Probability & Statistical Modelling

AQ077-3-2-PSMOD and VE1

Hypothesis Testing
 Topic & Structure of The Lesson
 Introduction to hypothesis testing
 Hypothesis testing on population mean
 Hypothesis testing on population proportion
 Test of Goodness-of-fit
 Test of Independency

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Learning Outcomes

At the end of this section, You should be able to:

 Explain the principles underlying hypothesis testing
 Structure a business decision situation about means
or proportions into the form of a test of a hypothesis
 Apply systematic testing procedures.
 Interpret hypothesis test results and draw conclusions

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Key Terms You Must Be Able To
Use
If you have mastered this topic, you should be able to use the following terms
correctly in your assignments and exams:
(Prepare your own list )

Hypothesis statement
Significance level
Test statistic
Decision rule
Critical value

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Hypothesis Testing
Introduction
 It is also called significance testing.
 It is used to test a belief or opinion by applying statistical
methods.
 In decision making, we make an assumption, called hypothesis,
then we collect some sample data, produce sample statistics and
use this information to decide how likely it is that our
hypothesized population parameter is true.
 Commonly used for testing sample means and proportion.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Hypothesis Testing
Null Hypothesis (H0):
• Is our assumption about parameter of
population.
• The hypothesis we wish to falsify.
• Assumed to be true until we can prove
otherwise.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 Alternative Hypothesis (H1):
• The opposite of null hypothesis.
• If H0 is rejected, then we accept H1.

Null Hypothesis Alternative Hypothesis

(H0) (H1)
=

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 Hypothesis Testing Common
Phrases

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 Hypothesis Testing
Steps:
1. State the claim and the alternative,
2. Choose a significance level or use the given one.
3. Draw the sampling distribution based on the
assumption that is true, and shade the area of
interest.
4. Calculate the test statistic. Find the p-value.
5. If the p-value is less than the significance level (p-
value < ), reject the null hypothesis.
6. If the p-value is greater than the significance level
(p-value>), do not reject the null hypothesis.
7. Write a statement to interpret the decision in the
context of the original claim.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 One-Tailed vs Two-Tailed Test

Two-Tailed Test Left-Tailed Test Right-Tailed Test

H0

H1

Rejection Region In both tails Left tail Right tail

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 One-Tailed vs Two-Tailed Test
One-tailed Test
 Is a significance test in which the null hypothesis can
be upset by values well above or below the mean but
not both.

Two-tailed test
 Is a significance test in which it will reject the null
hypothesis if the sample mean is significantly higher
or lower than hypothesized population mean.(i.e.
there are two rejection region)

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Quick Review Question
Determine whether the following is left-tailed, right-tailed, or
two tailed test.
1. To test whether or not a bank’s ATM is out of service for
an average of more than 8 hours per month.
2. To test if the mean length of experience of airport
security guards is different from 5 years.
3. To test if the mean length credit card debt of college
seniors is less than $1000.
4. To test if the mean amount of time spent per week
watching sports on television by all adult men is
different from 6 hours.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 Test Statistics
• When is given:

• When is not given, :

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 Test Statistics
• When is not given, :

• For proportion:

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 Example

A researcher reports that the average salary of

assistant professors is $42,000. A sample of 32
assistant professors has a mean salary of $43, 260.
At , test the claim that the average assistant
professors earn more than $42,000 a year. The
standard deviation of the population is $5230.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Quick Review Question
µ = Mean salary of assistant professors

Assumptions:
Ho :   42000 1. SRS
 2. Approx. Normal since n>30
H A :   42000 3. Independent: 10(32)=320

x Use a Z-Test since σ

z Pval  P ( x  43260)  P ( z  1.36)
 is known
n Pval  0.0869
43260  42000
z
5230
32
z  1.36 Fail to Reject the Ho since P-value (0.087)>α (0.05).
There is insufficient evidence to support the claim that
the mean salary of assistant professors is more than
$42000.
AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing
 Quick Review Question

The director of admissions at a large university

advises parents of incoming students about the
cost of textbooks during a typical semester. A
sample of 100 students enrolled in the university
indicated a sample mean cost of $315.40 with a
sample standard deviation of $43.20. Using the
0.01 level of significance, is there evidence that
the population mean is above $300?

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 Quick Review Question

The TIV Telephone Company provides long-distance

telephone service in an area. According to the company’s
records, the average length of all long-distance calls placed
was 12.44 minutes. A sample of 150 such calls placed
through this company produced a mean length of 13.71
minutes. The standard deviation of all such calls is 2.65
minutes. Using the 2% significance level, can you conclude
that the mean length of all current long-distance calls is
different from 12.44 minutes?

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 Quick Review Question

A waterworks officer insists that the average daily

household water use in a certain country is at least
400 gallons on average. To check this claim, a
random sample of 25 households was checked.
The average of those sampled was 380 with a
standard deviation of 62. Is this consistent with the
officer’s claim? Test at 2.5% significance level.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 Quick Review Question

A car manufacturer claims that at least 10.2% of

their new cars will be returned back to them due to
some technical problem. A random sample of 125
cars was selected and 10 of them are considered
default. Using 5% significance level, would you
agree with the manufacturer’s claim?

