Sie sind auf Seite 1von 38

Chapter 07 - Hypothesis Testing

CHAPTER 7

HYPOTHESIS TESTING
7-1.

H0: p = 0.8
H1: p 0.8

7-2.

H0: 200
H1: > 200

7-3.

H0: 12
H1: > 12

7-4.

H0: $49,160
H1: < $49,160

7-5.

H0: $3.75
H1: > $3.75

7-6.

The power of a test is the probability that a false null hypothesis will be detected by the test. It is
important if the ramifications of accepting a false null hypothesis outweigh the costs of rejecting
the null hypothesis.

7-7.

Since and are inversely related and since power = (1 - ), then the level of significance and
the power of the test are directly related to each other: as increases, power increases.

7-8.

Increasing the level of significance.

7-9.

a) Null: there are no weapons Alternative: there are weapons


b) Type I error: rejecting the null hypothesis and concluding there are weapons when there are no
weapons.
Type II error: accepting the null hypothesis of no weapons when weapons are present.
c) Type II error would be more costly in terms of endangering lives. Type I would just be an
inconvenience for the passenger.
d) Largest value of possible.
e) If the sensitivity of the detectors is increased, it should result in more Type I errors and less
Type II errors.
f) If is to be increased, then the sensitivity of the detectors should also be increased to
compensate for the higher .

7-10.

Plan for a large sample.

7-11.

a) left tailed H1: < 10


b) right tailed H1: p > 0.5
c) left tailed H1: < 100
d) right tailed H1: > 20
e) two-tailed H1: p 0.22

7-1

Chapter 07 - Hypothesis Testing

f) right tailed H1: > 50


g) two-tailed H1: 2 140
7-12.

a) p= .5 - .4599 = 0.0401
b) p = 0.9599
c) p = .0802

7-13.

a) to the left tail


b) to the right tail
c) either to the left or to the right tail

7-14. A test statistic is a sample statistic computed from the data. Knowing the distribution of the test
statistic enables the construction of the rejection regions about the mean of the test statistic.
7-15.

7-16.

a) p value will decrease


b) p value increases
c) p value decreases
H0: = 31.5
H1: 31.5
x = 29.8
n = 100
s = 6.6
x 0
29.8 31.5
z=
=
= 2.575
6.6 / 10
s/ n
Reject H0 at

7-17.

= 0.05 (and almost at

H0: = 247
H1: 247
x = 250
n = 60
s = 12
x 0
250 247
z=
=
= 1.936
12 / 60
s/ n

= .01). The p-value is approximately 0.01

= 0.05, then also 0.01

(Using the template: Testing Population Mean.xls, sheet: sample stats)


Evidence
Sample size
Sample Mean

60
250

n
x-bar

Known; Normal Population or Sample Size >= 30

Population Stdev.
12
Test Statistic 1.9365 z

Do not reject H0 at
increase your .
7-18.

Null Hypothesis
H0: = 247

= .05 nor at

H0: = 5
H1: 5
= .05 n = 120
x = 2.3

p-value
0.0528

At an of
5%

= .01. (p-value = .0528). Maybe take a larger sample, or

s = 1.5

7-2

Chapter 07 - Hypothesis Testing

z=

2.3 5
1.5 / 120

= 19.72

Reject H0 at = .05 and at smaller levels. Average miles traveled per day is probably smaller
than 5. Changes in service may be necessary (the p-value is very small).
7-19.

H0: = 77
H1: 77
x = 84
n = 50
s = 28
(Using the template: Testing Population Mean.xls, sheet: sample stats)
Hypothesis Testing - Population Mean
Evidence
Sample size
Sample Mean

50
84

n
x-bar

Known; Normal Population or Sample Size >= 30

Population Stdev.
28
Test Statistic 1.7678 z
Null Hypothesis
H0: = 77
H0: 77
H0: 77

p-value
0.0771
0.9615
0.0385

At an of
5%

Reject

Reject H0 at the 0.10 level of (the p-value is small). Customer satisfaction has probably
improved. If a one tail hypothesis test is conducted, the level is reduced to 0.05.
7-20.

H0: = 4.3 H1: 4.3


x = 3.8
n = 50
s = 1.1
Reject H0 at any (p-value = 0.1454).
Evidence
Sample size
Sample Mean

50
3.8

= .05

n
x-bar

Known; Normal Population or Sample Size >= 30

Population Stdev.
1.1
Test Statistic -3.2141 z
Null Hypothesis
H0: = 4.3
H0: 4.3
H0: 4.3

p-value
0.0013
0.0007
0.9993

7-3

At an of
5%
Reject
Reject

Chapter 07 - Hypothesis Testing

7-4

Chapter 07 - Hypothesis Testing

7-21.

H0: =14.25 H1:


x = 16.50
n = 16
t (15) =

16.50 14.25
5.8 /

14.25
s = 5.8

7-23.

H0: = 452.8 H1: 452.8


x = 501.9 s = 65
n = 12
t (11) =

= 0.05

= 1.55

16

Right-hand critical point for t (15) and


Do not reject H0 (p-value > .10).
7-22.

