From patterns to meaning: a mixed-methods framework that integrates computational and qualitative text analysis

Abstract

Applications of automated text analysis techniques have gained significant attention in recent years. Automated text analysis methods are transforming research by offering valuable tools for robust textual analysis and data-driven theory-building. These methods’ impacts on the rigor of qualitative text analysis remain under-researched. We use an advanced automated text analysis method, structural topic model (STM), to assess whether or not a mixed-methods approach that combines automated with qualitative text analysis improves analytical rigor by focusing on four widely recognized criteria: credibility, transferability, dependability, and confirmability. Based on these criteria, our study shows how the combination of automated text analysis and qualitative text analysis can enhance analytical rigor. Ultimately, our research argues that automated techniques are powerful tools in a mixed-methods framework, effectively complementing qualitative-interpretive analysis without replacing the insights and understanding provided by human researchers. We conclude by discussing the implications for future studies employing this mixed-methods approach.