory approach. which is better suited to
explaining and interpreting relationships. the primary focus of thís model
is to initially explore a phenomenon. Morgan (1998) suggested that thís
design is appropriate to use when testing elements of an emergent theory
resulting from the qualitative phase and that it can also be used to gener-
alize qualitative findings to different samples. Símllarly; Morse (1991) cited
one purpose for selecting this approach: to determine the distribution of a

212 Designing Research
phenomenon within a chosen population. Finally, the sequential
exploratory strategy is often discussed as the procedure of choice when a
researcher needs to develop an instrument because existing instruments
are inadequate or not available. Using a three-phase approach, the researcher
fírst gathers qualitative data and analyzes it (Phase 1), and uses the analy-
sis to development an instrument (Phase 2) that ís subsequently admínís-
tered to a sample of a population (Phase 3; Creswell & Plano Clark, 2007).
The sequential exploratory strategy has many of the same advantages
as the sequential explanatory model. Its two-phase approach (qualitative
research followed by quantitative research) makes ít easy to implement
and straightforward to describe and report. It ís useful to a researcher who
wants to explore a phenomenon but also wants to expand on the qualita-
tive findings. This model ís especially advantageous when a researcher ís
building a new instrument. In addition, this model could make a largely
qualitative study more palatable to an adviser, cornmittee, or research
community well versed in quantitative research and that may be unfamíl-
iar with the qualitative approaches. As with the sequential explanatory
approach, the sequential exploratory model requíres' a substantiallength
of time to complete both data collection phases, which can be a drawback
for some research situations. In addition, the researcher has to make some
key decisions about which findings from the ínitíal qualitative phase wíll be
focused on in the subsequent quantitative phase (e.g., one theme, compar-
isons among groups, multiple themes).
Sequential Transformative Sfrategy
This final sequential approach has two distinct data collection
phases, one foIlowing the other as in the first two strategies described (see
Figure 10.2c). The sequential transformative strategy is a two-phase
project with a theoreticallens (e.g., gender, race, social science theory)
overlaying the sequential procedures. It too has an initial phase (either
quantitative or qualitative) followed by a second phase (either qualitative
or quantitative) that builds on the earlier phase. The theoretical lens is
introduced in the introduction to a proposal, shapes a directional
research question aimed at exploring a problem (e.g., inequality, díscrím-
ination, injustice), creates sensitivity to collecting data from marginalized
or underrepresented groups, and en<\,>with a call for action. In this design
the researcher may use either method in the first phase of research, and
the weight can be given to either or distributed evenly to both phases. The
mixing is connected as in all sequential designs. Unlike the sequential
exploratory and explanatory approaches, the sequential transformative
model has a theoretical perspective to guide the study. The aim of this the-
oretical perspective, whether it be a conceptual framework. a specific ide-
ology, or advocacy, is more important in guiding the study than the use of
methods alone.
j

J
Mixed Methods Procedures 213
The purpose of a sequential transformative strategy is to best serve the
theoretical perspective of the researcher. By usíng two phases, a sequential
transformative researcher may be able to give voice to diverse perspectives.
to better advocate for participants. or to better understand a phenomenon
or process that is changing as a result of being studied.
The sequential transformative model shares the methodological
strengths and weaknesses of the other two sequential approaches. Its use
of distinct phases facilitates its implementation. description. and sharing
of results, although it also requires the time to complete two data collec-
tion phases. More important. this design places mixed methods research
within a transformative framework. Therefore. this strategy may be more
appealing and acceptable to those researchers already using a transforma-
tive framework within one distinct methodology, such as qualitative
research. Unfortunately. because little has been written to date on this
approach. one weakness is that there is little guidance on how to use the
transformative vision to guide the methods. Líkewíse, as with all sequen-
tíal strategies. key decisions need to be made about what findings in the
first phase will be the focus of the second phase.
