We examine the bias resulting from temporal and spatial aggregation of weather variables in environmental economics. In order to include temporally and/or spatially continuous environmental variables (such as temperature and precipitation), many studies discritize them. The finer the scale of discrization chosen, the more difficult it can be to obtain a complete and reliable data set. Studies performed at very fine scales often find tighter and more dramatic relationships between variables such as temperature and income per capita. We examine this question by repeating the same empirical study at various temporal and spatial scales and comparing the resulting parameter estimates.