4. The following is the regression output for the following equation, which was estimated for a large sample of people:
log(wage) = α + β1Schooling + β2Age + β3Female + β4Non-white
Variable Coefficient estimate (β-hat) Standard error
Schooling (years) .134 .024
Age (years) 0.009 0.002
Female -0.021 0.001
Non-white -0.014 0.003
(a) Calculate the t-statistics and discuss which variables are statistically significant.
(b) What do the estimates imply is the percent wage increase associated with an additional year of schooling?
(c) Why might the estimated effects of some of the variables be overstated? Provide some examples of what might lead to the overstatement.
(d) What if one could fully deal with the problems in (c) by adding variables to the estimation equation? Is there any reason to believe that the resulting estimates of the effect of schooling would then be understated?
(e) Explain why estimating the following regression equation might fail to capture the sheepskin effect.
Examine a regression output for the following equation.