Abstract
Case study has a tradition of collecting multiple forms of data—qualitative and quantitative—to gain a more complete understanding of the case. Case study integrates well with mixed methods, which seeks a more complete understanding through the integration of qualitative and quantitative research. We identify and characterize “mixed methods–case study designs” as mixed methods studies with a nested case study and “case study–mixed methods designs” as case studies with nested mixed methods. Based on a review of published research integrating mixed methods and case study designs, we describe key methodological features and discuss four exemplar interdisciplinary studies.
Both case study and mixed methods, alone, are popular approaches to research and evaluation. Mixed methods research combines qualitative and quantitative research in a study, or closely related series of studies, through the collection, analysis, and integration of qualitative and quantitative data (Creswell, 2015a). Mixed methods is growing in its adoption worldwide as evidenced by continual increases in mixed methods dissertations (McKim, 2017) and funded mixed methods studies (Coyle et al., 2016), two key leading indicators of its use. Case study is a commonly used, though at times underappreciated, approach to research and evaluation in many disciplines. Case study involves the investigation of one or more real-life cases to capture its complexity and details (Yin, 2014). Investigators have applied case study to a range of fields that are too numerous to list here but include political science (Gerring, 2004), education (Harland, 2014), health research (Hyett, Kenny, & Dickson-Swift, 2014), sport and leisure (Van Tuyckom & Bracke, 2014), and information technology (Cooper, 2000).
Case study experts have suggested integrating qualitative and quantitative research in investigating the case (Yin, 2014), and mixed methods experts have suggested mixed methods case study as a complex design (Creswell & Plano Clark, 2018). Investigators are increasingly combining case studies and mixed methods, which, if conducted systematically and thoughtfully, can yield a more complete understanding. Simply, a mixed methods case study can “enable you to address broader or more complicated research questions than case studies alone” (Yin, 2014, p. 67). Nevertheless, guidance on designing and conducting mixed methods case studies remains relatively limited.
Two Innovative Approaches to Integrated Mixed Methods and Case Study Designs: MM-CS and CS-MM
To address the need for guidance in this research approach, we advance an innovative distinction to conceptualize the research design used when integrating case studies and mixed methods. Researchers have at least two primary designs available (Figure 1):
In a mixed methods–case study design (MM-CS), researchers employ a parent mixed methods study that includes a nested case study for the qualitative component.
In a case study–mixed methods design (CS-MM), researchers employ a parent case study that includes a nested mixed methods design.
Researchers have at least two general design choices when integrating case study and mixed methods designs.
Stake (2005) has noted that case study is not tied to particular methods but is determined by the choice to explore a case. Case study research has a long tradition of collecting both qualitative and quantitative data to gain a more complete understanding of the case (Stake, 1995; Yin, 2014). In this regard, it may be an ideal example of one mixed methods research approach. Nevertheless, researchers using a mixed methods design and a case study for the qualitative component will likely benefit from understanding features of rigorous case study research. Similarly, researchers conducting a case study that uses qualitative and quantitative methods can benefit from recent innovations in mixed methods research, such as considering the mixed methods design and achieving meaningful integration of the two forms of data to yield new inferences and a more complete understanding.
Intent of the Review of Articles
Examining how contemporary research is being conducted in peer-reviewed, published literature is one way to develop guidance for future research. Writing about the development of qualitative inquiry, Morse (2011) eloquently noted with regard to how researchers shape the methodology, “We all do it: What we approve in the review process, what we publish, what we use to teach, and the way we conduct our research will determine our future” (p. 1019). Therefore, we undertook a review of published mixed methods case studies to understand the use of the design and provide guidance for other researchers. A brief explanation is needed to explain our focus on case study designs and mixed methods. A core characteristic of mixed methods research is the use of persuasive and rigorous quantitative and qualitative data collection and analysis (Creswell & Plano Clark, 2018). We believe it is equally important to identify and follow rigorous procedures for the quantitative design operating within a mixed methods study. However, our experience indicates that the qualitative design within mixed methods research often receives less attention. Therefore, our focus on integrating a promising qualitative design, case study, in mixed methods is meant to present the major approaches, discuss the value-added, and provide recommendations for researchers interested in the approach.
