In this seminar–the first in Statistical Horizons' Distinguished Speaker Series–Professor Gelman will share his perspectives on the use of multilevel modeling to address general challenges of design and analysis in a world of nonrandom samples, varying treatment effects, and noisy data.

The seminar will include a mix of examples, methods, discussion, and time to ask questions, as Professor Gelman explores such issues as:

• Using a conceptual framework for thinking about inference and generalization in the context of multiple sources of variation.
• Multilevel modeling as it applies to sampling and causal inference.
• Open questions in multilevel modeling and generalization.