Concurrent Triangulation Strategy
The concurrent triangulation approach is probably the most famil-
iar of the six majar mixed methods models (see Figure 10.3a). In a con-
current triangulation approach. the researcher collects both quantitative
and qualitative data concurrently and then compares the two databases to
determine if there is convergence. differences. or some combination. Some
authors refer to thís comparison as confirmation, disconjirmation, cross-
voluiaiion, or corroboraiion (Greene, Caracelli. & Graharn, 1989; Morgan,
1998; Steckler, Mcl.eroy, Goodman. Bird, & McCormick. 1992). This
model generally uses separate quantitative and qualitative methods as
a means to offset the weaknesses inherent within one method with the
strengths of the other (or conversely, the strength of one adds to the
strength of the other). In this approach, the quantitative and qualitative
data collection is concurrent, happening in one phase of the research
study. Ideally, the weight is equal between the two methods. but often in
practíce, priority may be given to one or the other. The mixing during this
approach. usually found in an interpretation or discussion section, is to
actually merge the data (i.e.. transform one type of data to~he other type
of data so that they can easily be compared) or integrate or compare the
results of two databases side by side in a discussion. This síde-by-side
íntegratíon is often seen in published mixed methods studies in which a dis-
cussion section first provides quantitative statistical results followed by
qualitative quotes that support or disconfirm the quantitative results.
This traditional mixed methods model is advantageous because it is famil-
iar to most researchers and can result in well-validated and substantiated

214 Designing Research
fmdings. I find that most researchers when they fírst consider mixed meth-
ods employ this model of gathering both quantitative and qualitative data
and comparing the two data sources. In addition, the concurrent data col-
lection results in a shorter data collection time period as compared to one
of the sequential approaches because both the qualitative and quantitative
data are gathered at one time at the research site.
This model also has a number of limitations. It requires great effort and
expertise to adequately study a phenomenon with two separate methods.
It also can be difficult to compare the results of two analyses using data of
different forms. In addition, a researcher may be unclear how to resolve
discrepancies that arise in comparing the results, although procedures are
emerging in the literature, such as conducting additional data collection
to resolve the discrepancy, revisiting the original database, gaining new
insight from the disparity of the data, or developing a new project that
addresses the discrepancy (Creswell & Plano Clark, 2007).
Concurrent Embedded Strafegy
Like the concurrent triangulation approach, the concurrent embed-
ded strategy of mixed methods can be identified by íts use of one data col-
lection phase, during which both quantitative and qualitative data are
collected simultaneously (see Figure 10.3b). Unlike the traditional trian-
gulation model, a concurrent ~q¡.jJ~ddedapproach has a prímary method
_~h?t guídes the project and a secondary database that provides a support-
ing role in. theprocedures. Given less priority, the secondary method
(quantitative or qualitative) is em.b_edded,or !!..~ted, within the predomi-
nant method (qualitative or quantitative). This embedding may mean that
the secondary method addresses a different question than the primary
method (e.g., in an experiment, the quantitative data addresses the out-
comes expected from the treatments while the qualitative data explores the
processes experienced by individuals in the treatment groups) or seeks
information at a different level of analysis (the analogy to hierarchical
analysis in quantitative research is helpful in conceptualizing these
levels-see Tashakkori and Teddlie, 1998). The mixing of the data from
the two methods is often to integrate the information and }::o·mpareone
data source with the other, typically accomplíshed in a díscússíon section
of a study. However, the data may also .QPtbe compared but reside side by
side as two different pictures that provide an overall composite assessment
of the problem. This would be the case when the researcher uses this
approach to assess different research questions or different levels in an
organization. Similar to the other approaches, an explicit theoretical per-
spective can be used in this model, typically to inform the primary method.
The concurrent embedded model may be used to serve a variety of pur-
poses. Often, this model is used so that a researcher can gain broader per-
spectives as a result of using the different methods as opposed to using the

Mixed Methods Procedures 215
predominant method alone. For example, Morse (1991) noted that a pri-
marily qualitative design could embed some quantítatíve data to enrich the
description of the sample participants. Likewise, she described how qualí-
tative data could be used to describe an aspect of a quantitative study that
cannot be quantified. In addition, a concurrent embedded model may be
employed when a researcher chooses to utilize different methods to study
different groups or levels. For example, if an organization is being studied,
employees could be studied quantitatively, managers could be interviewed
qualitatively, entire divisions could be analyzed with quantitative data, and
so forth. Tashakkori and Teddlie (1998) described this approach as a
multilevel designo Finally, one method could be used within a framework of
the other method, such as if a researcher designed and conducted an exper-
íment to examine treatment outcomes but used case study methodology to
study how participants in the study experienced the treatmentprocedures.
