Week 12 and 13

Week 12 & 13: Chapter 8
Hypothesis Testing
C
h
a
p
t
e
r
Tests of Hypotheses
Section 8.1: The Elements of a Test of Hypotheses
Section 8.2: Formulating Hypotheses and Setting Up the
rejection Region
Section 8.3: Test of Hypothesis about a Population Mean:
Normal (z) Statistic
Section 8.4: Observed Significance Levels: p-Values
Section 8.5: Test of a Hypothesis about a Population Mean:
Students t-Statistic
Section 8.6: Large-Sample Test of a Hypothesis about a
Population Proportion
8
2
C
h
a
p
t
e
r
Hypotheses
Hypothesis:
A hypothesis is a statement that something is true.
Example: The statement the mean weight of all bags of
pretzels packaged differs from the advertised weights of
454 grams is a hypothesis.
Statistical Hypothesis:
A statistical hypothesis is a conjecture about a population
parameter.
Hypotheses are always in terms of the parameter (eg., , ,
etc. ) NOT the statistic (eg. x , p , etc.)
C
h
a
p
t
e
r
Terminology
Null Hypothesis ( )
The hypothesis to be tested
If the original claim includes equality (, =, ), it is the

null hypothesis.
If the original claim does not include equality (<, , >),

then the null hypothesis is the complement of the
original claim.
The null hypothesis always includes the equal sign.
8
4
C
h
a
p
t
e
r
Terminology
Alternative Hypothesis ( or .)
A statement which is true if the null hypothesis is false.
Determines the type of test used (left-tail, right-tail, or

two-tail)
Also called research hypothesis.
8
5
C
h
a
p
t
e
r
The form of a Null Hypothesis

The form of a null hypothesis is:
0 : =
(. . 0 : = 21)
where the hypothesized value is a specific number
determined by the problem context.
The alternative hypothesis will have one of the

following forms:
: >
(e. g. : > 21 )
: <
(. . : < 21 )
:
(. . : 21 )
6
C
h
a
p
t
e
r
The form of a Null Hypothesis

The null hypothesis
is usually stated as
an equality
0: = 0
: < 0
: 0
: > 0
the alternative is an inequality.
8
7
C
h
a
p
t
e
r
Identifying Hypotheses
Accidents Involving Teen Drivers
Teenagers (age 15 to 20) make up 7% of the driving

population
14% of accidents studied involved teenage drivers

Does the study provide convincing
evidence that the proportion of
accidents involving teenage drivers
differs from .07, the proportion of
teens in the driving population?
Use = .05.
C
h
a
p
t
e
r
Accidents Involving Teen Drivers
Let p represent the proportion of accidents involving teenage
drivers.
0 : = 0.07; the proportion of accidents involving teenage

drivers is equal to the proportion of teens in the driving
population.
: 0.07; the proportion of

accidents involving teenage
drivers is not equal to the
proportion of teens in the
driving population.
8
9
C
h
a
p
t
e
r
Cholesterol in Children
Cholesterol levels in children is normally distributed
=15
= 190
A sample of 100 children yields

sample mean cholesterol of 196.2.
Do these children have mean
cholesterol levels higher than the
national average at a significance
level of = 0.01?
10
C
h
a
p
t
e
r
Cholesterol in Children
0 = 190; the average cholesterol level of children is

equal to the average cholesterol level of the nations
population.
> 190; the average cholesterol level of children is

higher than the average cholesterol level of the nations
population.
8
11
C
h
a
p
t
e
r
Statistical Test
A statistical test is:
a. Left-tailed if 1 states that
the parameter is less than
the value claimed in 0
b. Right-tailed if 1 states that
the parameter is greater
than the value claimed in 0
c. Two-tailed if 1 states that
the parameter is different
from (or not equal to) the
value claimed in 0
12
C
h
a
p
t
e
r
Scenarios for the Null and