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 Non-parametric Tests
Introduction
 The significance tests covered so far depend, to
greater or less extent, on the assumption, or
presence of the normal distribution
 They are also concerned with the parameters of
the distribution e.g. mean, proportion. Hence
given the mean of parametric tests.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Non-parametric Tests

However, non-parametric or distribution free tests

may be used for the following cases:
1. the data are not normal, or
2. contain extreme values or
3. not enough is known to be able to make any
assumption about the type of distribution.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Advantages of Non-Parametric
Tests
 No assumptions need to be made about the
underlying distribution.
 They can be used on data ranked in some order.
 Mathematic concepts are simpler than for
parametric tests.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Disadvantages of Non-
Parametric Tests
 They are less discriminating than parametric
tests. I.e. they are more prone to error and
less powerful.
 Although simple, the arithmetic may take a long
time

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Chi-square (2) Distribution
 Is used when it is wished to compare an actual,
observed distribution with a hypothesized, or
expected distribution.
 Often referred to as a ‘goodness of fit’ test.
 Chi Square:
2

 
O  E
where E
O = the observed frequency of any value
E = the expected frequency of any value

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Chi-square (2) Distribution
• Our question of interest is “Are the two
variables independent?”
• This question is set up using the following
hypothesis statements:
• Null Hypothesis: The two categorical
variables are independent.
• Alternative Hypothesis: The two categorical
variables are dependent.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Chi-square (2) Distribution

 The obtained value from the formula is compared with the

value from 2 table for a given significance level and the
number of degrees of freedom.

Degrees of freedom, df = (Rows - 1)(Columns – 1)

 If 2 (calculated) is more than 2 (table), the null

hypothesis is rejected.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 How to read Chi-square table?

• The chi-square distribution assumes nonnegative

values only, and these are denoted by the symbol X2
(read as chi-square).

• E.g. 1: Find the value of X2 for 7 degrees of freedom and an area

of 0.10 in the chi-square distribution.

• E.g. 2: Find the value of X2 for 12 degrees of freedom and an

area of 0.05 in the chi-square distribution.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 Chi-square (2) Distribution

• Expected frequency,

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Example
The following data concerning industrial accidents
and absentees classified according to the types of
employee.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 1. State the null and alternative hypothesis.
2. Find degree of freedom and (table) given the
significance level is 5%.
3. Find expected frequency and chi square.
4. Determine whether you reject or do not reject H0 if
the significance level is 5%.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Example (Answer)
1. H0: The absent due to accidents is independent from the
types of employees.
H1: The absent due to accidents is dependent
from the types of employees.
2. Degree of freedom df = (3 - 1)(2 – 1) = 2.
(table) = 5.991

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 3. Expected (E) =
Men Women Juvenile Total
s
Up to 1 month 26 16 8 50
(E=25) (E=15.625) (E=9.375)
1 month or longer 14 9 7 30
(E=15) (E=9.375) (E=5.625)
Total 40 25 15 80

Type of
O E (O – E)2
employee
Up to 1 month Men 26 25 1.000 0.040
Women 16 15.63 0.141 0.009
Juveniles 8 9.375 1.891 0.202
1 month or
Men 14 15 1.000 0.067
longer
Women 9 9.375 0.141 0.015
Juveniles 7 5.625 1.891 0.336

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 2

 
 O  E
 0.668
E

4. Since 0.668 < 5.99, we fail to reject H0.

Therefore the absent due to accidents is
independent from the types of employees.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 Quick Review Question

A random sample of 400 households is classified by two characteristics:

Whether they own a colour television and by what type of householders

(ie: owner-occupier, private tenant, council tenant). The results
investigations are:
Actual Frequencies
Owner Council Private TOTAL
occupier tenant tenant
Colour TV 150 60 20
No Colour TV 45 68 57
TOTAL

Test at 5% level of significance for the relationship between the two

classifications.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 Quick Review Question
A random sample of 400 households is classified by two
characteristics: Whether they own a colour television and by what
type of householders (ie: owner-occupier, private tenant, council
tenant). The results investigations are:
Actual Frequencies
Owner Council Private
occupier tenant tenant
Colour TV 150 60 20
No Colour TV 45 68 57

Test at 5% level of significance for the relationship between the

two classifications.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

Slide ‹#› of 9
 Follow Up Assignment

 Tutorial Questions

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Summary of Main Teaching Points

 Hypothesis or significant testing is testing a

belief or opinion by statistical methods.
 Significance levels are complementary concepts
to confidence limits.
 The Null hypothesis usually assumes there is no
difference between observed and believed values.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

  A one-tailed test is concerned with only one tail of
the distribution I.e. a difference in one direction only.
 The Chi-squared (2) distribution is used for
comparing an actual distribution with an expected
distribution.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

  A shorthand way of expressing the level of
significance for a test is 20.05, i.e. a 2 test using a
5% level of significance, t0.01 I.e. a t test using a 1%
level of significance and so on.

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Summary of Main Teaching Points

Hypothesis testing on population mean and proportion

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 Question and Answer Session

Q&A

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing

 What we will cover next

 Decision Making Techniques

AQ077-3-2 Probability and Statistical Modelling Hypothesis Testing