This mixed methods model is attractíve for several reasons. A researcher
is able to collect the two types of data simultaneously, during a single data
collectíon phase. It provides a study with the advantages of both quantita-
tive and qualitative data. In addition, by using the two different methods in
this fashion, a researcher can g..aiQ.-p~rsp~ct~vesfrom the different types of
data or from different levels within the study.
There are also límíjatíons to consider when choosing this approach. The
data need to be transformed in some way so that they can be.íntegrated
within the analysis phase of the research. In addition, if the two databases
are compared, discrepancies may occur that need to be resolved. Because
the two methods are unequal in their priority, this approach also results in
unequal evidence within a study, which may be a disadvantage when inter-
preting the fmal results.
Concurrent Transformative Strategy
As with the sequential transformative model, the concurrent trans-
formative approach is guided by the researcher's use of a specific theo-
retical perspective as well as the concurrent collection of both quantitative
and qualitative data (se e Figure 10.3c). This perspective can be based
on ideologies such as critical theory, advocacy, participatory research, or a
conceptual or theoretical framework. This perspective is reflected in the
purpose or research questions of the study. It is the driving force behind
all methodological choices, such as defining the problem, identifying the
design and data sources, analyzing, interpreting, and reporting results.
The choice of a concurrent model, whether it is triangulation or embedded
design, is made to facilitate this perspective. For example, the design may
have one method embedded in the other so that diverse participants are
given a voice in the change process of an organization. It may involve a tri-
angulation of quantitative and qualitative data to best converge informa-
tion to provide evidence for an inequality of policies in an organization.

216 Designing Research
Thus, the concurrent transformative model may take on the design
features of either a triangulation or an embedded approach (the two types
of data are collected at the same time during one data collection phase
and may have equal or unequal priority). The mixing of the data would be
through merging, connecting, or embedding the data. Because the con-
current transformative model shares features with the triangulation and
embedded approaches, it also shares their specific strengths and weak-
nesses. However, this model has the added advantage of positioning mixed
methods research within a transformative framework, which may make it
especially appealing to those qualitative or quantitative researchers already
using a transformative framework to guide their inquiry.
Choosing a Mixed Methods Strategy
Proposal developers need to convey the specific strategy for mixed meth-
ods data collection they plan to use. Further, they need to advance a visual
figure that presents the procedures they plan to use. Figures 10.2 and 10.3
provide some useful models for guidance. Here are some research tips
about how to select a mixed methods strategy:
• Use the information in Figure 10.1 to assess the aspects that you will
be working with in your mixed methods procedures, and then identify one
of the six approaches discussed in this chapter as the primary design for
your pro po sed study. Provide a working defmition for this design, along
with a visual model and a rationale as to why it is a useful design for you
to uso
• Consider the amount of time you have to collect data. Concurrent
approaches are less time consuming because both qualitative and quanti-
tative data are collected at the same time in the same visit to the field.
• Remember that the collection and analysis of both quantitative
and qualitative data is a rigorous, time-consuming process. WheIlti~_e is
a problem, I encourage students to think about an embedded model of
~~srgn. This model emphasizes a major primary form -of data collection
(e.g., surveys), and it can include a minor secondary form of data collec-
tion (e.g., a few interviews with some q.f the participants who completed
the surveys). The fact that both forms of data are not equal in size and rigor
enables the study to be reduced in scope and manageable for the time and
resources available.
• Consider using the explanatory sequential approach. It is a favorite
of many students, especially those who have little experience with qualita-
tive research and substantial background in quantitative research. In this
approach, an initial quantitative data collection is followed by a secondary
qualitative data collection to follow up the quantitative results.
~- ~_L_

I
1
I
I
I
J
Mixed Methods Procedures 217
• Study published articles that use different approaches and determine
which one makes the most sense to you. Creswell and Plano Clark (2007)
include four complete journal articles so that readers can examine the
detail of studies employing different forms of designs.
