Applied Statistics in Business &

Economics

David P. Doane and Lori E. Seward

Vũ Võ
[email protected]

9-1
 Chapter 9
One-Sample Hypothesis Testing

Chapter Contents
9.1 Logic of Hypothesis Testing
9.2 Type I and Type II Errors
9.3 Decision Rules and Critical Values
9.4 Testing a Mean: Known Population Variance
9.5 Testing a Mean: Unknown Population Variance
9.6 Testing a Proportion
9.7 Power Curves and OC Curves (Optional)
9.8 Tests for One Variance (Optional)

9-2
 Chapter 9
Chapter Learning Objectives (LOs)

LO9-1: Know the steps in testing hypotheses and define H0 and H1.
LO9-2: Define Type I error, Type II error, and power.
LO9-3: Formulate a null and alternative hypothesis for μ or π.
LO9-4: Explain decision rules, critical values, and rejection regions.
LO9-5: Perform a hypothesis test for a mean with known σ using z.
LO9-6: Use tables or Excel to find the p-value in tests of μ.

9-3
 Chapter 9
Chapter Learning Objectives (LOs),
continued

LO9-7: Perform a hypothesis test for a mean with unknown σ using t.

LO9-8: Perform a hypothesis test for a proportion and find the p-value.
LO9-9: Check whether normality may be assumed in testing a proportion.
LO9-10: Interpret a power curve or OC curve (optional).
LO9-11: Perform a hypothesis test for a variance (optional).

9-4
 Chapter 9
9.1 Logic of Hypothesis Testing
LO9-1: Know the steps in testing hypotheses and define
H0 and H1.
• A business analyst asks questions, makes assumptions, and proposes
testable theories about the values of key parameters of the business
operating environment.
• Each assumption is tested against observed data.
• If an assumption has not been disproved, in spite of rigorous efforts to do
so, the business may operate under the belief that the statement is true.
• The analyst states the assumption, called a hypothesis, in a format that can
be tested using well-known statistical procedures.
• The hypothesis is compared with sample data to determine if the data are
consistent or inconsistent with the hypothesis.
• When the data are found to be inconsistent (i.e., in conflict) with the
hypothesis, the hypothesis is either discarded or reformulated.

9-5
 Chapter 9
LO9-1: Know the steps in testing hypotheses and define
H0 and H1 (continued).

• The process of hypothesis testing can be an iterative process,

as illustrated in Figure 9.1 on the next slide.
• However, not every hypothesis has to be retested continuously.
• Some questions are relevant only at a point in time, such as
asking whether or not consumers under age 25 prefer a new
Coke flavor to an existing flavor.
• If a carefully designed marketing study provides a clear-cut
answer, no more testing is required.
• On the other hand, clinical testing of new drugs may go on for
years, at different hospitals and with different types of patients.
• There is plenty of work for data analysts who continually look for
changes in customer satisfaction, shifts in buying patterns, and
trends in warranty claims.

9-6
 Chapter 9
LO9-1: Know the steps in testing hypotheses and define
H0 and H1 (continued, 2).

Copyright ©2019 McGraw-Hill Education. All rights reserved. No reproduction or distribution without the
prior written consent of McGraw-Hill Education. 9-7
 Chapter 9
LO9-1: Know the steps in testing hypotheses and define
H0 and H1 (continued, 3).

Who Uses Hypothesis Testing?

• All business managers need at least a basic understanding of
hypothesis testing because managers often interact with
specialists, read technical reports, and then make
recommendations on key financial or strategic decisions based on
statistical evidence.
• A confidence interval sometimes gives enough information to
make a decision.
• Knowing the 95 percent range of likely values for a key decision
parameter (e.g., the proportion of repeat customers under age
30) may be all you need.
• This chapter extends the idea of confidence intervals by showing
how to test a sample against a benchmark, and how to assess the
risk of incorrect decisions.
Copyright ©2019 McGraw-Hill Education. All rights reserved. No reproduction or distribution without the
prior written consent of McGraw-Hill Education. 9-8
 Chapter 9
LO9-1: Know the steps in testing hypotheses and define
H0 and H1 (continued, 4).

Steps in Hypothesis Testing

• Step 1: State the hypothesis to be tested.

• Step 2: Specify what level of consistency with the data will lead to
rejection of the hypothesis. This is called the decision rule.
• Step 3: Collect data and calculate necessary statistics to test the
hypothesis.
• Step 4: Make a decision. Should the hypothesis be rejected or
not?
• Step 5: Take action based on the decision.