Use

501.9 452.8
65 / 12

=.05 is 2.131 (and for

Use

= .10, t = 1.753).

= .05

= 2.617

The right-hand critical point for t (11) at = 0.05 is 2.201, and for
H0 at = .05. The p-value is between 0.02 and 0.05.

H0: = 27

x = 24.53

H1:

27 n = 43

s = 4.954

= 0.02 it is 2.718. Reject

= .05

Critical value for right side: 1.96


24.53 27

= 3.269
4.954 / 43

z =

Reject H0, average staff age is under 27 years old. p-value = .0011
7-24.

H0: = 0
H1:
x = 3.1
n = 13
t (12) =

3.1 0
1/

13

0
s =1

= 11.177

The p-value is very small. Reject H0. Inclusion in the index likely increases average stock return (at
least initially).
Evidence
Sample size
Sample Mean
Sample Stdev.

13
3.1
1

n
x-bar
s

Unknown; Population Normal


Test Statistic 11.1772 t
Null Hypothesis
H0: = 0

7-25.

H0: = 102.5
H1: 102.5
x = 107 s = 10
n = 25
Try

p-value
0.0000

At an of
5%
Reject

= .05 and .01

7-5

Chapter 07 - Hypothesis Testing

t (24) =

107 102.5
10 /

25

= 2.25

The critical point at = .05 for t (24) is 2.064; the critical point at = .01 = 2.797. Therefore,
the results are significant at = .05, and not significant at = .01. Reject H0 at = .05, but
not at = .01.
7-26.

H0: = 2.5 H1: 2.5


= 0.8
x = 2.3
n = 20
Use z because the population is normal and
z=

2.3 2.5
0 .8 /

20

is known.

= 1.118

Do not reject H0 (p-value = 0.2636).


7-27.

H0: = 12.4 H1: 12.4


n = 100 x = 14.1
use = .05
(Using the template: Testing Population Mean.xls, sheet: sample stats)
Hypothesis Testing - Population Mean
Evidence
Sample size
Sample Mean

100
14.1

n
x-bar

Known; Normal Population or Sample Size >= 30

Population Stdev.
2.6
Test Statistic 6.5385 z
Null Hypothesis
H0: = 12.4
H0: 12.4
H0: 12.4

Reject H0 at

7-28.

p-value
0.0000
1.0000
0.0000

.39818 .42
(.42)(.58) / 550

= 1.037

Do not reject H0 (p-value = 0.2998).


7-29.

H0: p = 0.12

Reject

= .05 (p-value = 0.000).

H0: p = 0.42
H1: p 0.42
= 219/550 = 0.39818 use
x = 219 p
n = 550
z=

At an of
5%
Reject

H1: p

0.12

7-6

= .01

Chapter 07 - Hypothesis Testing

x = 17 use

n = 100

.17 .12

z=

= .05

= 1.539

(.12)(.88) / 100

Do not reject H0 (p-value = 0.1238).


7-30.

H0: 250
H1: < 250
x = 235 use = .05 and .01
n = 30
(Using the template: Testing Population Mean.xls, sheet: sample stats)
Hypothesis Testing - Population Mean
Evidence
Sample size
Sample Mean

30
235

n
x-bar

Known; Normal Population or Sample Size >= 30

Population Stdev.
85
Test Statistic -0.9666 z
Null Hypothesis
H0: = 250
H0: 250
H0: 250

Do not reject H0 at either


7-31.

= .05 and

H0: p = 0.56
H1: p 0.56
= 0.596 use
n = 500
x = 298 p
.596 .56

z=

(.56)(.44) / 500

p-value
0.3338
0.1669
0.8331

At an of
5%

= .01

= .01

= 1.622

Do not reject H0 (p-value = 0.1048).


7-32.

H0: u

7%

H1: u < 7%

s = 2.9225 x = 2.405

n = 41

= .01

2.405 7.00

= 10.07
2.9225 / 41

z=

Reject H0, Americans do not save enough.


7-33.

H0: = 0
H1: 0
n = 24 x = 0.12 s = 0.2 use
t (23) =

.12 0
.2 /

24

= .05

= 2.939 > 2.069 (the critical value of t (23) for

7-7

= .05, two-tailed test).

Chapter 07 - Hypothesis Testing

Reject H0. This industrys stocks probably have positive excess returns in the period in question.
7-34.

H0: = 3.3
H1: 3.3
x = 2.8 s = 1.7
n = 120
(Using the template: Testing Population Mean.xls, sheet: sample stats)
Hypothesis Testing - Population Mean
Evidence
Sample size
Sample Mean

120
2.8

n
x-bar

Known; Normal Population or Sample Size >= 30

Population Stdev.
1.7
Test Statistic -3.2219 z
Null Hypothesis
H0: = 3.3
H0: 3.3
H0: 3.3

p-value
0.0013
0.0006
0.9994

At an of
5%
Reject
Reject

Reject H0 (p-value is very small).

7-35.

H0: 1.5M H1: > 1.5M


x = 2.3M s = 0.5M
n = 100
z=

2.3 1.5
0.5 / 100

= 16.0

Stongly reject H0 (p-value is very small).


7-36.