• Find a published mixed methods journal article that uses your
design and introduce it to your adviser and faculty committee so that they
have a working model for the approach you plan to use in your study. Since
we are at the early stage of adopting mixed methods research in many
fields, a published example of research in your field will help to create both
legitimacy for mixed methods research and the idea that it is a feasible
approach to research for graduate cornmittees or other audiences.
DATA COLLECTION PROCEDURES
Although the visual model and the discussion about the specific strategies
in a pro pos al provide a picture of the procedures, it is helpful to discuss the
specific types of data to be collected. It is also important to identify the sam-
pling strategies and the approaches used to establish validity of the data.
• Identify and be specific about the type of data-both quantitative
and qualitative-that will be collected during the proposed study. Refer to
Table 1.3, which shows both quantitative and qualitative data. Data differs
in terms of open-ended versus closed-ended responses. Some forms of
data, such as interviews and observations, can be either quantitative or
qualitative depending on how open (qualitative) or closed (quantitative)
the response options might be in an interview or a checklist for an obser-
vation. Although reducing information to numbers is the approach used
in quantitative research, it is also used in qualitative research.
G Recognize that quantitative data often involve random sampling, so
that each individual has an equal probability of being selected, and the
sample can be generalized to the larger population. In qualitative data
collection, purposeful sampling is used so that individuals are selected
because they have experienced the central phenomenon. Rigorous sam-
pling procedures need to be conveyed in a proposal for the quantitative and
qualitatíve data collection. In addition, Teddlie aad Yu (2007) have devel-
oped a typology of five types of mixed methods sampling which relates
sampling to the strategies for mixed methods that I have discussed:
• Basic strategies that involve combining quantitative and qualita-
tive sampling (e.g., stratified purposeful sampling, purposive ran-
dom sampling)
• Sequential sampling, in which the sampling from the fírst phase or
strand informs the second phase or strand

218 Designing Research
• Concurrent sampling. in which quantitative probability and
qualitative purposeful sampling are combined as independent
sampling procedures or jointly (e.g .. a survey with both closed-
ended and open-ended responses)
• Multilevel samplíng, in which sampling occurs in two or more
levels or units of analysis
• Sampling using any combination of the foregoing strategies
• Include detailed procedures in your visual model. For example, in a
sequential explanatory model, the general procedures are higher on the
page and the detailed procedures below them, as shown in Figure 10.2a.
However, the figure can be detailed even further. For exarnple, a discussion
of this approach might include describing the use of survey data collection
followed by both descriptive and inferential data analysis in the fírst phase.
Then qualitative observations and coding and thematic analysis within an
ethnographic design might be mentioned for the second phase.
DATA ANALYSIS ANO VALlDATION PROCEDURES
Data analysis in mixed methods research relates to the type of research
strategy chosen for the procedures. Thus, in a proposal, the procedures
need to be identified within the designo However. analysis occurs both
within the quantitative (descriptive and inferential numeric analysis) and
the qualitative (description and thematic text or image analysis) approach
and often between the two approaches. For example, some of the more
popular mixed methods data analysis approaches are the following (see
Caracelli & Greene, 1993; Creswell & Plano Clark, 2007; Tashakkori &
Teddlíe, 1998):
• Data transformation: In the concurrent strategies. a researcher may
quantify the qualitative data. This involves creating codes and themes
qualitatively. then counting the number of times they occur in the text
data (or possibly the extent of talk about acode or theme by counting lines
or sentences). This quantification of qualitative data then enables a
researcher to compare quantitative results with the qualitative data.
Alternatively. an inquirer may qualily quantitative data. For instance, in a
factor analysis of data from a scale on an instrumento the researcher may
create factors or themes that then can be compared with themes from the
qualitative database.
• Explore outliers: In a sequential modelo an analysis of quantitative
data in the [írst phase can yield extreme or outlier cases. Follow-up quali-
tativ~. igtc:!rvi~ws with these outlier cases can pro vide insight about why
they diverged from the quantitative sample.
