1-1
Statistics for economists
Dr. Ahmed El-Kotory
Lecture(3)
1-2
Course content
Sampling Methods and the Central Limit Theorem
Estimation and Confidence Intervals
One-Sample Tests of Hypothesis
Two-Sample Tests of Hypothesis
Using Minitab 18
Analysis of Variance
Package
Linear Regression.
Index Numbers
Course content
1-3
Revision
Statistical Inference
Testing Hypotheses Estimation
Almost, no research field or applied study is
devoid of a hypothesis Or more, the researcher Confidence Point
wants, through his studies, to test its validity or
check it out. Interval
Hypothesis (Claim) It is an opinion about the Find only one
value of population parameter, Which value from sample
Unknown The sample data is relied upon for data instead of the
the purpose of validation That claim or deny it ̅± 𝛂 × unknown
𝟐 population
parameter
1- state the hypotheses ̅± 𝛂 ×
𝟐
The 2- calculate The test Statistic
steps ̅± 𝛂 ×
3- find The critical value (Table) 𝟐
4- The conclusion (decision)
1-4
Revision The Confidence Interval (C.I) for the population mean
≥ 0 < 30
̅± × ̅± ×
̅± × ̅± ×
The population is normal, or not The population must be normal.
VIP normal.
• The margin of error (the maximum error of estimation):
𝝈 𝑺 𝑺
𝟏
𝜶 × Or E= 𝟏
𝜶 × Or E= (𝒏 𝟏
𝜶
, ) ×
𝟐 𝒏 𝟐 𝒏 𝟐 𝒏
• The confidence level (C.L) is denoted by ( − ) %
• Lower Confidence Limit: = ̅−E • Upper Confidence Limit: = ̅+
• Some values of :
𝟎.𝟗𝟓 𝟎.𝟗𝟕𝟓 𝟎.𝟗𝟗 𝟎.𝟗𝟗𝟓
1.645 1.96 2.326 2.576 from the " " table
1-5
Steps and forms of hypothesis tests about the population mean
The test may take one of the following forms, depending on
the claim (hypothesis) made by a person.
Someone claims that Someone claims that Someone claims that
Second year students Second year students Second year students
have low average grades, have average grades, have high average grades,
No more than 60 degrees equal to 80 degrees Not less than 90 degrees
A person claims a value for the average of all students, i.e. population mean
Right side Tail test two-sided Tail test Left side Tail test
The claim is false if the The claim is false if the The claim is false if the
value is greater than 60 value is less than or value is less than 90
greater than 80
1-6
VIP Left tailed Test Two tailed Test Right tailed Test
1- state the
hypotheses
2- calculate The test 𝟎 𝟎
𝝈 𝒐𝒓 𝒔 OR 𝒔
Statistic
𝒏 𝒏
(-) ( ) , (+)
3- find The critical
Or Or Or
value (Table)
(-) ( ) ( , ) (+)
4- The conclusion Reject 𝑯𝟎 𝒊𝒇 𝒁∗ > critical
(decision) Accept (can't Reject) 𝑯𝟎 𝒊𝒇 𝒁∗ ≤ critical
𝟎 the null hypothesis (Contains the equal sign)
𝟏 the alternative hypothesis (doesn't Contains the equal sign)
𝟎 the hypothesized value
When finding critical values We use if it is a one-sided test
We use /2 if it is a two-sided test
1-7
≥ 0 < 30
𝟎
𝒔
𝒏
The population is normal, or The population must be
not normal. normal.
1- < 30
2-
3- population must be normal
1-8
Example (1):
In a random sample of size 36 drawn from a normal population
( , ) it is found that the sample mean = 24. Test the hypothesis
that the population mean is:
(a) Greater than 22 at significance level 1%.
(b) Equals 22 at significance level 1%.
(c) At least 25 at significance level 1%.
Solution:
(a) Greater than 22 at significance level 1%.
1- state the hypotheses
The claim Right tailed Test
1-9
2- calculate The test Statistic
We use because the test is
one-sided and (+) because it's right
3- find The critical value (Table) +
4- The conclusion
>
we reject and accept , i.e. accept the claim that the
population mean is greater than 22
1-10
1 2
Select Summarized Data
The value in null
Select to perform test
Select the sign of alternative
1-11
That is your alternative
The test Statistic Z*
How do you make the decision?
1-12
Decision Making
When we use Minitab we can make a decision using P_Value
reject
What is the P_Value
The area under the
distribution curve
and to the right of P_Value
the absolute value of
the test statistic
1-13
b) Equals 22 at significance level 1%.
1- state the hypotheses The claim
Two tailed Test
2- calculate The test Statistic
We use because the test is Two
tailed and ( ) because it's Two tailed
3- find The critical value (Table)
4- The conclusion
>
we reject and accept , i.e. reject the claim that the population
mean equals to 22
1-14
1
2
1-15
reject
reject
1-16
c) At least 25 at significance level 1%.