H0: p 0.17 H1: p > 0.17


= 381/2,000 = 0.1905 use
p
n = 2,000 x = 381
z=

.1905 .17
(.17)(.83) / 2,000

= .01

= 2.44 > 2.326

Reject H0 (p-value = .0073). The reported figure of 17% is probably not correct.

7-8

Chapter 07 - Hypothesis Testing

Evidence

Assumption
Both np and n(1-p) >= 5

Sample size 2000 n


#Successes 381 x
Sample Proportion 0.1905 p-hat
Test statistic 2.4406 z

Hull Hypothesis
H0: p = 0.17
H0: p >= 0.17
H0: p <= 0.17

7-37.

H0:
n = 100
z=

Reject

125 H1: < 125


x = 121 s = 2

121 125
2/

At an of
1%

p-value
0.0147
0.9927
0.0073

100

= 20

Reject H0 (p-value is very small). Replace all tires.


7-38.

Problem 7-38:
(Use template: testing population mean.xls; sheet:sample stats)
H0: =7
H1: 7
Hypothesis Testing - Population Mean
Evidence
Sample size
Sample Mean
Sample Stdev.

100
4
3

n
x-bar
s

Unknown; Population Normal


Test Statistic -10.0000 t
Null Hypothesis
H0: = 7
H0: 7
H0: 7

p-value
0.0000
0.0000
1.0000

At an of
5%
Reject
Reject

Reject the null hypothesis: meetings are less than 7 minutes late in starting.
7-39.

H0: p .45 H1: p < .45


= 49/125 = .392
n = 125 p
z=

.392 .45
(.45)(.55) / 125

= 1.303

7-9

Chapter 07 - Hypothesis Testing

Do not reject H0 (p-value = 0.0962). There is no strong evidence here that the program works in
reducing the chance of a midlife crisis.
7-40.

H0: p 0.11 H1: p < 0.11


= 421/3,500 = 0.1203
n = 3,500 x = 421 p
z=

.1203 .11
(.11)(.89) / 3,500

= +1.945

= .1203 > .11, we need not consider the test statistic value above. Do not reject H0
Since p
(p-value = 0.9741 > .05). There is no evidence that the unemployment rate has been reduced.

Evidence

Assumption
Both np and n(1-p) >= 5

Sample size 3500 n


#Successes 421 x
Sample Proportion 0.1203 p-hat
Test statistic 1.9448 z

Hull Hypothesis
H0: p = 0.11
H0: p >= 0.11

7-41.

At an of
5%

p-value
0.0518
0.9741

H0: 2.5 H1: > 2.5


n = 100 x = 5.2 s = 2.8
z=

5.2 2.5
2.8 / 100

= 9.643

Reject H0 (p-value is very small). There is probable cause for action.


7-42.

(Use template: testing population mean.xls; sheet:sample stats)


H0: 56 H1: < 56

7-10

Chapter 07 - Hypothesis Testing

Hypothesis Testing - Population Mean


Evidence
Sample size
Sample Mean

30
50

n
x-bar

Known; Normal Population or Sample Size >= 30

Population Stdev.
8
Test Statistic -4.1079 z
Null Hypothesis
H0: = 56
H0: 56
H0: 56

p-value
0.0000
0.0000
1.0000

At an of
5%
Reject
Reject

Reject the null hypothesis: the average temperature is less than 56.
7-43.

(Use template: testing population proportion.xls; sheet:normal)


H0: p = 0.75 H1: p 0.75
z-Test for Population Proportion
Evidence

Assumption
Both np and n(1-p) >= 5

Sample size 100 n


#Successes 61 x
Sample Proportion 0.6100 p-hat
Test statistic-3.2332z

Hull Hypothesis
H0: p = 0.75
H0: p >= 0.75
H0: p <= 0.75

p-value
0.0012
0.0006
0.9994

At an of
5%
Reject
Reject

Reject the null hypothesis: the percentage of trades is significantly less than 75%.
7-44.

H0: p = .60 H1: p .60


= 130/250 = .52 Two-tailed test:
p
n = 250 x = 130
p p 0
.52 .6
z=
=
= 2.582
(.6)(.4) / 250
p0 q0 / n

p-value = 2(.5 .4951) = 2(.0049) = 0.0098. Reject H0 at = .01. The proportion of frequent
business travelers who believe that daily service is important has probably decreased.

7-11

Chapter 07 - Hypothesis Testing

7-45.

Standing start:
H0: 5.28 H1: > 5.28
x = 5.8 s = 1.9
n = 100
5.8 5.28
z=
= 2.737
1.9 / 100
(Template: Testing Population Means.xls, sheet: sample stats)
Evidence
Sample size
Sample Mean

100
5.8

n
x-bar

Known; Normal Population or Sample Size >= 30

Population Stdev.
1.9
Test Statistic 2.7368 z
Null Hypothesis
H0: = 5.28
H0: 5.28
H0: 5.28

p-value
0.0062
0.9969
0.0031

At an of
5%
Reject
Reject

p-value = 0.0031. Reject H0.


