See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/276268085

Mixed Methods Designs: A Review of Strategies for Blending Quantitative and

Qualitative Methodologies

Article · July 2007

CITATIONS READS

24 5,945

1 author:

Kathryn Pole
University of Texas at Arlington
38 PUBLICATIONS 157 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Preparation of middle level teachers for social justice advocacy. View project

All content following this page was uploaded by Kathryn Pole on 13 May 2015.

The user has requested enhancement of the downloaded file.

 Mixed Method Designs: A Review of Strategies for Blending
Quantitative and Qualitative Methodologies
Kathryn Pole
Saint Louis University

Abstract
The historical debate surrounding quantitative and qualitative research paradigms has been at times
rather passionate. Arguments for and against methodologies often have centered on the philosophical
differences regarding issues such as generalizability, epistemology, and authentic representation of the
phenomena under research. More recently, however, considerable focus has shifted to discussion on how
mixed methods research can be performed and used effectively. Generally speaking, mixed methods can
be conceptualized as the use or blending of research methods from both quantitative and qualitative tradi-
tions. There exists considerable complexity in how these methods may be used together. Nevertheless,
increasing numbers of researchers are embracing the concept of mixed methods, and published articles
using mixed methods are more common than they once were. Given the ongoing development of mixed
methods research, the purpose of this paper is to provide a succinct summary of the most prominent
mixed methods research paradigms. Papers such as this one are needed to communicate current status of
the field in a concise manner for applied researchers.

The historical debate surrounding quantitative and exemplified, for example, in regression models of analysis
qualitative methodologies and research paradigms has been (Maxwell & Loomis, 2003).
at times rather passionate. Arguments for and against these In the 1950s through the 1970s, researchers began noting
methodologies often have centered on the philosophical dif- perceived difficulties associated with this objectivist stance
ferences regarding issues such as generalizability, epistemol- (Lincoln & Guba, 1985). For example, are there objective
ogy, and authentic representation of the phenomena under truths that exist outside of human experience and under-
research (see e.g., Howe, 1988; Reichardt & Rallis, 1994). standing? Is truth universal? As attention shifted to some of
More recently, however, considerable focus has shifted the problems, there was increasing rejection of the tenets of
to discussion on how mixed methods research can be per- objectivist inquiry.
formed and used effectively (Caracelli & Greene, 1993; This rejection gave rise to the interpretive movement.
Creswell, 2003; Tashakkori & Teddlie, 2003). Generally Interpretive researchers believe there are multiple realities,
speaking, mixed methods can be conceptualized as the use and that individuals perceive, understand, experience, and
or blending of research methods from both quantitative and make sense of reality in different ways depending on an
qualitative traditions. There exists considerable complexity individual’s unique background and experiences. In the
in how these methods may be used together. Nevertheless, 1970’s, interpretive researchers first began to express that
increasing numbers of researchers are embracing the con- what is learned from a study is related to the assumptions and
cept of mixed methods, and published research using mixed perspectives investigators bring to the study. According to
methods is more common than it once was. interpretivists, there is no single reality, because knowledge
Historical Perspectives is subjective and culture-bound. Interpretive researchers
work with qualitative, non-quantifiable data, including rich
Historically, educational research was conducted in accounts of social phenomena, contextualized narratives, and
such a way that if one asked a good question and operated the use of rhetorical techniques (Kamberelis & Dimitriadis,
with scientific methodology, there could be an answer that 2005). Qualitative paradigms are grounded in process theory,
was reliable, replicable, and generalizable. This position is which deals with events and the processes that connect them,
termed objectivism, and those operating in this paradigm and how events influence each other (Maxwell & Loomis,
ask questions that can be measured quantitatively. In this 2003). The aim is to understand social phenomena from the
traditional, quantitative view, science is seen as the way to perspective of the participants.
knowledge, and the way to understand phenomena so that As qualitative methodology rapidly gained in popularity
they can be predicted and controlled (Scientific research in with some researchers, they began to engage in the so-called
education, 2002). By the use of deductive reasoning, objectiv- paradigm wars (Gage, 1989) with each side of the quantita-
ist researchers pose hypotheses that can be tested. This kind tive/qualitative argument criticizing the others’ methods,
of research is characterized by an objective and dispassionate procedures, and validity of outcomes. These paradigm wars
stance, with the researcher usually playing a neutral, observer served to polarize the two sides of the disagreement.
role, in the study. It is grounded in variance theory, which
deals with variables and the correlations among them, and is Meanwhile, by the 1930s, some researchers, particularly
in the field of sociology, quietly began working in ways

