AI Insight
This article addresses limitations in mixed methods research by proposing an expanded framework for meta-inferences that goes beyond simply comparing quantitative and qualitative results. The authors demonstrate how meta-inferences can be used to generate actionable program recommendations and implications, particularly in evaluation contexts. They provide three examples from an ongoing prevention program evaluation showing how joint displays can integrate multiple data types to produce practical insights for program improvement.
Why it matters
This approach enhances the practical utility of program evaluations by providing evaluators with structured methods to translate mixed methods findings into concrete recommendations. By expanding the conceptualization of meta-inferences, the framework gives researchers more tools to address complex research questions and deliver real-world actionable insights that can directly inform program decisions and improvements.
Understand the Science
Integration is a defining feature of mixed methods research, and the generation of meta-inferences is the overarching outcome of a mixed methods study. Despite their role in demonstrating the added value of mixed methods research, meta-inferences are frequently limited to determining fit between the results, through a comparison of quantitative and qualitative results. This article builds on a small but growing literature that has expanded conceptualization of meta-inferences by proposing an additional type that generates program recommendations. We advocate enlarging this framework to include other types of meta-inferences to give researchers additional options for addressing research questions. Thus, the aim of this paper is to demonstrate an underutilized type of meta-inference that generates program implications and recommendations. Because the goal of evaluation is the utilization of results to make program decisions, this example provides an opportunity to showcase the value of meta-inferences as applied, real world, actionable insights. We provide three examples of generating meta-inferences that focus on identifying implications and program recommendations using joint displays from an ongoing evaluation of a prevention program.