Hidden Harmony

It is difficult to reach a consensus on how to measure a social construct such as the harmony level of a group of people. To circumvent the difficulty, we develop ``latent'' binary quantile regression for settings in which the binary regressand is unobserved and proxied by multiple indicators (e.g., separate reports of whether a group is harmonious). We demonstrate how to identify and estimate parameters for conditional quantiles of the hidden outcome (e.g., group harmony level), prove the strong consistency of the estimator, and run Monte Carlo experiments to verify its finite-sample performance. Based on such a general-purpose method, we use high-quality survey data to uncover factors affecting the harmony levels within college dormitory rooms. The application is NASA-inspired: a scientific understanding of small group harmony improves the selection process for long-duration spaceflight crews. Among other findings, we discover that sleeping schedule discordance damages relationship