Volume 20, Number 4 · Fall 2007 Mid-Western Educational Researcher 1

that combined quantitative and qualitative methods (e.g. a quantitative data collection procedure (e.g., an interview
The Hawthorne Studies, Mayo, 1933). These early mixed and test score) or research method (e.g., an ethnography
methodologists seemed unaware that they were doing any- and an experiment). Though mixed method research relies
thing unusual (Teddlie & Tashakkori, 2003), and did not on qualitative and quantitative data collection and analysis,
name their methodology. They used methods appropriate they often are parallel without much real mixing, and the
for their questions, and it was not until the “paradigm wars” questions they ask and inferences they make are often either
that researchers began questioning the appropriate use of qualitative or quantitative in nature, as opposed to blended
methodology blending. (Teddlie & Tashakkori, 2003).
There are distinct philosophical assumptions that shape Mixed model research, by comparison, occurs in several
the way researchers approach problems and collect and or all stages of a study, in sequential or concurrent phases,
analyze data. Quantitatively oriented researchers believe that including questions, methods, data collection and analysis,
scientifically based rules and laws shape the social world and the inference process. Underlying mixed model research
as they shape the physical world. Researchers can discover is the assumption that it is possible to have two worldviews,
these rules and laws, and then apply them objectively to an- or paradigms, mixed throughout a single research project.
swer questions and predict behavior. Qualitatively oriented There may be multiple research questions, each grounded in
researchers view an individual and the world as so intercon- a distinct paradigm, and there might be multiple inferences
nected that one does not exist without the other. The only way relating to different worldviews (Teddlie & Tashakkori,
to understand human behavior is to focus on the meanings 2003). Additionally, mixed model research can involve a team
that events have for the participants by looking at what people of researchers from different disciplines who bring different
think, feel, and do in a comprehensive way. theoretical and analytic perspectives to the analysis of a single
Data gathered through quantitative methods has some- problem (e.g. ,Green & Harker, 1988).
times been described as more objective and accurate because Mixed method research, then, differs from what is called
it is collected using standardized methods, can be replicated, multimethod research. Multimethod designs are those that use
and analyzed using statistical procedures. Qualitative is more than one method, but are restricted to one worldview
sometimes seen as less accurate and reliable. This distinction (e.g., qualitative/qualitative, or quantitative/quantitative
is too simplistic. Either approach may or may not satisfy the methods).
requirements of systematic rigor. Quantitative researchers are At the present time, researchers in social and behavioral
becoming increasingly aware that some of their data may not sciences can be roughly classified into three groups. There are
be accurate and valid. Respondents may not understand the those who are quantitatively oriented, conducting research
meaning of questions to which they respond, people’s ability in empirical ways using statistical analysis and deductively
to remember events is faulty, and it is difficult to control for arriving at conclusions. Qualitatively oriented researchers
human experiences. On the other hand, qualitative researchers rely on more subjective construction of reality to arrive at
have developed better techniques for classifying and analyz- descriptions of phenomena, and their work is influenced by
ing descriptive data. It is also increasingly recognized that all the theory they are using.
data collection, quantitative or qualitative, operates within a
cultural context and is affected by the biases and beliefs of the The third group is the mixed methodologists. As Teddlie
data collectors. As Onwuegbuzie (personal communication, & Tashakkori (2003) claim, they are neither traditional (quan-
January 30, 2005) noted, “Everything starts out qualitative.” titative) nor revolutionary (qualitative). Those researchers
The topic of the research, test design, interview questions, and using mixed methods tend to fit more closely with qualita-
choice of wording are all reflective of the researcher. tive worldviews, including the belief that there are multiple
realities that are dependent upon the individual, but they
Current Status answer questions by combining qualitative and quantitative
methods in various ways, in parallel, concurrent, or sequen-
Broadly defined, mixed method design is research design tial order.
that involves both quantitative and qualitative data in either
single study or in multiple studies in a sustained program of Rationale for Continued Use
inquiry (Creswell, 2003). As the field has evolved, there have
been inconsistencies and confusion in the way various terms Mixed methods approaches can sometimes be superior to
that relate to mixed method research have been defined. Ted- single method designs. Mixed methods research can answer
dlie and Tashakkori (2003) proposed mixed methods designs questions that the other single paradigms cannot. The methods
as the cover term that describes the use of both qualitative researchers use depends on the nature of the questions being
and quantitative data collection procedures and research asked. Certain questions cannot be answered by quantitative
methods, and includes mixed methods research and mixed methodology, and others cannot be answered by qualitative
model research. studies. Researchers can combine approaches so that one veri-
fies the findings of the other, one can serve as the groundwork
Mixed method research studies use qualitative and for the other, and the approaches may complement each other
quantitative data collection and analysis in the methods to explore different aspects of the same question.
part of the study. These studies have both a qualitative and