Braking: H0: 3.10 H1: > 3.10
x = 3.21 s = 0.6
n = 100
3.21 3.10
z=
= 1.83
0.6 / 100
Evidence
Sample size
Sample Mean

100
3.21

n
x-bar

Known; Normal Population or Sample Size >= 30

Population Stdev.
0.6
Test Statistic 1.8333 z
Null Hypothesis
H0: = 3.1
H0: 3.1
H0: 3.1

p-value
0.0668
0.9666
0.0334

p-value = 0.0334. Reject H0.


7-46.

H0: = 25.2 H1: 25.2


x = 26.1 s = 3.2
n = 115

7-12

At an of
5%

Reject

Chapter 07 - Hypothesis Testing

z=

26.1 25.2
3.2 /

115

= 3.016

p-value = 2(.5 - .4987) = 0.0026. Reject H0. The new model probably has a higher average electric
output.
(Template: Testing Difference in Means.xls, sheet: z-test from stats)
Evidence
Sample size
Sample Mean

115
26.1

n
x-bar

Known; Normal Population or Sample Size >= 30

Population Stdev.
3.2
Test Statistic 3.0161 z
Null Hypothesis
H0: = 25.2

7-47.

p-value
0.0026

At an of
5%
Reject

H0: 110 H1: > 110


= .05 n = 80 = 30
We need the power at 1 = 120:
C = 0 + 1.645 / n = 110 + 1.645(30)/ 80 = 115.5175

C 120

Power = P( X > C | = 120) = P Z >


/ n

= P
Z >

115.5175 120

= P(Z > 1.336) = .4092 + .5 = 0.9092.


30 / 80

Good power at 1 = 120.


7-48.

H0: p 60 H1: p< 60


= .05
n = 60

7-13

Chapter 07 - Hypothesis Testing

(Using the template: Testing Population Proportion.xls, sheet: Normal)


z-Test for Population Proportion
Evidence

Assumption
Both np and n(1-p) >= 5

Sample size 60 n
#Successes 21 x
Sample Proportion 0.3500 p-hat
Test statistic-3.9528z

Hull Hypothesis
H0: p = 0.6
H0: p >= 0.6
H0: p <= 0.6

p-value
0.0001
0.0000
1.0000

At an of
5%
Reject
Reject

Reject H0. p-value = 0.000


7-49.

H0: 25000 H1: > 25000


n = 100 = .05 x = 45600
s = 20000
(Using template: Testing Population Mean.xls, sheet: Sample Stats)
Evidence
Sample size
Sample Mean

100
45600

n
x-bar

Known; Normal Population or Sample Size >= 30


Population Stdev. 20000
Test Statistic 10.3000 z
Null Hypothesis
H0: = 25000
H0: 25000
H0: 25000

p-value
0.0000
1.0000
0.0000

At an of
5%
Reject
Reject

Reject H0. p-value = 0.000


7-50.

H0: p = .45 H1: p .45


n = 500 = .01
C1, 2 = .45 2.576 (.45)(.55) / 500 = 0.507, 0.393
> C1 | p = 0.5)
The power at p1 = 0.5 is: P( P

Z >
= P

= P(Z > 0.313) = .5 0.1228 = 0.3772


(.5)(.5) / 500

.507 0.5

7-14

Chapter 07 - Hypothesis Testing

< C 2 | p = 0.4)
The power at p1 = 0.4 is: P( P

Z<
= P

= P(Z < 0.3195) = .5 0.1253 = 0.3747


(.4)(.6) / 500

.393 0.4

Both are about 0.37.


7-51.

H0: u 246 H1: u < 246

n = 60

s = 20

x = 239

= .01

239 246

= 2.711
20 / 60

Z =

Reject H0. The engine does not provide the stated horsepower.
(Template: Testing Difference in Means.xls, sheet: z-test from stats)
Evidence
Sample size
Sample Mean

60
239

n
x-bar

Known; Normal Population or Sample Size >= 30

Population Stdev.
20
Test Statistic -2.7111 z
Null Hypothesis
H0: = 246
H0: 246

p-value
0.0067
0.0034

At an of
1%
Reject
Reject

7-52.

We can increase the power without changing n by increasing the level of significance used,

7-53.

H0: = 13
H1: 13
n = 50 x = 11
s= 6
(Using template: Testing Population Mean.xls, sheet: Sample Stats)
Evidence
Sample size
Sample Mean

50
11

n
x-bar

Known; Normal Population or Sample Size >= 30

Population Stdev.
6
Test Statistic -2.3570 z
Null Hypothesis
H0: = 13
H0: 13
H0: 13

p-value
0.0184
0.0092
0.9908

7-15

At an of
5%
Reject
Reject

Chapter 07 - Hypothesis Testing

At

7-54.

= .05, reject H0. The proportion is probably less than 13%.

p-value = 0.0184

H0: p .35 H1: p > .35


= 1,367/3.850 = .355
p
n = 3,850 x = 1,367
p p 0
z=
= .650
p0 q0 / n
Do not reject H0. p-value = .5 .2450 = 0.2550. There is no evidence in these sampling results
that the market share was higher than 35%.

7-55.

