a.

The estimated intercept (A) can be found by looking at the regression equation:

Yi = A + 0.180X

Since the intercept is the value of Y when X is equal to zero, we can set X to zero and solve for A:

A = Yi - 0.180X

Since we don't have the value of Yi or X for the entire population, we use the estimated values from the
sample. From the given information, we know that the estimated intercept (A) is 2.174 thousand dollars.

b. The standard error of the estimated slope (B) is given as 2.174 in parenthesis. However, this is actually
the standard error of the intercept, not the slope. The standard error of the estimated slope can be
calculated using the formula:

B = se / (sqrt(sum((Xi - X̄ )^2)))

where se is the standard error of the regression, Xi is the ith observation of the independent variable X,
X̄ is the sample mean of X, and n is the sample size. Since we don't have the actual data, we cannot
calculate the standard error of the estimated slope.

c. The economic interpretation of the estimated slope is that for a one percent increase in the
percentage of males 18 or older who are high school graduates, the mean income of males who are 18
years of age or older increases by 0.180 thousand dollars. The positive sign of the coefficient indicates
that there is a positive relationship between the percentage of high school graduates and mean income.
This is what we would expect from economic theory and common sense, as education is often
associated with higher earnings potential.

2. a. The sample regression function (SRF) is given by:

Ŷi = α + β1Xi1 + β2Xi2

where Ŷi is the predicted value of Y for the ith observation, Xi1 and Xi2 are the values of the
independent variables for the ith observation, and α, β1, and β2 are the estimated coefficients.

b. The formula for estimating the coefficients α, β1, and β2 using Ordinary Least Squares (OLS) is derived
by minimizing the sum of squared residuals (SSR), which is the sum of the squared differences between
the actual values of Y and the predicted values of Y. The OLS estimator for β1 is given by:

β1 = ∑(Xi1 - X̄ 1)(Yi - Ȳ) / ∑(Xi1 - X̄ 1)^2

where Xi1 is the ith observation of the independent variable X1, X̄ 1 is the sample mean of X1, Yi is the ith
observation of the dependent variable Y, and Ȳ is the sample mean of Y. Similarly, the OLS estimator for
β2 is given by:
β2 = ∑(Xi2 - X̄ 2)(Yi - Ȳ) / ∑(Xi2 - X̄ 2)^2

where Xi2 is the ith observation of the independent variable X2, X̄ 2 is the sample mean of X2, Yi is the ith
observation of the dependent variable Y, and Ȳ is the sample mean of Y. The OLS estimator for α is given
by:

α = Ȳ - β1X̄ 1 - β2X̄ 2

where X̄ 1 and X̄ 2 are the sample means of X1 and X2, respectively.

c. The variances of α, β1, and β2 can be formulated using the following formulas:

Var(α) = σ^2[1/n + X̄ 1^2/∑(Xi1 - X̄1)^2 + X̄2^2/∑(Xi2 - X̄ 2)^2]

Var(β1) = σ^2 / ∑(Xi1 - X̄ 1)^2

Var(β2) = σ^2 / ∑(Xi2 - X̄ 2)^2

where σ^2 is the variance of the error term ui, and n is the sample size. These formulas show that the
variances of the coefficients depend on the variance of the error term and the sample size, as well as the
variances of the independent variables. The larger the variance of the error term, the larger the
variances of the coefficients, which means that the estimates are less precise. Similarly, the smaller the
sample size, the larger the variances of the coefficients, which means that the estimates are less reliable.

3. a. The mean of Y can be found by summing up the values of Y and dividing by the total number of
observations:

mean of Y = (15 + 10 + 18 + 12 + 20) / 5 = 15

The mean of X1 can be found by summing up the values of X1 and dividing by the total number of
observations:

mean of X1 = (10 + 6 + 12 + 7 + 15) / 5 = 10

The mean of X2 can be found by summing up the values of X2 and dividing by the total number of
observations:

mean of X2 = (6 + 4 + 7 + 5 + 8) / 5 = 6

b. The sum of squares of deviations from the respective means are:

∑(Y - mean of Y)^2 = (15 - 15)^2 + (10 - 15)^2 + (18 - 15)^2 + (12 - 15)^2 + (20 - 15)^2 = 110

∑(X1 - mean of X1)^2 = (10 - 10)^2 + (6 - 10)^2 + (12 - 10)^2 + (7 - 10)^2 + (15 - 10)^2 = 50