Alternative Hypotheses
Null Hypothesis
Alternative Hypotheses & Type of Test
Claim about or
historical value of
You believe that

is less than the
value stated in 0
You believe that

is more than
the value stated
in 0
You believe that

is different
from the value
stated in 0
: = 0
1 : < 0
1 : > 0
1 : 0
Left-tailed test
Right-tailed test
Two-tailed test
8
13
C
h
a
p
t
e
r
Test Statistic
If the test statistic has
a high probability when
0 is true, then 0 is
not rejected.
If the test statistic has

a (very) low probability
when 0 is true, then
0 is rejected.
14
C
h
a
p
t
e
r
Errors in Hypothesis Testing

A Jury Trial
Null hypothesis: Defendant is innocent.
Alternative hypothesis: Defendant is guilty
8
15
C
h
a
p
t
e
r
Courtroom Analogy
Potential Choices and Errors
Choice 1: We cannot rule out that defendant is

innocent, so he or she is set free without penalty.
Potential error: A criminal has been erroneously freed.
Choice 2: We believe enough

evidence to conclude the
defendant is guilty.
Potential error: An innocent
person is falsely convicted and
guilty party remains free.
16
C
h
a
p
t
e
r
8
Courtroom Analogy
Each trial actually has 4 potential decisions two are

correct decisions, two are errors.
Possible decisions are based on:

the evidence of the defendants innocence or guilt;
the decision that the jury makes based on the evidence
Defendant is Actually
Jurys
Decision
Innocent
Guilty
Not Guilty
Correct
Error
Guilty
Worse Error
Correct
17
C
h
a
p
t
e
r
Hypothesis Testing
The reasoning of Hypothesis Testing is Similar
Each test has 4 potential decisions two are
correct decisions, two are errors.
Possible decisions are based on:
the reality about the null hypothesis;
your decision based on the evidence from the sample.
Null Hypothesis is Actually
Your
Decision
True
False
Dont Reject
the Null
Correct
Type II Error
Reject the
Null
Type I Error
Correct
18
C
h
a
p
t
e
r
Type I & Type II Errors

Type I error:
We reject 0 when in fact 0 is true.
Type II error:
We fail to reject 0 when in fact 0 is false.
8
19
C
h
a
p
t
e
r
Errors and their Consequences

A current cancer treatment has a remission rate of
40%. Is the new treatment more eective?
0 = .4 . : > .4
Type I error: you conclude that the
new treatment is more eective
than the current treatment when it
really isnt.
Type II error: you conclude that the
new treatment is not more
eective than the current
treatment when it really is.
20
C
h
a
p
t
e
r
Steps to do Hypothesis Testing

1. Label the parameter
2. Formulate the null and alternative hypotheses
3. Identify the test statistic and explain why you would use it.
4. State the level of significance
5. Describe the rejection region
6. Calculate the test statistic
7. Decide whether or not to reject the null hypothesis
8. Provide a conclusion in the context of the problem and that
answers the original research question
21
Week 12 & 13: Chapter 8.3-8.4

z-tests and p-values
22
C
h
a
p
t
e
r
Tests of Hypotheses
Rejection Region
8
23
C
h
a
p
t
e
r
Steps to do Hypothesis Testing

1. Label the parameter
2. Formulate the null and alternative hypotheses
3. Identify the test and its conditions
4. State the level of significance
5. Describe the rejection region
6. Calculate the test statistic
7. Decide whether or not to reject the null hypothesis
8. Provide a conclusion in the context of the problem and that
answers the original research question
24
C
h
a
p
t
e
r
Procedure
Label the parameter:
Let be the population mean.
Formulate the Hypotheses:
Null Hypothesis
0 : = 0 , a specified value; (in words)
Alternative hypothesis
: > 0 , a specified value; (in words)
: < 0 , a specified value; (in words)

: 0 , a specified value; (in words)
25
C
h
a
p
t
e
r
Procedure
Identify the test and its conditions:
Test
Large sample z-test for
Conditions
A random sample is selected from the target population.
The sample size is large.
State the level of significance:
= ??? (usually given, if not choose =.05)
26
C
h
a
p
t
e
r
Procedure
Describe the rejection region:
> if upper tailed test (or : > 0 )

