Hypothesis Testing for Two

Population Parameter
Glyzel Grace M. Francisco
STAT1200: Management Science
2nd Semester, 2022-2023

CENTRAL LUZON STATE UNIVERSITY

 DEPARTMENT of
STATISTICS
Learning Outcomes

At the end of this lesson, you would be able to:

1. Perform z-Test for Comparing Two Means from Independent Populations

2. Perform t-Test for Comparing Two Means from Independent Samples

3. Perform t-Test for Comparing Two Means when the Samples are Dependent

4. Perform F-test in Comparing the Difference Between Two Variances

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 2

 DEPARTMENT of
STATISTICS Hypothesis Testing for Two Populations

• It is used when researchers wish to compare two sample means, using

experimental and control groups.

• Main interest is on the difference between the two populations.

o Independent samples
o Dependent samples

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 3

 DEPARTMENT of
STATISTICS Hypothesis Testing for Two Populations
Dependent Samples
• paired/matched samples
• one sample is related to the other
• occurs when we get two values from each subject or we get one value
from each of two subjects sharing the same characteristic
Independent Samples
• samples selected from one population are not related to the samples
selected from the other population

Assumptions:
• Randomly selected samples
• Each of the two populations must be normally distributed
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 4
 DEPARTMENT of
STATISTICS
Hypothesis Testing for Two Populations
1. The effectiveness of Prilosec for treating heartburn is tested by measuring
gastric acid secretion in patients before and after the drug treatment.
(Dependent/paired samples)

2. The effectiveness of the Weight Watchers diet is tested in an experiment, and

for each subject, the weight before the diet and the weight after the diet are
recorded.
(Dependent/paired samples)

3. The effectiveness of a flu vaccine is tested by treating one group of subjects

with the vaccine while another group of subjects is given placebos.
(Independent samples)
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 5
 DEPARTMENT of
STATISTICS Hypothesis Testing for Two Populations
One Sample z-test (𝝈 is known)
One Population
One Samples One Sample t-test (𝝈 is unknown)
Parameter
z-test for a Proportion
Paired Sample z-test
Hypothesis (𝜎1 𝑎𝑛𝑑 𝜎2 are known)
Testing Dependent Samples
Paired Sample t-test
(𝜎1 𝑎𝑛𝑑 𝜎2 are unknown)
Two Population Independent Sample z-test
Parameter (𝜎1 𝑎𝑛𝑑 𝜎2 are known
Independent Sample t-test

Homogeneity
(𝜎1 𝑎𝑛𝑑 𝜎2 are unknown and 𝜎1 = 𝜎2 )

Test for the

of variance
Independent Samples
Independent Sample t-test
(𝜎1 𝑎𝑛𝑑 𝜎2 are unknown and 𝜎1 ≠ 𝜎2 )

z-test for two proportions

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 6

 DEPARTMENT of
STATISTICS Hypothesis Testing for Two Populations

Two Samples Perform

Homogeneity of
Variance

Are the two Are the

samples
Yes variances
Yes Independent Samples Test
independent? equal? (Equal Variances)

No
No
Independent Samples Test
Dependent or Paired
(Unequal Variances)
Samples Test
(z or t-test)

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 7

 DEPARTMENT of
STATISTICS Hypothesis Testing: Dependent Samples

𝑑ҧ − 𝜇𝑑
𝑡𝑐 = 𝑠
𝑑
ൗ 𝑛

Where: 𝑑ҧ = mean of the differences for the paired sample data

𝜇𝑑 = mean of the differences 𝑑 for the pop’n of paired data
𝑠𝑑 = std. dev’n of the differences 𝑑 for the pop’n of paired data
n = no. of pairs of data
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 8
 DEPARTMENT of
STATISTICS Dependent Samples: Example 1
Mental measurements of young children are often made by giving them blocks
and telling them to build a tower as tall as possible. One experiment of block
building was repeated a month later, with the times (in seconds) listed in the
accompanying table. Use a 0.01 level of significance to test the claim that there
is no difference between the two times.

