Proposal for novel mixed method research integrating the epistemology and methodology of Bayesian statistics

Abstract

This study proposes a novel mixed method research approach that integrates the epistemology and methodology of Bayesian statistics (including application in AI). To reframe each research paradigm: quantitative research (frequency theory) is positioned as hypothesis testing by deduction (theory driven); qualitative research as hypothesis generating via abduction and self-validation through qualitative induction (data interactive); and Bayesian statistics as hypothesis selection through quantitative induction (data driven). The proposed mixed method approach involves the following steps: 1) Designing the overall research framework and building the database using qualitative research; 2) Applying Bayesian statistics to the database to identify the most plausible model and narrow down key hypotheses and variables; 3) Testing these key hypotheses through quantitative research and summarizing the overall conclusions; 4) Feeding significant findings back into the research project to iteratively refine and restart the process.