The term mixed methods research is generally used to refer to research that combines quantitative and qualitative research approaches and methods in the same study. Some researchers include studies that combine different quantitative methods, or different qualitative methods, but the term multimethod research is more commonly used for these. Many prominent mixed methods researchers add that such studies should involve an actual integration of the results of the two methods, rather than simply being separate strands of a study with no real interaction. This entry explains the differences between qualitative and quantitative research and describes the history of mixed methods research, key issues in its development, important concepts and strategies in its use, and current controversies in the field.
There is no agreement on a precise distinction between quantitative and qualitative research; both include quite diverse approaches and methods, and there are multiple differences between the two, none of which are entirely definitive. The simplest (and common) distinction, that quantitative [Page 1070]research involves numbers and qualitative involves only words, is clearly inadequate; both fields use numbers (although quantitative research relies much more heavily on them), and the numbers/words distinction fails to capture other commonly invoked differences between these approaches, including the use of artificial versus natural settings, a primary reliance on deductive versus inductive strategies, and a positivist versus constructivist epistemology. None of these distinctions adequately capture the diversity of strategies within each approach, and none definitively distinguish the two approaches.
However, the distinction is extremely meaningful to researchers in both communities and was central to the development of mixed methods research. A different way of distinguishing the two approaches, in terms of their strategy for explanation, was proposed by the evaluation researcher Lawrence Mohr and may be helpful in clarifying the differences. Mohr identified two types of explanation, which he termed variance theory and process theory. Others had earlier presented similar distinctions but had not developed them as systematically. Variance theory is based on the concept of a variable, a property of something that can vary, and can be measured or categorized. This concept is fundamental to quantitative research; essentially, all such research involves the creation and correlation of different variables, or comparison of the values of particular variables, across persons or other units of analysis. The use of variables allows precision in counting or measuring social phenomena, determining differences between individuals or groups on particular variables, and identifying relationships between variables.
Qualitative research, in contrast, makes very little use of variance theory; although some qualitative researchers use the term variable, they do not employ it in the same way as quantitative researchers. Instead, they focus on describing the phenomena studied (behavior, meaning, experience, and social organization) in a specific context and understanding the processes (physical or mental) that connect these phenomena, thus being labeled as process theory by Mohr.
These differences are fundamental to understanding the explicit development of mixed methods research because conflicts between advocates for the two approaches were intrinsic to this development and because the complementarity between quantitative and qualitative research, in terms of their strengths and limitations, provides the main rationale for combining the two approaches. Quantitative research is better at answering “what” and “how much” questions, such as, “Did this educational program make a difference in academic achievement for these students, and how much of a difference?” Qualitative research is better at answering “how” and “why” questions, such as, “How was the program experienced and understood by participants, and how did this shape their responses; how was it influenced by the particular context in which it occurred; and how and why did it achieve these results?” The answers to both types of questions are important for policy and practice, and a mixed methods study is much more capable of answering both.
In 1959, the explicit emergence of mixed methods research has been traced to work by Donald Campbell and Donald Fiske on what they called the multitrait–multimethod matrix, followed by increasing discussion of the possibility of combining quantitative and qualitative methods in the 1970s and 1980s. However, the actual use and integration of quantitative and qualitative methods and data have a much longer history, particularly in the natural sciences, although the latter are almost never addressed in the mixed methods literature. In the physical sciences, the joint use of both methods can be found at least as far back as Galileo’s telescopic observations in the early 1600s, which combined visual description (e.g., of the topography of the moon, and the phases of Venus) with mathematical calculation and measurement. Similarly, field research in geology involves both descriptive observation and quantitative measurement. In biology, the work of ethologists such as Karl von Frisch, Konrad Lorentz, Niko Tinbergen, and Jane Goodall has also integrated qualitative observation and description with numerical data.