Our goal was not to deconstruct published mixed methods case studies but to focus on describing the methodological features of the research based on our analysis. Case study and mixed methods are reasonably established methodologies among themselves. However, certain adaptations, such as sampling strategies or presentation of results, may be needed when the two are combined. We endeavored to understand these complexities through a review and analysis of recent, published empirical research using mixed methods case studies. We have identified 81 articles to examine contemporary use of mixed methods case study.
What Is Mixed Methods?
Mixed methods is the process of integrating quantitative and qualitative research to more completely address a study’s purpose or research questions (Plano Clark & Ivankova, 2016). Scholars have advanced numerous designs and typologies for use in mixed methods, generally categorized into types of designs based on the level of integration—fully or partially mixed, the sequence of qualitative and quantitative components—concurrent or sequential, and the emphasis of components—equal or dominant (Leech & Onwuegbuzie, 2009). For the purpose of this article, we will categorize research based on the Creswell and Plano Clark (2011) mixed methods design typology because of its simplicity and popularity. Creswell and Plano Clark identified three core designs: explanatory sequential, exploratory sequential, and convergent. In explanatory sequential designs, the study begins with quantitative data collection and analysis and then proceeds to a follow-up qualitative phase for the purpose of explaining the results of the initial quantitative phase. An option for the follow-up phase is to conduct a case study. In exploratory sequential designs, the study begins with a qualitative exploration followed by a quantitative phase to test findings or attempt to generalize with a larger quantitative sample. Finally, in convergent designs, the study involves a single phase in which qualitative and quantitative data are collected, analyzed, and integrated typically for the purpose of comparing or relating results from the two forms of research. Researchers often begin with core designs and then add components, such as combining core designs with philosophical worldviews (e.g., participatory research, social justice) or other purposes (e.g., randomized controlled trials, evaluation, case studies, instrument development) in complex mixed methods designs (Creswell, 2015a).
Integration is a critical and defining feature of mixed methods that may occur at many levels of the research process. In this study, we focused on common approaches to the integration of methods: merging, connecting, and building as outlined by Fetters, Curry, and Creswell (2013). When merging, investigators compare or relate qualitative and quantitative results, such as juxtaposing the two side by side or examining patterns of themes with respect to statistics. In connecting, investigators use results from one type of research to inform the sampling of the other (e.g., based on quantitative results from a survey, determine cases to sample for a qualitative follow-up to best explain the original quantitative results). To employ building integration, investigators use the results of one type of research to inform data collection procedures of the other (e.g., based on qualitative findings, systematically develop an assessment instrument). Researchers may use more than one type of integration through the multiple phases of the research process. Based on these fundamentals, we now turn to understanding of the case study and examine the extent these integration features can be found and how they are utilized.
What Is Case Study?
The intent of a case study is to conduct an analysis and develop an in-depth understanding of a phenomenon (i.e., the case) within a real-world context (Yin, 2014). The case itself is well-defined through identifying its focus and unit of analysis, bounded with the goal to “fence” (Merriam, 2009, p. 40) in what you intend to study (e.g., student financial aid in a specific university, provider relationships in a particular clinic, leadership in a certain organization). Several methodological features are common in case studies: the collection and analysis of multiple sources of data, evidence of key understanding analysis or clarification of theoretical propositions, and a final report of the case description and themes (Creswell & Poth, 2018; Yin, 2014).
The selection of the case or cases depends on the research questions and purpose. Case study also brings design choices as summarized in Table 1. Cases are typically selected because they represent a phenomenon of interest in what Stake (1995) termed instrumental case studies. For example, Bullough (2015) examined differences and similarities between male and female Head Start teachers by selecting a male/female team in a classroom. The focus of the case was not on teachers themselves but on understanding gender differences and similarities, hence the instrumental case study approach. Another option is to select an intrinsic case that represents a unique or important situation, making the case itself of primary interest (Stake, 1995). Intrinsic case studies may be particularly applicable to program evaluations. For example, Kelley, Snell, and Bingham (2015) conducted a qualitative intrinsic case study of a peer recovery support (PRS) program for substance abuse among American Indian reservation communities. An intrinsic case study was a good fit because of the uniqueness of the program, as they described it, “the characteristics and support of PRS differ from those in non-American Indian communities” (Kelley et al., 2015, p. 274).
Case Study Design Choices.
