ECON 151 Lecture Notes - Lecture 4: Randomized Experiment, Intelligence Quotient
Document Summary
: error term, part of wage which is not explained by schooling. Schoolingi means that each individual in the dataset has a unique schooling value. We thus have as much equations as individuals in the dataset. The goal is thus to solve the set of n equations (n=total number of i individuals) with two unknown, 0 1. 0 = wage when schooling is 0. 1 = slope (econometrics = science of estimating this slope) Assumption that there"s no correlation between error term and schooling. Assumption of ols model = zero conditional mean assumption. Ability makes the relation between wage and schooling spurious. Schooling is endogenous : correlated with both wage and something else not included in the model (here ability). We need to "control" for ability consider schooling at a constant. To reduce bias, any variable correlated with both wage and schooling must be included in the regression. Among other, ability is a cause of bias but