CHILD 1 2 3 4 5 6 7 8 9
1ST TRIAL 30 19 19 23 29 78 42 20 12
2ND TRIAL 30 6 14 8 14 52 14 22 17
d 0 13 5 15 15 26 28 -2 -5
Difference between the 1st and the 2nd trial

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 9

 DEPARTMENT of
STATISTICS Dependent Samples: Example 1
Claim: there is no difference between the two times (𝝁𝒅 = 𝟎 or 𝝁𝟏 − 𝝁𝟐 = 𝟎)
𝑯𝒐 Since it contain the
CHILD 1st Trial 2nd Trial d 𝑑2
• α = 0.01, condition of equality
02 = 0
1 30 30 0
• 𝜇𝑑 = 0
2 19 6 13 132 = 169
• 𝑑ҧ = 10.5556
3 19 14 5 52 = 25
• 𝑠𝑑 = 11.8860
4 23 8 15 152 = 225
• n=9, df=n-1=8
5 29 14 15 152 = 225

6 78 52 26 262 = 676 σ 𝑑𝑖 95
𝑑ҧ = = = 𝟏𝟎. 𝟓𝟓𝟓𝟔
7 42 14 28 282 = 784 𝑛 9
(σ 𝑑𝑖 )2 (95)2
8 20 22 -2 (−2)2 = 4 σ 𝑑𝑖2 − 2133 − 9
𝑠𝑑 2 = 𝑛 = = 141.2778
9 12 17 -5 (−5)2 = 25 𝑛−1 8
σ 𝑑𝑖 = 95 σ 𝑑𝑖2 =2133
𝑠𝑑 = 141.2778 = 𝟏𝟏. 𝟖𝟖𝟔𝟎
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 10
 DEPARTMENT of
STATISTICS Dependent Samples: Example 1
Claim: there is no difference between the two times (𝝁𝒅 = 𝟎 or 𝝁𝟏 − 𝝁𝟐 = 𝟎)

1. 𝐻𝑜 : 𝝁𝒅 = 𝟎 𝐻𝑎 : 𝝁𝒅 ≠ 𝟎 𝑯𝒐
𝑜𝑟 𝐻𝑜 : 𝝁𝟏 − 𝝁𝟐 = 𝟎 𝐻𝑎 : 𝝁𝟏 − 𝝁𝟐 ≠ 𝟎 Since it contain the
condition of equality

2. test-statistic: t-test (since 𝜎 is unknown)

tail of distribution: two-tailed
Tail of distribution Null hypothesis (Ho) Alternative hypothesis (Ha)
Two-tailed 𝜇𝑑 = 0 𝜇𝑑 ≠ 0
Right-tailed 𝜇𝑑 ≤ 0 𝜇𝑑 > 0
Left-tailed 𝜇𝑑 ≥ 0 𝜇𝑑 < 0
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 11
 DEPARTMENT of
STATISTICS Dependent Samples: Example 1
3. Decision Rule
Reject 𝐻𝑜 if 𝑡𝑐 > 𝑡(𝛼/2,𝑑𝑓) (see slide 8 for your reference)
Acceptance
Given: α = 0.01, n=9, df=n-1=8 Region 0.01
= 0.005
𝑡(𝛼/2,𝑑𝑓) = 𝑡(0.01/2,8) = 𝟑. 𝟑𝟓𝟓(see next slide) 0.01 2
= 0.005
Reject 𝑯𝒐 if 𝒕𝒄 > 𝟑. 𝟑𝟓𝟓 2
−𝑡(0.005,8) = −3.355 𝑡(0.005,8) = 3.355

4. Computation Rejection Rejection

Given: α = 0.01, 𝜇𝑑 = 0, 𝑑ҧ = 10.5556, Region Region
𝑠𝑑 = 11.8860, n=9

𝑑ҧ − 𝜇𝑑 10.5556 − 0
𝑡𝑐 = 𝑠 = = 𝟐. 𝟔𝟔𝟒𝟐
𝑑
ൗ 𝑛 11.8860ൗ
9

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 12

 DEPARTMENT of
STATISTICS Dependent Samples:
Alpha Example 1
tail

df α = 0.01
Tail: two-tailed
df = n−1 = 8

𝑡(0.01/2,8) = 3.355

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 13

 DEPARTMENT of
STATISTICS Dependent Samples: Example 1
5. Decision
Since 2.6642 > 3.355, we fail to reject 𝐻𝑜 Acceptance
Region

−𝑡(0.005,8) = −3.355 𝑡(0.005,8) = 3.355

Rejection Rejection
Region Region
6. Conclusion
At 1% level of significance, there is no sufficient evidence to warrant
rejection of the claim that there is no difference between the times of
building blocks before and after a month.
(see next slide for the wording of final conclusion)

