Data‐driven or structure‐driven: Toward a process model of data behavior in humanities research

Abstract The emergence of digital humanities has transformed how scholars engage with research materials, yet data behaviors of humanities scholars remain underexplored compared to those in science and social science fields. This study develops a comprehensive process model of data behavior in humanities research and examines contextual factors shaping these behaviors. Through in‐depth interviews with 27 humanities scholars across multiple disciplines, we identified two distinct research approaches—data‐driven and structure‐driven—that fundamentally shape how scholars interact with data throughout their research processes. Data‐driven scholars aim to discover patterns from comprehensive data collection, while structure‐driven scholars engage in iterative interpretation and re‐structuring informed by selective data. Our findings reveal diverse data behaviors including processing, analysis, and sharing practices that vary significantly between these approaches. We identified three categories of contextual factors influencing these behaviors, including internal factors, social and organizational factors, and infrastructural factors. The study contributes a comprehensive model and reveals that humanities scholars' data behaviors are embedded in their ways of sensemaking. The study offers implications for research infrastructure development, tool design, and academic reward systems that can better support evolving scholarly practices in the age of AI.