In the social sciences, the deliberate integration of qualitative interviewing and observation with survey data and social statistics dates at least from [Page 1071]W. E. B. DuBois’s The Philadelphia Negro (1899), and such integration continued in classic sociological works such as Middletown (1929), Yankee City (1941), and many other studies through the 1960s, although becoming less common with the rise of statistical methods and quantitative research. In anthropology, quantitative data collection and analysis have frequently been integrated with ethnographic fieldwork since the 1920s, and psychologists such as Leon Festinger and Stanley Milgram also combined both methods in their research. There was a widespread recognition that both quantitative and qualitative methods had limitations and that combining the two could provide important benefits.
However, such combinations were not seen as a specific type of research, and within these studies, conflicts between proponents of the two approaches were largely absent; the lead researchers were typically involved in collecting and analyzing both types of data. Such conflicts became prominent after about 1970, due in part to the increasing dominance of quantitative research in prestige and funding, and led to what has been called the “paradigm wars” of the 1980s and 1990s.
The idea of a paradigm, popularized by Thomas Kuhn’s influential The Structure of Scientific Revolutions, became a key issue in the conflict between quantitative and qualitative research in the 1980s. In his 1969 postscript to this work, Kuhn described a paradigm as “the entire constellation of beliefs, values, techniques, and so on shared by the members of a given community.” However, within the social sciences, this term came to refer mainly to the philosophical and ethical presuppositions of the different approaches, which were assumed to be foundational for each approach. Quantitative research was generally seen as based on positivism or postpositivism, which emphasized objective measurement and researcher neutrality. Qualitative research was claimed by many of its proponents to be based on constructivism (the view that reality was socially constructed, rather than being an objective entity), critical theory (incorporating ethical values and working against the oppression of disempowered groups), and/or postmodernism as alternative paradigms to positivism and postpositivism.
This entry cannot discuss in detail these paradigm debates, but they have played an important role in the development of mixed methods research since the 1980s. Prominent qualitative researchers such as Egon Guba and Yvonna Lincoln, adapting Kuhn’s idea of the “incommensurability” of paradigms, argued that qualitative and quantitative research, being based on different paradigms, were therefore incompatible and could never legitimately be combined in a single study.
Although earlier presentations of mixed methods focused mainly on combining data collection and analysis methods, the paradigm wars forced proponents of combining methods to address the broader issues involved in combining research approaches and not simply methods. In response, some mixed methods researchers chose to simply ignore philosophical debates and do whatever they believed worked to produce useful results, a stance that was variously labeled “pragmatic” or “a-paradigmatic.” Others claimed that philosophical pragmatism, as developed by John Dewey and others, resolved these issues and was thus the appropriate paradigm for mixed methods research. This led some proponents to claim that mixed methods research was itself a third paradigm, in addition to quantitative and qualitative research. Still others argued that multiple philosophical stances could be employed in mixed methods research and that paradigms, and not simply methods, could be mixed.
These different positions continued to be argued into the 20th century but generally less vituperatively. However, some quantitative researchers still treated qualitative research as less than fully scientific, a tendency that often characterized the movement promoting what has been called evidence-based research, with randomized controlled trials as the gold standard. Similarly, some qualitative researchers viewed mixed methods research with suspicion, seeing it as simply “positivism in drag.”
Despite these disagreements, mixed methods research had become well established by 2000, [Page 1072]with textbooks appearing as early as 1989 and proliferating after 2000. The first edition of the SAGE Handbook of Mixed Methods in Social and Behavioral Research was published in 2003 and the second (entirely new) edition in 2010; the Journal of Mixed Methods Research was founded in 2007, and the Mixed Methods International Research Association was established in 2014. Mixed methods studies are now commonly published in top-ranked peer-reviewed journals in the social sciences and have been funded by many governmental and nonprofit agencies. Courses in mixed methods research are now often given in universities and through various research organizations. However, there are still ongoing debates about important issues.
