Question

In 2001 Kenkel and Terza published an article in which they investigated the impact on an individual's alcohol consumption of a physician's advice to reduce drinking. The simplified model below was estimated with their data set (although the econometric techniques used in the article are more complex than those we have discussed in this class).

DRINKS= 13.00 + 11.36 Advice-.20 EDUC+ 2.85 DIVSEP+14.20 UNEMP

(5.37) (-.65) (1.11) (2.75)

N = 500, R² = 0.377, t-Statistics in parentheses

DRINKS = drinks consumed by the individual in the last two weeks

ADVICE = 1 if a physician had advised the individual to cut back on drinking alcohol, 0 otherwise

EDUC = years of schooling of the individual

DIVSEP = 1 if the individual is divorced or separated, 0 otherwise

UNEMP = 1 if the individual is unemployed, 0 otherwise

14. Which of the following best reflects the interpretation of the coefficient of determination, R2?

a.      37.7 percent of the variation in DRINKS is due to random (unexplained) variations.

b.     37.7 percent of the variation in DRINKS is explained by variation in the independent variables.

c.      37.7 percent of people listen to the advice of a physician in regards to their drinking behaviors.

d.     37.7 percent of the people in this sample drink.

15. Which of the following best reflects the interpretation of the coefficient on the variable EDUC?

a.      For every additional year of schooling, the number of DRINKS decreases 0.2, ceteris paribus.

b.     People with more education drink 20% less than those with less education, ceteris paribus.

c.      For every additional year of schooling, people drink 20 percent less, ceteris paribus.

d.     Consuming one additional DRINK (every 2 weeks) results in people completing .2 years less education, ceteris paribus.

16. Which of the following best reflects the interpretation of the coefficient on the variable UNEMP?

a.      14.20 percent of the people in this sample were unemployed, while 2.75 percent were heavy drinkers, ceteris paribus.

b.     Consuming one additional DRINK (every 2 weeks) results in people being 14.20 percent more likely to be unemployed, ceteris paribus.

c.      An unemployed person, on average, consumed 14.20 percent more drinks (in the last 2 weeks) than an employed person, ceteris paribus.

d.     An unemployed person, on average, consumed 14.20 more drinks (in the last 2 weeks) than an employed person, ceteris paribus.

17. Conduct two-tailed hypothesis tests on the significance of each of the regression coefficients. Use the 5% level of significance and 120 degrees of freedom. What is the t-Critical value for these tests?

a.      3.374

b.     1.658

c.      1.980

d.     2.398

18. Conduct two-tailed hypothesis tests on the significance of each of the regression coefficients. Use the 5% level of significance and 120 degrees of freedom. Which coefficients are significantly different from 0?

a.      The coefficients on EDUC and DIVSEP only

b.     None of the coefficients are significant at the 5% level.

c.      All four coefficients are significant at the 5% level.

d.     The coefficients on ADVICE and UNEMP only

Answer 1