Search

Homophily is both a principle for social group formation with like-minded people as well as a mechanism for social interactions. Recent years have seen a growing body of management research on homophily particularly on large-scale social media and digital platforms. However, the predominant traditional qualitative and quantitative methods employed face validity issues and/or are not well-suited for big social data. There are scant guidelines for applying computational methods to specific research domains concerning descriptive patterns, explanatory mechanisms, or predictive indicators of homophily. To fill this research gap, this paper offers a structured review of the emerging literature on computational social science approaches to homophily with a particular emphasis on their relevance, appropriateness, and importance to information systems research. We derive a research taxonomy for homophily and offer methodological reflections and recommendations to help inform future research. [ABSTRACT FROM AUTHOR]

This research brings together a systematic literature review, topic modelling, and qualitative analysis in a novel, mixed method approach to synthesize research on workplace wellness interventions and offer directions for future research. First, we conducted a systematic literature review to create a collection, or corpus, of 5,674 research publications on wellness intervention research. Next, topic modelling, a quantitative machine learning technique that identifies key topics, was deployed to extract and investigate trends within the corpus. Finally, abstracts were qualitatively coded according to a framework that examined the trends across: 1) which journal outlets and disciplines are publishing workplace wellness intervention research; 2) the different types of wellness interventions that are studied in workplaces; 3) dependent variables measured as outcomes of wellness interventions, and any reported mechanisms or program theory linking the intervention to the outcomes; 4) the methodologies utilized by researchers and their corresponding rigour; and 5) the context the research takes place within. Interestingly, despite the growing recognition that wellness is a multi-dimensional concept, 'wellness' is increasingly operationalized using primarily psychological outcomes. Other trends depicted a lack of consistency in research design, program theory, and results making it difficult to specify how or why a program is beneficial for a specific population and setting and the extent to which results were reliable. [ABSTRACT FROM AUTHOR]