Structural topic modeling as a mixed methods research design: a study on employer size and labor market outcomes for vulnerable groups

Abstract

AbstractObtaining and maintaining steady employment can be challenging for people from vulnerable groups. Previous research has focused on the relationship between employer size and employment outcomes for these groups, but the findings have been inconsistent. To clarify this relationship, the current study uses structural topic modeling, a mixed methods research design, to disclose and explain factors behind the association between employer size and labor market outcomes for people from vulnerable groups. The data consist of qualitative interview transcripts concerning the hiring and inclusion of people from vulnerable groups. These were quantitized and analyzed using structural topic modeling. The goals were to investigate topical content and prevalence according to employer size, to provide a comprehensive guide for model estimation and interpretation, and to highlight the wide applicability of this method in social science research. Model estimation resulted in a model with five topics: training, practicalities of the inclusion processes, recruitment, contexts of inclusion, and work demands. The analysis revealed that topical prevalence differed between employers according to size. Thus, these estimated topics can provide evidence as to why the association between employer size and labor market outcomes for vulnerable groups varies across studies––different employers highlight different aspects of work inclusion. The article further demonstrates the strengths and limitations of using structural topic modeling as a mixed methods research design.