Jaggia1e PPT Ch09

1/14/2019
Learning objectives
9.1 Define the null hypothesis and the alternative
hypothesis.
CHAPTER 9 9.2 Distinguish between Type I and Type II errors.
9.3 Conduct a hypothesis test using the p-value
approach.
Hypothesis testing 9.4 Conduct a hypothesis test using the critical value
approach.
9.5 Differentiate between the test statistics for the
population mean.
9.6 Specify the test statistic for the population
proportion.
Copyright © 2016 McGraw-Hill Education (Australia) Pty Ltd Copyright © 2016 McGraw-Hill Education (Australia) Pty Ltd
Jaggia, Essentials of Business Statistics, 1e 9-1 Jaggia, Essentials of Business Statistics, 1e 9-2
Updated by: Azizur Rahman Updated by: Azizur Rahman
1 2
University students’ study habits University students’ study habits

• Are today’s university students studying hard or • Susan Knight, a dean at a Perth university, wonders
hardly studying? whether the findings are applicable to students at
• Findings of the Australian Survey of Student her university.
Engagement in 2012 indicate that students on average • She randomly selected 35 students to ask about
spend around 11 hours per week preparing for class their average study time per week. Using these
and studying (www.acer.edu.au). results, Susan wants to:
• Only 15% of students spend more than 20 hours each – Determine whether the mean study time of students at her
week on study. Most students are kept busy with work, university is less than 20 hours per week
caring for family and extracurricular activities. – Determine whether the mean study time of students at her
university differs from the national average of 11 hours per
• Around 67% of students are in paid employment, which week.
reduces their time for study.
continued
3 4
LO 9.1 Introduction to hypothesis

Introduction to hypothesis testing
testing
LO 9.1 Define the null hypothesis and the alternative
hypothesis. • In statistics we use sample information to make
inferences regarding the unknown population
• Hypothesis tests resolve conflicts between two
parameters of interest.
competing opinions (hypotheses).
• We conduct hypothesis tests to determine whether
• In a hypothesis test, define
sample evidence contradicts H0.
– H0, the null hypothesis, the presumed default state of
nature or status quo • On the basis of sample information, we either
▪ Any statement including one of the three signs ‘=’, ‘≤’ or – Reject the null hypothesis
‘≥’ is valid for the null hypothesis ▪ Sample evidence is inconsistent with H0
– HA, the alternative hypothesis, a contradiction of the – Do not reject the null hypothesis
default state of nature or status quo ▪ Sample evidence is not inconsistent with H0
▪ The alternative hypothesis is then specified with a ‘≠’, ▪ We do not have enough evidence to reject H0.
‘>’ or ‘<’ sign.
continued continued
5 6
1
1/14/2019
LO 9.1 Introduction to hypothesis LO 9.1 Introduction to hypothesis

testing testing
• One-tailed versus two-tailed tests • One-tailed versus two-tailed tests
– Two-tailed test: Reject H0 on either side of the – One-tailed test: Reject H0 only on one side of the
hypothesised value of the population parameter. hypothesised value of the population parameter.
For example For example

H0: m = m0 versus HA: m ≠ m0 H0: m < m0 versus HA: m > m0 (right-tailed test)
H0: p = p0 versus HA: p ≠ p0 H0: m > m0 versus HA: m < m0 (left-tailed test)
▪ The ‘≠’ symbol in HA indicates that both tail areas of the ▪ The inequality in HA determines which tail area will be used to
distribution will be used to make the decision regarding the make the decision regarding the rejection of H0.
rejection of H0.
continued continued
7 8
LO 9.1 Introduction to hypothesis LO 9.1 Introduction to hypothesis

testing testing
• Three steps to formulate hypotheses • Example: A trade group predicts that back-to-school
spending will average $606.40 per family this year. A
1. Identify the relevant population parameter of interest; different economic model is needed if the prediction
for example, m or p
is wrong.
2. Determine whether it is a one-tailed or two-tailed test 1. Parameter of interest is m.
3. Include some form of the equality sign in H0 and use 2. We want to determine whether m differs from $606.40
HA to establish a claim. (i.e. ≠ $606.40); it is a two-tailed test.
3. H0: m = 606.40 versus HA: m ≠ 606.40.
continued continued
9 10
LO 9.1 Introduction to hypothesis