Copyright ©2019 McGraw-Hill Education. All rights reserved. No reproduction or distribution without the
prior written consent of McGraw-Hill Education. 9-9
 Chapter 9
LO9-1: Know the steps in testing hypotheses and define
H0 and H1 (continued, 5).

Step 1: State the Hypothesis

• Hypotheses are a pair of mutually exclusive,

collectively exhaustive statements about some fact
about a population.
• One statement or the other must be true, but they
cannot both be true.
• H0: Null Hypothesis
H1: Alternative Hypothesis
• These two statements are hypotheses because the
truth is unknown.
Copyright ©2019 McGraw-Hill Education. All rights reserved. No reproduction or distribution without the
prior written consent of McGraw-Hill Education. 9-10
 Chapter 9
LO9-1: Know the steps in testing hypotheses and define
H0 and H1 (continued, 6).

Step 1: State the Hypothesis (continued)

• Efforts will be made to reject the null hypothesis (or the status
quo).
• H0 must be stated in a precise way, so that it can be tested
against empirical evidence from a sample.
• If H0 is rejected, we tentatively conclude H1 to be the case.
• H0 is sometimes called the maintained hypothesis.
• H1 is called the action alternative because action may be required
if we reject H0 in favor of H1.
• We cannot accept a null hypothesis, we can only fail to reject it.
• The null hypothesis is assumed true and a contradiction is sought.

9-11
 Chapter 9
LO9-1: Know the steps in testing hypotheses and define
H0 and H1 (continued, 7).

Step 2: Specify the Decision Rule

• Before collecting data to compare against the hypothesis, the

researcher must specify how the evidence will be used to
reach a decision about the null hypothesis.
• For example, in our legal system, the evidence presented by
the prosecutor must convince a jury “beyond a reasonable
doubt” that the defendant is not innocent.

9-12
 Chapter 9
LO9-1: Know the steps in testing hypotheses and define
H0 and H1 (continued, 8).

Steps 3 and 4: Data Collection and Decision Making

• Much of the critical work in hypothesis testing takes

place during steps 1 and 2.
• Once the hypotheses and decision rule have been
clearly articulated, the process of data collection, while
occasionally time-consuming, is straightforward.
• We compare the data with the hypothesis, using the
decision rule, and decide to reject or not reject the null
hypothesis.

9-13
 Chapter 9
LO9-1: Know the steps in testing hypotheses and define
H0 and H1 (continued, 9).

Step 5: Take Action Based on Decision

• This last step of taking action requires experience and expertise on

the part of the decision maker.
• Suppose the evidence presented at a trial convinces a jury that the
defendant is not innocent. What punishment should the judge
impose?
• Or suppose the blood sample of an athlete shows steroid use.
What fine should the athletic commission impose? Should the
athlete be banned from competing?
• Appropriate action for the decision should relate back to the
purpose of conducting the hypothesis test in the first place.

9-14
 Chapter 9
9.2 Type I and Type II Errors
LO9-2: Define Type I error, Type II error, and power.

• Our ability to collect evidence can be limited by our tools and by

time and financial resources.
• On occasion we will be making a decision about the null hypothesis
that could be wrong.
• Consequently, our decision rule will be based on the levels of risk of
making a wrong decision.
• We can allow more risk or less risk by changing the threshold of the
decision rule.
• It is possible to make an incorrect decision regarding the null
hypothesis. Either the null hypothesis is true or it is false.
• We have two possible choices concerning the null hypothesis:
We either reject H0 or fail to reject H0.

9-15
 Chapter 9
LO9-2: Define Type I error, Type II error, and power
(continued).

Types of Error

• Type I error: Rejecting the null hypothesis when it is

true. This occurs with probability α (level of
significance).
• Type II error: Failure to reject the null hypothesis when
it is false. This occurs with probability β.

9-16
 Chapter 9
LO9-2: Define Type I error, Type II error, and power
(continued, 2).

Consequences of Type I and Type II Errors

• The consequences of these two errors are quite different, and the
costs are borne by different parties.
• Example: Type I error is convicting an innocent defendant, so the
costs are borne by the defendant. Type II error is failing to convict a
guilty defendant, so the costs are borne by society if the guilty
person returns to the streets.
• Firms are increasingly wary of Type II error (failing to recall a
product as soon as sample evidence begins to indicate potential
problems).