1- state the hypotheses The claim
Left tailed Test
2- calculate The test Statistic
𝟎
𝝈 𝟒 𝟒 -1.5
𝒏 𝟔
𝟑𝟔 We use because the test is one
tailed and ( ) because it's Left tailed
3- find The critical value (Table)
4- The conclusion
<
we accept (can't Reject) and reject , i.e. accept the claim that
the population mean is at least 25
1-17
1
2
1-18
That is your alternative
The test Statistic Z*
1-19
Example (2):
Consider the following null and alternative hypotheses:
: ≥ 40 : < 40 A random sample of 64
observations taken from this population produced a sample mean of
38.9. The population standard deviation is known to be 6. At 5%
significance level, would you reject the null hypothesis?
Solution:
1- state the hypotheses
Left tailed Test
2- calculate The test Statistic
𝟎
𝝈 𝟔 -1.47
𝒏 𝟔𝟒
1-20
We use because the test is one
tailed and ( ) because it's Left tailed
3- find The critical value (Table)
4- The conclusion
<
we accept (can't Reject) and reject
i.e the population mean at least 40 at significance level 5%
1-21
Example (3):
In a random sample of the life time of 100 car batteries of
a known brand, it was found that the sample mean is 30
months, and the sample variance is 400. Test that the
population mean is smaller than 35 month, at = .
Solution:
1- state the hypotheses
The claim Left tailed
2- calculate The test Statistic
𝟎
𝑺 𝟐𝟎 -2.5
𝒏 𝟏𝟎𝟎
1-22
We use because the test is one
tailed and ( ) because it's Left tailed
3- find The critical value (Table)
4- The conclusion
>
we reject and accept , i.e. accept the claim that the
population mean is smaller than 35. (the life time average for this
kind of batteries less than 35 month)
1-23
Note: when we use computer no need for Approximation
1
2
1-24
1-25
Example (4):
The management of some industry is considering a new
method of assembling its products. The present method
requires 41.3 minutes, on average, to assemble one unit.
The mean assembly time for a random sample of 100 units
using the new method is 40 minutes and the sample
standard deviation is 5 minutes. Can we conclude that the
new method is better than the present one, at significance
level 0.05?
Solution: The new method will be better than the present one if it takes less time
For new method 𝟎 For present method = 41.3
The claim Left tailed
1-26
2- calculate The test Statistic
𝟎
𝑺 𝟓 -2.6
𝒏 𝟏𝟎𝟎
We use because the test is one
tailed and ( ) because it's Left tailed
3- find The critical value (Table)
4- The conclusion
>
we reject and accept , i.e. accept the claim that the new
method is better than the present.
1-27
Example (5):
A major car manufacturer wants to determine whether a
new engine meets new air pollution standards. The mean
emission of all engines of this type ( ) must be less than 20
parts per million of carbon. A random sample of 49 engines
was drawn for testing purpose, and the emission level of
each determined giving a sample mean of 18 and a sample
standard deviation of 5.
(1) At significance level = . , do the data supply
sufficient evidence to allow the manufacturer of that type
of engines meets the pollution standard?
(2) Construct a 95% confidence interval for the mean
emission level of all engines of that type.
Solution:
1-28
a) Test at significance level 1%.
1- state the hypotheses
The claim Left tailed Test
2- calculate The test Statistic
𝟎
𝒔 𝟓 -2.8
𝒏 𝟒𝟗
We use because the test is one
tailed and ( ) because it's Left tailed
3- find The critical value (Table)
4- The conclusion
>
we Reject and accept , i.e. accept the claim that this type of
engines meets the pollution standard
1-29
2) a 95% confidence interval for the mean emission level of all engines of that type
Note that both
limits less
Thus we are 95% confident that the population than 20
mean lies between 16.6 and 19.4. ??????
1-30
Example (6):
The past records of a supermarket show that its customers
spend an average of L.E 65 per visit with standard deviation
L.E 24. Recently the management of the store initiated a
promotional campaign according to which each customer
receives coupons depending on how much he spends. These
coupons allow the customer, on his next visit, to purchase some
of the store's products for reduced prices. The management
expects that as a result of his campaign, the customers should
be encouraged to spend more at the store. To check whether
this is true, the following data is collected on the amount of
money spent by a sample of 9 customers on their visit to the
store after the initiation of the campaign.
88 69 78 106 75 110 51 78 74
Assuming the normality, evaluate the effect of the promotional
campaign at level 5%.
1-31
1- state the hypotheses
The claim Right tailed Test
Spend more
2- calculate The test Statistic
1-32
We use because the test is
one-sided and (+) because it's right
3- find The critical value (Table) +
4- The conclusion >
we reject and accept , i.e. accept the claim that the
customers spend more.
