Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

The Shape of Agency: Designing for Personal Agency in Qualitative Data Analysis

Luka Ugaya Mazza, Plinio Morita, James R. Wallace

TL;DR

Qualitative researchers face the dual challenge of expanding data access while maintaining ownership and autonomy over analysis. The authors adopt a design science approach to build a data-visualization prototype that enables computational thematic analysis without programming, using personal agency as a design lens. They identify core factors—trust, delegation, and guided choice—that shape adoption, and show how an interactive, transparent interface can preserve researcher autonomy while leveraging automation. The work provides practical guidance for designing visualization-centered, human-centered AI tools that augment qualitative analysis rather than replace researchers, with implications for HCI practice and future interactive ML research.

Abstract

Computational thematic analysis is rapidly emerging as a method of using large text corpora to understand the lived experience of people across the continuum of health care: patients, practitioners, and everyone in between. However, many qualitative researchers do not have the necessary programming skills to write machine learning code on their own, but also seek to maintain ownership, intimacy, and control over their analysis. In this work we explore the use of data visualizations to foster researcher agency and make computational thematic analysis more accessible to domain experts. We used a design science research approach to develop a datavis prototype over four phases: (1) problem comprehension, (2) specifying needs and requirements, (3) prototype development, and (4) feedback on the prototype. We show that qualitative researchers have a wide range of cognitive needs when conducting data analysis and place high importance upon choices and freedom, wanting to feel autonomy over their own research and not be replaced or hindered by AI.

The Shape of Agency: Designing for Personal Agency in Qualitative Data Analysis

TL;DR

Qualitative researchers face the dual challenge of expanding data access while maintaining ownership and autonomy over analysis. The authors adopt a design science approach to build a data-visualization prototype that enables computational thematic analysis without programming, using personal agency as a design lens. They identify core factors—trust, delegation, and guided choice—that shape adoption, and show how an interactive, transparent interface can preserve researcher autonomy while leveraging automation. The work provides practical guidance for designing visualization-centered, human-centered AI tools that augment qualitative analysis rather than replace researchers, with implications for HCI practice and future interactive ML research.

Abstract

Computational thematic analysis is rapidly emerging as a method of using large text corpora to understand the lived experience of people across the continuum of health care: patients, practitioners, and everyone in between. However, many qualitative researchers do not have the necessary programming skills to write machine learning code on their own, but also seek to maintain ownership, intimacy, and control over their analysis. In this work we explore the use of data visualizations to foster researcher agency and make computational thematic analysis more accessible to domain experts. We used a design science research approach to develop a datavis prototype over four phases: (1) problem comprehension, (2) specifying needs and requirements, (3) prototype development, and (4) feedback on the prototype. We show that qualitative researchers have a wide range of cognitive needs when conducting data analysis and place high importance upon choices and freedom, wanting to feel autonomy over their own research and not be replaced or hindered by AI.

Paper Structure

This paper contains 26 sections, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Agency
  4. Overview of Research Process
  5. Specifying Needs and Requirements
  6. Recruitment Criteria
  7. Interview protocol
  8. Analysis Method
  9. Results
  10. Trust
  11. Delegation
  12. Guidance
  13. Development of Digital Paper Prototype
  14. Feedback on Prototype
  15. Autonomy, Delegation, and Transparency
  16. ...and 11 more sections

Figures (3)

  • Figure 1: Our Design Science Research process followed a series of four phases: 1) Problem Comprehension, 2) Specifying user needs and requirements, 3) Development of prototype, and 4) Testing of prototype.
  • Figure 2: Participant Sketches captured during our interviews. Participants used a variety of techniques to group (1,2) and mind map (3,4) during their qualitative research practices.
  • Figure 3: Our prototype interface. We used this prototype as a design probe, and discussed how it would support qualitative research with our participants.