∑(X2 - mean of X2)^2 = (6 - 6)^2 + (4 - 6)^2 + (7 - 6)^2 + (5 - 6)^2 + (8 - 6)^2 = 10

∑X1X2 = 106 + 64 + 127 + 75 + 15*8 = 356

∑X1Y = 1015 + 610 + 1218 + 712 + 15*20 = 825

∑X2Y = 615 + 410 + 718 + 512 + 8*20 = 540

c. The sums of products between X1 and Y, X2 and Y, and X1 and X2 are:

∑X1Y = 825

∑X2Y = 540

∑X1X2 = 356

d. The parameter estimates of α, β1, and β2 can be calculated using the following formulas:

β1 = ∑(X1 - mean of X1)(Y - mean of Y) / ∑(X1 - mean of X1)^2

= (10 - 10)(15 - 15) + (6 - 10)(10 - 15) + (12 - 10)(18 - 15) + (7 - 10)(12 - 15) + (15 - 10)(20 - 15) / [(10 -
10)^2 + (6 - 10)^2 + (12 - 10)^2 + (7 - 10)^2 + (15 - 10)^2]

= 15/5 = 3

β2 = ∑(X2 - mean of X2)(Y - mean of Y) / ∑(X2 - mean of X2)^2

= (6 - 6)(15 - 15) + (4 - 6)(10 - 15) + (7 - 6)(18 - 15) + (5 - 6)(12 - 15) + (8 - 6)(20 - 15) / [(6 - 6)^2 + (4 - 6)^2
+ (7 - 6)^2 + (5 - 6)^2 + (8 - 6)^2]

= 30/10 = 3

α = mean of Y - β1(mean of X1) - β2(mean of X2)

= 15 - 3(10) - 3(6) = -3

Therefore, the parameter estimates are:

α = -3, β1 = 3, β2 = 3

e. The estimated sample multiple regression model is:

Ŷ = -3 + 3X1 + 3X2

f. The interpretation of the result is that for a one-unit increase in X1, holding X2 constant, the predicted
value of Y increases by 3 units. Similarly, for a one-unit increase in X2, holding X1 constant, the predicted
value of Y increases by 3 units. The intercept of -3 represents the predicted value of Y when both X1 and
X2 are equal to zero.
g. To compute the predicted value of annual salaries when years of education = 6 and year of experience
= 13, we substitute X1 = 6 and X2 = 13 into the estimated regression equation:

Ŷ = -3 + 3(6) + 3(13) = 45

Therefore, the predicted value of annual salaries is 45 thousand Birr.

h. The analyst's claim that the model is likely to suffer from heteroskedasticity problem is valid because
the variance of the error term ui is not constant for all values of X1 and X2. Heteroskedasticity occurs
when the variance of the error term is not constant across the range of values of the independent
variables. This can lead to biased and inefficient estimates of the coefficients, as well as incorrect
standard errors and hypothesis tests. In this case, the analyst is correct because the variance of the error
term is not known to be constant, and there is no reason to believe that it is

4. a. To find the mean annual salary of teachers in private schools, we set D1 = 0 in the regression
equation:

Y = 13,269.45 - 1673.23(0) + 3.07X = 13,269.45 + 3.07X

b. To find the mean annual salary of teachers in public schools, we set D1 = 1 in the regression equation:

Y = 13,269.45 - 1673.23(1) + 3.07X = 10,596.22 + 3.07X

c. To calculate the change in a school teacher's annual salary, on average, as school expenditure per
student goes up by a thousand ETB, we look at the coefficient of X in the regression equation, which is
3.07. This means that for every 1,000 ETB increase in school expenditure per student, a school teacher's
annual salary is expected to increase by 3,070 ETB, on average.

d. To test whether the annual salary of a private school teacher is on average higher than that of a public
school teacher, we can compare the coefficients of D1 in the regression equation. The coefficient of D1
is -1673.23, which means that the average annual salary of a public school teacher is expected to be
1,673.23 ETB lower than that of a private school teacher, holding school expenditure per student
constant. The P-value associated with this coefficient can be used to test whether this difference is
statistically significant. If the P-value is less than the significance level (e.g., 0.05), we can conclude that
there is evidence to suggest that the annual salary of a private school teacher is on average higher than
that of a public school teacher. However, the P-value is not provided in the given information, so we
cannot determine whether this difference is statistically significant.