(
1
t
e
e
"r
i:
q
r
q
1
s
¡:;
(:
e
1
d
f(
V
o
J<
ti
( t
e

Mixed Methods Procedures 219
• Instrument development: In a sequential approach, obtain themes
and specir.t~tat~ments from participants in an initial qualítatíve data col-
lection. In the n~xt pJ?:,!-~~,usethe~_~ª!erp.ertt~ as specific items and the
themes forscales to create a survey instrument that is grounded in
the views of the participants. A third, fínal phase might be to validate the
instrument with a large sample representative of a population.
• Examine multiple levels: In a concurrent embedded model, conduct
a survey at one level (e.g., with familíes) to gather quantitative results
about a sample. At the same time, collect qualitative interviews (e.g., with
individuals) to explore the phenomenon with specific individuals in the
families.
• Create a matrix: When comparing data in a concurrent type of
approach, combine information from both the quantitative and qualitative
data collection into a matrix. The horizontal axis of this matrix could be a
quantitative categorical variable (e.g., type of provider-nurse, physician,
and medical assistant) and the vertical axis would be the qualitative data
(e.g., five themes about caring relationships between providers and
patients). Information in the cells could be either quotes from the qualíta-
tive data, counts of the number of codes from the qualitative data, or some
combination. In this way, the matrix would present an analysis of the
combined qualitative and quantitative data. Qualitative computer soft-
ware programs provide matrix output capabilities for the mixed methods
researcher (see Chapter 9).
Another aspect of data analysis in mixed methods research to describe
in a proposal is the series of steps taken to check the validity of both the
quantitative data and the accuracy of the qualitative findings. Writers on
mixed methods advocate for the use of validity procedures for both the
quantitative and qualitative phases of the study (Tashakkori & Teddlie,
1998). The proposal writer discusses the validity and reliability of the
scores from past uses of instruments employed in the study. In addition,
potential threats to internal validity for experiments and surveys are noted
(see Chapter 8). For the qualitative data, the strategies that will be used to
check the accuracy of the findings need to be mentioned (see Chapter 9).
These may include triangulating data sources, member checking, detailed
description, or other approaches. •
An emerging field of study is to consider how validity might be different
for mixed methods studies than for a quantitative or a qualitative study.
Writers have begun to develop a bilingual nomenclature for mixed meth-
ods research and have called validity Iegitimation (Onwuegbuzie &
[ohnson, 2006, p. 55). The legitimation of the mixed methods study
relates to many phases of the research process, from philosophical issues
(e.g., are the philosophical positions blended into a usable form?) to infer-
ences drawn (e.g., yield high-quality inferences) and to the value of the

i
220 Designing Research
study for consumers (see Onwuegbuzie & [ohnson, 2006). For individuals
writing a mixed methods research pro pos al, consider the types of validity
associated with the quantitative component (see Chapter 8), validity
related to the qualitative strand (see Chapter 9), and any validity issues
that might arise that relate to the mixed methods approach. Validity issues
in mixed methods research may relate to the types of strategies discussed
in this chapter. These may relate to sample selection, sample size, follow up
on contradictory results, bias in data collection, inadequate procedures, or
the use of conflicting research questions (see Creswell & Plano Clark, 2007
for a discussion of these ideas).
REPORT PRESENTATION STRUCTURE
The structure for the report, like the data analysis, follows the type of strat-
egy chosen for the proposed study. Because mixed methods studies may
not be familiar to audiences, it is helpful to provide .some guidance as to
how to structure the final reporto
• For a sequen ti al study, mixed methods researchers typically organize
the report of procedures into quantitative data collection and quantitative
data analysis followed by qualitative data and collection and analysis.
Then, in the conclusions or interpretation phase of the study, the
researcher cornments on how the qualitative flndíngs helped to elaborate
on or extend the quantitative results. Alternatively, the qualitative data
collection and analysis could come fírst, foHowed by the quantitative data
collectíon and analysis. In either structure, the writer typically presents
the project a~JWºJ,li.~tIDctphases, with sepárate .headíngs for each phase.