Yin (2014) has identified four types of case study designs, based on whether it is a single- or multiple-case design and whether it is a holistic (single unit of analysis) or embedded (multiple units of analysis) design. The first choice is between a single case or multiple cases. Investigators might also select multiple cases using a variation of case study design that allows for cross-case comparisons. As illustrated in Figure 2, the second choice is between a holistic case study design that uses global-level unit of analysis (e.g., a program, a school, a clinic) or an embedded case study design that involves units of analysis that come from multiple levels, such as selecting clinics (i.e., the embedded units) within a larger intervention or selecting schools within a district. Consider a case study of a hospital’s approach to providing cardiac arrest care. It may involve individual care team staff as one level of analysis and also the hospital’s overall culture with regard to quality improvement, leadership, education, and training as another level of analysis.
The units of analysis for a holistic and embedded case study design.
Integrating Case Study and Mixed Methods Designs
As noted, researchers have at least two overarching approaches to integrate mixed methods and case study designs. We define a mixed methods–case study design as a research design that employs a “parent” mixed methods design and uses case study for the qualitative component. For example, Ivankova and Stick (2007) researched persistence in an online doctoral program using an explanatory sequential mixed methods design. They began with a quantitative survey to identify factors that predicted persistence—remaining in the program. Based on those results, they purposefully selected four individuals (i.e., the cases) and employed a follow-up qualitative multiple case study “to explain why certain external and internal factors, tested in the first phase, were significant predictors of students’ persistence in the program” (Ivankova & Stick, 2007, p. 97). To apply the case study design types, the qualitative component can be considered a holistic, multiple-case design (Yin, 2014) because of the focus on a single unit of analysis (i.e., the student) and the selection of four cases. Transparency about the qualitative design used within this mixed methods study gave readers a strong sense of the systematic procedures guiding the inquiry. In brief, it was clear that the qualitative phase was not a mere afterthought but robust inquiry.
As a second possibility, we define a case study–mixed methods design as a research design that employs a “parent” case study design and uses mixed methods by collecting, analyzing, and integrating qualitative and quantitative data. For example, Scammon et al. (2013) examined primary care practice transformation within a network of 10 university-owned community clinics using a convergent mixed methods multiple case study design. The complex nature of practice transformation required multiple types of data, so the researchers implemented mixed methods within the case study. Within the case study, they integrated qualitative data (e.g., documents, clinic observations, employee interviews, leader interviews, patient focus groups) and quantitative data (e.g., employee surveys, cost data, operational and clinic data, patient satisfaction surveys) to understand the practice transformation and inform future strategic planning (Scammon et al., 2013). This case study could be considered an embedded, multiple-case design because the data came from multiple levels from patients to employees to the clinic level and because 10 cases were included. The application of a mixed method design provided clarity to the methods of the case study and how the researchers integrated the data sources for a more complete understanding.
Using this background knowledge of mixed methods, case study, and the potential for mixed methods case studies, we undertook a careful review of recently published mixed methods case studies. Our overall purpose of the review was to categorize and identify methodological features to inform recommendations for conducting mixed methods case studies.
Method of Analysis and Review
We conducted a methodological review study, sometimes termed mixed methods research prevalence studies (Molina-Azorin & Fetters, 2016), by selecting 81 peer-reviewed articles to examine contemporary, state-of-the art use of mixed methods case study published in 2011 to 2016. As a caveat, in this article we do not consider teaching cases, which refer to descriptions of situations, organizations, or events meant to illustrate a real-life scenario for teaching purposes (Yin, 2014). Rather, we focused on empirical mixed methods case studies with a research purpose.
Search Strategy
We began the search for articles using Academic Search Premier because it is a multidisciplinary literature database that searches journals across all major disciplines and several databases (e.g., PsychINFO). Search terms were mixed methods AND case study appearing in the abstract for the years 2011 to 2016. It yielded 1,294 articles. We repeated the search in the Educational Resources Information Center (ERIC), which yielded 392 articles, and in PubMed, which yielded 278 health-related articles. The additional searches beyond Academic Search Premier yielded mostly duplicates, which we filtered out. We read each abstract to ensure that (1) the article indicated both qualitative and quantitative data collection and analysis and (2) the use of case study research for either the qualitative component or the use of a parent case study with nested mixed methods research. This review yielded 109 articles. While analyzing the articles, we eliminated 14 that did not include case studies but referenced teaching cases, 8 because they did not report empirical research that was actually conducted but were study protocols of planned research or methodological articles, 4 that were clearly not mixed methods, and 2 that did not contain enough detail to discern what methods were used. A final corpus of 81 articles remained for analysis and broadly represent topics of social sciences, health science, evaluation, and business or management. The full list is available as a supplemental table.
