Mixed Methods Research
Abstract and Keywords
Mixed methods research integrates both qualitative and quantitative methods into a single study to produce a more inclusive and expansive understanding of a topic. This article defines mixed methods in social work research, and discusses design notation, language, popular mixed methods designs, and data integration. Using mixed methods provides an opportunity for social workers to take advantage of the strengths of both qualitative and quantitative approaches and to offset their weaknesses. It is important that social workers engaged in mixed methods research maximize the interpretation of their findings and articulate the advantages of using mixed methods over qualitative or quantitative methods alone. Given the unique features of the profession, it is imperative that social workers carve out a distinctive mixed methods niche for social work researchers and practitioners.
Keywords: mixed methods, research, qualitative methods, quantitative methods, convergent parallel design, explanatory sequential design, exploratory sequential design, embedded design, transformative design, multiphase design
This article provides an in-depth overview of mixed methods research in social work. Though using mixed methods in social work research is not new, it has grown in utility and popularity in recent years (Bronstein & Kovacs, 2013; Chaumba, 2013; Haight & Bidwell, 2016). This article provides a synopsis of some basic concepts germane to the application of mixed methods research in social work. First, mixed methods in social work research is defined. Next, the nuts and bolts of mixed methods research are covered, including design notation, language, and popular mixed methods designs. Finally, the actual “mixing” (or, data integration) in mixed methods research is reviewed, and examples of the types of mixing techniques most often used by mixed methods researchers are described. Before defining mixed methods research in social work, it is important to first contextualize how research is understood and conveyed by social work scholars, which is through the use of the knowledge-level continuum.
Teachers of social work research and evaluation underscore the importance of the knowledge-level continuum in understanding how to conduct research in social work (Grinnell & Unrau, 2014; Krysik & Finn, 2013; Rubin & Babbie, 2013). The knowledge-level continuum suggests that there are three types of research (i.e., exploratory, descriptive, and explanatory), with each serving as a precursor for the next. With research at the exploratory level, interest in a topic is usually stimulated because little is known, and there is a need to discover concepts about the topic to move knowledge about the topic forward. After exploratory information is generated, descriptive information is needed to grasp a broader understanding of the concepts that were developed at the exploratory level. Descriptive research can involve numeric and/or text data and can determine how many people, places, and things are involved in the study; what their characteristics are; and what processes occur and how often. After descriptive information is understood, there needs to be research conducted at the explanatory level. Explanatory research helps researchers focus on what is happening between and within concepts by determining causation and predictions. Explanatory research not only can help explain the relationship between variables, but also can be used to test social work theories.
Understanding the knowledge-level continuum is important to understanding how mixed methods can be applied to social work research (Watkins & Gioia, 2015). This is because the purpose of social work research is to deepen understanding of the social injustices and inequalities of marginalized groups so that a plan of action can be devised to help them improve their social and economic situations. The knowledge level continuum provides a path to accomplishing this goal. Whether social workers acquire knowledge that is exploratory, descriptive, explanatory, or a combination of the three, they can use that knowledge to enact change and make a meaningful impact. Often, in determining how to answer research questions, different types of knowledge need to be obtained in order to deepen understanding. This is where mixed methods can be used to advance social work research and practice.
What Are Mixed Methods in Social Work Research?
Watkins and Gioia (2015) suggested that mixed methods in the social work profession are a paradigm, a methodology, and a philosophical underpinning. This is primarily because both mixed methods and social work are deeply rooted in a philosophical movement called pragmatism. Pragmatism is a philosophical worldview that underscores the consequences of research, the importance of the question asked, and the use of multiple methods of data collection to inform research problems (Haight & Bidwell, 2016; Hesse-Biber, 2010; Watkins & Gioia, 2015). By this definition alone, it is clear that neither mixed methods nor social work is for the faint at heart. But the successful execution of both, collectively, promises to provide a more complete, holistic picture of the problems that matter to social work scholars. This is because mixed methods allows social workers to make contributions to the profession by collecting and analyzing quantitative data, which evokes the deductive, or “top down,” approach to social work research, as well as qualitative data, which evokes the inductive, or “bottom up,” approach to research (Grinnell & Unrau, 2014). Social work research and evaluation texts teach us that the top down approach moves from theory to hypothesis to data to confirm or contradict the theory. The bottom up approach uses respondents’ views to build broader themes and generate a theory that connects the themes (Grinnell & Unrau, 2014; Krysik & Finn, 2013; Rubin & Babbie, 2013).
Mixed methods in social work models many of the general mixed methods strategies adopted from other disciplines and professions, such as educational psychology (Creswell, 2015; Creswell & Plano Clark, 2011), sociology (Small, 2011), public health (Curry & Nunez-Smith, 2015; Stewart, Makwarimba, Barnfather, Letourneau, & Neufeld, 2008), and nursing (Sandelowski, 2014). Despite this, however, the past decade has shown an increase in the number of social work researchers who have disaggregated the techniques of other disciplines and aggregated them into guides that are practical and more specific to the social work profession (Bronstein & Kovacs, 2013; Chaumba, 2013; Haight & Bidwell, 2016; Watkins & Gioia, 2015). In an effort to ensure that social workers have a place in the past, present, and future of mixed methods research, some of the leading mixed methods researchers in the social work profession, such as Watkins and Gioia, have provided a definition of mixed methods in the context of social work research:
Mixed methods in social work research is the rigorous and epistemological application and integration of qualitative and quantitative research approaches to draw interpretations based on the combined strengths of both approaches for the purposes of influencing social work research, practice, and policy. (2015, p. 15)
The growth and popularity of mixed methods research in social work are evident. Similarly, the social work profession is beginning to think more about the advantages of using mixed methods not only in social work practice, but also in research. For example, using mixed methods improves the validity of social work research due to the triangulation of multiple methods, it affords an opportunity for the researcher to take advantage of the strengths of both qualitative and quantitative approaches, and it recognizes its alignment with social work principles to examine social and organizational mechanisms holistically (Hesse-Biber, 2010; Watkins & Gioia, 2015). So, not only does the integration of qualitative and quantitative methods “improve lives for individuals, groups, families, organizations, and communities; and make the world a more just place for all” (Bronstein & Kovacs, 2013, p. 359) but also incorporating mixed methods into research and evaluation efforts for social work can advance the science of social work as a discipline and as a profession.
