Measuring women’s agency with women’s own narratives through open-ended interviews is considered by many to be the measurement gold-standard. Yet data processing of such qualitative data relies on hand-coding, presenting limitations related to budget, time, and discretion of annotators. We propose a text-analysis measure of women’s agency, drawing on text-to-speech transcripts of women’s answers to open-ended questions. This method adapts the Franzosi (1994, 2004) methodology for analyzing semi-structured interviews, attributing an automated agency score to women’s narratives about how they make decisions, how they live their life and how this compares to their preferences. We recommend the use of this tool in the design, monitoring, and evaluation of development programs aimed at improving women’s empowerment.
Link to: tool content and guidelines, CTO file, and statistical annex
Duration: This tool takes on average 9.2 minutes to be implemented.
Permitted use of the tool: all users are free to use the tool with citation: “MAGNET (2023). Algorithmic Assessment of Agency. http://magnet.ifpri.info/algorithmic-assessment-of-agency/”
Tool Sample
Topics: major household purchase, household income.
- Now I would like to ask you questions about the last time that a decision about a major household purchase was made in your household. I will be recording the conversation if that’s ok. I’ll give you a moment to think of a recent example.
- Have you thought of a recent major expenditure?
- If yes, continue.
- If no, let respondent think.
- What was it?
- Tell me about how the decision was made?
Measurement properties
- Geographies Tested: India, Uganda
- Populations included: Female
- Age range: Adults, Adolescents
For details on these testing metrics, visit our scoring methodology page.
International Food Policy Research Institute