Analysis Procedures
We applied an a priori codebook of general information, case study features, and mixed methods features, as outlined in Table 2. In addition to the a priori codes, we noted additional methodological features that emerged through reading the articles. We compiled the coding with descriptive statistics and analyzed for trends and patterns.
Codebook Used to Code Corpus of Articles Integrating Mixed Methods and Case Study Designs.
Note. MM-CS = mixed methods–case study; CS-MM = case study–mixed methods.
Results
The topics in the reviewed articles included social science (n = 22), health (n = 17), management or business (n = 16), education (n = 14), evaluation (n = 11), and natural sciences (n = 2). The number exceeded our total of 81 because many articles spanned disciplines. One of the articles self-identified as an intrinsic case study, and five identified as instrumental case studies. Although not specified, all remaining articles can be classified as intrinsic case studies. The majority of articles (83%) reflected CS-MM designs that consisted of case studies using nested mixed methods data collection and analysis. However, the integration of the components was typically unclear or not thoroughly addressed. Table 3 summarizes and compares mixed methods features reported for MM-CS and CS-MM articles we examined.
Mixed Methods and Case Study Features Reported (N = 81).
Note. Four studies included multiple strategies for integration.
Case Study–Mixed Methods Design
In our corpus, 67 (83%) articles represented CS-MM designs. Among these articles, we subclassified the core mixed methods design, which further modifies the CS-MM design. We categorized 93% of the articles as convergent though authors actually specified the mixed methods design in 22% of the articles. In two thirds of the CS-MM design articles, the approach to integrating qualitative and qualitative research was unclear. The remaining one third reflected merging integration by bringing results together to compare or relate, and two articles also reflected connecting from one source to sampling for the other. Presentation of integrated qualitative and quantitative results appeared in narrative fashion in 66% of the articles, most often the discussion section, and 7% included a visual joint display. In general, these articles demonstrated fewer mixed methods features, which may reflect the primary case study orientation. However, it indicates that researchers may benefit from employing recent advances in mixed methods methodology, such as procedures for integrating, ways to clearly represent integrated results, and consideration of the value added of mixed methods.
Regarding case study features of the CS-MM designs, 36% reflected a multiple (or collective) case study with the remaining a single case. The most common sampling approach was purposive sampling (54%) often involving multiple levels and units of analysis, and the second most common was other nonprobability sampling, such as snowball or convenience sampling (25%). Authors mentioned saturation in 6% of the articles. Data collection typically involved individual interviews (69%) and open-ended survey items (30%). Given the emphasis of multiple sources of data as a key characteristic of case study (Creswell & Poth, 2018; Stake, 1995; Yin, 2014), we examined whether studies included more than one type of data and found that 25% involved more than one type of qualitative data collection. In some, but not all articles, the researchers relied on quantitative sources as another form of evidence. The analysis approach was most often some type of thematic (65%) or content analysis (16%) method.
Mixed Methods–Case Study Design
In a minority of articles (17%), we found studies that were predominately mixed methods designs that nested a case study for the qualitative component. Interestingly, no predominately health sciences articles fell into the MM-CS design though some were primarily educational research in health contexts. In 50% of MM-CS articles, the authors employed an explanatory sequential core design and noted that the initial quantitative results guided the section of cases. For the MM-CS design articles, the integration approach was clear in about 70% of studies, relative to 33% in the CS-MM design articles. For the MM-CS design articles, the approach to integration was evenly distributed between merging and connecting, with two articles demonstrating more than one integration strategy. The reporting of integrated mixed methods results was clear in most of the articles and typically presented in a narrative way (86%). Two articles included a visual joint display of integrated results. Six (43%) of the MM-CS design articles used a multiple case study design.
In most articles, the description of case study methodology consisted of identifying and perhaps bounding the case, yet the remaining discussion of the qualitative method reflected a more general descriptive qualitative design. The frequencies of sampling approaches for MM-CS design articles was similar to the CS-MM design articles, with most using a type of purposive or other nonprobability approach. Three articles (21%) indicated theoretical sampling, which borrows from grounded theory. Data collection for the MM-CS design articles was primarily through interviews (64%) or open-ended survey questions (29%), and 43% included more than one qualitative data source. Analysis was primarily thematic, though 28% did not identify a specific analysis method. Taking full advantage of case study methodology (e.g., developing case descriptions, specifying theoretical propositions) may enhance researcher’s ability to gain a more complete understanding of their research questions.
