S&W, Chapter 5

Solution Exercises
 Exercise 5.1
෣
𝑇𝑒𝑠𝑡𝑆𝑐𝑜𝑟𝑒 = 640.3 − 4.93 ∙ 𝐶𝑆 𝑅2 = 0.11 𝑆𝐸𝑅 = 8.7
(23.5) (2.02)

(a) The 95% confidence interval for 𝛽1 is −4.93 ± 1.96 ∙ 2.02 , that is −8.889; −0.9708

෡1 −0
𝛽 −4.93
(b) Calculate the t-statistic: 𝑡 = ෡1 ) = = −2.4406
𝑆𝐸(𝛽 2.02
The p-value for the test 𝐻0 : 𝛽1 = 0 𝑣𝑠. 𝐻1 : 𝛽1 ≠ 0 is:
𝑝 − 𝑣𝑎𝑙𝑢𝑒 = 2𝜑 − 𝑡 𝑎𝑐𝑡 = 2𝜑 −2.4406 = 2 ∙ 0.0073 = 0.0146
The p-value is less than 0.05, so we can reject the null hypothesis at the 5% s.l.
The p-value is larger than 0.01, so we cannot reject the null hypothesis at the 1% s.l.
 Exercise 5.1 (ctd.)
෡1 +5
𝛽 0.07
(c) Calculate the t-statistic: 𝑡 = ෡1 ) = 2.02 = 0.0346
𝑆𝐸(𝛽
The p-value for the test 𝐻0 : 𝛽1 = −5 𝑣𝑠. 𝐻1 : 𝛽1 ≠ −5 is:
𝑝 − 𝑣𝑎𝑙𝑢𝑒 = 2𝜑 − 𝑡 𝑎𝑐𝑡 = 2𝜑 −0.0346 = 2 ∙ 0.4840 = 0.968
The p-value is larger than 0.10, so we cannot reject the null hypothesis at the 10%, 5%
or 1% s.l. Because 𝛽1 = −5 is not rejected at the 5% level, this value is contained in the
95% confidence interval.

(d) The 90% confidence interval for 𝛽0 is 640.3 ± 1.645 ∙ 23.5 , that is 601.64; 678.96
 Exercise 5.2
෣ = 10.73 + 1.78 ∙ 𝑀𝑎𝑙𝑒
𝑊𝑎𝑔𝑒 𝑅2 = 0.09 𝑆𝐸𝑅 = 3.8
(0.16) (0.29)

(a) The estimated gender gap is $1.78 per hour.

(b) 𝐻0 : 𝛽1 = 0 𝑣𝑠. 𝐻1 : 𝛽1 ≠ 0
෡1 −0
𝛽 1.78
Calculate the t-statistic: 𝑡 = ෡1 ) = = 6.1379
𝑆𝐸(𝛽 0.29
The p-value for the test 𝐻0 : 𝛽1 = 0 𝑣𝑠. 𝐻1 : 𝛽1 ≠ 0 is:
𝑝 − 𝑣𝑎𝑙𝑢𝑒 = 2𝜑 − 𝑡 𝑎𝑐𝑡 = 2𝜑 −6.1379 = 2 ∙ 0.0000 = 0.0000
The p-value is less than 0.01, so we can reject the null hypothesis that there is no
gender gap at the 1% s.l.
(c) The 95% confidence interval for the gender gap 𝛽1 is 1.2116; 2.3484

(d) The sample average wage of women is 𝛽መ0 = $10.73/ℎ𝑜𝑢𝑟.