Research design is the central issue for mixed methods research because the conception of design has been quite different in quantitative and qualitative research and even between different types of quantitative research. In experimental research, design has usually referred to particular types of research strategies, such as randomized experiments and different forms of quasi-experimental and single-subject research. Nonexperimental quantitative research has not usually been conceptualized in this way, although categorization in terms of the types of statistical analysis employed, such as structural equation modeling and hierarchical linear modeling, is common. Qualitative research, in contrast, lacks any explicit design categories; although works on qualitative research often distinguish between different approaches to research, such as grounded theory, phenomenology, and narrative research, these typically include philosophical and theoretical stances as well as methodological ones and aren’t usually thought of as designs.
The dominant conceptualization of design in mixed methods research has been similar to that in experimental research, of defining specific types of mixed methods studies, based on criteria such as the order in which the methods are used, the relative dominance of the different methods, the degree of integration of the methods and results, and the purposes for which the methods are combined. However, this approach has been criticized by some mixed methods researchers, and alternative conceptions of design have gained increased recognition.
The most prominent of these alternatives has been termed an interactive, systemic, or dynamic approach that sees design as the relationships and mutual influences of the different components of a research study. These components include the study’s goals, conceptual framework or theory, research questions, methods, and validity issues. The research questions are seen as the center or hub of the system, and influence and are influenced by all of the other components. This model is much closer to qualitative conceptions of research, in which design is an inductive and flexible aspect of a study that can adapt to unexpected developments or results.
The mixed methods community has been divided on the nature of appropriate research questions for mixed methods research. Some authors have argued that mixed methods studies require (in addition to possible quantitative and qualitative research questions) specifically mixed methods questions, ones that can be answered only by integrating qualitative and quantitative data. Others do not believe that this is necessary, seeing many questions as potentially or partially answerable by either qualitative or quantitative methods and arguing that the linking of qualitative and quantitative questions, in creating a broader and more inclusive understanding of the phenomena studied, is a legitimate goal of mixed methods research.
Closely connected to the different conceptions of design, there has also been disagreement over the relationship between research questions and research methods. For some researchers, the research questions are the primary and determining component, and the methods must follow from this (a view more characteristic of quantitative research). For others (and this is implied by the interactive concept of design), the research questions, though fundamentally important, need to be responsive to how the methods actually play out in [Page 1073]practice, and to unexpected findings, validity threats, or theories that emerge during the study.
As noted earlier, the goal of integrating qualitative and quantitative methods and data, rather than keeping them as separate strands of a study, has typically been seen as a defining feature of mixed methods research. However, the ways in which this integration can be accomplished have been inadequately studied and theorized, and researchers have received little guidance in how to do this.
Earlier mixed methods studies, those prior to the emergence of mixed methods as an explicit type of research, typically emphasized what later became known as triangulation—the use of one approach to test or confirm the results of the other. However, some of these studies also combined methods to provide a broader and more in-depth understanding of the phenomena studied, based on the complementarity of the two approaches in focusing on different aspects of these phenomena. For example, a quantitative approach could be used to rigorously measure the effect of a new educational program on student achievement, and a qualitative approach to understand how the program was perceived by teachers and students and the ways in which it was implemented in the settings studied.
The paradigm debates of the 1970s and 1980s problematized both of these approaches, raising the issue of whether such philosophically divergent approaches could in fact be combined in these ways. As this debate waned, however, additional purposes for combining methods emerged. In some studies, the two methods were not used concurrently but sequentially. This allowed the use of one method to develop the second method—for example, by using qualitative interviews or observations to develop a survey questionnaire or by further exploring the results of a quantitative survey through focus groups. Such sequential designs have become a prominent part of mixed methods. However, attempts to use the concurrent versus sequential distinction as a basis for a typology of mixed methods designs do not accommodate many studies in which the relationship is more complex than this—for example, iterative designs in which there is alternation or partial overlap in time between different methods or in which the sequencing of approaches in data collection and data analysis is different. Despite this proliferation of purposes and strategies for integrating methods, there has still been little explicit theorization of how such integration is done. Until this occurs, the most productive way to understand the integration of methods is to read accounts of exemplary mixed methods studies.