H0: p .43 H1: p > .43


= 0.458
p
n = 5,500 x = 2,521
p p 0
z=
= 4.249
p0 q0 / n
Reject H0. Small p-value (< .00003). The executive should talk about the p-value of this test. There
is strong evidence that GMs market share is over 43%

7-56.

H0: = 6
H1: 6
n = 120 x = 7.2 s= 3.5
(Using template: Testing Population Mean.xls, sheet: Sample Stats)
Evidence
Sample size
Sample Mean

120
7.2

n
x-bar

Known; Normal Population or Sample Size >= 30

Population Stdev.
3.5
Test Statistic 3.7558 z
Null Hypothesis
H0: = 6
H0: 6
H0: 6

p-value
0.0002
0.9999
0.0001

At an of
5%
Reject
Reject

Reject H0. The number of years in a house is more than 6 years. p-value= 0.0002
7-57.

a) Type I error: Keeping the beach closed when it could have been opened
Type II error: Opening the beach when it should have been closed.
b) More costly error? Depends on whose costs. Merchants would find Type I error more costly.
Health officials would find Type II error more costly.
c) = 0.05

7-58.

a) and b) will result in an increase in the power of a test, other things being equal.

7-16

Chapter 07 - Hypothesis Testing

7-59.

b) will result in an increase in the p-value only if the sample mean is less than .
c) will result in an increase in the p-value.

7-60.

b) will result in an increase in the p-value.

7-61.

should be reduced and/or the sample size should be increased. Reducing , reduces the
probability of a Type I error. Increasing the sample size also reduces the Type I error.

7-62.

Statements a), b), c), and f) are all true; just different ways of saying the same thing. d) and e) are
clearly false: for d) it is a Type I error that will be committed and for e) there is still the possibility
of a Type I error being committed.

7-63.

It is important to know the power of a test so that we know our probability of rejecting H0 when it
should indeed be rejected, for given values of the parameter under H1.

7.64.

7-65.

The level of is set before the test as our maximum allowed probability of a Type I error. The pvalue is the actual significance level of the test. It is the smallest at which H0 can be rejected.
H0: p .05 H1: p < .05
= .0212922
n = 1,362 x = 29 p
p p 0
z=
= 4.86
p0 q0 / n
Reject H0; very small p-value. The proportion of women in top management is probably less than
5%.
Suppose = .05. Find the power at p 1 = .04:
C = .05 1.645

(.05)(.95)
= 0.0402854
1362

Power = P(Z < .0537) = .5 + .0214 = 0.5214


Evidence
Sample size 1362
#Successes
29
Sample Proportion 0.0213
Test statistic -4.8612

Hull Hypothesis
H0: p = 0.05
H0: p >= 0.05

Assumption
Both np and n(1-p) >= 5
n
x
p-hat
z
At an of
5%
Reject
Reject

p-value
0.0000
0.0000

7-17

Chapter 07 - Hypothesis Testing

H0: p

0.05

Sample Size

7-66.

Assumption:
Both np and n(1-p) are >= 5

1362
5%

When p = 0.04
P(Type II Error) 0.4807
Power 0.5193

H0: p 21 H1: p < 21


= 4.8
p
= .05
n = 250 x = 12
(Using template: Testing Population Proportion.xls, sheet: Normal)
Evidence

Assumption
Both np and n(1-p) >= 5

Sample size 250 n


#Successes
12
x
Sample Proportion 0.0480 p-hat
Test statistic -6.2887 z

Hull Hypothesis
H0: p = 0.21
H0: p >= 0.21
H0: p <= 0.21

p-value
0.0000
0.0000
1.0000

At an of
5%
Reject
Reject

Reject H0. Vacancy rate is less than 21%. p-value = 0.000


7-67.

H0: 2 156 H1: 2 > 156


n = 25 s 2 = 175 = .05

2
( 24 )

( n 1) s

2
0

24(175)
= 26.923
156

The critical value for = .05 is 36.4150, therefore we cannot reject H0 and the p-value is > .10.
There is no evidence that the variance is above 156. Maybe we need to take a larger sample.
7-68.

H0: p .95 H1: p < .95


= .92
p
n = 1,500 x = 1,380
z=

.92 .95
(.95)(.05) / 1,500

= 5.33

Reject H0. There is evidence that the proportion of rental units decreased from 0.95 during the
period in question. (The p-value is very small.)

7-18

Chapter 07 - Hypothesis Testing

(Template: Testing Population Proportion.xls, sheet: normal)


Evidence

Assumption
Both np and n(1-p) >= 5

Sample size
#Successes
Sample Proportion
Test statistic

1500
1380
0.9200
-5.3311

Hull Hypothesis
H0: p = 0.95
H0: p >= 0.95

7-69.

At an of
5%
Reject
Reject

p-value
0.0000
0.0000

H0: 3.1 H1: > 3.1


n = 21 x = 4.38 s = 5.55
t (20) =

4.38 3.1
5.55 /

21

= 1.06

Do not reject H0 even at


more off-time.
7-70.

n
x
p-hat
z

= .10 (p-value = 0.15). No evidence that increased benefits lead to

H0: p .30 H1: p > .30


= 59/166 = .3554
N = 1,000 n = 166 x = 59 = .10 p
n/N = 166/1,000 = .166 > .05, therefore, we need to use the f.p.c.f.

p p 0
z=

.0554216

= 1.705
p0 q0 N n =
.0324979

n N 1

Reject H0 at = .05; p-value = .5 .4559 = 0.0441. There is some evidence that over 30% of the
Fortune top 1,000 use the multiple scenario approach.
7-71.

