If an economist is trying to predict the outcome from particular causes and effects, what is the best tool(s) for her to use in order to test out her economic model?