• In a concurrent study, the quantitative and qualitative data collection
may be presented in separate sections, but the analysis and interpretation
combines the two forms of data to seek convergence or similarities among
the results. The structure of this type of mixed methods study does not
make a clear distinction between the quantitative and qualitative phases.
• In a transformative study, the structure typically involves advancing
the advocacy issue in the beginning and then using either the sequen tial
or concurrent structure as a means of organizing the content. In the end,
a separate section may advance an agenda for change or reform that has
developed as a result of the research.
EXAMPLES OF MIXED METHODS PROCEDURES
illustrations foHow of mixed methods studies that use both the sequential
and concurrent strategies and procedures.
1
I
j

Mixed Methods Procedures 221
Example 10.1 A Sequential Strategy of Inquiry
Kushman (1992) studied two types of teacher workplace commítment=-,
organizational cornmitment and commítment to student learníng-c-ln .(53
urban elementary and middle schools. He posed a two-phase' eXplmat'ory .'
sequential mixed methods study, as presented in the purpose statement:
The central premise of this study was that organizational cornrnltrnent.ond: '.
commitment to student learning address distinct but equollv import-dnt,
teacher attitudes for an organizationally effective schooL an 'idea that h~S',
some support in the literature but requires further empirical validation, .. ':',."
Phase 1 was a quanntative study that looked at statistical relationships betWeen' ',
teacher commitment and organizational antecedents and outcomes in ele~ ,
mentary and middle schools. Following this macrolevel an~lysis; Phasé 2
looked within specific schools, using qualitative/case st~9Y methods to better '
understand the dynamics of teacher commitment.
(Kushman, 1992, p.13)
This purpose statement illustrates the combination of a purpose with fue
rationale for mixing ("to better understand") as well as fue specifictypes of data
collected during fue study. The introduction focused on fue need to examine
organizational cornmitment and cornmitment to student learning,leading to a
priority for fue quantitative approach. This priority was further illustrated in
sections defíníng organizational cornmitment and cornmitment to student
learning and the use of extensive literature to document these two concepts. A
conceptual framework then followed, complete wíth a visual model, and
research questions were posed to explore relationships. This provided a theo-
reticallead for the quantitative phase of the study (Morse, 1991). The imple-
mentation was QUAN--¿qual in this two-phase study. The author presented
results in two phases, with the frrst-the quantitative results-displaying and
discussing correlations. regressíons, and two-way ANOVAs.Then the case
study results were presented in terrns of themes and subthemes supported by
quotations. The mixing of the quantitative results and qualitative findings
occurred in the fmal discussion, in which the researcher highlighted fue quan-
titative results and the complexities that surfaced from the qualitative results.
The author did not use a theoretical perspective as a lens in the study.
~"'.J&lC5ifa4I~"T'=."". """"' ••••• """"''''''''''MCO' •••• •••Z:a.''''_'''''·••• _ •••.••••••••••••••• ._ ••••••••••
Example 10.2 A Concurrent Strategy of Inquiry
In 1993, Hossler and Vesper conducted a study examining the factor s assocí-
ated with paren tal savings for children attending higher education campuses.
(Continued)

222 Designing Research
(Contínued) " . '¡,
: : " ~." -. -. ,. , ;
usuig íoÍi.gitudinaldáta;cblléCtedfró~ studentsand parents over a 3-year
períod, thda~Üíbts~xciniliiédf~étótsili.8st str6ó.gly·a~sbciated withpar~ntal ,,,:"""'i,:"'<"-""';':"~"-'" ':,i: -, .. ~<.~\: , ;.. ':~'.. ...:.:'O',~.' '".' ',:,. ,',.:.' ..•. , .•. .: .Ó: ••• :_ c Óv Ó, ,." .:",.
savíngs for postseéoJid.aiye<iü~ation;,T~efrr~stilts.fOtiríq that párental 'súp-> ,
. po'rt;eddéati9~~ e~péci:,iiti9ris;(aQ.~krib*t~dgé. or-colleg~ costS~w~re mlPor- .... '. ." _ ••.• .' '. • r·· .-. • Ó: ~'.: . .. : ... _- " -. ; - ~-,". -;..~;. .."( . . . ".',','
'. tant .factors. . Most' ímportant; .for ..our-purposes» the authors collected ..