1-33
1
1-34
1-35
Example (7):
An auto batteries producer claims that the mean life of his
batteries is at least 36 months. A consumer protection agency
tested a random sample of 16 such batteries to check the
claim. It found that the mean life of the batteries in this
sample is 34.5 months with standard deviation 2 months. At 5
% significance level, test the producer's claim, assuming that
the life time has approximately normal distribution.
Solution:
1- state the hypotheses The claim
Left tailed
2- calculate The test Statistic
normal distribution
1-36
= -3
We use because the test is one
tailed and ( ) because it's Left tailed
3- find The critical value (Table)
4- The conclusion
>
we reject and accept , i.e. reject the claim that the life time is
at least 36 months.
1-37
1-38
Example (8):
A past study claims that students spend an average of 26
hours a week on certain activities. A researcher wanted to
test this claim. He took a sample of 5 students and collected
the time they spend on these activities. The data (in hours)
are as follows: 26 21 20 25 28 Assume that the time spent by
all students is normally distributed. Can we conclude that the
claim of the study is true, at significance level 5%?
Solution:
1- state the hypotheses The claim
Two tailed
2- calculate The test Statistic
population is normal
1-39
-1.32
We use because the test is Two
tailed and ( ) because it's Two tailed
3- find The critical value (Table)
4- The conclusion
<
we accept and reject , i.e. accept the claim that the students
spend an average 26 hours
1-40
Homework
Question (1): In a random sample of the life time of 100 car batteries
of a known brand, it was found that the sample mean is 30 months,
and the sample variance is 400. Test that the population mean is
smaller than 35 at =0.05.
Question (2): A government agency has received a complaint that
boxes supposed to be containing 500 grams of a certain product
actually contain less. A simple random sample of 36 such boxes is
selected from various stores to clarify this issue, gave a sample mean
480 grams. Assume that the weights of the boxes are normally
distributed with standard deviation 30 grams. Test the consumer
complaint at 0.01 level of significance.
1-41
Question (3): A machine is adjusted to produce pipes with mean
diameter 5 cm and standard deviation 4 mm. A sample of 25 pipes
from the production of this machine was selected at random and
examined for quality. The mean diameter was found to be 4.8 cm.
Does this information provide enough evidence to claim that the
machine is not working up to the standards at significance level 5%?
Question (4): In a random sample of 16 observations it was found
that ̅ = 2= . Test the claim that the population mean
is bigger mean than 47 at = . .
Question (5): A certain machine is designed to produce a metal bars
with average length 45 cm. The production supervisor claims that this
machine needs to maintenance. To investigate a random sample of 9
units were selected and the length of each is measured as follows: 42,
39, 45, 43, 40, 41, 44, 46, 47 Assume that the length of all units
produced by this machine is normally distributed. Do you agree with
the production supervisor claim? Use significance level 0.01.
1-42
Question (6): A contract between an iron and steel company and a
raw iron supplier states that the amount of ash and dust must not
exceed 80 Kg/ton with standard deviation 30Kg/ton. The quality
control department of this company has recently examined a random
sample of 25 different tons of raw iron and found that the average
weight of ash and dust in this sample was 100Kg/ton. As a result, a
report is raised to the top management indicating that this supplier is
not meeting the contact commitments, and a certain action should be
taken. At 5% significance level, do you accept this conclusion of the
quality control department? (Assuming the normality).
Question (7): Assume that = , ̅ = , 2= . Test the claim
that the population mean is larger than 25 at =0.05. Construct the
95% C.I for . Assume the normality.
1-43
Question (8): The present time pilot take to complete a certain
corrective action is 6.8 second on average. The aviation
administration is considering a new training program for improving
the performance of pilot. A random sample of 10 pilots was tested in
a flight simulator after two months of training and the time for each
to complete the target corrective action was measured in seconds to
give the follows: the sample mean =6 and the sample standard
deviation = 0.91287. Assuming the normality: Test the effectiveness
of the training program, using a significance level 0.05.
1-44
Question (9): An expert in education claims that his new method
for teaching Math for primary school students has a significantly
larger mean score than the old method whose mean was 60. To
study this claim, a 4 random sample of 25 students was drawn and
the new method was applied for a year. The result of the final exam
showed that the sample mean score is 64 and the sample variance
is 125. Do the following, Assuming the normality: a) Test the
previous claim at =0.05 b) Test the claim that the two methods
are not significantly different at = . .
Question (10): The experts in an agricultural research center
claim that they produced a new brand of wheat seeds which has a
significantly larger productivity than the old brand whose mean
productivity is 20. A random sample of 10 similar acres was
drawn and planted with the new seeds. The sample mean and
variance are 24 and 40 respectively. Assume the normality to test
the expert’s claim at significance level 0.05.
1-45
T Table
1-46