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 14

 DEPARTMENT of
STATISTICS

Claim: 𝝁𝒅 = 𝟎
Decision: Fail to reject 𝑯𝒐
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 15
 DEPARTMENT of
STATISTICS Dependent Samples : Example 2
The management of Discount Furniture stores designed an
incentive plan for salespeople. To evaluate the innovation plan, 8 sales
person were selected at random, and their weekly incomes before and
after the plan were recorded. Was there a significant increase in the
average salesperson’s weekly income due to the innovative incentive plan?
Use the α = 0.05.
Salesperson Before After Salesperson Before After
A 320 340 E 625 631
B 290 285 F 560 560
C 421 475 G 360 365
D 510 510 H 431 431
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 16
 DEPARTMENT of
STATISTICS Dependent Samples: Example 2
Since it does not contain
Claim: there is a significant increase in the average salesperson’s the condition of equality
weekly income due to the innovative incentive plan (𝝁𝒅 < 𝟎 𝑜𝑟 𝝁𝟏 < 𝝁𝟐
𝑜𝑟 𝝁𝟏 − 𝝁𝟐 < 𝟎)
SALESPERSON Before After d 𝑑2
• α = 0.05,
A 320 340 -20 (−20)2 = 400
290 285
• 𝜇𝑑 = 0 𝑯𝒂
B 5 52 = 256
• 𝑠𝑑 = 19.2799
C 421 475 (−54)2 = 2916 Why <?
-54 • n=8, df=n-1=7 Because the claim is “there
D 510 510 02 = 0
0 −80 is an increase after the
E 625 631
-6 (−6)2 = 36 𝑑ҧ = = −𝟏𝟎 plan”, therefore,
8 𝜇2 (𝑎𝑓𝑡𝑒𝑟) > 𝜇1 (𝐵𝑒𝑓𝑜𝑟𝑒)
F 560 560 02 = 0
0
(σ 𝑑𝑖 )2 (80)2
G 360 365 (−5)2 = 25 σ 𝑑𝑖2 − 3402 − 8
-5 𝑠𝑑 2 = 𝑛 = = 371.7143
H 431 431 02 = 0 𝑛−1 7
0
σ 𝑑𝑖 = −80 σ 𝑑𝑖2 =3402 𝑠𝑑 = 371.7143 = 𝟏𝟗. 𝟐𝟕𝟗𝟗

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 17

 DEPARTMENT of
STATISTICS Dependent Samples: Example 2
Claim: there is a significant increase in the average salesperson’s (𝝁𝒅 < 𝟎 𝑜𝑟 𝝁𝟏 < 𝝁𝟐
weekly income due to the innovative incentive plan
𝑜𝑟 𝝁𝟏 − 𝝁𝟐 < 𝟎)
1. 𝐻𝑜 : 𝝁𝒅 ≥ 𝟎 𝐻𝑎 : 𝝁𝒅 < 𝟎
𝑜𝑟 𝐻𝑜 : 𝝁𝟏 ≥ 𝝁𝟐 𝐻𝑎 : 𝝁𝟏 < 𝝁𝟐 𝑯𝒂
𝑜𝑟 𝐻𝑜 : 𝝁𝟏 − 𝝁𝟐 ≥ 𝟎 𝐻𝑎 : 𝝁𝟏 − 𝝁𝟐 < 𝟎
Since it does not contain
2. test-statistic: t-test (since 𝜎 is unknown) the condition of equality
tail of distribution: left-tailed

Tail of distribution Null hypothesis (Ho) Alternative hypothesis (Ha)

Two-tailed 𝜇𝑑 = 0 𝜇𝑑 ≠ 0
Right-tailed 𝜇𝑑 ≤ 0 𝜇𝑑 > 0
Left-tailed 𝜇𝑑 ≥ 0 𝜇𝑑 < 0

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 18

 DEPARTMENT of
STATISTICS Dependent Samples: Example 2
3. Decision Rule
Reject 𝐻𝑜 if 𝑡𝑐 < −𝑡(𝛼,𝑑𝑓) (see slide 8 for your reference)
Acceptance
Given: α = 0.05, n=8, df=n-1=7 Region
𝑡(𝛼,𝑑𝑓) = 𝑡(0.05,7) = −𝟏. 𝟖𝟗𝟓 (see next slide)
0.05
Reject 𝑯𝒐 if 𝒕𝒄 < −𝟏. 𝟖𝟗𝟓
−𝑡(0.005,7) = −1.895