9-17
 Chapter 9
LO9-2: Define Type I error, Type II error, and power
(continued, 3).

Probability of Type I and Type II Errors

• The probability of a Type I error (rejecting a true null hypothesis)

is denoted α (the lowercase Greek letter “alpha”).
• Statisticians refer to α as the level of significance.
• The probability of a Type II error (not rejecting a false hypothesis)
is denoted β (the lowercase Greek letter “beta”), as shown
in Table 9.1 on the next slide.

9-18
 Chapter 9
LO9-2: Define Type I error, Type II error, and power
(continued, 4).

Probability of Type I and Type II Errors (continued)

• If we choose α = .05, we expect to commit a Type I error about five

times in 100.
• β cannot be chosen in advance because it depends on α and the
sample size.
• A small β is desirable, other things being equal.
9-19
 Chapter 9
LO9-2: Define Type I error, Type II error, and power
(continued, 5).
Power of a Test
• The power of a test is the probability that a false
hypothesis will be rejected (as it should be).
• Power equals 1 − β and is the complement of Type II error.
• Reducing β would correspondingly increase power (usually
accomplished by increasing the sample size).
• Larger samples lead to increased power, which is why clinical
trials often involve thousands of people.

9-20
 Chapter 9
LO9-2: Define Type I error, Type II error, and power
(continued, 6).

Relationship Between α and β

• Both a small α and a small β are desirable.

• For a given type of test and fixed sample size, there is
a trade-off between α and β.
• The larger critical value needed to reduce α risk makes
it harder to reject H0, thereby increasing β risk.
• Both α and β can be reduced simultaneously only by
increasing the sample size.

9-21
 Chapter 9
9.3 Decision Rules and Critical Values

LO9-3: Formulate a null and alternative hypothesis for μ or π.

• A statistical hypothesis is a statement about the value of

a population parameter.
• A hypothesis test is a decision between two competing
mutually exclusive and collectively exhaustive
hypotheses about the value of the parameter.
• When testing a mean we can choose between three
tests.

9-22
 Chapter 9
LO9-3: Formulate a null and alternative hypothesis
for μ or π (continued).

Where Do We Get μ0 (or π0)?

• For a mean (or proportion), the value of μ0 (or π0) that we are testing is
a benchmark based on past experience, an industry standard, a target,
or a product specification.
• The value of μ0 (or π0) does not come from a sample.

One-Tailed and Two-Tailed Tests

9-23
 Chapter 9
LO9-3: Formulate a null and alternative hypothesis
for μ or π (continued, 2).

The application will dictate which of the three

alternatives is appropriate. The direction of the test is
indicated by which way the inequality symbol points
in H1:

> indicates a right-tailed test

< indicates a left-tailed test
≠ indicates a two-tailed test

9-24
 Chapter 9
LO9-4: Explain decision rules, critical values, and rejection
regions.

Decision Rule

• A test statistic shows how far the sample estimate is from

its expected value, in terms of its own standard error.
• The decision rule uses the known sampling distribution of
the test statistic to establish the critical value that divides
the sampling distribution into two regions.
• Reject H0 if the test statistic lies in the rejection region.

9-25
 Chapter 9
LO9-4: Explain decision rules, critical values, and rejection
regions (continued).
When to Use a One- or Two-Sided Test

• A two-sided hypothesis test ( ≠ ) is used when direction

(< or >) is of no interest to the decision maker
• A one-sided hypothesis test is used when
• the consequences of rejecting H0 are asymmetric,
or
• where one tail of the distribution is of special
importance to the researcher.
• Rejection in a two-sided test guarantees rejection in a
one-sided test, other things being equal.

9-26
 Chapter 9
LO9-4: Explain decision rules, critical values, and rejection
regions (continued, 2).
Decision Rule for Two-Tailed Test

• Reject H0 if the test statistic < left-tail critical value or if

the test statistic > right-tail critical value.

9-27
 Chapter 9
LO9-4: Explain decision rules, critical values, and rejection
regions (continued, 3).
Decision Rule for Left-Tailed Test
• Reject H0 if the test statistic < left-tail critical value.

Figure 9.2

9-28
 Chapter 9
LO9-4: Explain decision rules, critical values, and rejection
regions (continued, 4).
Decision Rule for Right-Tailed Test
• Reject H0 if the test statistic > right-tail critical value.