Use of mixed methods in social work is not a new phenomenon (Chaumba, 2013; Haight & Bidwell, 2016), yet mixed methods experts agree that many of the forefathers and foremothers of mixed methods were not formally trained. In fact, it is through their efforts and the skills they sought, that more training opportunities for mixed methods currently exist (Gutterman et al., 2017). Experts in mixed methods pride themselves on the innovation and uniqueness of their method incorporation and integration techniques (Bryman, 2006; Curry & Nunez-Smith, 2015; Hesse-Biber, 2010; Maxwell & Loomis, 2003; Nastasi, Hitchcock, & Brown, 2010; O’Cathain, 2010; Onwuegbuzie & Combs, 2010). Despite this, there are certain characteristics of mixed methods that cross-discipline scholars agree on, such as design notation and language. In order to speak the language of mixed methods researchers, it is important for social work researchers to learn the language and to be able to apply it to mixed methods projects in the social work profession.
The Nuts and Bolts: Mixed Methods Design Notation and Language
The purpose of mixed methods notation is to shorten method-specific phrases and vernacular for use in mixed methods research designs. Early developers of mixed methods adopted a language and created notation to use when designing and implementing their projects. For example, Morse (1991) was one of the first to frame a notation system for mixed methods, which helped to solidify mixed methods as an approach, apart from qualitative and quantitative methods alone, with its own standards, procedures, and notations. In this notation system, everything is shortened to help readers grasp the overall concepts and language of mixed methods. The key to understanding the notation system is to not be confused by the semantics (i.e., the meaning) but rather to understand what the notation represents: a calculated formula for designing, implementing, and reporting mixed methods research (Creswell, 2015; Watkins & Gioia, 2015). For example, in the mixed methods notation system, a quantitative phase is represented using quan and the qualitative phase is represented using qual. When uppercase letters are use to describe the quantitative and qualitative phases (i.e., QUAN and QUAL), this means that the single method will be treated as the priority in the mixed methods study; subsequently, lowercase quan and qual notations are used to indicate which method phases will be treated as a supplement, or to complement, the priority phase (Creswell, 2015; Onwuegbuzie & Combs, 2010).
The plus sign (+) is a common symbol used in quantitative research, but when the plus sign is used in mixed methods research it means that the quantitative and qualitative methods will occur simultaneously. The arrow (→) is also a common notation used in mixed methods and usually means that the two method phases will occur sequentially. When parentheses are used, it usually means that the researcher wants to embed one phase (either qualitative or quantitative) within another. This often occurs when a mixed methods procedure is conducted in the context of a theoretical or program-objective framework (e.g., intersectionality, or a social justice framework). Brackets—[ ]—are used when a mixed methods project is implemented within a single study or within a series of studies. Table 1 illustrates these mixed methods notations and their meanings.
Table 1. Mixed Methods Research Design Notation System
Meaning of notation
A shorthand way to write “quantitative” and “qualitative.”
Capitalizing all letters indicates which phases are prioritized in the design.
Lower-casing all letters indicates which phases are secondary or supplemental in the design.
The methods occur concurrently.
The methods occur in a sequence.
The method in the parentheses is embedded within a larger design (or procedure) or mixed within a theoretical framework.
A mixed methods design is embedded in another. One study is usually embedded within a single study or within a series of studies.
Knowing the language of mixed methods enthusiasts is important because it streamlines the communication between colleagues within the social work profession, as well as with other scholars (e.g., those not in social work) who are engaged in mixed methods research. When discussing the type of language used by mixed methods researchers, it is best to begin with an element of mixed methods research that tends to give people the most trouble: the actual “mixing” of the methods. Here, how qualitative and quantitative methods can be mixed in a few different ways is presented. Arguably, the three most popular ways to mix methods are: (1) connecting the data, (2) merging the data, or (3) embedding the data.
In mixed methods research, connecting the data involves analyzing one type of data, then using the results from that analysis to inform the subsequent data-collection process (Creswell & Plano Clark, 2011; Watkins & Gioia, 2015). In other words, connecting the data attaches the results from one phase of mixed methods research (first type of method) with the data collection of the second phase of mixed methods research (second type of method). Connecting the data in a mixed methods study tends to be sequential and the completion of the second phase is contingent upon the success of the first phase (Creswell, 2015).
Merging the data involves combining qualitative data (in text or images) with quantitative data (in numeric form). Upon first glance, this process may appear intimidating, but it is quite an enriched experience that requires time, patience, and a little creativity. An example of merging the data is using a table or spreadsheet to demonstrate differences among qualitative categories (e.g., concepts) and quantitative ratings (e.g., on a scale).
