ECON 151 Lecture Notes - Lecture 10: Internal Validity, Parental Investment, Twin Study
Document Summary
Controlling for ability y = + s + a + u with s schooling and a ability. If we omit a, according to frisch and waugh: E( ys ) = + as where as is bias (bias = correlation between ability and schooling and correlation between ability and income) We can try to capture a using a list of addition control variables. When ability increase by 1 std deviation, salary increases by 7. 9% Sd = standardized results, normalize distribution, allows you to standardize results. Even when controlling for ability, only able to explain 18. 6% of the things that affect income 81% remains unexplained by the variables in the regression. Things that could be missing - education type, gender, quality of education. Fixed effect estimator = first differences same for more than 2 _____ Structural equation yi1 = si1 + zi1 + xi + i + 1i (1) yi2 = si2 +zi2 + xi + i + 2i (2)