Rejection
Region
4. Computation
Given: α = 0.05, 𝜇𝑑 = 0, 𝑑ҧ = 10, 𝑠𝑑 = 19.2799,
n = 8, df = n-1 = 7

𝑑ҧ − 𝜇𝑑 −10 − 0
𝑡𝑐 = 𝑠 = = −𝟏. 𝟒𝟔𝟕𝟎
𝑑
ൗ 𝑛 19.2799ൗ
8
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 19
 DEPARTMENT of
STATISTICS
Alpha
Dependent Samples:
tail Example 2

df
α = 0.05
Tail: one-tailed
df = n−1 = 7

𝑡(0.05,7) = 1.895

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 20

 DEPARTMENT of
STATISTICS Dependent Samples: Example 2
5. Decision
Since −1.4670 < -1.895, we fail to reject 𝐻𝑜
Acceptance
Region

0.05

−𝑡(0.005,7) = −1.895

Rejection
Region
6. Conclusion
At 5% level of significance, there is no sufficient sample evidence to support
the claim that there is a significant increase in the average salesperson’s
weekly income due to the innovative incentive plan
(see next slide for the wording of final conclusion)

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 21

 DEPARTMENT of
STATISTICS

Claim: 𝝁𝒅 < 𝟎 ✓
Decision: Fail to reject 𝑯𝒐
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 22
 DEPARTMENT of
STATISTICS Hypothesis Testing: Independent Samples

2
𝑠12 𝑠22
𝑛1 + 𝑛2
𝑑𝑓 = 2 2
𝑠12 𝑠22
𝑛1 𝑛2
+𝑛 −1
Type equation here. 𝑛1 − 1 2
*round up to whole number
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 23
 DEPARTMENT of
STATISTICS Hypothesis Testing: Independent Samples

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 24

 DEPARTMENT of
STATISTICS Components of Hypothesis Testing

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 25

 DEPARTMENT of
STATISTICS Independent Samples : Example 1

A survey found that the average hotel room rate in New Orleans
is $88.42, while it is $80.61 in Phoenix. Assume that the data
were obtained from two samples of 50 hotels each and that the
population standard deviations were $5.62 and $4.83,
respectively. At α = 0.05, can it be concluded that there is a
significant difference in the hotel room rates?

Claim: there is a significant difference in the hotel room rates

α = 0.05, 𝑥ҧ1 (𝑁𝑒𝑤 𝑂𝑟𝑙𝑒𝑎𝑛𝑠) = $88.42, 𝑥ҧ2 (𝑃ℎ𝑜𝑒𝑛𝑖𝑥) = $80.61,
𝜎1 = $5.62, 𝜎2 = $4.83, 𝑛1 = 50, 𝑛2 = 50

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 26

 DEPARTMENT of
STATISTICS Independent Samples: Example 1
Claim: there is a significant difference in the hotel room rates (𝝁𝒅 ≠ 𝟎 𝑜𝑟 𝝁𝟏 ≠ 𝝁𝟐
𝑜𝑟 𝝁𝟏 − 𝝁𝟐 ≠ 𝟎)
1. 𝐻𝑜 : 𝝁𝒅 = 𝟎 𝐻𝑎 : 𝝁𝒅 ≠ 𝟎
𝑜𝑟 𝐻𝑜 : 𝝁𝟏 = 𝝁𝟐 𝐻𝑎 : 𝝁𝟏 ≠ 𝝁𝟐
𝑜𝑟 𝐻𝑜 : 𝝁𝟏 − 𝝁𝟐 = 𝟎 𝐻𝑎 : 𝝁𝟏 − 𝝁𝟐 ≠ 𝟎 𝑯𝒂
Since it does not contain
2. test-statistic: z-test (since 𝜎 is known) the condition of equality
tail of distribution: two-tailed

Tail of distribution Null hypothesis (Ho) Alternative hypothesis (Ha)