9-29
 Chapter 9
LO9-4: Explain decision rules, critical values, and rejection
regions (continued, 5).
Type I Error

• A reasonably small level of significance α is desirable, other

things being equal.
• Chosen in advance, common choices for a are
.10, .05, .025, .01 and .005
(i.e., 10%, 5%, 2.5%, 1% and .5%).
• The α risk is the area under the tail(s) of the sampling distribution.
• In a two-sided test, the α risk is split with α/2 in each tail since
there are two ways to reject H0.

9-30
 Chapter 9
9.4 Testing a Mean: Known Population
Variance
LO9-5: Perform a hypothesis test for a mean with
known σ using z.
• A test statistic measures the difference between a given
sample mean and a benchmark µ0 in terms of the standard
error of the mean.
• The test statistic is the “standardized score” of the sample
statistic.
• When testing μ with a known σ, the test statistic is a z-score.
• Once we have collected our sample, we calculate a value of
the test statistic using the sample mean and then compare it
against the critical value of z.
• We will refer to the calculated value of the test statistic as zcalc.

9-31
 Chapter 9
LO9-5: Perform a hypothesis test for a mean with
known σ using z (continued).

If the true mean of the population is μ0, then the value of a particular
sample mean ̅ calculated from our sample should be near μ0 and
therefore the test statistic should be near zero.
9-32
 Chapter 9
LO9-5: Perform a hypothesis test for a mean with
known σ using z (continued, 2).
• The test statistic is compared with a critical value from a table.
• The critical value is the boundary between two regions (reject H0, do not
reject H0) in the decision rule.
• The critical value shows the range of values for the test statistic that
would be expected by chance if the null hypothesis were true.

9-33
 Chapter 9
LO9-5: Perform a hypothesis test for a mean with
known σ using z (continued, 3).

Example: Paper Manufacturing

Testing the Hypothesis

• Step 1: State the hypotheses.

For example, H0: μ ≤ 216 mm
H1: μ > 216 mm.
• Step 2: Specify the decision
rule. For example, for = .05
for the right-tail area, reject H0
if zcalc > 1.645, otherwise do
not reject H0.

9-34
 Chapter 9
LO9-5: Perform a hypothesis test for a mean with
known σ using z (continued, 4).
Example: Paper Manufacturing (continued):
Testing the Hypothesis

• Step 3: Collect sample data and calculate the test statistic.

If H0 is true, then the test statistic should be near 0 because the
sample mean should be near μ0. The value of the test statistic is:

• Step 4: Make the decision.

The test statistic falls in the right rejection region, so we reject
the null hypothesis H0: μ  216 and conclude the alternative
hypothesis H1: μ > 216 at the 5% level of significance.
9-35
 Chapter 9
LO9-5: Perform a hypothesis test for a mean with
known σ using z (continued, 5).
Example: Paper Manufacturing (continued):

Testing the Hypothesis

• Step 5: Take action.

Now that we have concluded that the process is producing paper
with an average width greater than the specification, it is time to
adjust the manufacturing process to bring the average width back
to specification. Our course of action could be to readjust the
machine settings or it could be time to re-sharpen the cutting tools.
At this point it is the responsibility of the process engineers to
determine the best course of action.

9-36
 Chapter 9
LO9-6: Use tables or Excel to find the p-value in tests of μ.

p-Value Method

• The p-value is the probability of the sample result (or one more
extreme) assuming that H0 is true.
• The p-value can be obtained using Excel’s cumulative standard
normal function.
• The p-value can also be obtained from Appendix C-2.
• Using the p-value, we reject H0 if p-value < α.
• Whether we use the critical value approach or the p-value approach,
our decision about the null hypothesis will be the same.

9-37
 Chapter 9
LO9-6: Use tables or Excel to find the p-value in
tests of μ, (continued).

p-Value Method (continued)

• What is a p-value? A sample statistic is a random
variable that may differ from the hypothesized value
merely by chance, so we do not expect the sample to
agree exactly with H0.
• The p-value is the probability of obtaining a test statistic
as extreme as the one observed, assuming that the null
hypothesis is true.
• A large p-value (near 1.00) tends to support H0, while a
small p-value (near 0.00) tends to contradict H0.
• If the p-value is less than the chosen level of
significance (α), then we conclude that the null
hypothesis is false.
9-38
 Chapter 9
LO9-6: Use tables or Excel to find the p-value in
tests of μ, (continued, 2).

