This paper estimates determinants of long-run growth rates of GDP per capita in a cross section of countries. We propose a novel Measurement Error Model Averaging (MEMA) approach that accounts for measurement error in international income data as well as model uncertainty. Estimating the model using eight vintages of the Penn World Tables (PWT) together with other proposed growth determinants, we identify 18 variables related to economic growth. In addition, the results are robust to allowing for outliers in the form of heteroscedastic model errors.