Another popular way to mix methods is by embedding the data, which involves implanting a lower-priority method within a larger, primary mixed methods design. Embedding the data can sometimes give researchers new to mixed methods the most trouble. But here is an example that hopefully explains how embedding data works: the qualitative data of a project can occasionally help explain the quantitative data, or vice versa. Therefore, sometimes, researchers will decide that they will embed a qualitative phase of a project within an existing mixed methods study. Many times, the mixed methods study may have already started, but not until the research team gets deeper into the study do they realize how beneficial it would be to conduct an additional qualitative phase of the study to help maximize their understanding of the mixed methods study already in progress!
So, connecting the data, merging the data, and embedding the data are just three ways to “mix” data in a mixed methods study. But certainly, many more methods exist, and perhaps readers may one day develop another way to mix qualitative and quantitative data. Just remember that, regardless of how the data are mixed, mixed methods research involves combining qualitative and quantitative data within the context of a single research study to achieve a primary goal.
Types of Mixed Methods Designs
Six mixed methods designs appear to be the most popular among researchers. In this section, the more popular designs are described: the convergent parallel design, the explanatory sequential design, the exploratory sequential design, the embedded design, the transformative design, and the multiphase design (Creswell, 2015; Creswell & Plano Clark, 2011; Watkins & Gioia, 2015). Creswell (2015) designated the first three designs the “basic” designs (because any one of the designs can be the foundation of every mixed methods study) and the last three the more “advanced” designs (because they extend the conceptualization and implementation of a mixed methods study beyond that of the basic designs). Each mixed methods design is described below. The section concludes with some important decisions that should be taken into consideration when choosing a mixed methods design.
Convergent Parallel Design
The convergent parallel design is one of the most popular mixed methods designs because the researcher collects quantitative and qualitative data concurrently, analyzes the two data sets separately, and then mixes the qualitative and quantitative data phases by merging the results during interpretation and/or data analysis. Researchers will choose a convergent parallel design when they want to gain a more complete understanding from the qualitative and quantitative data, have plans to corroborate results from different methods, or want to compare multiple levels within a single system. The convergent parallel design is the design more often used by novice mixed methods scholars because of its numerous advantages. For example, it usually involves collecting both types of data in one visit to the field, both types of data have equal value for understanding the research problem, and it is the epitome of team science because one can include colleagues who have either quantitative or qualitative research skills on the same mixed methods team. A convergent parallel design is depicted in Figure 1.
With the convergent parallel design, both single-method phases are performed during one phase of the study. This saves time, making the convergent parallel design both intuitive and efficient. Again, the convergent parallel design is also team-friendly. Each phase can be collected and analyzed independently, using approaches that are relevant to each data type. Though probably the most popular mixed methods design, the convergent parallel design is not without its challenges. For example, like the other designs, it requires substantial effort and expertise. Researchers sometimes have to be prepared to resolve issues related to the differing samples and sample sizes. Similarly, merging two sets of very different data is not easy, although a way to address this concern is to be sure that the designs of the qualitative and quantitative studies address the same concepts. Also, while some may enjoy this challenge, many times researchers panic once they discover that their quantitative and qualitative results do not fit nicely together. This is not necessarily a disadvantage and, in fact, some researchers have even found that it expands their thinking about a certain topic of interest. Such inconsistencies provide new insight on the topic, but the differences can be difficult to resolve and may require collecting more data, which requires more time and resources. Despite the challenges, novice social workers almost always choose the convergent parallel design for their first mixed methods project.
Explanatory Sequential Design
For the explanatory sequential design, researchers first collect and analyze quantitative data, and then they collect and analyze qualitative data. This is because in an explanatory sequential design, the quantitative results shape the qualitative research questions, sampling, and data collection. The explanatory sequential design begins with a quantitative phase; so, the research problem and purpose call for greater emphasis to be placed on the quantitative features of the study. The explanatory sequential design uses qualitative data to help explain the quantitative results. Figure 2 is an illustration of the explanatory sequential design. The researcher moves from quantitative data collection and analysis to results that he or she would then follow up with qualitative data collection and analysis. Interpretation of the findings from an explanatory sequential design occurs at the end of the study.
Some scholars suggest that the explanatory sequential design should be used only under certain conditions, mainly, if the research problem and question are quantitatively oriented. Explanatory sequential design can also be used if the researcher is quantitatively oriented, or if he or she knows important variables and instruments are available. Usually, variables can be tested and instruments can be disseminated during the quantitative phase of the mixed methods study, and the qualitative phase is used to confirm, complement, or explain the quantitative phase. The explanatory sequential design may be especially attractive to quantitative researchers, as it allows them to complete the quantitative phase first, before the qualitative phase. Exploring, qualitatively, a concept that emerges from the quantitative data is an important advantage of the explanatory sequential design. Yet, often implementing the two separate phases can require time and result in potential delays from the institutional review board (IRB) when challenges arise from the lack of detail involved during the qualitative second phase. Other activities associated with the explanatory sequential design include deciding what quantitative results to follow up, criteria for selecting participants, and how to contact participants for a second round of data collection.
Exploratory Sequential Design
If quantitative data are not a priority, social workers may prefer the exploratory sequential design, which involves collecting qualitative data, analyzing it, then using the results to build the subsequent quantitative phase. The two method phases of an exploratory sequential design are connected when the qualitative results are used to shape the quantitative phase. This can occur by using the qualitative results to help specify research questions and variables, to develop an instrument, or to generate a typology to be explored (Creswell & Plano Clark, 2011; Watkins & Gioia, 2015). The exploratory sequential design is popular because it allows the researcher to explore concepts related to the research topic when there are still variables, theories, and/or hypotheses that are unknown. Instrument development and/or assessing whether qualitative themes generalize to a population are two other outcomes of an exploratory sequential design. Figure 3 illustrates what the exploratory sequential design looks like. It begins with a qualitative phase that includes data collection and analyses, and is followed by the point of interface in the data mixing, which is where the researchers use results from the qualitative phase to develop an instrument, identify variables, or state propositions for testing based on an emergent theory.