2 Mid-Western Educational Researcher Volume 20, Number 4 · Fall 2007

 An advantage of mixed methods research is that it en- order for us to understand the differences (Mertens, 2003).
ables the researcher to simultaneously answer confirmatory This paradigm assumes that repression (racial, gender,
and exploratory questions. A researcher can confirm an effect ethnic, disability, etc.) is at the root of social problems, and
on a phenomenon by statistical analysis of quantitative data, asks questions such as, “When teachers are not sensitive to
and then explore the reasons behind the observed effect by cultural diversity in their classroom, what is the impact on
using field research, case study data, or surveys (Tashakkori achievement and future options for the student?” Answers
& Teddlie, 2003a). A researcher may also use qualitative to questions that transformative-emancipatory researchers
methodology to generate theory, and quantitative methods ask are framed in the importance that culture and repression
to test that theory. of culture play on society, and have, as a goal, to improve
Mixed methods research can provide for stronger infer- conditions for the group being studied.
ences because the data are looked at from multiple perspec- The multiple paradigm position simply states that re-
tives. One method can provide greater depth, the other greater searchers use the methods that are most likely to answer their
breadth, and together they confirm or complement each other. questions. The methods vary according to the study at hand,
For example, quantitative data may be used to measure the and fit generally into one of four models based on simulta-
success of an intervention, and qualitative data used to explain neous/sequential mixing (Creswell, Plano-Clark, Gutmann,
the process of the intervention. Mixed methods are useful & Hanson, 2003). Qualitative data can be used as a base to
when they give better opportunities to answer the research help develop quantitative measures and tools; quantitative
questions of interest, and when they help the researcher data can be used to elaborate a qualitative study; qualitative
evaluate the “goodness” of their answers (Tashakkori & methods can be used to help explain quantitative findings;
Teddlie, 2003, p. 14). and qualitative and quantitative methods can be used equally
and in parallel to arrive at the study results. The form of the
Common Models for Conducting model is dependent upon the questions being asked, and re-
Mixed Methods Research searchers reflectively choose methods. In addition, when data
Mixed methods research takes on different forms, de- analysis occurs is dependent upon the questions and model
pending on the researcher and the questions being asked. chosen (Onwuegbuzie & Teddlie, 2003). In models where
Three of the most common approaches are pragmatism, qualitative and quantitative data are gathered at the same time,
transformative-emancipatory, and the multiple paradigm the analysis of data from each may also occur concurrently,
position (Tashakkori & Teddlie, 2003a). either during the course of the study or after all the data is
gathered. In models where methods are used sequentially,
Pragmatism is considered a dialectical stance (Tashak- the data from the first model will be analyzed prior to the
kori & Teddlie, 2003b, p. 706), in which contradictory ideas collection of the data from the subsequent method.
are sought and played with. It rejects concepts like “truth”
and “reality,” and instead focuses on “what works” regard- Using Mixed Methods Designs
ing the research question. Researchers intentionally engage
in multiple sets of paradigms, rather than making either/ An awareness of the theoretical drive of the project is im-
or choices, and “examine the tensions that emerge from portant (Morse, 2003), as it affects how the research questions
the juxtaposition of these multiple diverse perspectives” are addressed and how the study is designed. If the purpose
(Tashakkori & Teddlie, 2003a, p. 677). One of the reasons of a study is to describe or find meaning, the methods will
that pragmatism is the most common paradigm in mixed generally be qualitative, with a focus on things that provide
methods research is because it fits in applied settings where thick narrative descriptions. If the purpose is to confirm, as
there are complex social phenomena. Pragmatic researchers in theory-testing, the methods will usually be quantitative.
consider the question to be more important than the method The direction of the theoretical drive has consequences on
used to answer the question or the paradigm that shapes the study design issues. For example, qualitative data is usually
method (Maxcy, 2003). They use a broad array of techniques, gathered in small sample sizes, while quantitative data usually
selecting based upon the question at hand rather than a sense means larger sample sizes. Qualitative samples are usually
of superiority of techniques. Questions that might be studied purposefully selected based on the needs of the study, and
under a pragmatic paradigm include, “What are the reasons don’t meet the assumptions that shape quantitative stud-
that Strategy A is more effective than Strategy B?” Research- ies (i.e. randomization). The researcher will need to make
ers answering such a question would use quantitative data choices that reconcile these issues.
such as test scores and demographics, and qualitative data Each methodology relies on base assumptions that guide
such as field notes and interviews, in a blended concurrent the collection and analysis of data. Qualitative data is gathered
way in order to arrive at answers. in different, more subjective ways, than quantitative data. In
The second most common mixed methods paradigm order to quantify qualitative data, the researcher would need
is known as transformative-emancipatory. Transformative- to assure that quantitative assumptions have been met, includ-
emancipatory researchers hold that there are diverse views ing such things as, “Were all the participants asked the same
in social realities, but those views need to be placed in a questions in the same ways?” Conversely, researchers working
social, political, historical and economic value system in with quantitative data may be tempted to analyze notes that