< if lower tailed test (or : < 0 )
< 2 > 2 if two-tailed test (or : 0 )
Rejection regions for common values of for large sample

-test
Lower
Tailed
Upper
Tailed
Two tailed
= .10
z - 1.28
z > 1.28
z-1.645 or z > 1.645
= .05
z - 1.645
z > 1.645
z -1.96 or z > 1.96
= .01
z - 2.33
z > 2.33
z-2.575 or z > 2.575

27
C
h
a
p
t
e
r
Rejection Regions
Example
=.05 (Upper Tailed Test)
=.01 (Two Tailed Test)
8
28
C
h
a
p
t
e
r
Procedure
Calculate the test statistic:
x 0 x 0
zcal
x
/ n
Provide the decision and conclusion:
If the calculated value of the proposed test

statistic belongs to the rejection region, we reject
0 ; otherwise we fail to reject 0 .
Write the conclusion in the context.

29
C
h
a
p
t
e
r
How to write the conclusion?

Conclusions are based on the original claim, which
may be the null or alternative hypotheses.
Original Claim
Decision
H0
Reject
Ha
Support
Reject H0
sufficient
We conclude that there is

sufficient evidence at %
level of significance to
reject the claim that

support the claim that
Fail to reject H0
insufficient

insufficient evidence at %

support the claim that
30
C
h
a
p
t
e
r
Example: Response Time
n=100 rats
Sample mean = 1.05 seconds
Standard deviation = .5 second
Control mean = 1.2 seconds
Does the mean response

time for drug-injected rats
differ from 1.2 seconds at
= .01.
31
C
h
a
p
t
e
r

Let represent the mean response time for druginjected rats.
Hypotheses:
: = 1.2; that is, mean response time is 1.2

seconds.
: 1.2; that is, mean response time less than
1.2 or greater than 1.2 seconds.
8
32
C
h
a
p
t
e
r

Test:
We will perform a large sample z-test for the mean
response time for drug-injected rats.
Conditions:
Assume that 100 rats are selected randomly.
The sample size n=100 is large.
8
33
C
h
a
p
t
e
r

Level of Significance:
= 0.01
Rejection Region:
< 2.575 > 2.575
Calculation of test statistic:
zcal
x 0
x 0 1.05 1.2
3.0
/ n .5 / 100
34
C
h
a
p
t
e
r

Decision and Conclusion:
Since = 3.0 is less than 2.572 we

reject the null hypothesis.
Conclusion ???
35
C
h
a
p
t
e
r
How to write the conclusion

Conclusions are based on the original claim, which
may be the null or alternative hypotheses.
Original Claim
Decision
0
Reject
Support
Reject 0
sufficient


sufficient evidence at % level
of significance to support the
claim that
Fail to reject 0
insufficient


level of significance to support
the claim that
36
C
h
a
p
t
e
r

Since = 3.0 is less than 2.572 we reject

the null hypothesis.
We conclude that there is sufficient evidence at

1% level of significance to support the claim that
the mean response time for drug-injected rats
differ from the control mean of 1.2 seconds.
37
C
h
a
p
t
e
r
Example: Internet Use
n=676 parents of Canadian teens

Sample mean = 6.5
Sample standard deviation = 8.6
Do the sample data
provide convincing
evidence that the mean
number of hours that
parents think their teens
spend online is less than
10 hours per week at
=.05?
38
C
h
a
p
t
e
r

Let = mean number of hours per week that
parents think their Canadian teens spend online.
Hypotheses:
: = 10 ; the mean number of hours that

parents think their teens spend online is equal to
10 hours per week.
: < 10 ; the mean number of hours that
parents think their teens spend online is less than
10 hours per week.
39
C
h
a
p
t
e
r