Two-tailed 𝜇𝑑 = 0 𝜇𝑑 ≠ 0
Right-tailed 𝜇𝑑 ≤ 0 𝜇𝑑 > 0
Left-tailed 𝜇𝑑 ≥ 0 𝜇𝑑 < 0

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 27

 DEPARTMENT of
STATISTICS Independent Samples: Example 1
3. Decision Rule
Reject 𝐻𝑜 if 𝑧𝑐 > 𝑧(𝛼/2) (see slide 23 for your reference)
Acceptance
𝑧(𝛼/2) = 𝑧(0.05/2) = 𝟏. 𝟗𝟔 (see next slide) Region 0.05
= 0.025
Reject 𝑯𝒐 if 𝒛𝒄 > 𝟏. 𝟗𝟔 0.05 0.5-0.025 0.5-0.025 2
= 0.025 = 0.475 = 0.475
2
𝑧(0.05/2) = −1.96 𝑧(0.05/2) = 1.96
Rejection Rejection
4. Computation Region Region

Given: α = 0.05, 𝑥ҧ1 (𝑁𝑒𝑤 𝑂𝑟𝑙𝑒𝑎𝑛𝑠) = $88.42, 𝑥ҧ2 (𝑃ℎ𝑜𝑒𝑛𝑖𝑥) = $80.61,

𝜎1 = $5.62, 𝜎2 = $4.83, 𝑛1 = 50, 𝑛2 = 50
(𝑥ҧ1 − 𝑥ҧ2 ) − (𝜇1 − 𝜇2 ) 88.42 − 80.61 − (0)
𝑧𝑐 = = = 𝟕. 𝟒𝟓𝟐𝟒
𝜎12 𝜎22 5.622 4.832
+
+𝑛 50 50
𝑛1 2

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 28

 DEPARTMENT of
STATISTICS Independent Samples:
Example 1
𝐹𝑖𝑛𝑑 𝑡ℎ𝑒 𝑧 𝑣𝑎𝑙𝑢𝑒 𝑤ℎ𝑒𝑛 𝑡ℎ𝑒
𝑎𝑟𝑒𝑎 𝑖𝑠 0.475

𝑍0.05/2 = 𝟏. 𝟗𝟔

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 29

 DEPARTMENT of
STATISTICS Independent Samples: Example 1
5. Decision
Since |7.4524| > 1.96, we reject 𝐻𝑜 Acceptance
Region
0.5-0.025 0.5-0.025
= 0.475 = 0.475

𝑧(0.05/2) = −1.96 𝑧(0.05/2) = 1.96

Rejection Rejection
Region Region

6. Conclusion
At 5% level of significance, the sample data support the claim that
there is a significant difference in the hotel room rates
(see next slide for the wording of final conclusion)

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 30

 DEPARTMENT of
STATISTICS

✓ ✓

Claim: 𝝁𝒅 ≠ 𝟎
Decision: Reject 𝑯𝒐
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 31
 DEPARTMENT of
STATISTICS Independent Samples : Example 2
In a packing plant, a machine packs cartons with jars. It is supposed that a new
machine will pack faster on the average than the machine currently used. To test
that hypothesis, the times it takes each machine to pack ten cartons are
recorded. The results, in seconds, are shown in the following table.
Statistics New Machine Old Machine
Sample size 10 64
Mean 42.14 43.23
Variance 0.47 0.56
Do the data provide sufficient evidence to conclude that, on the average, the
new machine packs faster? Perform the hypothesis test at the 10% level of
significance assuming equal population variances.
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 32
 DEPARTMENT of
STATISTICS Independent Samples: Example 2
Claim: the new machine packs faster (𝝁𝒅 < 𝟎 𝑜𝑟 𝝁𝟏 < 𝝁𝟐 𝑜𝑟 𝝁𝟏 − 𝝁𝟐 < 𝟎)

1. 𝐻𝑜 : 𝝁𝒅 ≥ 𝟎 𝐻𝑎 : 𝝁𝒅 < 𝟎
𝑜𝑟 𝐻𝑜 : 𝝁𝟏 ≥ 𝝁𝟐 𝐻𝑎 : 𝝁𝟏 < 𝝁𝟐 𝑯𝒂
𝑜𝑟 𝐻𝑜 : 𝝁𝟏 − 𝝁𝟐 ≥ 𝟎 𝐻𝑎 : 𝝁𝟏 − 𝝁𝟐 < 𝟎 Since it does not contain
the condition of equality
2. test-statistic: t-test (since 𝜎 is unknown) – Assume equal variances
tail of distribution: left-tailed

Tail of distribution Null hypothesis (Ho) Alternative hypothesis (Ha)