Illustrative Articles Integrating Mixed Methods and Case Study Designs
We purposefully selected four articles to illustrate innovative, published research that integrated mixed methods with case studies and spanned disciplines of behavioral, social, and health sciences. We do not claim that any of the four mixed methods case studies are perfect. Our goal for describing these case studies is to give the reader a better sense of applied research involving the approach. The first illustration is a mixed methods case study in the health sciences with a cross-cultural focus on an underserved community. The second illustration is from education and reports mixed methods case study of a unique program to engage higher education faculty in assessment. The third illustration leveraged a mixed methods instrumental case study to address a methodological research question. The final study featured is an example of a multiple case study and includes an innovative visual joint display to represent integration.
Case Study–Mixed Methods to Examine Cross-Cultural Health Care
Little, Motohara, Miyazaki, Arato, and Fetters (2013) conducted a CS-MM evaluation, following a convergent mixed methods design. Their purpose was to evaluate a program to provide prenatal group visits for individuals with limited English proficiency. The case in this single intrinsic case study was defined and bound by participants in the prenatal group visit program. Quantitative data sources included a health questionnaire and a pregnancy distress questionnaire (n = 42). Qualitative data arose from in-depth interviews with patients after delivery of their child (n =20). Integration occurred through merging and presenting a narrative discussion of the quantitative scores with related qualitative themes to more comprehensively evaluate the program.
Case Study–Mixed Methods to Examine Assessment Culture in Higher Education
To examine the phenomena of leadership and culture within higher education assessment, Guetterman and Mitchell (2016) conducted a CS-MM research study of a faculty inquiry program to promote meaningful use of student learning assessment evidence. The case was bound by the inquiry program and its faculty participants. The authors employed a convergent mixed methods design, nested within a case study. The article also illustrates the innovation of including a conceptual model into a CS-MM design. The conceptual model from Kezar (2013) was based on the influence of culture, leadership, and organizational policies on assessment practices and guided the intervention, data collection, data analysis, and interpretation. Quantitative data collection consisted of instruments with evidence of validity and reliability to measure faculty assessment attitudes and knowledge, organizational characteristics, and information characteristics. Qualitative data sources included open-ended survey items, narrative responses from participants, and posters (as documents) developed by participants in the program. Integration occurred through merging qualitative themes with related faculty survey responses in a narrative discussion. In this article, the authors included a section in the results, labeled mixed methods integration, in which they discussed their results using a merging approach to compare qualitative and quantitative results. Reflecting on what they learned, they noted, “. . . qualitative data then revealed the contextual nuances, indicating that, when faculty members view assessment beyond the scope of meeting accreditation, they see additional possibilities and learn from the assessment in other ways” (Guetterman & Mitchell, 2016, p. 54). The research provided the authors a better understanding of what worked well and not so well in the program to develop recommended best practices for universities organizing for assessment.
Case Study–Mixed Methods to Investigate a Methodological Question
Onwuegbuzie and Leech (2010) used a CS-MM design to examine generalizations beyond the sample made in qualitative research by analyzing published articles with a journal, The Qualitative Report. This exemplar is innovative in many ways, including the application of CS-MM to a methodological topic and the use of a sequential mixed methods design. They began with quantitative analysis and then used the results to inform the qualitative analysis, which is consistent with Creswell and Plano Clark’s (2018) explanatory sequential design. Onwuegbuzie and Leech (2010) identified their research as an instrumental case study because their focus was gaining insight into generalization practices rather than the cases themselves. Onwuegbuzie and Leech (2010) described their mixed methods design as fully mixed and explained it as “qualitative and quantitative research approaches were mixed within and across several stages (i.e., research formulation, planning, and implementation) of the research process” (p. 884). The case study design was an embedded, multiple-case study design with two levels: the journal and the articles within the journal. Quantitative data focused on the frequency with which articles contained generalizations, and qualitative analysis focused on cross-case comparisons of the extent or level of generalizations made.