Use of the exploratory sequential design can be maximized under certain conditions, specifically, when (a) the researcher and research problem are qualitatively oriented, (b) the researcher has time to conduct two phases, (c) important variables are unknown and instruments are not available, (d) the researcher has limited resources and needs to collect and analyze one data type at a time, and (e) new questions have emerged from qualitative results. Some mixed methods enthusiasts believe that the exploratory sequential design is straightforward to design, implement, and report. Also, the quantitative component can make the qualitative approach feel more acceptable to audiences who may be more quantitatively-oriented research. A potential product from an exploratory sequential design is a scale or larger survey instrument.
The exploratory sequential design is in and of itself an emergent design, though it is not without its challenges. For instance, if the qualitative phase of an exploratory sequential design is an ethnographic study (which may require a lot of time to complete), then implementation of the mixed methods design might be prolonged. Another challenge may arise when deciding which of the qualitative findings to use for the quantitative phase of the study.
If a researcher wants to collect and analyze additional quantitative and qualitative data within a quantitative research study, or a qualitative research study, or a procedure, then using an embedded design would be the best approach to a mixed methods project. The collection and analysis of the second data set can occur before, during, and/or after the collection and analysis of the first set of data. The embedded design differs from the other mixed methods designs because it addresses different questions that call for different methods to augment a mixed method study. This improvement may contain specific aspects of the mixed methods study, such as participant recruitment procedures, a program implementation process, or participants’ reactions to a program. See Figure 4 for an example of an embedded design.
Unlike in the other mixed methods designs, the embedded phase can occur before, during, or after the primary phase of the study. The timing decision is derived from the purpose of the supplemental data within the context of the larger mixed methods design. As a process, the embedded design includes five steps: (1) planning the overall mixed methods design, (2) determining the reason why qualitative (or quantitative) embedded data need to be included, (3) collecting and analyzing qualitative (or quantitative) embedded data to enhance the mixed methods design, (4) collecting and analyzing quantitative (qualitative) outcome data for the remainder of the design, and (5) interpreting how the embedded data enhance the mixed methods procedures and/or understanding of the study outcomes (Creswell & Plano Clark, 2011; Watkins & Gioia, 2015).
Typically, researchers who have expertise with the primary, or dominant, design of a mixed methods study choose the embedded design. Also, embedded designs are chosen when researchers have little experience with the supplemental (second) method, resources determine the need to place equal priority on both methods, and there is a need for the supplemental method and the data that emerge. The embedded design may be an attractive option for social workers because it may require less time and fewer resources, particularly if some resources from the primary method are used to collect data for the supplementary method. Likewise, the embedded design will probably improve the larger design with supplemental data, and it is akin to the social work team science often implemented in social work research. Of course, social workers who decide to use an embedded design will need expertise in the primary method used. Methodologically, some find the purpose and the timing of the supplemental data collection difficult to specify and the results difficult to integrate because the supplemental data can be treated as peripheral data unrelated to the project, rather than as supplemental data that will enhance the study findings and tell a complete story about the data sources. Given the nature of social work research and practice, the embedded design is of particular interest given its popularity with interventions. Therefore, it may be a natural extension for many social workers interested in mixed methods to simply embed another data phase within a current one.
The transformative design, because of its social justice undertones, is an appropriate design for social workers interested in mixed methods research. The transformative design has other names, such as the feminist lens, disability lens, and socioeconomic class lens, which have implications for social work research and practice (Creswell, 2015; Mertens, 2009, 2013; Watkins & Gioia, 2015). Applying a transformative design to a mixed methods study involves using a theory-based framework to advance the inquiry needs of marginalized populations. Then, the data are collected and analyzed concurrently or sequentially, depending on which of the basic designs is used as the core design. The purpose of the transformative design is to conduct research that is “change oriented” and seeks to advance social justice causes. A philosophical assumption of the transformative design is that the transformative paradigm (Mertens, 2009, 2013) provides the overarching assumptions behind the change-oriented nature of the outcomes of the transformative design. This paradigm acts as an umbrella to the project and involves implementing transformative constructs at each step of the research process. Social workers in particular seem drawn to the transformative design because political action, empowerment, collaboration, and change-oriented research perspectives are at its core (Creswell & Plano Clark, 2011; Haight & Bidwell, 2016; Mertens, 2013; Watkins & Gioia, 2015). Interestingly, the transformative design looks just like any of the basic designs (i.e., convergent parallel design, the explanatory sequential, or the exploratory sequential design). The only difference is that the transformative paradigm and theoretical lens used by the researcher have pervasive influence throughout the entire research process. This difference is usually illustrated by a dotted line, to represent the lens through which each step of the mixed methods process is viewed (see Figure 5). During each step, the transformative paradigm plays a role, and there is a consistent focus on the social justice sensitivity and application of each phase of this design.
The transformative design is chosen by social workers who want to address issues of social justice (Cornelius & Harrington, 2014), who want to focus on the needs of underrepresented or marginalized populations, who have a good working knowledge of social justice theoretical frameworks, and who can conduct the study without further marginalizing the population of interest. There are several noteworthy advantages of using a transformative mixed methods design. For instance, it helps to empower individuals and to bring about change, participants often play a participatory role in the study, and it may produce results useful to community members and credible to stakeholders. Despite these advantages, a few disadvantages exist. For example, there are not a lot of transformative design models to support one’s work and to use as guidance while developing a transformative design. Unfortunately, meager guidance may translate into needing to justify the use of the approach before it can be applied. By virtue of how the transformative design is implemented, social workers must be prepared to build trusting relationships with the study participants and plan to conduct the research in a culturally sensitive way. The transformative design assumes that any theoretical foundation can be used as the lens through which the mixed methods work is completed. Figure 5 is an example of a mixed methods design that incorporates a feminist lens.
