Mixed methods research is based on systematically combining qualitative and quantitative research methods. Mixed methods have a long history of use in sociology and, perhaps to a lesser degree, in anthropology and are thus relevant to the study of social and cultural foundations of education. This entry describes several principles of mixed methods research; contrasts qualitative and quantitative data; notes several paradigms attributed to mixed methods research; sketches the process of mixed methods research design; details one kind of design; and provides a simple, yet illustrative, example.
Several Principles Of Mixed Methods Research
The potential complexity of mixed methods research makes it necessary here to merely present several principles of mixed methods research. A first principle is to mix quantitative and qualitative methods in a way that makes their strengths complementary and their weaknesses nonoverlapping. Mixing alone is not enough.
Assume that a typical research study involves three stages: specifying research objectives, collecting data, and analyzing the data. In mixed methods research, there is a qualitative phase and a quantitative phase; within each phase, the methods used are of the same type, either qualitative or quantitative. The two phases can be carried out sequentially or concurrently.
Practitioners sometimes assume that the methods selected for data analysis dictate the type of data that must be collected. A second principle is that how data are collected is not logically dependent on how data are analyzed. If the phases of a mixed methods research study are sequential, which is often the case, then the data analysis stage of one phase will serve as input to the data collection stage of the other; that is, the qualitative (quantitative) phase will provide input to the quantitative (qualitative) phase.
Quality control procedures should ensure that the data collected are not contaminated. Once the data are collected, it is necessary to ask how accurately and how reliably they represent the phenomena of interest. A third (obvious) principle is that if the data collected are of poor quality, no amount of subsequent analysis can extract meaningful results.
Data: Both Qualitative And Quantitative
Oddly enough, many accounts of mixed methods do not explicitly distinguish between quantitative data and qualitative data. The former are most broadly defined as categorical in nature: that is, each datum is assigned to a category and can thus be counted. Qualitative data can be defined as textual in nature; written and spoken words, utterances, and even images can be regarded as qualitative data. Although this distinction is helpful, the illustrative example given below suggests that it is not as precise as one might think.
Combatants during the “paradigm wars” of educational research history could declare their allegiance monosyllabically: QUAL or QUAN. For some methodologists, and many practitioners, the two paradigms were incommensurable. Those accepting the possibility of mixing advanced either foundational or pragmatic justifications of mixed methods research. The widening acceptance of pragmatism in mixing has led to a wide array of mixed methods research designs.
Designing Designs
The phrase “research design” refers to a plan for addressing a set of research questions. In the context of mixed methods research, such a plan identifies the emphasis given to the phases and their ordering in time. If the qualitative (quantitative) phase is emphasized and is first, we have a QUAL → quan (QUAN → qual) design. Other combinations are possible depending on whether either paradigm is emphasized and whether the phases are carried out concurrently. Despite their usefulness, such taxonomies can distract researchers from the creative act of designing, which involves recognizing how purposes, conceptual frameworks, methods, issues of validity, and research questions interact.
The purposes of a study can be practical, intellectual, and personal. If the purpose of a study is to test a theory, this will constrain the kind of research question that the study can address. If the study purpose is to determine how strongly two variables are related, this will also constrain the types of admissible research questions. The purpose(s) of a study also constrain the kinds of conceptual framework that can inform the study.
Some theory or collection of theories provides a conceptual framework for the study. Broadly speaking, the conceptual framework for a QUAL study is likely to involve a process-oriented theory, whereas a QUAN study is likely to involve a variance-oriented theory. Whereas process-oriented theories aim at causal explanation, variance-oriented theories aim at causal description.
There is, of course, the question of which methods and procedures will actually be used. Will the data collected be categorical or textual? Will the data be analyzed using qualitative or quantitative methods? Will qualitative (quantitative) data be “quantitized” (“qualitized”)? How these questions are answered can affect the accuracy/reliability of the results produced in response to the research questions of the study.
Unfortunately, the conclusions of a study can be wrong. Various types of validity (and, therefore, invalidity) have been enumerated for both QUAL and QUAN research methods. These standards can be summarized in the form of questions: Do the data meet the minimum criteria of acceptability/trustworthiness? Do the data adequately represent the theoretical phenomena being studied? Do the inferences made based on the data meet minimum standards of credibility/ defensibility? The standards appropriate to each phase should be considered separately.
Research questions mediate the other four components of a mixed methods research design. By itself, a research question pertaining to the variability of the data may not shed much light on the processes that generated the data. Similarly, a research question pertaining to the meaning that an individual assigns to an utterance may by itself do little to explain why some individuals generate more utterances than others. For a sequential mixed methods research study, which is the focus of this entry, the research questions for the second phase should logically presuppose those of the first phase.