Example: Paper Manufacturing (continued)

Two-Tail Test of Hypothesis

Step 1: State the hypotheses.

For a two-tailed test, the hypotheses are:

H0: μ = 216 mm (product mean is what it is supposed to be)

H1: μ ≠ 216 mm (product mean is not what it is supposed to be)

9-39
 Chapter 9
LO9-6: Use tables or Excel to find the p-value in
tests of μ, (continued, 3).

Example: Paper Manufacturing (continued)

Two-Tail Test of Hypothesis

Step 2: Specify the decision rule.

We will use the same α = .05 as in the right-tailed test. But for a two-tailed
test, we split the risk of Type I error by putting α/2 = .05/2 =
.025 in each tail. For α = .05 in a two-tailed test, the critical value is
z.025 = ±1.96, so the decision rule is:

Reject H0 if zcalc > +1.96 or if zcalc < −1.96.

Otherwise, do not reject H0.

9-40
 Chapter 9
LO9-6: Use tables or Excel to find the p-value in
tests of μ, (continued, 4).
Example: Paper Manufacturing (continued)
Two-Tail Test of Hypothesis
The decision rule is illustrated in Figure 9.6.

9-41
 Chapter 9
LO9-6: Use tables or Excel to find the p-value in
tests of μ, (continued, 5).

Example: Paper Manufacturing (continued)

Two-Tail Test of Hypothesis

Step 3: Calculate the test statistic.

The test statistic is unaffected by the hypotheses or the level of

significance. The value of the test statistic is the same as for the one-
tailed test:

9-42
 Chapter 9
LO9-6: Use tables or Excel to find the p-value in
tests of μ, (continued, 6).

Example: Paper Manufacturing (continued)

Two-Tail Test of Hypothesis
Step 4: Make the decision.

Because the test statistic falls in the right tail of the rejection region, we reject
the null hypothesis H0: μ = 216 and conclude H1: μ ≠ 216 at the 5 percent level
of significance.

Another way to say this is that the sample mean differs significantly from the
desired specification at α = .05 in a two-tailed test. Note that this decision is
rather a close one because the test statistic just barely falls into the rejection
region.

9-43
 Chapter 9
LO9-6: Use tables or Excel to find the p-value in
tests of μ, (continued, 7).

Example: Paper Manufacturing (continued)

Two-Tail Test of Hypothesis

Step 5: Take action.

An adjustment is needed, such as changing the cutting tool settings.

Now it is up to the process engineers to choose the best course of
action.

9-44
 Chapter 9
LO9-6: Use tables or Excel to find the p-value in
tests of μ, (continued, 8).
Testing the Hypothesis
Using the p-Value Approach
In a two-tailed test, the decision rule using the p-value is the same as in a one-tailed
test.

Reject H0 if p-value < α. Otherwise, do not reject H0.

The difference between a one-tailed and a two-tailed test is how we obtain the p-
value. Because we allow rejection in either the left or the right tail in a two-tailed
test, the level of significance, α, is divided equally between the two tails to establish
the rejection region.

In order to fairly evaluate the p-value against α, we must now double the tail area.
The p-value in this two-tailed test is 2 × P(zcalc > 2.152) = 2 × .0157 = .0314. This says
that in a two-tailed test, a result as extreme as 2.152 would arise about 3.14 percent
of the time by chance alone if the null hypothesis were true.
9-45
 Chapter 9
LO9-6: Use tables or Excel to find the p-value in
tests of μ, (continued, 9).
Testing the Hypothesis P-value = 0.0314 < α = 0.05,
Using the p-Value Approach so the null hypothesis is
rejected.

9-46
 Chapter 9
LO9-6: Use tables or Excel to find the p-value in
tests of μ, (continued, 10).

Analogy to Confidence Intervals

• A two-tailed hypothesis test at the 5% level of

significance (α = .05) is exactly equivalent to asking
whether the 95% confidence interval for the mean
includes the hypothesized mean.
• If the confidence interval includes the hypothesized
mean, then we cannot reject the null hypothesis.

9-47
 Chapter 9
9.5 Testing a Mean: Unknown
Population Variance
LO9-7: Perform a hypothesis test for a mean with
unknown σ using t.
Using Student’s t

If the population variance σ2 must be estimated from the

sample, the hypothesis testing procedure is modified. There is a
loss of information when s replaces σ in the formulas, and it is
no longer appropriate to use the normal distribution. However,
the basic hypothesis testing steps are the same.