The sample average wage of men is 𝛽መ0 + 𝛽መ1 = $10.73 + $1.78 = 12.51/ℎ𝑜𝑢𝑟.
 Exercise 5.2 (ctd.)
(e) The binary variable regression model relating wages to gender can be written as either

𝑊𝑎𝑔𝑒 = 𝛽0 + 𝛽1 𝑀𝑎𝑙𝑒 + 𝑢𝑖 or 𝑊𝑎𝑔𝑒 = 𝛾0 + 𝛾1 𝐹𝑒𝑚𝑎𝑙𝑒 + 𝑣𝑖

In the first regression equation, Male equals 1 for men and 0 for women; 𝛽0 is the
population mean of wages for women and 𝛽0 + 𝛽1 is the population mean of wages
for men. In the second regression equation, Female equals 1 for women and 0 for
men; 𝛾0 is the population mean of wages for men and 𝛾0 + 𝛾1 is the population mean
of wages for women. We have the following relationship for the coefficients in the two
regression equations: 𝛾0 = 𝛽0 + 𝛽1 and 𝛾0 + 𝛾1 = 𝛽0

Given the coefficient estimates 𝛽መ0 and 𝛽መ1 , we have

𝛾ො0 = 𝛽መ0 + 𝛽መ1 = 12.51 and 𝛾ො1 = 𝛽መ0 − 𝛾ො0 = −𝛽መ1 = −1.78

Due to the relationship among coefficient estimates, for each individual observation,
the OLS residual is the same under the two regression equations: 𝑢ො 𝑖 = 𝑣ො𝑖 . Thus the
sum of squared residuals 𝑆𝑆𝑅 = σ𝑛𝑖=1 𝑢ො 𝑖 ² is the same under the two regressions. This
𝑆𝑆𝑅 1/2 𝑆𝑆𝑅
implies that both 𝑆𝐸𝑅 = and 𝑅² = 1 − 𝑇𝑆𝑆 are unchanged.
𝑛−1

In summary, in regression Wages on Female, we will get

෣ = 12.51 − 1.78𝐹𝑒𝑚𝑎𝑙𝑒, 𝑅2 = 0.09 𝑆𝐸𝑅 = 3.8
𝑊𝑎𝑔𝑒𝑠
 Exercise 5.2e (simpler)

෣ = 10.73 + 1.78 ∙ 𝑀𝑎𝑙𝑒

𝑊𝑎𝑔𝑒 𝑅2 = 0.09 𝑆𝐸𝑅 = 3.8
(0.16) (0.29)

෣ = 10.73 + 1.78 ∙ 1 − 𝐹𝑒𝑚𝑎𝑙𝑒

𝑊𝑎𝑔𝑒
= 10.73 + 1.78 − 1.78 ∙ 𝐹𝑒𝑚𝑎𝑙𝑒
= 12.51 − 1.78 ∙ 𝐹𝑒𝑚𝑎𝑙𝑒
 Exercise 5.3
෣ = −79.24 + 4.16 ∙ 𝐻𝑒𝑖𝑔ℎ𝑡
𝑊𝑒𝑖𝑔ℎ𝑡 𝑅2 = 0.72 𝑆𝐸𝑅 = 12.6
(3.42) (0.42)

The 95% confidence interval is 2 ∗ 4.16 ± 1.96 ∙ 0.42 = 6.6736; 9.9664

 Exercise 5.4

෣ = −7.29 + 1.93𝑒𝑑𝑢𝑐
𝐸𝑎𝑟𝑛𝑖𝑛𝑔𝑠
1.1 (0.08)

(a) −7.29 + 1.93 ∙ 16 = $23.59/ℎ𝑜𝑢𝑟

(b) The wage is expected to increase from $15.87/hour to $19.73/hour or by

$3.86/hour.

(c) The increase in wages for college education is 4𝛽1 . This, the counselor’s assertion is
that 𝛽1 = 2.5. The t-statistic for this null hypothesis is:
1.93 − 2.5
𝑡= = −7.125
0.08
Which has a p-value of 0.000 (<0.01). The counselor’s assertion can thus be rejected
at the 1% significance level.
A 95% confidence interval for 4𝛽1 is 4 ∙ 1.93 ± 1.96 ∙ 0.08 = [7.09; 8,35]. This
confidence interval does not include the counselor’s assertion of a gain of $10. So, the
assertion is not consistent with the regression results.
 Exercise 5.5
෣
𝑇𝑒𝑠𝑡𝑆𝑐𝑜𝑟𝑒 = 918.0 + 13.9 × 𝑆𝑚𝑎𝑙𝑙 𝑅² = 0.01 𝑆𝐸𝑅 = 74.6
1.6 (2.5)

(a) The estimated gain from being in a small class is 13.9 points. This is equal to
approximately 1/5 of the standard deviation in test scores, a moderate increase.