The purpose of the multiphase design is to address a set of incremental questions that advance one overall programmatic objective (Watkins & Gioia, 2015). The steps of a multiphase design involve examining an overall objective, implementing an iteration of connected quantitative and/or qualitative studies, and then building each new study on what was learned previously. The multiphase design allows for each single-method study to address a specific set of research questions that evolves to address a larger program objective (Creswell, 2015). The steps within a given study phase, or sequence of studies, may mirror the procedures for implementing one or more of the basic mixed methods designs. Researchers employing a multiphase design should be sure to state the research questions for each study, which both contribute to the overall program of inquiry and build upon what has been acquired in previous phases, and design procedures that build on the earlier findings and results. Figure 6 illustrates a multiphase design.
Researchers often choose the multiphase design if they need to complete a long-term objective with multiple qualitative and quantitative studies. Also, having experience in large-scale or longitudinal research, having sufficient resources and funding, having a team that includes practitioners and researchers, and anticipating that emergent questions will arise at different stages are all reasons why social workers should consider a multiphase mixed methods design. The multiphase design is flexible and allows social workers to address interconnected questions. Furthermore, social workers may find the multiphase design easier to publish because each individual study can be published along the way, all the while contributing to the larger program objective. Like the embedded design, the multiphase design also fits well with program-evaluation studies, as well as studies that need to unfold via multiple studies over multiple years.
Challenges of the multiphase design include the anticipation of any of the challenges associated with the other designs. Just like the other designs, multiphase designs need sufficient resources, time, and effort. Unfortunately, this need may present challenges for those in social services agencies that offer social work services. Success of multiphase designs often means that social workers should be working on a team. Challenges can also occur when deciding how to connect individual studies of a multiphase design in a meaningful way. Translating research into practice is important to the social work profession, but it can be time-consuming and resource-intensive, especially if social workers work for agencies that value research (Bronstein & Kovacs, 2013; Chaumba, 2013; Cornelius & Harrington, 2014; Watkins & Gioia, 2015). Finally, because the protection of human subjects is essential in all research, multiphase studies may require multiple IRB applications or amendments before, during, and after implementation.
Important Considerations When Choosing a Mixed Methods Design
While it may be tempting to select a mixed methods design based on its perceived sophistication, there are some aspects of the overall mixed methods study that can assist social workers in determining which mixed methods design is right for them. For example, finalizing important aspects of a mixed methods study, such as the mixed methods research question, the priority of the phases, the timing of the phases, and the level of interaction between the phases, can help streamline the process for selecting an appropriate mixed methods design.
One of the first clues as to which mixed methods design should be selected for a project is the mixed methods research question (Creswell, 2015; Plano Clark & Badiee, 2010). During the early stages of a mixed methods study, many researchers engage in an iterative process that involves writing the mixed methods research question and then studying mixed methods designs to see which one would be most appropriate for answering the mixed methods research question (Plano Clark & Badiee, 2010; Watkins & Gioia, 2015). When writing mixed methods research questions, it is almost second nature to think about how your research question can be addressed using a specific mixed methods design. For sample mixed methods research questions that are aligned with the six mixed methods designs, the reader should see Chapter 2 in the text by Watkins and Gioia (2015).
The priority of the quantitative and qualitative phases is another important consideration when choosing a mixed methods study design. In this context, priority refers to the relative importance (i.e., weighting) of the quantitative and qualitative methods for answering the mixed methods research question. When it comes to weighting, there are at least three options. The first option involves giving the two methods equal priority so that they both play an equally important role in addressing the research question. The second weighting option employs the use of a quantitative priority, where a greater emphasis is placed on the quantitative methods and the qualitative methods are used in a secondary, more supplemental role. The final weighting option involves using a qualitative priority, where a greater emphasis is placed on the qualitative methods and the quantitative methods are used in a secondary, more accompanying role (Creswell, 2015; Creswell & Plano Clark, 2011; Watkins & Gioia, 2015).
The third consideration in a mixed methods study that could influence the design choice is the timing of the quantitative and qualitative phases. Timing is important to mixed methods researchers because it refers to the sequential relationship between the quantitative and qualitative phases within a study (Watkins & Gioia, 2015). In other words, timing refers to deciding when to plan, collect, analyze, and integrate the quantitative and qualitative phases of the study so that they help address the research question and overall study goal. Previous mixed methods researchers have designated three categories of phase timing, with each having a distinct role and responsibility depending on whether the qualitative or the quantitative phase is prioritized: concurrent timing, sequential timing, and multiphase combination timing. Concurrent timing occurs when the study implements both the quantitative and qualitative methods during a single phase, with the collection and analysis of both types of data occurring at the same time. Sequential timing is when the study uses the two methods in two distinct phases, with the collection and analysis of one type of data (e.g., qualitative) occurring after the collection and analysis of the other type (e.g., quantitative). Multiphase combination timing occurs when the study is implemented over multiple phases that include sequential and/or concurrent timing over an entire project time line (Creswell, 2015). Studies conducted over three or more phases or those that combine both concurrent and sequential elements within one program are examples of multiphase combination timing.