Test:
We will perform a large sample z-test for the mean
number of hours per week that parents think their
Canadian teens spend online.
Conditions:
It is reported that a sample of 676 parents of
Canadian teens were selected randomly.
The sample size n=676 is large.
8
40
C
h
a
p
t
e
r

Level of Significance:
= 0.05
Rejection Region:
< 1.645
Calculation of test statistic:
We are given: = 676, = 6.5, = 8.6
zcal
x 0
x 0
6.5 10
10.58
/ n 8.6 / 676
41
C
h
a
p
t
e
r

Since = 10.58 is less than 1.645, that is,
= 10.58 belongs to the rejection region, so
we reject the null hypothesis.
We conclude that there is sufficient evidence at

5% level of significance to support the claim that
the mean number of hours that parents think their
teens spend online is less than 10 hours per
week.
42
C
h
a
p
t
e
r
Terminology
The -Value (Short for Probability value)
The probability of obtaining a test statistic from the
sampling distribution that is as extreme or more extreme (as
specified by ) than the observed test statistic (computed
from the sample data) under the assumption that 0 is true.
Calculated instead of the rejection region
Decision based on the -value
8
43
C
h
a
p
t
e
r
Determining the -value when

the test statistic is normal
Upper-tailed test:
: > 0 (hypothesized value)
8
44
C
h
a
p
t
e
r

Lower-tailed test:
: < 0 (hypothesized value)

P-value = area in lower tail = P(z zcal)
8
45
C
h
a
p
t
e
r

Two-tailed test:
: 0 (hypothesized value)
P-value = sum of area in two tails
= P(z - zcal or z zcal) = 2 P(z | zcal |)
8
46
C
h
a
p
t
e
r
Decision rules based on the -value
If the p-value < , we reject 0
If the p-value , we do not reject 0
8
47
Week 12 & 13: Chapter 8.5

t-test for a Population Mean
48
C
h
a
p
t
e
r
Tests of Hypotheses
Rejection Region
8
49
C
h
a
p
t
e
r
Using a t-test for hypothesis testing

Label the parameter
Let be the population mean.
Hypotheses
Null Hypothesis
0 : = 0 a specified value; (in words)
Alternative Hypothesis
: > 0 a specified value; (in words)
: < 0 a specified value; (in words)
: 0 a specified value; (in words)
50
C
h
a
p
t
e
r

Test
We will perform a small sample t-test for the population
mean.
Required Conditions
The sample was randomly selected from the population of
interest or there is some other indication that it was
representative (implying randomness).
The original population is known to be normal.
Population standard deviation () is unknown.
Sample size is small.
51
C
h
a
p
t
e
r

Level of Significance
=? ? ? (usually given, if not choose = .05)
Rejection Region
<
>
if two-tailed test (or : 0 )
Calculation of test statistic

x 0
tcal
s/ n
~ t( n 1) df
52
C
h
a
p
t
e
r

Decision and conclusion
If the calculated value of the proposed test statistic belongs
to the rejection region, we reject H0; otherwise we fail to
reject H0.
8
53
C
h
a
p
t
e
r
Finding -Values for a Test
: >
2. Lower-tailed test
: <
3. Two-tailed test
= P (t > tcal)
1. Upper-tailed test
= P (t < tcal)
= 2 *P (t > |tcal|)
Decision Rule
If < , we reject 0
54
C
h
a
p
t
e
r
Example: 40lb bags of Dog Food

A random sample of 10 bags of Dogspal dog food:
37.25
38.25
40.10
40.50
41.25
39.45
37.00
39.25
38.00
40.75
Conduct a test of significance

to test the claim that Dogspal
40lb bags have less than 40.5
pounds of dog food in their
bags at a significance level of
=.05
55
C
h
a
p
t
e
r

