Professional Documents
Culture Documents
4.1. (a) The predicted average test score is
that is, the regression predicts that test scores will fall by 23.28 points.
(c) Using the formula for b̂ 0 in Equation (4.8), we know the sample average of the
(d) Use the formula for the standard error of the regression (SER) in Equation (4.19)
to get the sum of squared residuals:
Use the formula for R 2 in Equation (4.16) to get the total sum of squares:
SSR 12961
TSS = 2
= = 14088.
1− R 1− 0.08
sY = sY2 = 11.9.
(a) Substituting Height = 70, 65, and 74 inches into the equation, the predicted
weights are 176.39, 156.69, and 192.15 pounds.
(c) We have the following relations: 1 in = 2.54 cm and 1 lb = 0.4536 kg. Suppose
the regression equation in the centimeter-kilogram space is
! = ˆ + ˆ Height .
Weight γ0 γ1
4.3. (a) The coefficient 9.6 shows the marginal effect of Age on AWE; that is, AWE is
expected to increase by $9.6 for each additional year of age. 696.7 is the intercept
of the regression line. It determines the overall level of the line.
(b) SER is in the same units as the dependent variable (Y, or AWE in this
example). Thus SER is measures in dollars per week.
(e) No. The oldest worker in the sample is 65 years old. 99 years is far outside the
range of the sample data.
(f) No. The distribution of earning is positively skewed and has kurtosis larger
than the normal.
(g) bˆ0 = Y - bˆ1 X , so that Y = bˆ0 + bˆ1 X . Thus the sample mean of AWE is 696.7
4.4. (a) ( R - R f ) = b ( Rm - R f ) + u ,
4.5. (a) ui represents factors other than time that influence the student’s performance
on the exam including amount of time studying, aptitude for the material, and so
forth. Some students will have studied more than average, other less; some
students will have higher than average aptitude for the subject, others lower, and
so forth.
(c) (2) is satisfied if this year’s class is typical of other classes, that is, students in
this year’s class can be viewed as random draws from the population of students
that enroll in the class. (3) is satisfied because 0 £ Yi £ 100 and Xi can take on
only two values (90 and 120).
(d) (i) 49 + 0.24 ´ 90 = 70.6; 49 + 0.24 ´120 = 77.8; 49 + 0.24 ´150 = 85.0
4.7. The expectation of bˆ0 is obtained by taking expectations of both sides of Equation
(4.8):
é æç 1 n ö÷ ù
E ( bˆ0 ) = E (Y - bˆ1 X ) = E ê ç b 0 + b1 X + å ui ÷ - bˆ1 X ú
ë çè n i =1 ÷ø û
n
1
= b 0 + E ( b1 - bˆ1 ) X + å E (ui )
n i =1
= b0
where the third equality in the above equation has used the facts that E(ui) = 0 and
E[( b̂1 −b1) X ] = E[ X × (E( b̂1 −b1)| X )] = 0 because E[( b1 - bˆ1 ) | X ] = 0 (see text
equation (4.31).)
4.8. The only change is that the mean of bˆ0 is now b0 + 2. An easy way to see this is
The new regression error is (ui - 2) and the new intercept is (b0 + 2). All of the
assumptions of Key Concept 4.3 hold for this regression model.
4.9. (a) With bˆ1 = 0, bˆ0 = Y , and Yˆi = bˆ0 = Y . Thus ESS = 0 and R2 = 0.
(b) If R2 = 0, then ESS = 0, so that Ŷi = Y for all i. But, using the formula for β̂0 ,
Ŷi = β̂0 + β̂1 X i = Y − β̂1 ( X i − X ) , Thus Yˆi = Y for all i implies that bˆ1 = 0, or that
å
n
Xi is constant for all i. If Xi is constant for all i, then i -1
( X i - X )2 = 0 and b̂1
is undefined (see equation (4.7)).
4.10. (a) E(ui|X = 0) = 0 and E(ui|X = 1) = 0. (Xi, ui) are i.i.d. so that (Xi, Yi) are i.i.d.
(because Yi is a function of Xi and ui). Xi is bounded and so has finite fourth
moment; the fourth moment is non-zero because Pr(Xi = 0) and Pr(Xi = 1) are both
non-zero. Following calculations like those exercise 2.13, ui also has nonzero
finite fourth moment.
where the first equality follows because E[(Xi - µX)ui] = 0, and the second
equality follows from the law of iterated expectations.
E[( X i - µ X )ui |X i = 0]2 = 0.22 ´1, and E[( X i - µ X )ui |X i = 1]2 = (1 - 0.2) 2 ´ 4.
å
n
4.11. (a) The least squares objective function is i =1
(Yi - b1 X i )2. Differentiating with
(b) Following the same steps in (a) yields bˆ1 = åin=1 X i (Yi - 4)
åin=1 X i2
ë s X sY û
s XY s
(c) Because rXY = , rXY Y =
s X sY sX
n
1 n
s XY å
(n - 1) i =1
( X i - X )(Yi - Y ) å ( X i - X )(Yi - Y )
= n
= i =1 n = bˆ1
sX2 1
å(Xi - X )
(n - 1) i =1
2
åi =1
(Xi - X ) 2
4.13. The answer follows the derivations in Appendix 4.3 in “Large-Sample Normal
Distribution of the OLS Estimator.” In particular, the expression for ni is now ni =
(Xi − µX)kui, so that var(ni) = k3var[(Xi − µX)ui], and the term k2 carry through the rest
of the calculations.
4.14. Because bˆ0 = Y - bˆ1 X , Y = bˆ0 + b1 X . The sample regression line is y = β̂0 + β̂1x ,
4.15
(a.i) Because (Yoos, Xoos) are from some population as the in-sample observation,
(a.ii)
where the second line follows because Xoos is independent of ( β̂0 , β̂1 ) because they
depend only on the in-sample observations.
( ) ( )
û oos = u oos − ⎡⎢ β̂0 − β0 + β̂1 − β1 x oos ⎤⎥
⎣ ⎦
and the result follows by noting that the two terms on the right hand side of the equation
are independent, and therefore uncorrelated with each other.
(b.i) No. Because (Yoos, Xoos) are drawn from distribution different that the sample data,
then in general E(Yoos|Xoos = xoos) ≠ E(Y|X = xoos) = b0 + b1xoos.