Type I and Type II errors
testing
LO 9.2 Distinguish between Type I and Type II errors.
• Another example: A television research analyst
wishes to test a claim that more than 50% of • Type I error: Committed when we reject H0 when
households will tune in to an episode of a show. H0 is actually true.
Specify the null and alternative hypotheses to test – Occurs with probability a, which is chosen in advance.
the claim.
• Type II error: Committed when we do not reject H0
1. Parameter of interest is p since we are interested in the and H0 is actually false.
proportion of households
– Occurs with probability b . Power of the test = 1 − b .
2. Since the analyst wants to determine whether p > 0.50,
it is a one-tailed test • For a given sample size n, a decrease in a will
increase b , and vice versa.
3. H0: p < 0.50 versus HA: p > 0.50.
• Both a and b decrease as n increases.
continued
11 12
2
1/14/2019
LO 9.2 Type I and Type II errors LO 9.2 Type I and Type II errors
• This table illustrates the decisions that may be made • Example: Consider the following competing
when hypothesis testing hypotheses from a court of law.
H0: The accused person is innocent
HA: The accused person is guilty
– Consequences of Type I and Type II errors

▪ Type I error: Conclude that the accused is guilty when in
– Correct decisions reality, she is innocent.
▪ Reject H0 when H0 is false ▪ Type II error: Conclude that the accused is innocent when
in reality, she is guilty.
▪ Do not reject H0 when H0 is true
– Incorrect decisions
▪ Reject H0 when H0 is true (Type I error)
▪ Do not reject H0 when H0 is false (Type II error)
continued
13 14
Hypothesis test of the population LO 9.3 Hypothesis test of the

mean when  is known population mean when  is known
LO 9.3 Conduct a hypothesis test using the p-value • The p-value approach
approach.
– The value of the test statistic for the hypothesis test of the
• Hypothesis testing enables us to determine whether population mean m when the population standard deviation
the sample evidence is inconsistent with what is  is known is computed as
hypothesised under the null hypothesis H0. x − m0
z=
 n
• Basic principle: First assume that H0 is true and
then determine whether sample evidence contradicts where m0 is the hypothesised mean value.
this assumption. – p-value: the likelihood of obtaining a sample mean that is at
least as extreme as the one derived from the given sample,
• Two approaches to hypothesis testing under the assumption that the null hypothesis is true.
– The p-value approach
– The critical value approach
continued continued
15 16
LO 9.3 Hypothesis test of the LO 9.3 Hypothesis test of the

population mean when  is known population mean when  is known
• The p-value approach • The p-value approach
– Under the assumption that m = m0, the p-value is the – Determining the p-value depending on the specification of
likelihood of observing a sample mean that is at least as the competing hypotheses
extreme as the one derived from the given sample.
– The calculation of the p-value depends on the specification
of the alternative hypothesis.
Alternative hypothesis p-value

HA: μ > μ 0 Right-tail probability: P(Z ≥ z)
HA: μ < μ 0 Left-tail probability: P(Z ≤ z)
HA: μ ≠ μ 0 Two-tail probability: 2P(Z ≥ z) if z > 0, or 2P(Z ≤ z) if
z<0
– Decision rule: reject H0 if p-value < a.

continued continued
17 18
3
1/14/2019

• Four-step procedure using the p-value approach • Example: For the trade group’s prediction on back-to-
school spending, consider the following:
– Step 1. Specify the null and alternative hypotheses.
n = 30, x = 622.85 and  = 65
– Step 2. Specify the significance level.
– Step 1. State the hypotheses:
– Step 3. Calculate the value of the test statistic and the
H 0 : m = 606 .40
p-value. H A : m  606.40
– Step 4. State the conclusion and interpret the results. – Step 2. Specify the significance level:
a = 0.05
continued continued
19 20
LO 9.3 Hypothesis test of the Hypothesis test of the population

population mean when  is known mean when  is known
• Example: LO 9.4 Conduct a hypothesis test using the critical value
approach.
– Step 3. Calculate the test statistic and compute its p-value:
• Critical value approach
The value of the test statistic is
x − m 0 622.85 − 606.40 – Rejection region: a region of values such that if the test
z= = = 1.39
/ n 65 / 30 statistic falls into this region, then we reject H0.
▪ The location of this region is determined by HA.
The p-value is 2P(Z ≥ 1.39) = 2 × 0.0823 = 0.1646.
– Step 4. State the conclusion and interpret the results: – Critical value: a point that separates the rejection region
from the non-rejection region. It depends on HA.
Since 0.1646 > 0.05, we do not reject H0. Thus, at the 5%
significance level, we cannot conclude that average back-to-
school spending differs from $606.40 per family this year.
continued
21 22

