SODAS Data Discussion III (Fall 2026)
SODAS Data Discussion with Mikkel Stahlschmidt and Jeppe Johansen.
Copenhagen Center for Social Data Science (SODAS) aspirers to be a resource for all students and researchers at the Faculty of Social Sciences. We therefore invite researchers across the faculty to present ongoing research projects, project applications or just a loose idea that relates to the subject of social data science.
Two researchers will present their work. The rules are simple: Short research presentations of ten minutes are followed by twenty minutes of debate. No papers will be circulated beforehand, and the presentations cannot be longer than five slides.
Discussion 1
Presenter: Mikkel Stahlschmidt
Title: What Do We Gain (and Lose) from Responsive Survey Questions?
Abstract:
We implement a randomized controlled trial that assigns participants to one of four interview modes varying in interface and degree of adaptive probing: (i) a conventional survey with standardized follow-ups, (ii) a chat interface with the same standardized questions, (iii) a chat-based AI interview with a single adaptive follow-up probe per item, and (iv) a chat-based AI-interview allowing multiple adaptive follow-ups.
Given our setup, we can estimate the causal effects of interface and adaptivity. We evaluate outcomes along five dimensions of qualitative data quality: exposure, specificity, relevance, interpretative confidence, and interpretative gain. We also measure cross-interview comparability and attrition.
Discussion 2
Presenter: Jeppe Johansen
Title: LASCO: A Programming Language for Regulation and Legal Texts
Abstract:
Public administration increasingly runs on software, and jurisdictions such as Denmark now require legislation to be "digitisation-ready." Yet law is drafted in prose and implemented in code separately, often years apart and by different people, which both breeds errors, inconsistencies, and costly rework, and scales poorly as the volume of regulation grows.
This talk argues for developing the executable code of a rule alongside its text and introduces LASCO, a domain-specific functional programming language in which each provision retains its verbatim statutory text alongside its logic. The talk motivates how insights from programming languages can improve the way regulation is conceptualized and implemented, while focusing in depth on two challenges. First, how can legal texts that require interpretation be handled in a programming language? Second, how can a programming language for regulation be made maintainable, particularly when provisions are added, amended, or removed?