Two-tailed 𝜇𝑑 = 0 𝜇𝑑 ≠ 0
Right-tailed 𝜇𝑑 ≤ 0 𝜇𝑑 > 0
Left-tailed 𝜇𝑑 ≥ 0 𝜇𝑑 < 0

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 33

 DEPARTMENT of
STATISTICS Independent Samples: Example 2
3. Decision Rule
Reject 𝐻𝑜 if 𝑡𝑐 < −𝑡(𝛼,𝑑𝑓) (see slide 23 for your reference) Acceptance
Given: α = 0.10, 𝑛1 = 10, 𝑛2 = 64, Region

𝑑𝑓 = 𝑛1 + 𝑛2 − 2 = 10 + 64 − 2 = 72 0.05
−𝑡 𝛼,𝑑𝑓 = −𝑡 0.10,72 = −𝟏. 𝟐𝟖𝟐 (see next slide)
−𝑡(0.005,7) = −1.282
Reject 𝑯𝒐 if 𝒕𝒄 < −𝟏. 𝟐𝟖𝟐
Rejection
4. Computation Region
Given: 𝑥ҧ1 (𝑁𝑒𝑤 𝑚𝑎𝑐ℎ𝑖𝑛𝑒) = 42.14, 𝑥ҧ2 (𝑜𝑙𝑑 𝑚𝑎𝑐ℎ𝑖𝑛𝑒) = 43.23, 𝑠12 = 0.47, 𝑠22 = 0.56, 𝑛1 = 10, 𝑛2 = 64

(𝑥ҧ1 − 𝑥ҧ2 ) − (𝜇1 − 𝜇2 ) 42.14 − 43.23 − (0)

𝑡𝑐 = = = −𝟒. 𝟑𝟐𝟕𝟏
1 1 1 1
𝑠𝑝2 𝑛1 + 𝑛2 0.5488 10 + 64

2 2
2
𝑛1 − 1 𝑠1 + 𝑛2 − 1 𝑠2 10 − 1 0.47 + 64 − 1 0.56
𝑠𝑝 = = = 0.5488
𝑛1 + 𝑛2 − 2 10 + 64 − 2
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 34
 DEPARTMENT of
STATISTICS Independent Samples:
Alpha
tail Example 2
α = 0.10
Tail: one-tailed

df = 𝑛2 + 𝑛2 − 2
= 10 + 64 − 2 = 72

𝑡(0.10,72) = 1.282

df
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 35
 DEPARTMENT of
STATISTICS Independent Samples: Example 2
5. Decision
Since -4.3271 < -1.282, we reject 𝐻𝑜 Since the variances are
Acceptance
unequal, df=smaller between
Region
𝑛1 − 1 𝑎𝑛𝑑 𝑛2 − 1 (see slide
45 for your reference)

−𝑡(0.005,7) = −1.282
Rejection
Region

6. Conclusion
At 10% level of significance, the sample data support the claim
that the new machine packs faster
(see next slide for the wording of final conclusion)

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 36

 DEPARTMENT of
STATISTICS

✓
✓
Claim: 𝝁𝒅 < 𝟎
Decision: Reject 𝑯𝒐
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 37
 DEPARTMENT of
STATISTICS Independent Samples : Example 3
An experiment was conducted to test the effects of alcohol. The errors
were recorded in a test of visual and motor skills for a treatment group of
people who drank ethanol and another group given a placebo. The results
are shown in the accompanying table. Assuming unequal variances use a
0.05 significance level to test the claim that the two groups come from
populations with the same mean. Do these results support the common
belief that drinking is hazardous for drivers, pilots, ship captains and so on?
Statistics Treatment Group Placebo Group
Size 20 22
Mean 4.20 1.71
Standard Deviation 2.20 0.72
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 38
 DEPARTMENT of
STATISTICS Independent Samples: Example 3
Claim: drinking is hazardous for drivers, pilots, ship captains and so on (𝝁𝒅 > 𝟎 𝑜𝑟 𝝁𝟏 > 𝝁𝟐
𝑜𝑟 𝝁𝟏 − 𝝁𝟐 > 𝟎)
1. 𝐻𝑜 : 𝝁𝒅 ≤ 𝟎 𝐻𝑎 : 𝝁𝒅 > 𝟎
𝑜𝑟 𝐻𝑜 : 𝝁𝟏 ≤ 𝝁𝟐 𝐻𝑎 : 𝝁𝟏 > 𝝁𝟐
𝑜𝑟 𝐻𝑜 : 𝝁𝟏 − 𝝁𝟐 ≤ 𝟎 𝐻𝑎 : 𝝁𝟏 − 𝝁𝟐 > 𝟎 𝑯𝒐
Since it contains the
2. test-statistic: t-test (since 𝜎 is unknown) – Assume unequal variances condition of equality
tail of distribution: right-tailed