H0: 3.8 H1: < 3.8


n = 15 x = 2.1 s = 1.2

7-19

Chapter 07 - Hypothesis Testing

(Using template: Testing Population Mean.xls, sheet: Sample Stats)


Evidence
Sample size
Sample Mean
Sample Stdev.

15
2.1
1.2

n
x-bar
s

Unknown; Population Normal


Test Statistic -5.4867 t
Null Hypothesis
H0: = 3.8
H0: 3.8
H0: 3.8

p-value
0.0001
0.0000
1.0000

At an of
5%
Reject
Reject

Reject H0. The number of visitors is less than 3.8M. p-value = 0.000

7-72.

H0: = 125 H1: 125


n = 25 x = 151.96 s = 90.619
t (24) =

151.96 125
90.619 /

25

= 1.4875

We cannot reject H0 even at = .10. The critical points of t (24) are 1.711. There is no
evidence that the average price is different from $125 million.

7-73.

H0: p 0.25 H1: p > 0.25


= x/n = .31333
p
n = 300 x = 94
p p 0
z=
= 2.53
p0 q0 / n
Reject H0. There is evidence that the proportion of microcomputer owners interested in the
software package is over 25%. The company should probably market the product.
p-value = .5 .4943 = 0.0057.

7-74.

H0: p 0.20 H1: p < 0.20


= .1752
n = 5,000 x = 876 p
p p 0
z=
= 4.38
p0 q0 / n
The p-value is very small (p-value < .00003). Reject H0. There is strong evidence that the NSF
study overestimated the proportion of foreign scientists employed by American laboratories.

7-20

Chapter 07 - Hypothesis Testing

7-75.

H0: = 5.5
H1: 5.5
Do not reject H0 at 0.05 level of significance
(Using template: Testing Population Mean.xls, sheet: Sample Stats)
Evidence
Sample size
Sample
Mean

50

5.1

x-bar

Known; Normal Population or Sample Size >= 30


Population
Stdev. 2.3
Test Statistic -1.2298 z
At an of
Null Hypothesis
p-value
5%
H0: = 5.5
0.2188
7-76.

H0: =3
H1: 3
(Using template: Testing Population Mean.xls, sheet: Sample Stats)
Evidence
Sample size
Sample Mean
Sample Stdev.

20
4.6
2.1

n
x-bar
s

Unknown; Population Normal


Test Statistic 3.4073 t
Null Hypothesis
H0: = 3
H0: 3
H0: 3

Reject H0
7-77.

p-value
0.0030
0.9985
0.0015

At an of
5%
Reject
Reject

It takes longer than 3 seconds to print a page. p-value = 0.0030

H0: p 58%
H1: p < 58%
Do not reject H0 at 0.05 level of significance
(Using template: Testing Population Proportion.xls, sheet: Binomial)

7-21

Chapter 07 - Hypothesis Testing

Evidence

Assumption
Large Population

Sample size 70 n
#Successes 36.4 x
Sample Proportion 0.5200 p-hat

Null Hypothesis
H0: p >= 0.58

7-78.

p-value

At an of
5%

0.1604

H0: p 50
H1: p < 50
n = 700 x = 220
p = 31.4
(Using template: Testing Population Proportion.xls, sheet: Normal)
Evidence

Assumption
Both np and n(1-p) >= 5

Sample size 700 n


#Successes 220 x
Sample Proportion 0.3143 p-hat
Test statistic -9.8271 z

Hull Hypothesis
H0: p = 0.5
H0: p >= 0.5
H0: p <= 0.5

At an of
5%
Reject
Reject

p-value
0.0000
0.0000
1.0000

Reject H0 Ownership proportion is less than 50%. p-value = 0.000


7-79.

H0: 60
H1: > 60
Do not reject H0 at 0.05 level of significance
(Using template: Testing Population Mean.xls, sheet: Sample Data)
Evidence
Sample size
Sample Mean
Sample Stdev.

17
63.8824
13.4531

n
x-bar
s

Unknown; Population Normal


Test Statistic 1.1899 t
Null Hypothesis
H0: = 60
H0: 60
H0: 60

p-value
0.2515
0.8743
0.1257

Do not reject H0. p-value = 0.1257

7-22

At an of
5%

Chapter 07 - Hypothesis Testing

7-80.

H0: 40
H1: < 40
(Template: Testing Population Mean.xls, sheet: sample data)

Evidence
Sample size
Sample Mean
Sample Stdev.

20
53.95
24.2823

n
x-bar
s

Unknown; Population Normal


Test Statistic 2.5692 t
Null Hypothesis
H0: = 40
H0: 40
H0: 40

p-value
0.0188
0.9906
0.0094

At an of
5%
Reject
Reject

Do not reject H0
7-81.