• Critical value approach • Critical value approach
– Determining the critical value(s) depending on the
Alternative hypothesis Critical value
specification of the competing hypotheses
HA: μ > μ0 Right-tailed critical value is zα, where P(Z ≥ zα) = α.
HA: μ < μ0 Left-tailed critical value is –zα, where P(Z ≤ –zα) = α.
HA: μ ≠ μ0 Two-tailed critical values –zα/2 and zα/2, where
P(Z ≥ zα/2) = α/2 and P(Z ≤ –zα/2) = α/2.
– Decision rule: reject H0 if

▪ z > za for a right-tailed test
▪ z < −za for a left-tailed test
▪ z > za/2 or z < −za/2 for a two-tailed test.
continued continued
23 24
4
1/14/2019

• Four-step procedure using the critical value • Example: For the trade group’s prediction on back-
approach to-school spending, consider the following:
– Step 1. Specify the null and alternative hypotheses. n = 30, x = 622.85 and  = 65
– Step 1. State the hypotheses:

– Step 2. Specify the significance level and find the critical
value(s) for the test statistic. H 0 : m = 606 .40
H A : m  606.40
– Step 3. Calculate the value of the test statistic.
– Step 2. Specify the significance level and find the critical
– Step 4. State the conclusion and interpret the results. value(s):
α = 0.05
So, zα/2 = z0.025 = 1.96 and the critical values are –1.96
and 1.96. Thus, the decision rule is to reject H0 if z > 1.96
or z < –1.96.
continued continued
25 26

• Example: • Confidence intervals and two-tailed hypothesis
– Step 3. Calculate the test statistic: tests
The value of the test statistic is – Given the significance level a, we can use the sample data
to construct a 100(1 − a)% confidence interval for the
z=
x − m 0 622.85 − 606.40
= = 1.39
population mean m.
/ n 65 / 30
– Decision rule
▪ Reject H0 if the confidence interval does not contain the value
– Step 4. State the conclusion and interpret the results: of the hypothesised mean m0.
Since –1.96 < 1.39 < 1.96, we do not reject the null ▪ Do not reject H0 if the confidence interval does contain the
hypothesis. This is same conclusion as we reached using value of the hypothesised mean m0.
the p-value approach.
continued continued
27 28

• Implementing a two-tailed test using a confidence • Example: Recall that a research analyst wishes to
interval determine whether average back-to-school spending
– The general specification for a 100(1 − a)% confidence differs from $606.40.
interval of the population mean m when the population – Out of 30 randomly drawn households from a normally
standard deviation  is known is computed as distributed population, the standard deviation is $65 and
sample mean is $622.85.
x  za /2  n or 
 x − za /2  n , x + za /2  n
 – H0: m = 606.40, HA: m ≠ 606.40
– a = 0.05; The critical value
– Decision rule: reject H0 if is za/2 = z0.025 = ±1.96
▪ m0  x − za /2  n – Calculated 95% confidence
interval (CI) [599.59, 646.11]
▪ m0  x + za /2  n
– Since the hypothesised value
606.40 falls in CI, we do not reject H0
continued
29 30
5
1/14/2019

mean when  is unknown population mean when  is unknown
LO 9.5 Differentiate between the test statistics for the
population mean.
• Example: The critical value approach
– In the introductory case, Susan Knight wants to determine
• Test statistic for m when  is unknown whether the mean study time of students at her university
– When the population standard deviation  is unknown, is less than 20 hours per week. Consider the following:
the test statistic for testing the population mean m is n = 35, x = 16.37, s = 7.22
assumed to follow the tdf distribution with (n − 1) degrees – Step 1. State the hypotheses:
of freedom df.
H 0 : m  20 hours
x − m0 H A : m  20 hours
– The value of tdf is computed as tdf = Thus, m0 = 20.
s n
– Step 2. Specify the significance level and critical value(s):
a = 0.05. Based on HA, a left-tailed test with
df = n – 1 = 34, the critical value is –ta,df = –t0.05,34 = –1.691.
continued continued
31 32

population mean when  is unknown population mean when  is unknown
• Example: • Example: The p-value approach
– Step 3. Calculate the test statistic: – Recall that Susan Knight wants to see whether the mean
study time of students at her university differs from 11 hours
x − m 0 16.37 − 20 per week. Consider the following: n = 35, x = 16.37, s = 7.22
t34 = = = −2.9744
s / n 7.22 / 35 – Step 1. H0: m = 11 hours, HA: m ≠ 11 hours
– Step 4. State the conclusion and interpret the results: – Step 2. Specify the significance level: a = 0.05
Since –2.9744 < –1.691, we reject the null hypothesis. – Step 3. Calculate the test statistic and the p-value:
At the 5% significance level, we conclude that average x − m 0 16.37 − 11
study time at the university is less than 20 hours per week. t34 = = = 4.40
s / n 7.22 / 35
▪ Since t34 = 4.40 > 0, the p-value for a two-tailed test is
2P(T34 > t34). From the tdf table for df = 34, we find the exact
probability P(T34 > 4.40) cannot be determined.
continued continued
33 34