'üiformation. frp~·p·¡Jeh·ts··gnd··~ttid~I;lt~:·Ó]{i8i.s~~eYs and fróhi.s6. ínter- '.'
.vi,~:~e~tJ.~:i?i~~~~:~i~;~~~rJ,iMr~~J~:~i:...':JIl 9'\ ~tf9r;t:tó.~~red.,Ij!;jqt.or\p':arKh~d.I~9ving,th,l~.artict~:exarn!!I~s'par¡:¡ntal:";
.' . sdvingbehatloiS'; lli¡rig's+J¿j~ntd'nd'~d~~rit:dafO'frdm,a longltudinal~fÚdY>·", ..'\<¡....., . ':_.'";}'<::.~:',«: -~.::''.,}.!~:'~:-,7: ",/';" :1,<':':' ,},t>¡~: ,':r''';~;, ·7¡-· ~:-,:.~~.-;';:'.~':, ~'i\ \ ",: '...::::,,_,:,,\;~.\ .. -:.-.: ,. ',,. )\' .',: '\:", ; ,"
. employir'lgmultiple slJ!Y~Ysóve'r¡(l,thr~\3~y~arperiod,.logistic regre,sslon.wos :..":.
. .used té ideritit{ .-fhefahór~: ~~st~~tro~gly:;s~¿~i;ted ~ith pa~e~tal'~a~i~g .' :;
, 'i6f' Pbst~~8~ntia'rYe'dw~'at¡6ri:.·fn:bdditi8'r;:;~ih~t9·hfsgai.n~dfróm:;héint~r~:::(:'
...~),~~~~1~i~~~~ltt(y~r;;~~~),t~~:i~.:
'. ,':. " .,.;:".:,"';,.",.:.;.;;/, ..;.::;,;/:.:;:,/')),:..;h:Y",;.J~~~il~t'.~,Y:,s\~r,.~:~~(b.::,~4,1),::~.::.,
The actual data was collected from 182 student and parent participants
from surveys over a 4-year period of time and from 56 students and their
parents in interviews. From the purpose statement, we can see that they
collected data concurrently as an implementation strategy. They provide
extensive discussion of the quantitative analysis of the survey data,
including a discussion about the measurement of variables and the details
of the logistic regression data analysís. They also mention the limitations
of the quantitative analysis and specific t-test and regression results. In
contrast, they devote one page to the qualitative data analysis and note
briefly the themes that occurred in the discussion. The weight in this
mixed methods study was assigned to quantitative data collection and
analysis, and the notation for the study would be QUAN + qual. The míx-
ing of the two data sources occurred in a section titled "Díscussion of
Survey and Interview Results" (p. 1~5), at the interpretation stage of the
research process. In this section, they compared the importance of the fac-
tors explaining parental savings for the quantitative results, on one hand,
with the findings frorn the interview data on the other. Similar to Example
10.1, no theoreticallens guided the study, although the article began with
the literature on econometric studies and research on college choice and
ended with an "Augmented Model of Parental Savíngs." Thus, we might
characterize the use of theory in this study as inductive (as in qualitative
inquiry), drawn from the literature (as in quantitative research), and ulti-
mately as generated during the process of research.
1

Mixed Methods Procedures 223
Example 10.3 A Transformative Strategy of Inquiry
A feminist lens was used in a transformative triangulation míxed methods
study by Bhopal (2000). She was interested in examíníng whether theories
oi patriarchy apply to South Asian women (from India, Pakistan and
Bangladesh) living in East London. Since these women often have arranged
marriages and are given dowries, she assumed that forms of patriarchy exist
for them that do not exist for White women in Britain. Her overall aim was to
"buíld up a detailed knowledge of women's lives, their feelíngs towards their
own roles and what they actually did in the home, theír attitudes towards
arranged marriages, dowries, domestíc labour and domes tic fínance" (p. 70).
She studied 60 women using qualitative and quantitative methods,·
(to) investigate the varying significance of difference through which women
experienced patriarchy ... and to present accurate information regarding
the number of women who experienced different forms of patriarchy and test
for the strength of associations between different influences of potrlorcnv,
(Bhopol, "2000, p.68)
She found that education had a significant ímpact upon the women's líves.
