Formal mixed methods research designs are relatively new in social science research. According to Creswell and Plano Clark (2010), a mixed methods design collects and analyzes both quantitative and qualitative data and mixes the analyses one or more of three ways: (1) the datasets can be merged into a cohesive whole, (2) the results of one can build on the other, or (3) one dataset might be embedded in the other. Furthermore, Morse (2003) points out that mixed methods designs characteristically integrate methods that are not normally used together, such as embedding open-ended questions within Likert scale instruments. By using different types of data and analyses in a study, researchers can gain a greater depth of understanding than by using either method on its own, or, as Jick (1979) states, “Where there is convergence, confidence in the results grows considerably…However, where divergent results emerge, alternative, and likely more complex, explanations are generated” (p. 608).
Creswell and Plano Clark (2010) contend that mixed methods designs can be very effective because of the possibility of triangulating data and results. For example, if the qualitative analysis of interview transcripts can be used to corroborate the quantitative results of a survey, then the researcher has a stronger base of evidence upon which to build an argument, which can increase the validity of the mixed results. Additionally, if the qualitative and quantitative analyses yield contradictory findings, the researcher may uncover hidden complexities or be able to formulate new research hypotheses to resolve the contradiction.
Visual model of the research design.
Due to the complexity of many mixed methods designs, Creswell and Plano Clark (2010) recommend that researchers provide a visual model of their particular design. The design used in this investigation, as shown in Figure 6, was a 2-phase QUAN/QUAL concurrent triangulation model (Creswell, 2009, p. 213). The rationale for using the mixed methods approach is that the results of the two forms of data analysis could be compared and merged into an integrated analysis which would be stronger than if either a quantitative or qualitative analysis was performed in isolation. This comparison of analyses is known as triangulation or sometimes as a convergent design (Creswell & Plano Clark, 2010).
Figure 6. Visual Model of Research Method