Mixed Methods–Case Study of Health Clinics That Incorporates a Visual Joint Display
Shaw et al. (2013) evaluated a quality improvement intervention consisting of learning collaborative and facilitated team meetings developed to improve cancer screening rates in clinics. Although not described as such, the study involved case study nested within a convergent mixed methods intervention design. The investigators conducted a randomized trial but found no statistically significant improvement in screening rates. However, qualitative research provided insight into implementation practices, leadership, and staff experiences. Specifically, they developed descriptive case summaries of the intervention group. The investigators integrated through merging the qualitative implementation characteristics and quantitative screening rate results. This article includes an exemplar joint display, which is a way to represent mixed methods integration through visual means (Guetterman, Fetters, & Creswell, 2015). The joint display included a row to represent each of the 12 clinics (i.e., cases) along with columns of qualitatively derived implementation characteristics. Each cell indicated the degree (strong, moderate, weak) of that implementation characteristic. Next to the qualitative results, they presented a column with pre–post quantitative measures. Thus, in each row, the reader gains an understanding of each case both quantitatively and qualitatively. Their narrative discussion of qualitative findings was then organized around this display. The Shaw et al. (2013) research also illustrates connecting integration. Because high-performing implementation did not always equate to improved screening rates, they selected and developed three case studies to better explain the relationship between implementation characteristics and outcomes. Thus, the overall research represented a mixed methods study, but the investigators then employed a follow-up case study to explain results as seen in an explanatory sequential mixed methods design.
Discussion
Investigators integrating mixed methods and case study designs have options regarding the design, the qualitative and quantitative methods used, and the integration approach. Awareness of these options can help investigators better harness the value added from mixed methods case studies, provided these choices are made with careful attention to rationale and research questions.
One set of options at the design level involves whether to nest case studies within mixed methods or to nest mixed methods within a case study. Researchers conducting MM-CS might consider case study for the qualitative component and more clearly delineate its design. With increased calls for transparency in reporting research (American Educational Research Association, 2006), it is important to specify whether a qualitative approach (e.g., case study, grounded theory, phenomenology, narrative research, ethnography) was used for the qualitative component and procedures. Our results indicate that case study is a viable and productive option to integrate with the mixed methods designs. For CS-MM designs, considering the mixed methods design and related elements, such as the sequence of quantitative and qualitative data collection and analysis, may help researchers better address research questions and achieve objectives.
Not surprisingly, given the relative nascence of the field, our review indicates a general paucity of mixed methods features (e.g., consideration of the mixed methods design, integration, use of joint displays), especially for the CS-MM designs. For those conducting a case study and nesting mixed methods within it, the growing scientific advances around mixed methods (Creswell, 2015b) provide guidance for conducting rigorous qualitative and quantitative research and the integration of the two. Failure to achieve meaningful integration can affect the quality of inferences generated and simply does not harness the value added of mixed methods research. The mixed methods designs are not meant to restrict researchers but to help them plan for meaningful integration.
Another methodological option is whether to select one or more than one case. Availability of resources is a factor to consider, but when possible, selecting more than one case typically yields more insight (Yin, 2014). Conducting cross-case comparisons of sites (e.g., schools, clinics), investigators can gain a better understanding of common and unique features of each case. However, most studies in our review were based on a single case.
From a mixed methods perspective, integrating quantitative research into case studies can reveal broader trends, statistical relationships, and generalizable inferences as long as the study has adequate sampling and a logical design. Those realizations are only achievable through systematic integration of qualitative and quantitative methods to generate meaningful mixed methods meta-inferences. Table 4 summarizes the potential value added of integration for both CS-MM and MM-CS designs. Mixed methods designs rely on the assumption that integrating the two forms of research yields a more complete understanding. CS-MM research can benefit from systematic integration procedures. For example, merging qualitative and quantitative data within a case can yield insight into the extent to which the two confirm, contradict, or relate to one another. Or, as in Shaw et al. (2013), integration can yield more meaningful cross-case comparisons based on both process and outcome data. Another integration option is connecting as one form of data informs the sampling of the other. The results of a quantitative phase might inform case selection to best explain results as in the Ivankova and Stick (2007) MM-CS of persistence in doctoral education. Another possibility is to use mixed methods for the selection of the cases in a multiple case study. For instance, Sharp et al. (2012) conducted a longitudinal multisite mixed methods case study of statewide education policies. They used a series of qualitative and quantitative analyses to narrow their final case selection. Finally, building integration provides an option to develop instruments, identify variables, or inform an intervention. For example, an evaluator conducting a CS-MM of a program might benefit from using qualitative findings to build outcome measures and instruments. Finally, including visual joint displays to represent integration presents an opportunity to clearly communicate inferences and the value of integrating mixed methods and case study. Developing a joint display, such as a table of integrating results, can make the integration clearer for the reader.