The fourth and final consideration of a mixed methods study that influences the design choice is the level of interaction between the qualitative and quantitative phases. This is the extent to which the two method phases are either kept independent or interact with each other. When the two method phases are independent, the level of interaction is virtually nonexistent and the researcher keeps the quantitative and qualitative research questions, data collection, and data analysis separate. For independent levels of interaction, the researcher mixes the two phases at one time, when drawing conclusions during the overall interpretation at the end of the study. A more interactive level of interaction is when a direct interaction exists between the quantitative and qualitative method phases of the mixed methods study. This level of interaction can occur at different points in the research process and in a few different ways; the design and conduct of one phase of the study may depend on the results from the other phase. Similarly, the data from one phase (e.g., qualitative codes) may be transformed into the other type of data (e.g., quantitative frequencies), and then the different data sets are analyzed collectively.
Mixing in Mixed Methods Research: Data Integration at Its Best
While some social workers may find that choosing a mixed methods design is the most challenging aspect of their mixed methods study, the majority may argue that the biggest challenge faced by social work researchers is deciding how to integrate (and interpret) the findings from their mixed methods studies. Integrating and interpreting the findings from a mixed methods study is much like interpreting the findings from a single-method study. The only difference is that mixed methods studies require that you consider the place where the qualitative and quantitative data come together, or their integration (Bazeley, 2009; Fetters, Curry, & Creswell, 2013; Gutterman, Fetters, & Creswell, 2015). Another term for this is mixing. Previous mixed methods researchers have acknowledged the multiple time points in the research process where data integration can occur (Bazeley, 2009; Bryman, 2006; Creswell, 2015; Fetters, Curry, & Creswell, 2013; Watkins & Gioia, 2015). Below, four points in time as well as four types of data integration and interpretation processes that can occur for mixed methods designs that are sequential (i.e., explanatory sequential and exploratory sequential designs) and for designs that occur concurrently (i.e., convergent designs) are discussed.
Mixing Data at Different Points in Time
Determining where and how to mix the quantitative and qualitative phases of a mixed methods study is an important part of the data integration process. Mixing (also known as combining or integrating) is the precise interrelating of the study’s quantitative and qualitative data phases; it is the process by which social workers implement the independent or interactive relationships of a mixed methods study (Haight & Bidwell, 2016; Watkins & Gioia, 2015). In order to understand when and how mixing occurs, the point of interface and the mixing strategies must be explained (Creswell, 2015). The point of interface is where the quantitative and qualitative method phases meet. Previous mixed methods scholars have suggested that this can occur at four possible points in the research process, mainly while (1) designing the study, (2) collecting the data, (3) analyzing the data, and (4) interpreting the results (Creswell, 2015; Watkins & Gioia, 2015).
Mixing While Designing the Study
Mixing at the design level involves integrating the quantitative and qualitative phases of a mixed methods study while designing the research study and can include one of three strategies: embedded mixing, mixing within a theoretical framework, and mixing within a program-based objective framework . Embedded mixing involves embedding quantitative and qualitative methods within a design that is associated with one of these two methods (Creswell & Plano Clark, 2011; Watkins & Gioia, 2015). When mixing within a theoretical framework, quantitative and qualitative data are mixed within a transformative framework (such as feminism) or a substantive framework (such as social science theory) and the framework guides the study design (Chaumba, 2013). Mixing within a program-based objective framework involves mixing quantitative and qualitative phases of a mixed methods study within an overall program objective that guides the joining of multiple projects or studies in a multiphase project (Creswell, 2015; Watkins & Gioia, 2015).
Mixing While Collecting the Data
Mixing while collecting the data involves collecting either quantitative or qualitative data and then deciding that a second set of data needs to be collected. Then, by using a strategy called “connecting,” the researcher can use one method phase (e.g., quantitative) to build the data collection of the other method phase (e.g., qualitative). Mixing ends up occurring in a way that connects the two phases of data. Essentially, this connection uses the results of the first phase of the study to collect data for the second phase of the study (Creswell, 2015). Specifying research questions, selecting participants, and developing data-collection protocols or instruments operationalize this process. Integration at the data-collection phase involves bringing the qualitative and quantitative phases of the mixed methods study together while collecting the data. An example of mixing at the data-collection level is the use of data-collection tools that include both open-ended and closed-ended questions (Creswell, 2015; Grinnell & Unrau, 2014). The tools themselves may include both qualitative and quantitative measures. However, the data that are collected by the tools may be analyzed and interpreted separately.
Mixing While Analyzing the Data
To mix during data analysis, researchers must analyze the data from the quantitative strand and qualitative strands separately and then, using an interactive strategy of “merging,” the two sets of results are explicitly brought together through a combined analysis. Simply put, the researcher further analyzes the quantitative and qualitative results by relating them to each other in a medium that is conducive to comparisons and interpretations. Integration at the data analysis phase involves using a mixed methods design as a framework for integrating and interpreting the data that follow (Creswell, 2015; Fetters, Curry, & Creswell, 2013; Watkins & Gioia, 2015). For example, using an explanatory sequential design will require collecting quantitative data first, analyzing the data, and then making decisions about how the qualitative data will be collected based on the results from the quantitative data. In this regard, the process of integration acts as an explanatory sequential design that is, indeed, sequential. Instead, if the example uses a convergent design, then the integration and interpretation of the data will need to happen simultaneously (Creswell & Plano Clark, 2011). The analysis and interpretation of the qualitative and quantitative strands will occur concurrently, after both the qualitative and quantitative data have been collected.
Mixing While Interpreting the Data
Mixing during the interpretation phase happens when the quantitative and qualitative phases are mixed at the end of the mixed methods study. This usually involves drawing conclusions or inferences that reflect what was learned from combining results from the two method phases of the mixed methods study and comparing or synthesizing the results in a discussion. All mixed methods designs should reflect on what was learned by the combination of methods in the final interpretation (Creswell & Plano Clark, 2011). For mixed methods designs that keep the two strands independent (such as the convergent parallel design), mixing during the interpretation of the study is the only point in the research process where the two phases interact.