Tail of distribution Null hypothesis (Ho) Alternative hypothesis (Ha)

Two-tailed 𝜇𝑑 = 0 𝜇𝑑 ≠ 0
Right-tailed 𝜇𝑑 ≤ 0 𝜇𝑑 > 0
Left-tailed 𝜇𝑑 ≥ 0 𝜇𝑑 < 0

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 39

 DEPARTMENT of
STATISTICS Independent Samples: Example 3
3. Decision Rule Since the variances are unequal,
2
Reject 𝐻𝑜 if 𝑡𝑐 > 𝑡(𝛼,𝑑𝑓) (see slide 23 for your reference) 𝑠12 𝑠22
𝑛1 + 𝑛2
Given: α = 0.05, 𝑠1 = 2.20, 𝑠2 = 0.72, 𝑛1 = 20, 𝑛2 = 22 𝑑𝑓 = 2 2
𝑠12 𝑠22
2
2
2.20 2
0.72 𝑛1 𝑛2
20 + 22 0.070524 𝑛1 − 1 + 𝑛2 − 1
𝑑𝑓 = 2 2 = = 22.74 ≈ 23 *round up to whole number
2.202 0.722 0.003109 (see slide 23 for your reference)
20 22
+
20 − 1 22 − 1

𝑡(𝛼,𝑑𝑓) = 𝑡(0.05,23) = 𝟏. 𝟕𝟏𝟒 (see next slide) Acceptance

Region 0.05
Reject 𝑯𝒐 if𝒕𝒄 > 𝟏. 𝟕𝟏𝟒

𝑡(0.05,23) = 1.714

Rejection
Region
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 40
 DEPARTMENT of
STATISTICS
Alpha
Independent Samples:
tail Example 3
α = 0.05
Tail: one-tailed

df = 23

𝑡(0.05,72) = 1.714
df

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 41

 DEPARTMENT of
STATISTICS Independent Samples: Example 3
Given: 𝑥ҧ1 (𝑡𝑟𝑒𝑎𝑡𝑚𝑒𝑛𝑡) = 4.20,
4. Computation 𝑥ҧ2 (𝑝𝑙𝑎𝑐𝑒𝑏𝑜) = 1.71,
(𝑥ҧ1 − 𝑥ҧ2 ) − (𝜇1 − 𝜇2 ) 4.20 − 1.71 − (0) 𝑠1 = 2.20, 𝑠2 = 0.72,
𝑡𝑐 = = = 𝟒. 𝟖𝟑𝟏𝟗
2
𝑠1 𝑠2 2 2.202 0.722 𝑛1 = 20, 𝑛2 = 22
20 + 22
𝑛1 + 𝑛2

Acceptance
5. Decision Region 0.05
Since 4.8319 > 1.714, we reject 𝐻𝑜

𝑡(0.05,23) = 1.714

Rejection
6. Conclusion Region
At 5% level of significance, the sample data support the claim that drinking is
hazardous for drivers, pilots, ship captains and so on
(see next slide for the wording of final conclusion)
GGMFRANCISCO HYPOTHESIS TESTING 42
 DEPARTMENT of
STATISTICS

✓
✓
Claim: 𝝁𝒅 > 𝟎
Decision: Reject 𝑯𝒐
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 43
 DEPARTMENT of
STATISTICS Comparing Two Variances
• If two independent samples are selected from two normally distributed
populations in which the population variances are equal (𝜎12 = 𝜎22 )and if
𝑠12
the sample variances 𝑠12 and 𝑠22 are compared as , the sampling
𝑠22
distribution of the variances is called the F distribution.
Characteristics of the F distribution:
1. The values of F cannot be negative, because variances are always positive
or zero.
2. The distribution is positively skewed.
3. The mean value of F is approximately equal to 1.
4. The F distribution is a family of curves based on the degrees of freedom
of the variance of the numerator and the degrees of freedom of the
variance of the denominator.
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 44
 DEPARTMENT of
STATISTICS Comparing Two Variances
𝑠12
Formula for the F Test: 𝐹= 𝑠22
where the larger of the two variances is placed in the numerator regardless
of the subscripts.
Remarks:
1. The F test has two values for the degrees of freedom: that of the
numerator, 𝑛1 − 1, and that of the denominator, 𝑛2 − 1, where 𝑛1 is the
sample size from which the larger variance was obtained.
2. When you are finding the F test value, the larger of the variances is
placed in the numerator of the F formula; this is not necessarily the
variance of the larger of the two sample sizes.
3. Decision Rule: Reject Ho if F Test > F critical value