Volume 20, Number 4 · Fall 2007 Mid-Western Educational Researcher 3

 respondents write in margins of survey instruments. Because Howe, K. R. (1988). Against the quantitative-qualitative
the survey instrument was not designed to provide qualitative incompatibility theses, or dogmas die hard. Educational
data, and all respondents were not requested to write in the Researcher, 17, 10-16.
margins, this data cannot justifiably be used (Morse, 2003). Kamberelis, G., & Dimitriadis, G. (2005). Qualitative in-
Researchers using mixed methodology must take care to select quiry: Approaches to language and literacy research.
methodologies that serve the purpose and objective of the study. New York: Teacher College Press.
In studies with sequential designs for exploratory purposes, the Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry.
decision on how to analyze data may emerge as trends become Beverly Hills, CA: Sage.
evident in the study. For example, the initial, qualitative phase
Maxcy, S. J. (2003). Pragmatic threads in mixed methods
the study may point to themes that will then lead the researcher
research in the social sciences: The search for multiple
to the quantitative data-collection and analysis methods; or the
modes of inquiry and the end of the philosophy of formal-
initial, quantitative phase may lead the researcher to use a certain
ism. In A. Tashakkori & C. Teddlie (Eds.), Handbook of
qualitative methodology. In studies where quantitative and quali- mixed methods in social and behavioral research. Thou-
tative data are gathered and analyzed in a parallel phase design, sand Oaks, CA: Sage.
the decisions regarding data analysis would often be made at the
very start of the study, and the planned analysis would serve as Maxwell, J. A., & Loomis, D. M. (2003). Mixed methods
design: An alternative approach. In A. Tashakkori & C.
a guide for the collection of data. The purpose and objective of
Teddlie (Eds.), Handbook of mixed methods in social and
a study determine what kind of data is collected and how it is
behavioral research. Thousand Oaks, CA: Sage.
analyzed. Data must be treated in ways that fit the purpose of
the study (see Newman, et al., 2003). Mayo, E. (1933). The human problems of an industrialized
civilization. New York: McMillan.
Conclusion Mertens, D. M. (2003). Mixed methods and the politics of
The use of a combination of qualitative and quantita- human research: The transformative-emancipatory per-
spective. In A. Tashakkori & C. Teddlie (Eds.), Handbook
tive methodology can build on the strengths and neutralize
of mixed methods in social and behavioral research.
the limitations of either methodology used alone. There are
Thousand Oaks, CA: Sage.
advantages and disadvantages of each singular methodol-