1.
Label the parameter
2.
Formulate the hypotheses
3.
Identify the test and its conditions
4.
State the level of significance
5.
Describe the rejection region
6.
Calculate the test statistic
7.
State the decision and

conclusion
8
56
C
h
a
p
t
e
r

1. Label the target parameter
, the true mean weight of 40 lb bags
2. State the hypotheses:

o
0 : = 40.5; the mean of all 40 pound bags from

Dogspal is 40.5 pounds, as stated by the company.
: < 40.5; the mean of all 40 pound bags from

Dogspal is less than the 40.5 pounds stated by the
company.
8
57
C
h
a
p
t
e
r

3. State the test
o We will perform a one sample t- test for the population
mean.
4. Verify the required conditions:

We are told that the sample of 10 was a random
sample.
The distribution of weights packaged by Dogspal is
normal based on the boxplot we saw in Chapter 7.
? The standard deviation of weights packaged by Dogspal
is unknown.
Sample size n=10, is small.
58
C
h
a
p
t
e
r

5. State the level of
significance
o = 0.05
6. Define the rejection

region:
o < 1.833
7. Calculate the test statistic:

We are given:
= 10
= 39.18
= 1.4963
df n 1 10 1 9
tcal
x 0 39.18 40.5
2.79
s / n 1.4963 / 10
59
C
h
a
p
t
e
r

P-value
p-value = area of t-curve under 9 df to the left of -2.79
= area of t-curve under 9 df to the right of 2.79
=?
OR
p-value = < 2.79 = > 2.79 = ?
8
60
C
h
a
p
t
e
r
.010 p-value .025
8
2.79
61
C
h
a
p
t
e
r

P-value
p-value = area of t-curve under 9 df to the left of -2.79
= area of t-curve under 9 df to the right of 2.79
= . < <.
OR
p-value = < 2.79 = > 2.79 = . < <.
8
62
C
h
a
p
t
e
r

Decision and Conclusion
Decision Using Rejection Region
Since = 2.79 is less than 1.833 so we reject the null
hypothesis.
Decision Using p-value

We get .01 < < .025, that is, p-value is between .01
and .025, not inclusive.
Since < = .05, so we reject the null hypothesis.
We conclude that there is sufficient evidence at 5% level of

significance to support the claim that the mean of all 40
pound bags from Dogspal is less than the 40.5 pounds
stated by the company.
63
Week 12 & 13: Chapter 8.6

Large-Sample Test for a Population Proportion
64
C
h
a
p
t
e
r
Tests of Hypotheses
Rejection Region
8
65
C
h
a
p
t
e
r
Using a Large-Sample z-Test

Label the parameter
Let be the proportion of success.
Hypotheses
Null Hypothesis
0 : = 0 a specified value; (in words)
Alternative Hypothesis
: > 0 a specified value; (in words)
: < 0 a specified value; (in words)
: 0 a specified value; (in words)
66
C
h
a
p
t
e
r

Test
We will perform a large sample z-test for the population
proportion (p).
Required Conditions
o Random sample
o The sample size is large
(0 15, and (1 0 ) 15)
Verification of these assumptions makes it reasonable to
assume the approximate normality of the sampling distribution
of sample proportion,. Therefore, we can perform the z-test.
67
C
h
a
p
t
e
r

Level of Significance
=? (usually given, if not choose = .05)
Rejection Region
<
>
if two-tailed test (or : 0 )

zcal
p p0
N (0,1)
p0 (1 p0 )
n
68
C
h
a
p
t
e
r

Decision and conclusion
If the calculated value of the proposed test statistic belongs
to the rejection region, we reject 0 ; otherwise we fail to
reject 0 .
8
69
C
h
a
p
t
e
r
Determination of the -Value when

the test statistic is
Upper-tailed test
o : >
Lower-tailed test
= P (z > zcal)
= P (z < zcal)
o : <
Two-tailed test
o :
= 2 *P (z > |zcal |)
Decision Rule
If < , we reject 0
70
C
h
a
p
t
e
r
Example: Auto Accidents

