Regression Results. Robustness Checks As a reminder for how to interpret regress

Question

Regression Results. Robustness Checks As a reminder for how to interpret regressions using logarithmic transformations, below we reproduce part of Stock and Watson's Key Concept 8.2 Use this hint, and your knowledge about interpreting nonlinear regression models, to interpret the results in Table 4, and to answer questions 11 - 18 below. In the first specification, if AGE increases from 25 to 26, how are earnings expected to change? a. Increase by 35 cents b. Increase by 35 percent c. Increase by 80cents x This is an easy call, as the model in the "first specification" d. Increase by 80 percent i.e. the| the model in column 1, is linear - linear. To answer question 12, let's begin by looking at the model from the second column in equation form: AHE_hat = - 5.752 - 5.081*FEMALE + 0.798*AGE + 7.432*BACHELOR + 1.249*FEMALE*BACHELOR To find the predicted AHE for a male, set the value of the variable FEMALE to zero; the equation becomes: AHE - hat = - 5.752 + 0.798*AGE + 7.432*BACHELOR. Thus for men a college degree increases AHE b 7.432. In the second specification, is the effect of education greater for men than for women? Is the effect statistically significant? a. NO; No x b. No; Yes c. Yes; No d. Yes; Yes (answer continued here: To find predicted AHE for a female, set the value of FEMALE to 1: AHE - hat = - 5.752 - 5.081 + 0.798*AGE + 7.432*BACHELOR + 1.249*BACHELOR Thus if for a female if BACHELOR equals 1, AHE increase by 7.432 + 1.249, more than for a male, but the effect is insignificant

Explanation / Answer

We observe that the coefficient of interaction term female_times_bachelor is not significant (with b=1.240, SE=1.161) as it has no star above it indicating that effect of education is not grater for men as it is not statistically significant.

Related Questions

Navigate

• Browse (All)
• Browse R
• Subjects

---