13.9
(b) The t-statistic is 𝑡 𝑎𝑐𝑡 = 2.5 = 5.56, which has a p-value of 0.000. Thus the
null hypothesis is rejected at the 5% (and 1%) level.

(c) 13.9 ± 2.58 × 2.5 = 13.9 ± 6.45.

 Exercise 5.6
(a) The question asks whether the variability in test scores in large classes is the same
as the variability in small classes. It is hard to say. On the one hand, teachers in small
classes might be able to spend more time bringing all of the students along, reducing
the poor performance of particularly unprepared students. On the other hand, most
of the variability in test scores might be beyond the control of the teacher.

(b) The formula in 5.3 is valid for heteroskedasticity of homoskedasticity; thus

inferences are valid in either case.
 Exercise 5.7
𝑌෠ = 5.4 + 3.2𝑋 𝑅² = 0.26 𝑆𝐸𝑅 = 6.2
3.1 (1.5)
3.2
(a) The t-statistic is 1.5 = 2.13 with a p-value of 0.03; since the p-value is less than
0.05, the null hypothesis is rejected at the 5% level.

(b) 3.2 ± 1.96 × 1.5 = 3.2 ± 2.94

(c) Yes. If Y and X are independent, then 𝛽1 = 0; but this null hypothesis was rejected
at the 5% level in part (a).
Note: “the (𝑌𝑖 , 𝑋𝑖 ) are independently and identically distributed” (drawn
independently from the same bivariate population distribution) does not mean
that (the values of) 𝑌𝑖 and 𝑋𝑖 are independent (i.e., that there is no relationship
between 𝑌𝑖 and 𝑋𝑖 ).

(d) β1 would be rejected at the 5% level in 5% of the samples;

95% of the confidence intervals would contain the value 𝛽1 = 0.
 Exercise 5.8
(a) 43.2 ± 2.05 × 10.2 or 43.2 ± 20.91, where 2.05 is the 5% two-sided critical value
from the 𝑡28 distribution.

61.5−55
(b) The t-statistic is 𝑡 𝑎𝑐𝑡 = 7.4 = 0.88, which is less (in absolute value) than the
critical value of 2.05. Thus, the null hypothesis is not rejected at the 5% level.

(c) The one-sided 5% critical value is 1.70; 𝑡 𝑎𝑐𝑡 is less than this critical value, so that
the null hypothesis is not rejected at the 5% level.
 Exercise 5.15

Because the samples are independent, 𝛽መ𝑚,1 and 𝛽መ𝑤,1 are independent.

Thus 𝑣𝑎𝑟 𝛽መ𝑚,1 − 𝛽መ𝑤,1 = 𝑣𝑎𝑟 𝛽መ𝑚,1 + 𝑣𝑎𝑟(𝛽መ𝑤,1 ).

V𝑎𝑟 𝛽መ𝑚,1 is consistently estimated as [𝑆𝐸(𝛽መ𝑚,1 )]² and 𝑣𝑎𝑟(𝛽መ𝑤,1 ) is

2
consistently estimated as 𝑆𝐸 𝛽መ𝑤,1 ,
2
so that 𝑣𝑎𝑟 𝛽መ𝑚,1 − 𝛽መ𝑤,1 is consistently estimated by 𝑆𝐸 𝛽መ𝑚,1 +

[𝑆𝐸(𝛽መ𝑤,1 )]²,
and the result follows by noting the SE is the square root of the estimated
variance.