Moreover, her discussion presented how femínístmethodologíes ínformed her
study. She discussed how she addressed women's lives in their,:o~, terms, ,
using the language and categories in which women express theniselves. Her
research did not just involve women, it was for women. 'Her research also
involved putting herself into the process of productíon in whích sherriade
explicit her reasoning procedures and was self-reflexive about her own per-
ceptions and biases. In this way, her mixed methods study helped to eXPo.se
the lives of women and have a transformative effect in which the. "women
benefited from the research project" (p. 76).
The intent of the author was to provide a voice for women and to give a
more powerful voice to gender inequality. The quantitative data provided
the generalized patterns of participation while the qualitative data pro-
vided the personal narratives of women. The timing of the study was to
first collect survey data and then interview women to Iollow up and under-
stand their participation in more depth (an explanatory sequential
design). The weight of the qualitative and quantitative components was
equal, with the thought that both contribute to understanding the
research problem. The mixing was through connecting the results from
the quantitative survey and exploring these in more depth in the qualita-
tive phase. Because feminist theory was discussed throughout the article
with a focus on equality and giving voice to women, the study employed an
explícit theoretical feminist lens.

224 Designing Research
SUMMARY
In designing the procedures for a mixed methods study, begin by conveying the
nature of mixed methods research. This includes tracing its history, defining
it, and mentioning its applications in many fields of research. Then, state and
employ four criteria to select an appropriate mixed methods strategy. Indicate
the tímíng strategy for data collection (concurrent or sequential). Also state
weight or priority given to the quantitative or qualitative approach, such as
equal weight, or a priority to quantitative or qualitative data. Mention how
the data will be míxed, such as through merging the data, connecting the data
from one phase to another, or embedding a secondary source of data into a
larger, primary source. Finally, identify whether a theoreticallens or frame-
work willguide the study, such as a theory from the social sciences or a lens
from an advocacy perspective (e.g., femínísm, racial perspective).
Six strategies are organized around whether the data are collected sequen-
tially (explanatory and exploratory), concurrently (triangulation and nested),
or with a transformative lens (sequential or concurrent). Each model has
strengths and weaknesses, although the sequential approach is the easiest to
implement. Choice of strategy also can be presented in a figure in the research
proposal. Then, specific procedures can be related to the figure to help the
reader understand the flow of activities in a project. These include the types
of quantitative and qualitative data to be collected as well as the proce-
dures for data analysis. Typically, data analysis involves data transformation,
exploring outlíers, examining multiple levels, or creating matrices that
combine the quantitative results and the qualitative fmdings. Validity proce-
dures also need to be explicitly described. The fínal written report, because it
may be unfamiliar to audíences, can also be described in a proposal. Each of
the three types of strategies-sequential, concurrent, and transformative--
has a different structural approach to writing a mixed methods study.
1. Design a combined qualitative and quantitative study that
\ employs two phases sequentially. Discuss and provide a rationale for
~ why the phases are ordered in th~ sequence you pro pose.
! 2. Design a combined qualitative and quantitative study that gives I• priority to qualitative data collection and les s priority to quantitative i
I data collection. Discuss the approach to be taken in writing the intro- ~
j ductíon, the purpose statement, the research questíons, and the spe- ;
! cífic forms of data collection. ~
~ 3. Develop a visual figure and specific procedures that illustrate 1
j the use of a theoretícal lens, such as a ferníníst perspective. Use the ~
~ procedures of either a sequential or concurrent model for conducting ¡~ "
f: the study. Use appropriate notation in the figure. ~
~'/Q'.u;::~"'~Y"'~'">l~",.!/",..~,:r •.:!JC.~~...t'!.o*:J:·~J."'~J.r.uc~.~)·.~·,.~"v,"·.:r..l!'_v.t'r")7..i><»:."~"f.~'-~~'I\1'.-.wo.:.~,,:~~U''/"...(•. 1'J.,,. •..••.•.•.•.:,~~,,,;1.oJ,