The Methods-Inference Map: Visualizing the Interactions Between Methods and Inferences in Mixed Methods Research

Introduction

In its simplest form, mixed methods research involves two acts of data collection and data analysis, commonly called “research strands”, one including qualitative data collection and data analysis, and one including quantitative data collection and data analysis (Bryman, 2008; Creswell & Plano Clark, 2018). Consequently, mixed methods researchers must develop their design at two levels. They must design each individual research strand such that its research question can be answered using the chosen sample, methods of data collection, and methods of data analysis. They must design the study as a whole such that its individual research strands together answer the overall research question and achieve the overall research goal. Thus, mixed methods researchers must attend to both their individual research strands and their study as a whole. This makes mixed methods research a complex endeavour (Poth, 2018).

To get a better grasp of such complexity, mixed methods researchers frequently use visualizations to support combination and integration in all phases of their study (Shannon-Baker & Edwards, 2018). Combining or integrating the findings of individual research strands is supported by joint displays (Guetterman, Creswell et al., 2015; Guetterman, Fetters et al., 2015), visualizations that combine qualitative and quantitative elements (Creamer, 2020; Fetters, 2020; Plano Clark & Sanders, 2015; Wheeldon, 2010). Joint displays can also support combining and integrating research strands in data analysis (Bazeley, 2018; Fetters, 2020); interactive joint displays support the joint analysis of quantitative and qualitative data (Schoonenboom & Johnson, 2021; see also Chapter 25). For a discussion of using visuals to teach and learn about mixed methods research, see also Chapter 31.

Visualizations can also support mixed methods design. For example, Maxwell's (2013) interactive model of research design enables researchers to reflect the fit of their study's main components. For a further discussion, see also Chapter 2. Other visualizations show the study's design flow—that is, how the research strands follow after each other (Ivankova & Stick, 2007; Morse & Niehaus, 2009; Tashakkori & Teddlie, 2003). Furthermore, visualizations of a selection of elements for all separate research strands can be used to reflect the links between these elements (Bazeley, 2018; Miles & Huberman, 1994; Plano Clark & Sanders, 2015). Such visualizations have proven their worth; the Journal of Mixed Methods Research now requires submissions of an empirical study to include a joint display of its findings and a flow diagram of the data analysis and data collection procedures (Fetters & Molina-Azorin, 2019).

Visualizations differ in what they can and cannot show about a research design. Some visualizations [Page 115]present a static overview of the study, whereas others display a study's research flow. Some make a visual distinction between qualitative and quantitative components, whereas others do not. They also differ widely in what they show about the study's structure. Finally, only some visualizations display integration points between the qualitative and quantitative components, such as the meta-inference, the overall conclusion at the end of a mixed methods study. No visualization displays all these elements. Consequently, researchers who switch from an overall view of their study to the details of their separate quantitative and qualitative research strands also have to switch visualizations.

...