H0: = 40
H1: 40
Even at a two-tailed test, reject H0 p-value = 0.0188

7-82.

H0: 4.2%
H1: p < 4.2%
reject H0 at 0.01 level of significance
(Template: Testing Population Mean.xls, sheet: Sample stats)
Evidence
Sample size
Sample Mean
Sample Stdev.

50
1.4
0.8

n
x-bar
s

Unknown; Population Normal


Test Statistic -24.7487 t
Null Hypothesis
H0: 4.2

p-value

At an of
1%

0.0000

Reject

7-23

Chapter 07 - Hypothesis Testing

7-83. a) versus chart


Type I and Type II Error probabilities
Data

Assumption

H0:

56
2.16

Popn. Stdev.

Tabulation of when =

Either Normal Population


Or n > 30
1-Tail test

55

5.2%
30
40
50
60

0.3%
58%
43%
30%
20%

0.4%
55%
39%
27%
18%

0.5%
52%
36%
24%
16%

7-23

0.6%
49%
34%
22%
14%

0.8% 1.0%
45% 42%
30% 27%
19% 17%
12% 10%

1.2%
39%
25%
15%
9%

2.0%
31%
19%
11%
6%

4.0% 10.0%
22% 10%
12%
5%
6%
2%
3%
1%

Chapter 07 - Hypothesis Testing

b)
power = 0.4968
Power Curve for a Test
Assumption:
Either Normal Population
Or n >= 30

H0:

When =
55.5
P(Type II Error) 0.5032
Power 0.4968

56
Popn. Stdev. 2.16
Sample Size 50 n
Significance Level 5%

c) OC Curve not much to be gained going to sample size of 60


Operating Characteristic Curve
Assumption:
Either Normal Population
Or n >= 30

H0:

56
Sample
Size
Sample
Size

Popn. Stdev. 2.16


Significance Level 5%

7-24

50

n1

60

n2

Chapter 07 - Hypothesis Testing

Legend
n = 50
n = 60

7-84.

a) Power Curve

H0: p

Sample Size

0.25

1000
5%

Assumption:
Both np and n(1-p) are >= 5
n

When p = 0.22
P(Type II Error) 0.4087
Power 0.5913

7-25

Chapter 07 - Hypothesis Testing

b) OC Curve. Not much to be gained.

H0: p

0.25

Sample Size

n1
1000

Assumption:
Both np and n(1-p) are >= 5
n2
1200

5%

7-26

Chapter 07 - Hypothesis Testing

Legend

n = 1000
n = 1200

7-85.

a) versus chart

Data
H0:
Popn. Stdev.

30
0.52

Tabulation
of when
=

30.3

0.7%
30
40
50
60

0.3%
34%
18%
9%
4%

0.4%
31%
16%
8%
3%

0.5%
28%
14%
7%
3%

0.6%
26%
13%
6%
3%

7-27

0.8%
23%
11%
5%
2%

1.0%
20%
9%
4%
2%

1.2%
18%
8%
3%
1%

2.0%
13%
6%
2%
1%

4.0% 10.0%
8%
3%
3%
1%
1%
0%
0%
0%

Chapter 07 - Hypothesis Testing

7-28

Chapter 07 - Hypothesis Testing

b) Power = 0.6779

H0:

When =
30.2
P(Type II
Error) 0.3221
Power 0.6779

30
Popn. Stdev.
Sample Size
Significance
Level

0.52
30

5%

c) better to go with n = 60 for a Type II error of almost 0.


Operating Characteristic Curve
Assumption:
Either Normal Population
Or n >= 30

H0:

30
Sample
Size
Sample
Size

Popn. Stdev. 0.52


Significance Level 5%

7-29

30

n1

60

n2

Chapter 07 - Hypothesis Testing

Legend
n = 30
n = 60

7-86.

The probability of a Type I error is equal to , whether it is a two-tailed test or a one-tailed test.
The probability of making a Type I error declines as the sample mean moves further and further
away from .

7-87 (Use template: testing population mean.xls; sheet:sample size)


1. minimum sample size: n = 54
Sample Size Determination for Testing

H0:
Popn. Stdev.
Significance Level

248

5
10.00%

Probability of Type II Error


When =
250
Desired Maximum 5.00%
Required Sample Size

54

Assumption
Either Normal Population
Or n >= 30

7-30

Chapter 07 - Hypothesis Testing

2. OC curve:
Operating Characteristic Curve
Assumption:
Either Normal Population
Or n >= 30

H0:

248

Sample Size

Significance Level 10%

Sample Size

Popn. Stdev.

54

1.00
0.90

P(Accept H0)

0.80
0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00
246

247

248

249

250

251

252

Actual

3. sample sizes for various means and s:


Tabulation of required sample size

3.0%
4.0%
5.0%
6.0%
7.0%
8.0%

249
251
230
215
202
191
181

Actual
250
251
63
28
58
26
54
24
51
23
48
22
46
21

252
16
15
14
13
12
12

The range of sample sizes runs from a minimum of 12 to a maximum of 251.