ogy, but in combination, educational researchers are able to Newman, I., Ridenour, C. S., Newman, C., & DeMarco, G.
build stronger studies, which lead to better inferences, by M. P. (2003). A typology of research purposes and its
using mixed methods research designs. The understanding relationship to mixed methods. In A. Tashakkori & C.
that social phenomena is complex leads to an awareness that Teddlie (Eds.), Handbook of mixed methods in social and
studying these phenomena using multiple methods supports behavioral research. Thousand Oaks, CA: Sage.
the use of mixed methods research in education. Onwuegbuzie, A. J., & Teddlie, C. (2003). A framework for
analyzing data in mixed methods research. In A. Tashak-
References kori & C. Teddlie (Eds.), Handbook of mixed methods
in social and behavioral research. Thousand Oaks, CA:
Caracelli, V. W., & Greene, J. C. (1993). Data analysis strate- Sage.
gies for mixed-method evaluation designs. Educational Reichardt, C. S. & Rallis, S. F. (Eds.) (1994). The qualitative-
evaluation and policy analysis, 15(2), 195-207. quantitative debate: New perspectives. New Directions for
Committee on Scientific Principles for Education Research, Program Evaluation, No. 61. San Francisco: Jossey-Bass.
R. J. Shavelson & L. Towne, (Eds.) (2002). Scientific Tashakkori, A., & Teddlie, C. (2003a). The past and future
research in education. Washington, DC: National Acad- of mixed methods research: From data triangulation to
emies Press. mixed model designs. In A. Tashakkori & C. Teddlie
Creswell, J. W. (2003). Research design: Qualitative, quanti- (Eds.), Handbook of mixed method research in social and
tative, and mixed methods approaches (2nd ed.). Thousand behavioral research. Thousand Oaks, CA: Sage.
Oaks, CA: Sage. Tashakkori, A., & Teddlie, C. (Eds.). (2003b). Handbook of
Creswell, J. W., Plano-Clark, V. L., Gutmann, M. I., & Han- mixed methods in social and behavioral research. Thou-
son, W. E. (2003). Advanced mixed methods research sand Oaks, CA: Sage.
designs. In A. Tashakkori & C. Teddlie (Eds.), Handbook Teddlie, C., & Tashakkori, A. (2003). Major issues and con-
of mixed methods in social and behavioral research. troversies in the use of mixed methods in the social and
Thousand Oaks, CA: Sage. behavioral sciences. In A. Tashakkori & C. Teddlie (Eds.),
Gage, N. L. (1989). The paradigm wars and their aftermath: Handbook of mixed methods in social and behavioral
A “historical” sketch of research on teaching since 1989. research (pp. 3-50). Thousand Oaks, CA: Sage.
Educational Researcher, 18(7), 4-10.
Green, J. L., & Harker, J. O. (Eds.) (1988). Multiple per-
spective analysis of classroom discourse (Vol. XXVIII).
Norwood, NJ: Ablex Publishing.

4 Mid-Western Educational Researcher Volume 20, Number 4 · Fall 2007

View publication stats