ESRA 2027 sessions by theme
Back to Overview of Sessions
Human and Algorithmic Judgements: Survey Experiments for Measuring and Auditing Inequality |
| Coordinator 1 | Professor Maria Gabriella Campolo (University of Messina) |
| Coordinator 2 | Professor Giovanni Busetta (University of Messina) |
| Coordinator 3 | Professor Orietta Nicolis (University of Messina and University of Valparaíso) |
Session Details
Survey experiments have become a central tool for examining how individual characteristics and contextual cues shape decisions in labour markets, housing, education, family life, and access to services. At the same time, artificial intelligence systems are increasingly used to support, simulate, or automate similar evaluations. This creates a need for rigorous designs that compare human and algorithmic judgements and assess whether AI reproduces, amplifies, attenuates, or conceals social inequalities.
This session invites methodological and substantive contributions using factorial surveys, vignette and conjoint experiments, synthetic respondents, large language models, or other AI-based systems to study inequality and discrimination. We particularly welcome papers that compare human responses with algorithmic outputs; audit AI tools across gender, ethnicity, migration background, age, parenthood, social class, disability, or place; examine prompt sensitivity, model instability, replicability, and external validity; or combine survey experiments with administrative, qualitative, digital-trace, or field-experimental data.
Contributions may focus on recruitment, wages and career progression, housing allocation, migration, fertility and family decisions, education, wellbeing, digital exclusion, or other domains in which individuals and groups are evaluated. Papers addressing the methodological status of AI-generated responses are especially encouraged: under what conditions can synthetic respondents be used for questionnaire pretesting, hypothesis generation, benchmarking, or comparison with human survey data, and where do they fail?
The session aims to bring together survey methodologists, social statisticians, computational social scientists, and substantive inequality researchers. Its broader goal is to strengthen standards for the transparent, reproducible, and ethically accountable use of AI within survey-based research.