Value-Added of Integration in Case Study–Mixed Methods (CS–MM) and Mixed Methods–Case Study (MM–CS) Designs.
Adaptations of Case Study or Mixed Methods Approaches
Our review indicates that investigators conducting mixed methods case studies often mix units of analysis and sources, which challenges conventional advice to maintain parallel units of analysis in convergent designs (e.g., Creswell, 2015a; Lieber & Weisner, 2010). Perhaps, an adaption needed to conduct mixed methods nested within a case study is to relax the recommendation for parallel levels of analysis. After all, embedded case studies by definition have multiple levels of analysis to give a more complete understanding of the case at particularistic and global levels. Researchers conducting a convergent mixed methods study might consider a case study for the qualitative component (i.e., an MM-CS design) and remember that the case can be bound to mirror the quantitative component including embedded units of analysis.
Related to units of analysis, sampling may be a unique consideration in mixed methods case studies that requires adaptation. Sampling can be considered at two hierarchical levels: (1) choosing the case and (2) choosing participants or units within the case, as in the Onwuegbuzie and Leech (2010) study that involved the section of the journal at one level and then articles within it as a second level. Sampling becomes more complex and requires transparency at both levels. A rationale for selecting the case and strategy for sampling are both necessary when integrating mixed methods and case studies.
Finally, a potential adaptation involves the use of multiple data sources. For an MM-CS, collecting multiple sources of data is imperative for a robust case study (Yin, 2014), yet less than half of the MM-CS articles in our review included multiple qualitative data sources. The situation, however, may be somewhat different for mixed methods nested within a case study in which the multiple data sources can include qualitative and quantitative data. Our analysis revealed that 75% of CS-MM articles included only one qualitative source. An innovative application is to mix multiple quantitative and qualitative sources. Researchers can leverage the value added by mixed methods to gain a more comprehensive understanding of the case in ways that only qualitative sources might not.
Recommendations for Designing Mixed Methods Case Studies
Based on our analysis of features reported in the articles reviewed and cross-case comparisons of the four illustrative case studies, we recommend investigators to
Determine whether to conduct a parent case study that employs mixed methods within it or a parent mixed methods study that employs case study for the qualitative component.
Determine a rationale for selecting the case or multiple cases. Consider that multiple cases tend to yield more evidence.
Articulate a sampling rationale for selecting cases and use convenience or snowball methods with caution.
Describe the case and the boundaries—what is and what is not included.
Collect multiple sources of data.
Be certain to review (and perhaps cite) both mixed methods literature and case study literature.
Carefully consider your approach to mixed methods integration.
Clearly report integrated mixed methods results.
Conclusion
This review highlights several key issues. First, a key contribution of this study to the mixed methods literature is the clarification of the MM-CS and CS-MM designs. Second, it begins to address the gap in guidance for conducting mixed methods case studies. Case study scholars have suggested incorporating mixed methods, and mixed methods scholars have suggested conducting case studies, but little is written about the approach. Second, in our experience, reviewers and editors are increasingly requiring a qualitative design specified in mixed methods publications and proposals. We completely agree with requests to provide that detail and transparency but have found little written about how to conduct a mixed methods case study. Third, a compilation of recent mixed methods case studies and analysis of their features provides a basis to suggest best practices in conducting mixed methods case study. Fourth, the article presents four exemplar case studies to help investigators see this research approach in practice. Overall, we encourage researchers to become more aware of methodological choices and clearly document methodological procedures.
Research integrating mixed methods and case study will benefit from rigorous and sophisticated qualitative and quantitative methods. Perhaps related to its growth from the qualitative tradition, we found that many CS-MM articles reviewed did not address integration in detail. Qualitatively oriented researchers might consider collaborating with a mixed methods and quantitative researcher to leverage their skillsets.
Supplemental Material
Supplemental_Table_1 – Supplemental material for Two Methodological Approaches to the Integration of Mixed Methods and Case Study Designs: A Systematic Review
Supplemental material, Supplemental_Table_1 for Two Methodological Approaches to the Integration of Mixed Methods and Case Study Designs: A Systematic Review by Timothy C. Guetterman and Michael D. Fetters in American Behavioral Scientist
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Author Biographies
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
