This paper considers the possible effects of making inferences about individuals from aggregate data. It assumes a knowledge of regression analysis, and explores the utility of techniques designed to make the inferences in causal modelling more reliable, including a comparison between ecological regression models and ecological correlation.

Professor Langbein teaches quantitative methods, program evaluation, policy analysis, and public choice. Her research fields include: theories of bureaucratic discretion, productivity, principal-agent models, social capital, and cooperation in the workplace; theories of influence of interest groups in Congress and the bureaucracy; empirical applications in various policy areas, including the environment, education, defense, housing, criminal justice (death penalty and police), and corruption. Her articles have appeared in numerous journals on politics, economics, policy analysis and public administration. Her most recent publications examine the consequences of varying levels of discretion in federal agencies, and the World Bank's measurement of corruption in countries.