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 45

 DEPARTMENT of
STATISTICS Comparing Two Variances

Assumptions for Testing the Difference Between Two Variances

1. The samples must be random samples.

2. The populations from which the samples were obtained must be

normally distributed. (Note: The test should not be used when the
distributions depart from normality.)

3. The samples must be independent of one another.

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 46

 DEPARTMENT of
STATISTICS Comparing Two Variances

Steps in Performing F-test in Comparing the Difference Between Two

Variances:

Step 1: State the hypotheses and identify the claim.

Step 2: Find the critical value.
Step 3: Compute the test value.
Step 4: Make the decision.
Step 5: Draw conclusion.

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 47

 DEPARTMENT of
STATISTICS Comparing Two Variances: Example

A medical researcher wishes to see whether the variance of the heart

rates (in beats per minute) of smokers is different from the variance of
heart rates of people who do not smoke. Two samples are selected,
and the data are shown. Using ∝ = 0.01, is there enough evidence to
support the claim? Assume the variable is normally distributed.

Smokers Nonsmokers
Sample size 10 8
Variance 36 10

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 48

 DEPARTMENT of
STATISTICS Comparing Two Variances: Example
Claim: the variance of the heart rates (in beats per minute) (𝝈𝟐𝟏 ≠ 𝝈𝟐𝟐 )
of smokers is from the variance of heart rates of people who
do not smoke
𝑯𝒂
1. 𝐻𝑜 : 𝝈𝟐𝟏
= 𝝈𝟐𝟐 ≠𝐻𝑎 : 𝝈𝟐𝟏 𝝈𝟐𝟐
Since it does not contain
2. test-statistic: F-test the condition of equality
3. Decision Rule
Reject 𝐻𝑜 if 𝐹𝑐 > 𝐹𝛼,𝑑𝑓1 ,𝑑𝑓2 (see slide 45 for your reference)
Reject 𝑯𝒐 if 𝑭𝒄 > 𝟔. 𝟕𝟐 (see next slide)
4. Computation Given: 𝑠12 (𝑠𝑚𝑜𝑘𝑒𝑟) = 36, 𝑠22 (𝑛𝑜𝑛𝑠𝑚𝑜𝑘𝑒𝑟) = 10
𝑠12 36
𝐹𝑐 = 𝑠22
= 10
= 𝟑. 𝟔

Note: Take the largest variance and divide it by the smallest variance to get the F-value
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 49
 DEPARTMENT of
STATISTICS Comparing Two Population Variances:
Example 2

Numerator degrees of freedom

df = 𝑛1 − 1 = 10 − 1 = 9

Denominator degrees of freedom

df = 𝑛2 − 1 = 8 − 1 = 7

𝛼 = 0.01

𝐹0.10,9,7 = 𝟔. 𝟕𝟐

GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 50

 DEPARTMENT of
STATISTICS Comparing Two Population Variances: Example 2
5. Decision
Since 3.60> 6.72, we failed to reject 𝐻𝑜
6. Conclusion
At 1% level of significance, there is no sufficient sample evidence to
support the claim that the variance of the heart rates (in beats per minute)
of smokers is different from the variance of heart rates of people who do not
smoke. (see next slide for the wording of final conclusion)
or
For the Homogeneity of variance test conclusion (before conducting an independent t-test):
Since we failed to reject Ho, we will assume equal population variances.
Note: 𝐻𝑜 : 𝜎12 = 𝜎22 𝐻𝑎 : 𝜎12 ≠ 𝜎22
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 51
 DEPARTMENT of
STATISTICS

Claim: 𝝈𝟐𝟏 ≠ 𝝈𝟐𝟐

Decision: Fail to reject 𝑯𝒐
✓
GGMFRANCISCO HYPOTHESIS TESTING FOR TWO POPULATION PARAMETER | 52