Teenagers = 7% of the driving population
In a study of accidents:
n = 500 randomly selected accidents
Teenagers involved in sampled accidents = 14%
Does the study provide convincing

evidence that the proportion of
accidents involving teenage
drivers differs from .07, the
proportion of teens in the driving
population? Use =.05.
71
C
h
a
p
t
e
r

Let represent the proportion of accidents involving
teenage drivers.
Hypotheses:
o 0 : = 0.07; the proportion of accidents involving
teenage drivers is equal to the proportion of teens in the
driving population.
o : 0.07; the proportion of accidents involving
teenage drivers is not equal to the proportion of teens in
the driving population.
8
72
C
h
a
p
t
e
r

Test:
We will perform a large sample -test for the population
proportion ().
Conditions:
o The sample was a random sample of all accidents.
o 0 = 500 0.07 = 35 > 15, and
o 1 0 = 500 0.93 = 465 > 15
8
73
C
h
a
p
t
e
r

Level of significance
We are given, = .05
Rejection Region
< 1.96 or > 1.96

zcal
p p0
p0 (1 p0 )
n
.14 .07
6.13
.07 (1 .07 )
500
74
C
h
a
p
t
e
r

P-value
= 2 (area under the z-curve to the right of 6.13)
20=0
OR
= 2 ( > )
= 2 > 6.13
20
=0
75
C
h
a
p
t
e
r

Decision and Conclusion
Decision Using Rejection Region
Since = 6.13 is greater than 1.96 so we reject the
null hypothesis.
Decision Using p-value

Since = 0 < = .05, so we reject the null
hypothesis.
We conclude that there is sufficient evidence at 5% level of
significance to support the claim that the proportion of all
accidents involving teenage drivers is different the proportion
of teens in the driving population.
76

Week 12 and 13

Diunggah oleh

Informasi Dokumen

Deskripsi Asli:

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Week 12 and 13

Diunggah oleh

Hak Cipta:

Format Tersedia

Week 12 & 13: Chapter 8

The hypothesis to be tested

If the original claim includes equality (, =, ), it is the

If the original claim does not include equality (<, , >),

The null hypothesis always includes the equal sign.

A statement which is true if the null hypothesis is false.

Determines the type of test used (left-tail, right-tail, or

Also called research hypothesis.

The form of a Null Hypothesis

The alternative hypothesis will have one of the

The form of a Null Hypothesis

the alternative is an inequality.

Teenagers (age 15 to 20) make up 7% of the driving

14% of accidents studied involved teenage drivers

0 : = 0.07; the proportion of accidents involving teenage

: 0.07; the proportion of

A sample of 100 children yields

0 = 190; the average cholesterol level of children is

> 190; the average cholesterol level of children is

Scenarios for the Null and

Alternative Hypotheses & Type of Test

You believe that

You believe that

You believe that

If the test statistic has

Errors in Hypothesis Testing

Null hypothesis: Defendant is innocent.

Alternative hypothesis: Defendant is guilty

Choice 1: We cannot rule out that defendant is

Choice 2: We believe enough

Each trial actually has 4 potential decisions two are

Possible decisions are based on:

Type I & Type II Errors

Errors and their Consequences

Steps to do Hypothesis Testing

Week 12 & 13: Chapter 8.3-8.4

Steps to do Hypothesis Testing

Let be the population mean.

Formulate the Hypotheses:

: < 0 , a specified value; (in words)

State the level of significance:

= ??? (usually given, if not choose =.05)

> if upper tailed test (or : > 0 )

Rejection regions for common values of for large sample

z-1.645 or z > 1.645

z -1.96 or z > 1.96

z-2.575 or z > 2.575

=.01 (Two Tailed Test)

Provide the decision and conclusion:

If the calculated value of the proposed test

Write the conclusion in the context.

How to write the conclusion?

We conclude that there is

We conclude that there is

We conclude that there is

We conclude that there is

Example: Response Time

Does the mean response

Example: Response Time

Let represent the mean response time for druginjected rats.

: = 1.2; that is, mean response time is 1.2

Example: Response Time

The sample size n=100 is large.