Types of Data Integration and Pathways to Interpretation
There are at least four ways social workers can integrate quantitative and qualitative data: merging the data, explaining the data, building the data, and embedding the data (Bazeley, 2009; Creswell, 2015; Fetters, Curry, & Creswell, 2013; Gutterman, Fetters, & Creswell, 2015; Maxwell & Loomis, 2003).
Merging the Data
Data merging occurs most frequently when using a convergent parallel design. Thus, given the nature of this design, one can imagine that merging the data occurs at the end of a mixed methods project and involves bringing the qualitative and quantitative data phases together for comparison. This type of data integration accentuates the independent strengths of qualitative and quantitative approaches by allowing each to occur as a single-method study, with the goal of integrating the results of each single-method study after both phases are completed (Creswell, 2015; Creswell & Plano Clark, 2011). Social workers may find merging the data to be the most user-friendly option for data integration because it allows the qualitative and quantitative phases to operate apart from one another and not be brought together until the results from both are finalized.
Explaining the Data
When data are integrated to help explain one of the methods applied in a mixed methods study, it is called integrating the data for the purpose of explaining (Creswell & Plano Clark, 2011; Watkins & Gioia, 2015). An example of this is the explanatory sequential design, which involves collecting and analyzing quantitative data first, then using the results from the quantitative data to collect and analyze qualitative data. This way, the results of the quantitative data can be explicated. Mixed methods scholars in social work have affectionately called this process providing a voice behind the numbers (Watkins, 2012; Watkins & Gioia, 2015), and it may be especially useful when trying to develop program and evaluation strategies using large-sample survey data. In this case, social workers might decide to analyze the large-survey data first, determine the results from this analysis, and then use the findings from the large-survey data to create an in-depth (qualitative) interview questionnaire. This could help the research team gauge people’s reactions to the findings from the large survey. The qualitative data could then be used to frame a pilot program to improve the well-being of a target group.
Building the Data
Integrating qualitative and quantitative data for the purpose of building most often occurs with an exploratory sequential design. Building the data involves using the qualitative results of a mixed methods study to build a quantitative phase of the study. Thus, building the data often happens when researchers want to use text from their qualitative reports to develop a new instrument, discover new variables, or produce a new intervention or characteristics of an intervention (Gutterman, Fetters, & Creswell, 2015). This is accomplished by translating the words and concepts derived from the qualitative phase of the study into survey items. This type of process is particularly useful to social workers who want to gauge the reaction of a larger sample to some of the findings from the smaller sample (Watkins & Gioia, 2015). While this is a common goal of building the data, it is important to note that developing surveys is not easy, and should be done in partnership with experts in psychometrics to achieve appropriate question order, timing, layout, and wording, and to test for reliability and validity after repeated use.
Embedding the Data
Embedding occurs when the data collection and analysis are being linked at multiple points during the course of a mixed methods study. Unlike other types of data integration, embedding may involve any combination of connecting, building, or merging, but the hallmark is when qualitative data are repeatedly linked to quantitative data collection at multiple points (Fetters, Curry, & Creswell, 2013; Gutterman, Fetters, & Creswell, 2015). Embedding the data as a way to integrate qualitative and quantitative results may seem applicable to social work research because it allows the team to include an additional data component in an ongoing intervention that can help improve knowledge about a phenomenon of interest. An example is using a quantitative instrument to examine a regional policy and how it affects a local community and then using a qualitative instrument to assess how the local stakeholders roll out the policy within their jurisdiction.
This article presents mixed methods in social work research, as well as design notation, language, popular mixed methods designs, and data integration. Social workers interested in mixed methods research should familiarize themselves with the mixed methods designs, their strengths and weaknesses, and understand these characteristics in the context of the social justice problem under study. Similar to the process of preparing data for single-method studies, mixed methods studies involve systematic procedures that help to address the research question. Beyond preparing the data, organizing the data into concepts, developing the story, and maximizing rigor by validating conclusions, mixed methods data collection, analysis, and interpretation are more of an art than a science. As such, it remains imperative that social workers engaged in mixed methods research interpret the findings of their projects and articulate the advantages of using mixed methods over using a single method. Many features of mixed methods are common across all disciplines. Nevertheless, there are certain areas of mixed methods that warrant a closer look at the need to tailor certain strategies so that they adhere more closely to social work research norms and values.
