My dissertation utilizes Bayesian Model Averaging (BMA) to examine model uncertainty in cross-country datasets. When there is theoretical support for a large number of explanatory variables in an area of study, the exact nature of the "true" model is often unknown. BMA is able to test a large variety of different specifications in order to determine which potential regressors are most likely part of the underlying true model. The first chapter of the dissertation examines whether an Environmental Kuznets Curve (EKC) exists in sulfur dioxide emissions, and also tests the importance of certain macroeconomic, institutional, and climatic variables in determining sulfur dioxide concentrations. The study finds only weak evidence of an inverted U-shaped relationship between income and pollution that the EKC suggests. Once oversampling from industrialized countries is corrected for even a weak EKC fails to hold, and other macroeconomic variables (such as trade), institutional variables (such as executive constraints), and climatic variables (such as precipitation variation) have the most explanatory power. The second chapter also allows for model uncertainty and tests for parameter heterogeneity in several factors which have been theoretically and empirically linked to East Asian growth. Primary schooling and labor market flexibility are found to be particularly important for East Asian growth, while other growth determinants such as foreign direct investment and international trade do not appear to have a disparate effect on economic growth in the region compared to the rest of the world. The last chapter of the dissertation tests for multiple regimes in cross-country growth due to differences in initial conditions between countries. Specifically, the study employs BMA to allow for parameter heterogeneity in a large suite of explanatory variables due to differences in initial income and human capital. Initial income and human capital are found to be important