ECON 2001/2008: Economic Statistics Slot 2

Tutorial Test: August 10, 2016 Time: 40 min

STUDENT NUMBER: ______________________________________________________

General Instructions

1. The answers should be short and precise, written neatly and be clearly legible. (If
I cannot read, it is wrong.)
2. Make sure that your student number is on the copy of the test that is finally
submitted.
3. Answer all questions and show all workings and formulae in the spaces provided.
4. Round all your answers to two decimal places.

Good Luck!

Question 1

Suppose you were investigating the relationship between GDP and Government
spending (Gov) and you obtained the following regression model (standard errors are in
parentheses):

log(GDP) = 8.09 + 0.35 log(Gov)

(a) Calculate the t-statistic for the slope. [2 marks]

(b) Assume that the regression results are drawn from a large sample. Is the slope
statistically different from zero at 5% level of significance? Explain or show how you arrived
at you conclusion! [1+3 =4 marks]

Based on the rule of thumb, for large sample at =0.05 the critical value for t
= 2. (1 mark)

The calculated t-value < t-critical therefore we fail to reject the null hypothesis
(slope=0) (2 marks) and conclude that the slope is statistically not different
from zero (1 mark).

(c) State the Gauss-Markov theorem and explain its theoretical properties. [2+4=6 marks]
The OLS estimators have minimum variance compared to any class of linear
estimators, i.e, they are BLUE (2 marks). BLUE Property: Best Linear
Unbiased Estimators (4 marks for stating and explaining BLUE).
Question 2

In inferential statistics, conclusions are a function of some degree of error. Explain the two types of
errors that a researcher has to contend with. Also provide one example for each type of error.
[4+2 =6 marks]

A Type I error

occurs when the null hypothesis is true, but is rejected (false positive error) (2
marks and 1 mark for an appropriate example.

is asserting something as true when it is actually false.

a result that indicates a given condition has been fulfilled when it actually has not
been fulfilled

A Type II error

occurs when the null hypothesis is false, but erroneously fails to be rejected
(false negative error) (2 marks and 1 mark for an appropriate example).

is where a test result indicates that a condition failed, while it actually was
successful.

A Type II error is committed when we fail to believe a true hypothesis.

Question 3

Below is a table showing the amount of carats present in a cut diamond, Yi, and the weight of
the diamond stone before it is cut (Xi) in milligrams. Use the table to calculate:

Yi Xi
0.0 1.9
0.9 12.4
0.4 6.4
1.2 7.0
0.3 7.0
1.9 15
2.1 17
0.8 10

a) the expected amount of carats of diamond from a stone with a weight of 27 milligrams.
(Hint: use a regression model) [8 marks]
b) the sample standard deviation of X. [2 marks]
c) the sample skewness of X and provide an interpretation [1+2 =3
marks]
d) Give two reasons that could motivate your choice between a students test over a
standardized normal test? [4 marks]
Solution:

Y X x-xbar xY xsqr
0 1.9 -7.69 0 59.1361
0.9 12.4 2.81 2.529 7.8961
0.4 6.4 -3.19 -1.276 10.1761
1.2 7 -2.59 -3.108 6.7081
0.3 7 -2.59 -0.777 6.7081
1.9 15 5.41 10.279 29.2681
2.1 17 7.41 15.561 54.9081
0.8 10 0.41 0.328 0.1681
Total 7.6 76.7 -0.02 23.536 174.9688
Mean 0.95 9.5875

variance 25
std 5

(a)
23.54
2 = 2 = = 0.13 (2marks (formulae) + 2marks (solution))[4 marks]
174.97

1 = 2 = 0.95 0.13 9.59 = 0.3 [2 Marks]

Then, the value of Y when X is 40 is: = 1 + 2 = 0.3 + 0.13 27 = 3.21

[2 Marks]

( )2
(b) Sample standard deviation = =5 [1 (formulae)+1 mark (for the
1
solution)]

()3 60.5
(c) Then, Sk = = = 0.07( )
(1) 3 7124.97

The data is positively skewed, the mean is greater than the median/mode.
[1 (formulae & solution) + 2 marks (for the explanation)] Any other
method for calculating skewness will suffice, ie., Pearsons coefficient of
skewness.

(d) (1) the sample has fat tails or platykurtic, (2) we dont know the population
variance and (3) the sample size is small. i.e., less than 15. (Any 2 of the listed 3
reasons, 2 marks each)