population mean when  is unknown population mean when  is unknown
• Example: • Example:
– Step 3. Compute the p-value (continued): – Step 3. Compute the p-value (continued):
▪ Look up t34 = 4.40 in the t table to find the p-value.
▪ p-value < 0.01
▪ a = 0.05.
– Step 4. State the conclusion and interpret the results:
▪ Since the p-value < 0.01 is less than α = 0.05; we reject the
null hypothesis.
▪ Therefore, the mean study time of students at the university is
significantly different from the national average of 11 hours per
▪ Note that t34 = 4.40 is above 2.728, indicating that P(T34 ≥
week.
4.40) < 0.005.
▪ Since this is a two-tailed test, we multiply by two, which results
in 2P(T34 ≥ 4.40) < 0.01.
continued
35 36
6
1/14/2019

proportion population proportion
LO 9.6 Specify the test statistic for the population
proportion.
• Example: A popular magazine asserts that fewer
than 40% of Victorian households have changed
• Test statistic for p gas retailers to benefit from market competition.
– P can be approximated by a normal distribution if Consider the following: n = 180, x = 67, p0 = 0.4
np > 5 and n(1 − p) > 5.
– Step 1. H0: p > 0.40, HA: p < 0.40
– The test statistic for the hypothesis test of the population
– Step 2. Specify the significance level: a = 0.10
proportion p is assumed to follow the z distribution
– Step 3. Calculate the test statistic and p-value:
where p = x n
p − p0 and p0 is the hypothesised ▪ First verify that the sample is large enough
z=
p0 (1 − p0 ) n value of the population np0 = 180  0.4 = 72  5
proportion. n(1 − p0 ) = 180  0.6 = 108  5
▪ p equals 67/180 = 0.3722

continued continued
37 38

population proportion population proportion
• Example: • Example:
– Step 3. Calculate the test statistic and p-value (continued): – Step 4. State the conclusion and interpret the results:
p − p0 0.3722 − 0.4 ▪ p-value = 0.2236 > a = 0.10
z= = = −0.76
p0 (1 − p0 ) n 0.4(1 − 0.4 ) 180 ▪ Do not reject H0: p > 0.40; do not conclude HA: p < 0.40.
▪ This means the claim that fewer than 40% of households in
▪ Based on HA: p < 0.40, Victoria have switched their retailers because of escalating
this is a left-tailed test. gas prices is not justified by the sample data.
Compute the p-value as
P(Z < z) = P(Z < −0.76)
= 0.2236.
continued
39 40

Jaggia1e PPT Ch09

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Jaggia1e PPT Ch09

Transféré par

Droits d'auteur :

Formats disponibles

1/14/2019

University students’ study habits University students’ study habits

LO 9.1 Introduction to hypothesis

LO 9.1 Introduction to hypothesis LO 9.1 Introduction to hypothesis

For example For example

LO 9.1 Introduction to hypothesis LO 9.1 Introduction to hypothesis

3. H0: m = 606.40 versus HA: m ≠ 606.40.

LO 9.1 Introduction to hypothesis

– Consequences of Type I and Type II errors

Hypothesis test of the population LO 9.3 Hypothesis test of the

LO 9.3 Hypothesis test of the LO 9.3 Hypothesis test of the

Alternative hypothesis p-value

– Decision rule: reject H0 if p-value < a.

LO 9.3 Hypothesis test of the LO 9.3 Hypothesis test of the

LO 9.3 Hypothesis test of the Hypothesis test of the population

LO 9.4 Hypothesis test of the LO 9.4 Hypothesis test of the

– Decision rule: reject H0 if

LO 9.4 Hypothesis test of the LO 9.4 Hypothesis test of the

– Step 1. State the hypotheses:

LO 9.4 Hypothesis test of the LO 9.4 Hypothesis test of the

LO 9.4 Hypothesis test of the LO 9.4 Hypothesis test of the

Hypothesis test of the population LO 9.5 Hypothesis test of the

LO 9.5 Hypothesis test of the LO 9.5 Hypothesis test of the

LO 9.5 Hypothesis test of the LO 9.5 Hypothesis test of the

Hypothesis test of the population LO 9.6 Hypothesis test of the

▪ p equals 67/180 = 0.3722

LO 9.6 Hypothesis test of the LO 9.6 Hypothesis test of the

Vous aimerez peut-être aussi