2026-04: When Regression Isn’t Enough: Modeling Romantic Relationships Data

Why we picked it

**When Regression Isn’t Enough: Modeling Romantic Relationships Data** Regression is a core tool in data science, but it can break down when data are structured, such as repeated measures or when observations are nested within groups. In these cases, standard models can produce biased or misleading results. In this talk, Kristina introduces hierarchical linear models (HLMs) as a practical solution to these challenges. Using real data from research on romantic relationships, she will demonstrate how she tested the research question: "which types of shared relationship experiences are associated with heightened satisfaction for different types of individuals?" The session will cover the benefits of HLMs, how they extend from regression, and how to interpret their results. This talk is intended for those who want to deepen their understanding of HLMs, improve their interpretation of regression-based analyses, or simply want to hear about findings from recent relationship science. **Where