Bernard, H. H. R., & Ryan, G. W. (2009). Analyzing qualitative data: Systematic approaches. Thousand Oaks, CA: SAGE.Find this resource:
Miles, M. B., Huberman, A. M., & Saldaña, J. (2013). Qualitative data analysis: A methods sourcebook. Thousand Oaks, CA: SAGE.Find this resource:
Padgett, D. K. (2008). Qualitative methods in social work research (Vol. 36). Thousand Oaks, CA: SAGE.Find this resource:
Saldaña, J. (2012). The coding manual for qualitative researchers (No. 14). Thousand Oaks, CA: SAGE.Find this resource:
Silverman, D. (2011). Interpreting qualitative data. Thousand Oaks, CA: SAGE.Find this resource:
Watkins, D. C., Wharton, T., Mitchell, J. A., Matusko, N., & Kales, H. (2015). Perceptions and receptivity of non-spousal family support: A mixed methods study of psychological distress among older, church-going African American men. Journal of Mixed Methods Research. Published online before print, December 30, 2015.Find this resource:
Bazeley, P. (2009). Integrating data analyses in mixed methods research. Journal of Mixed Methods Research, 3(3), 203–206.Find this resource:
Bronstein, L. R., & Kovacs, P. J. (2013). Writing a mixed methods report in social work research. Research on Social Work Practice, 23(3), 354–360.Find this resource:
Bryman, A. (2006). Integrating quantitative and qualitative research: How is it done? Qualitative Research, 6(1), 97–113.Find this resource:
Chaumba, J. (2013). The use and value of mixed methods research in social work. Advances in Social Work, 14(2), 307–333.Find this resource:
Cornelius, L. J., & Harrington, D. (2014). A social justice approach to survey design and analysis. New York: Oxford University Press.Find this resource:
Creswell, J. W. (2015). A concise introduction to mixed methods research. Thousand Oaks, CA: SAGE.Find this resource:
Creswell, J. W., & Plano Clark, V. L. (2011). Designing and conducting mixed methods research (2d ed.). Thousand Oaks, CA: SAGE.Find this resource:
Curry, L., & Nunez-Smith, M. (2015). Mixed methods in health sciences research: A practical primer. Thousand Oaks, CA: SAGE.Find this resource:
Fetters, M. D., Curry, L. A., & Creswell, J. W. (2013). Achieving integration in mixed methods designs─Principles and practices. Health Services Research, 48(6pt2), 2134–2156.Find this resource:
Grinnell, R. M., Jr., & Unrau, Y. A. (2014). Social work research and evaluation: Foundations of evidence-based practice (10th ed.). New York: Oxford University Press.Find this resource:
Guetterman, T. C., Creswell, J. W., Wittink, M., Barg, F. K., Castro, F. G., Dahlberg, B., … Gallo, J. J. (2017). Development of a self-rated mixed methods skills assessment: The NIH mixed methods research training program for the health sciences. Journal of Continuing Education in the Health Professions, 37(2), 76–82.Find this resource:
Gutterman, T. C., Fetters, M. D., & Creswell, J. W. (2015). Integrating quantitative and qualitative results in health science mixed methods research through joint displays. Annals of Family Medicine, 13(6), 554–561.Find this resource:
Haight, W. L., & Bidwell, L. N. (2016). Mixed methods social work research: Integrating methodologies to strengthen practice and policy. Chicago: Lyceum.Find this resource:
Hesse-Biber, S. N. (2010). Mixed methods research: Merging theory with practice. New York: Guilford Press.Find this resource:
Krysik, J., & Finn, J. (2013). Research for effective social work practice (3d ed.). New York: Routledge.Find this resource:
Maxwell, J., & Loomis, D. (2003). Mixed methods design: An alternative approach. In A. Tashakkori & C. Teddlie (Eds.), The SAGE handbook of mixed methods in social & behavioral research (pp. 241–271). Thousand Oaks, CA: SAGE.Find this resource:
Mertens, D. M. (2009). Transformative research and evaluation. New York: Guilford Press.Find this resource:
Mertens, D. M. (2013). Emerging advances in mixed methods: Addressing social justice. Journal of Mixed Methods Research, 7(3), 215–218.Find this resource:
Morse, J. M. (1991). Approaches to qualitative-quantitative methodological triangulation. Nursing Research, 40(2), 120–123.Find this resource:
Nastasi, B. K., Hitchcock, J. H., & Brown, L. M. (2010). An inclusive framework for conceptualizing mixed methods design typologies: Moving toward fully integrated synergistic research models. In A. Tashakkori & C. Teddlie (Eds.), The SAGE handbook of mixed methods in social & behavioral research (2d ed., pp. 305–338). Thousand Oaks, CA: SAGE.Find this resource:
O’Cathain, A. (2010). Assessing the quality of mixed methods research: Toward a comprehensive framework. In A. Tashakkori & C. Teddlie (Eds.), The SAGE handbook of mixed methods in social & behavioral research (2d ed., pp. 531–558). Thousand Oaks, CA: SAGE.Find this resource:
Onwuegbuzie, A. J., & Combs, J. P. (2010). Emergent data analysis techniques in mixed methods research. In A. Tashakkori & C. Teddlie (Eds.), The SAGE handbook of mixed methods in social & behavioral research (2d ed., pp. 397–430). Thousand Oaks, CA: SAGE.Find this resource:
Plano Clark, V. L., & Badiee, M. (2010). Research questions in mixed methods research. In A. Tashakkori & C. Teddlie (Eds.), The SAGE handbook of mixed methods in social & behavioral research (2d ed., pp. 275–304). Thousand Oaks, CA: SAGE.Find this resource:
Rubin, A., & Babbie, E. (2013). Brooks/Cole empowerment series: Research methods for social work (8th ed.). Belmont, CA: Cengage Learning.Find this resource:
Sandelowski, M. (2014). Unmixing mixed-methods research. Research in Nursing & Health, 37(1), 3–8.Find this resource:
Small, M. L. (2011). How to conduct a mixed methods study: Recent trends in a rapidly growing literature. Annual Review of Sociology, 37(1), 57–86.Find this resource:
Stewart, M., Makwarimba, E., Barnfather, A., Letourneau, N., & Neufeld, A. (2008). Researching reducing health disparities: Mixed-methods approaches. Social Science & Medicine, 66(6), 1406–1417.Find this resource:
Watkins, D. C. (2012). Qualitative research: The importance of conducting research that doesn’t “count.” Health Promotion Practice, 13(2), 153–158.Find this resource:
Watkins, D. C., & Gioia, D. (2015). Mixed methods research. Pocket Guides to Social Work Research Methods Series. New York: Oxford University Press.Find this resource: