Investigating Tool Support for Decoding Conversations

Investigating Digital Tool Use for Analyzing Talk:
Challenges and Opportunities for Designing Transdisciplinary Tools

2019 - 2021
Human Computer Interaction Institute

Conceptualized and led the research project, conducting in-depth interviews with experienced in-person conversation researchers to investigate their analytical workflows, identify key challenges, and uncover actionable insights to inform tool development and enhance research effectiveness and efficiency.

[pdf (most recent v.)]

Researchers across diverse fields analyze face-to-face conversation data to uncover how language conveys meaning and to identify patterns tied to critical outcomes in health, education, and social interaction. Despite the availability of numerous transcription and analysis tools, little is known about how researchers actually use these tools in practice. Through in-depth interviews with 16 experienced researchers, we identified three core challenges: insufficient support for fluidly exploring data at different levels of granularity, the labor-intensive demands of manual annotation and transcription, and the absence of tools for uncovering broader, meaningful patterns beyond fine-grained, domain-specific analysis. Existing tools were described as either overly complex, with steep learning curves for detailed analysis, or overly simplistic, offering limited insights beyond basic quantitative metrics. Participants expressed a strong desire for automated solutions to streamline repetitive tasks and user-friendly tools that enable holistic, intuitive exploration of both verbal and nonverbal features. These findings underscore the need for user-centered tools that combine machine intelligence with human expertise, fostering a more effective and accessible ecosystem for conversation research across disciplines.