RMIT Classification: Trusted

Basic Econometrics

Solutions to Face-Face Class + Online and Post-Class Activities

Topic 1: Normal and Sampling Distribution

Question 1)
Assume Z N (0 ,1), what is the probability that
a. Z<1.58
b. Z>2.12
c. 1.58<Z<2.12

Solutions:
a. P(Z<1.58) = 0.9429
b. P(Z>2.12) = 1-P(Z<2.12) = 1-0.983 = 0.017
c. P(1.58<Z<2.12) = P(Z<2.12)-P(Z<1.58) = 0.983-0.9429=0.04

Since these Z values correspond to a standard normal distribution, you can look up
P(Z<1.58) and P(Z<2.12) in a set of standard normal tables (such as the ones in Table E2
included on Canvas). Doing so will give you areas under the standard normal curve of 0.9429
and 0.983─but note that for part b) you want P(Z>2.12), the upper tail, so you need to
deduct 0.983 from 1, the total area under the standard normal curve.

However, it is much easier to use a computer package to look up the probabilities. We can
use the pnorm function in R to do this.

The R code corresponding to above questions is:

Note that this code can either be entered interactively into the console window or typed
into the source window and then run
 RMIT Classification: Trusted
2

Question 2)

Tuan’s Trucking Company determined that, on an annual basis, the distance travelled per
truck is normally distributed with a mean of 100,000 kilometres and a standard deviation of
10, 000 kilometres.
a. Calculate the Z value for a truck that travels 80,000 and 120,000 kilometres
respectively in the year. Interpret the meaning.
b. What proportion of Tuan’s trucks can be expected to travel between 80,000 and
120, 000 kilometres in the year?
c. What percentage of Tuan’s trucks can be expected to travel either below 60,000 or
above 140,000 kilometres in the year?

Solutions:
X−μ 80,000−100,000
a. Z= = =−2, meaning the distance of 80,000 kilometres is 2 σ
σ 10,000
below the mean of 100,000 kilometres.

X−μ 120,000−100,000
Z= = =+ 2, meaning the distance of 120,000 kilometres is 2 σ
σ 10,000
above the mean of 100,000 kilometres.

b. P(80< X <120)=P(−2< Z <2)=0.9772−0.0228=0.9545

Thus 95.45% of Tuan’s trucks can be expected to travel between 80,000 and 120,
000 kilometres in the year.

Alternatively, you can pre-calculate the upper and lower z values and put these into
the pnorm function in R.

c. P( X< 60)+ P (X >140)=P( Z <−4)+ P(Z> 4)=0.0000633428

Thus, almost none of Tuan’s trucks can be expected to travel either below 60,000 or
above 140,000 kilometres in the year.

Alternatively,
 RMIT Classification: Trusted
3

Question 3)

Simulating sampling distribution [This is a very important question to help you understand
How the sampling distribution is derived and its two important properties]:

Referring to the example given in the lecture about people’s IQ, where IQ is normally
distributed with the mean ( μ) = 100, and std dev (σ ) = 15. Suppose each time, you select a
random sample of 5 people and record the sample mean. You repeat doing this for 10,000
times, implying that you will have 10,000 values for the sample mean.

Draw the histogram of the sample means and verify the two important properties of
sampling distribution of the sample mean.

Solutions: please refer to the R code provide below.

xbar<-NULL
mu=100
sigma=15
for (i in 1:10000) {
sampledata=rnorm(5,mu,sigma)
xbar[i]=mean(sampledata)
}
length(xbar)
# xbar
hist(xbar)

The histogram produced should look something like:

 RMIT Classification: Trusted
4

xbar will have: (i) a mean close to 100; and, (ii) a sample standard deviation that is much
smaller than 15/√5 (the population standard deviation)
 RMIT Classification: Trusted
5

Question 4)

Travel time on the bus between the Lotte Mart and the RMIT Saigon campus bus stops is
normally distributed with:
μ=8 and σ =2.
a. If you select a random sample of 25 bus trips, what is the probability that the sample
mean is between 6.9 and 8.2 minutes?
b. If you select a random sample of 100 bus trips, what is the probability that the
sample mean is between 6.9 and 8.2 minutes?
c. Explain the difference between the results of (a) and (b).

Solutions:
The key to answering this question is to note that x is following a normal distribution
with mean µ=8 and standard error σ=2/5=0.4

 2
X   0.4
(a) = n 25 ;
P(6.9 < X < 8.2) = P(– 2.75 < Z < 0.50) = 0.6915 – 0.0028 = 0.6887

(b) With a larger sample size of 100, the std error (se) becomes
sigma/sqrt(n)=2/10=0.2
 2
  0 .2
X n 100
= ;
P(6.9 < X < 8.2) = P(– 6.50 < Z < 1.00) = 0.8413 – 0.0000 = 0.8413

(c) With the sample size increasing from n = 25 to n = 100, more of sample means
will be closer to the distribution mean. The standard error of the sampling
distribution of size 100 is much smaller than that of size 25, so the likelihood
that the sample mean will fall within 0.2 minutes of the mean is much
higher for samples of size 100 (probability = 0.8413) than for samples of size
25 (probability = 0.6887).
 RMIT Classification: Trusted
6

Question 5

Download the gpa1.RData dataset from Canvas*

This dataset contains information on a sample of 141 US university undergraduate students

including their current GPA (Grade Point Average), called colGPA, and a binary variable
indicating whether the student owned a personal computer called PC.

i) Calculate the mean, median and standard deviation for the variables called age in this
dataset.

The mean of age is 20.89, the median is 21.0 and the standard deviation is 1.272.

ii) Create a bar chart for the distribution of personal computers Note: PC takes the value of 1
if a student has a personal computer and the value of 0 otherwise.

Will produce this ’tab’ in the source Window

and produce in the graph the Plots, tab of the bottom right-hand window.
 RMIT Classification: Trusted
7

If
If you don’t see this horizontal bar graph, be sure to click on “Plots”. If it then appears
‘shished’ you should adjust the size of this window or click the Zoom panel.