7-31

n1
n2

Chapter 07 - Hypothesis Testing

7-88 (Use template: testing population proportion.xls; sheet:sample size)


Sample Size Determination for Testing p
1. minimum sample size: n = 158

H0: p
Significance Level

0.03
5.00%

Chances of Type II Error


When p =
0.08
Desired Maximum 10.00%
Required Sample
size

158

2. OC curve:
OC Curve of a
Test for p

H0: p
0.03

Sample Size

n1
158

5%

Assumption:
Both np and n(1-p) are >= 5
n2

P(Accept H 0)

OC Curves of the Test


1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0%

5%

10%

15%

20%

Actual p

7-32

Chapter 07 - Hypothesis Testing

3. sample sizes for various percentage defectives and s:


Tabulation of Required Sample Size

8%
4%
6%
8%
12%

0.06
420
539
470
420
348

Actual p
0.07
0.08
256
176
331
229
287
198
256
176
211
144

0.09
130
170
147
130
106

0.1
101
133
114
101
82

Minimum sample sizes range from 82 to 420.


7-89.

H0: p = .60
n = 1000

H1: p .60
= .55 = 55
p

Evidence

Assumption
Both np and n(1-p) >= 5

Sample size 1000 n


#Successes 550 x
Sample Proportion 0.5500 p-hat
Test statistic -3.2275 z

Hull Hypothesis
H0: p = 0.6
H0: p >= 0.6
H0: p <= 0.6

At an of
5%
Reject
Reject

p-value
0.0012
0.0006
0.9994

Reject H0. The number of executives changing careers is less than 60%. p-value = 0.0012
Case 9: Tiresome Tires I
(Use template: testing population mean.xls; sheet:various)
1.
Type I and Type II Error probabilities
Data
H0:

Assumption

Popn. Stdev.

2800
20

Either Normal Population


Or n >= 30
1-Tail test

7-33

Chapter 07 - Hypothesis Testing

Tabulation
of when
=

2790

2.9%
30
40
60
80

1.0%
34%
20%
6%
2%

2.0%
25%
13%
3%
1%

2.5%
22%
11%
3%
1%

3.0%
20%
10%
2%
0%

3.5%
18%
9%
2%
0%

4.0%
16%
8%
2%
0%

4.5%
15%
7%
1%
0%

60%
50%

40%

30
40

30%

60
80

20%
10%
0%
0.0%

2.0%

4.0%

6.0%

8.0%

10.0%

12.0%

A sample size of approximately 60 tires is appropriate.


2) exact sample size is 64.

H0:
Popn. Stdev.
Significance Level

2800
20
5.00%

Probability of Type II Error


When =
2790
Desired Maximum 1.00%

Required Sample Size

64

Assumption
Either Normal Population
Or n >= 30

7-34

5.0%
14%
6%
1%
0%

6.0% 10.0%
12%
7%
5%
3%
1%
0%
0%
0%

Chapter 07 - Hypothesis Testing

Sensitivity analysis table:


Tabulation of required sample size

: 2788
1.0%
44
1.8%
39
2.6%
36
3.4%
34
4.2%
32
5.0%
31

Actual
2790
2792 2794
64
99
176
57
88
156
52
81
144
49
76
134
46
72
127
44
68
121

3) Power Curve:
Power Curve for a Test
Assumption:
Either Normal Population
Or n >= 30

H0:

When = 2790
P(Type II
Error) 0.0646
Power 0.9354

2800
Popn. Stdev. 20
Sample Size 40 n
Significance Level 5%

1.00
0.90
0.80
0.70

Power

0.60
0.50
0.40
0.30
0.20
0.10
0.00
2784

2786

2788

2790

2792

2794

Actual

2796

7-35

2798

2800

2802

Chapter 07 - Hypothesis Testing

4) Operating Characteristic Curve:


Operating Characteristic Curve
Assumption:
Either Normal Population
Or n >= 30

H0:

2800
Popn. Stdev.

20

Significance Level

5%

Sample
Size
Sample
Size

80

n1

40

n2

1.00
0.90
0.80

P(Accept H0)

0.70
0.60
0.50

Legend
n = 80
- - - - n = 40

0.40
0.30
0.20
0.10
0.00
2788

2790

2792

2794

2796

2798

2800

2802

2804

2806

Actual

5) Increasing has a similar effect as increasing the sample size in reducing the probability of a Type II
error. At = 0.05, the probability of a Type II error, , was 0.0646. At = 0.10, is reduced to 0.03,
which is still larger than the desired of 0.01. To get a desired of 0.01, the significance level would have
to increased to 20%.
H0:

When = 2790
P(Type II
Error) 0.0300
Power 0.9700

2800
Popn. Stdev.
Sample Size
Significance
Level

20
40

10%

7-36

Chapter 07 - Hypothesis Testing

1.00
0.90
0.80
0.70

Power

0.60
0.50
0.40
0.30
0.20
0.10
0.00
2786

2788

2790

2792

H0:

2794

2796

Actual

20
40

20%

2800

2802

When = 2790
P(Type II
Error) 0.0102
Power 0.9898

2800
Popn. Stdev.
Sample Size
Significance
Level

2798

7-37

Das könnte Ihnen auch gefallen