A Simple Design
The purpose of this section is to illustrate a relatively simple QUAN → qual design. In almost any small, task-oriented group, some members initiate conversation, and some respond to a conversational prompt, more often than others. Perhaps frequent initiators (responders) express particular kinds of speech acts such as commands (agreements). Perhaps there is a relationship between the social status of group members and the types of speech acts they employ in conversation. The remainder of this section is a sketch of a possible mixed methods research plan.
The purposes of the research are twofold. First, evaluate a model of social status and task participation in small, task-oriented discussion groups. Second, explore ways in which that model might be improved using conversational data. These purposes might be regarded as QUAN and qual, respectively.
The conceptual framework has two components: the theory of status characteristics and the theory of conversational analysis. The former describes a social psychological process by which people come to hold performance expectations. When people interact, they bring with them status characteristics such as education level, gender, age, and native language. People use status characteristics in evaluating their work and that of others; they form expectations for their performance and that of others. Status characteristics theory is most readily labeled as a kind of QUAN theory because of its quantitative form and previous testing in an experimental setting. Conversation analysis, which is usually regarded as a kind of QUAL theory, focuses on the conversational mechanisms by which turns are allocated in various types of speech interaction. Thus, the conceptual framework has both a QUAN and a qual component.
Several methods are employed. Previously collected conversational data from each of twenty-three group meetings will be used. In the QUAN phase, each actual conversation is compared to several simulated conversations. Each simulated conversation is generated using a computational model that reflects the status characteristics of the meeting participants. The computational model controlled by several quantitative parameters that shape how often simulated social ties form between simulated participants. With different parameters, each simulation of a meeting is likely to generate different simulated conversations. The “best” parameter values are those that result in simulated conversations “most like” the actual conversation. A simulated conversation that involves the same sequence of speakers addressing responders as the actual conversation is “most like” the actual conversation. An automatic search process finds the best model parameters for each meeting, which concludes the QUAN phase. In the qual phase, those conversational turns associated with the formation of a simulated social tie are examined to see which types of speech acts actually occurred.
The design makes use of speech data obtained from a naturalistic setting. Such data are often considered to have high causal validity because they have the potential to better represent the processes involved in the situation under study. Hence, the speech data provide a firm foundation for quantitative analysis. That analysis is based on a simulation model that itself generalizes findings of status characteristics theory previously obtained in experimental settings. Hence, the simulation provides a moderately high level of generalizability. Finally, because the (simulated) social networks evolve over (simulated) time, it is possible to represent for each meeting the process by which a status order emerges (in which a social tie has formed for each pair of actors) and a conversation develops within a small, task-oriented group. The design offers moderately high generalizability and high causal validity, which neither the QUAN nor the qual components can provide alone.
These four design components are organized around the following research questions. For each meeting, what parameter values produce the closest match between the actual conversation and simulated conversations? Does the model represent some meetings better than others? What types of speech acts accompany the formation of simulated social ties? Clearly, these research questions are linked to the purposes, conceptual framework, methods, and validity issues pertinent to the study.
Those who write mixed methods research plans must take care to state as early and as clearly as they can that mixed methods research is involved. Otherwise, the reader may make inferences about the clarity and value of the research design based on an assumption that the research methods involved are either quantitative or qualitative. Similarly, readers of research plans should be sensitive to the possibility that indicators of methods often seen in QUAL or QUAN research may simply be components of a mixed methods research design.
Pros And Cons
Of course, mixed research has its own weaknesses. It can be time consuming, expensive, and difficult for a single researcher to execute, and it can present difficulties concerning how to carry out the data analysis task in a phase-one type using data obtained in a phase of a different type. Set against these weaknesses is the claim that a broader and deeper set of research questions can be addressed and potentially answered. Methodological correctness aside, mixed research methods may also aid in the discovery of new meanings and new relationships, presumably one of the aims of those interested in the cultural and social foundations of education.
Bibliography:
- Abbott, A. (2004). Methods of discovery: Heuristics for the social sciences. New York: W. W. Norton.
- Johnson, B., & Christensen, L. (2004). Educational research: Quantitative, qualitative, and mixed approaches. Boston: Pearson.
- Maxwell, J. A., & Loomis, D. M. (2003). Mixed methods design: An alternative approach. In A. Tashakkori & C. Teddlie (Eds.), Handbook of mixed methods in social & behavioral research (pp. 241–272). Thousand Oaks, CA. Sage.
- Ragin, C., Nagel, J., & White, P. (2004). General guidance for developing qualitative research projects. In C. Ragin, J. Nagel, & P. White (Eds.), Workshop on scientific foundations of qualitative research (pp. 9–16). Washington, DC: National Science Foundation.
- Tashakkori, A., & Teddlie, C. (2003). Handbook of mixed methods in social & behavioral research. Thousand Oaks, CA: Sage.
This example Mixed Methods Research Essay is published for educational and informational purposes only. If you need a custom essay or research paper on this topic please use our writing services. EssayEmpire.com offers reliable custom essay writing services that can help you to receive high grades and impress your professors with the quality of each essay or research paper you hand in.