9-48
 Chapter 9
LO9-7: Perform a hypothesis test for a mean with
unknown σ using t (continued).
Using Student’s t (continued)
• When the population standard deviation σ is unknown and the
population may be assumed normal, the test statistic follows the
Student’s t distribution with n – 1 degrees of freedom.

9-49
 Chapter 9
LO9-7: Perform a hypothesis test for a mean with
unknown σ using t (continued, 2).

Example: Hot Chocolate

Testing the Hypothesis

• Step 1: State the hypotheses.

H0: μ = 142
H1: μ ≠ 142
• Step 2: Specify the decision rule.
For α = .10 for the two-tail test and with d.f. n – 1 = 24
– 1 = 23, reject H0 if tcalc > 1.714 or if tcalc < -1.714,
otherwise do not reject H0.

9-50
 Chapter 9
LO9-7: Perform a hypothesis test for a mean with
unknown σ using t (continued, 3).

Example: Hot Chocolate (continued)

Testing the Hypothesis

Critical
regions for
the test.

9-51
 Chapter 9
LO9-7: Perform a hypothesis test for a mean with
unknown σ using t (continued, 4).

Example: Hot Chocolate (continued)

Testing the Hypothesis

• Step 3: Collect sample data and calculate the test

statistic.
If H0 is true, then the test statistic should be near 0
because the sample mean should be near μ0. The value
of the test statistic is:

9-52
 Chapter 9
LO9-7: Perform a hypothesis test for a mean with
unknown σ using t (continued, 5).

Example: Hot Chocolate (continued)

Testing the Hypothesis

• Step 4: Since the test statistic lies within the range of

chance variation, we cannot reject the null hypothesis
H0: μ = 142.
• Note: The p-value = .13867. Since .13867 > .1 (level of
significance), do not reject the null hypothesis.

9-53
 Chapter 9
LO9-7: Perform a hypothesis test for a mean with
unknown σ using t (continued, 6).

Using the p-value (Excel)

9-54
 Chapter 9
LO9-7: Perform a hypothesis test for a mean with
unknown σ using t (continued, 7).

Using the p-value (Excel)

9-55
 Chapter 9
LO9-7: Perform a hypothesis test for a mean with
unknown σ using t (continued, 8).
Confidence Intervals versus Hypothesis Test
• A two-tailed hypothesis test at the 10% level of significance
(α = .10) is equivalent to a two-sided 90% confidence interval for
the mean.
• If the confidence interval does not include the hypothesized
mean, then we reject the null hypothesis.
• The 90% confidence interval for the mean is given next.

• Since μ = 142 lies within the 90 percent confidence interval

[140.677, 142.073], we cannot reject the hypothesis H0: μ = 142
at α = .10 in a two-tailed test.
9-56
 Chapter 9
9.6 Testing a Proportion

LO9-8: Perform a hypothesis test for a proportion and find

the p-value.

• To conduct a hypothesis test, we need to know

• the parameter being tested
• the sample statistic
• the sampling distribution of the sample statistic
• The sampling distribution tells us which test statistic to use.
• A sample proportion p estimates the population proportion π.
• Remember that for a large sample, p can be assumed to follow a
normal distribution. If so, the test statistic is z.
• Our rule is to assume normality if nπ0 ≥ 10 and n(1 − π0) ≥ 10.

9-57
 Chapter 9
LO9-8: Perform a hypothesis test for a proportion and find
the p-value (continued).

• The value of π0 that we are testing is a benchmark

such as past experience, an industry standard, or a
product specification.
• The value of π0 does not come from a sample.

9-58
 Chapter 9
LO9-8: Perform a hypothesis test for a proportion and find
the p-value (continued, 2).

Critical Value
• The test statistic is compared with a critical z value from a
table.
• The critical value shows the range of values for the test
statistic that would be expected by chance if the H0 were
true.
Example: Return Policy

Steps in Testing a Proportion

• Step 1: State the hypotheses.
For example, H0: π ≥ .13
H1: π < .13
9-59
 Chapter 9
LO9-8: Perform a hypothesis test for a proportion and find
the p-value (continued, 3).
Steps in Testing a Proportion (continued)
• Step 2: Specify the decision rule.
For α = .05 for a left-tail area, reject H0 if z < -1.645,
otherwise do not reject H0.

Figure 9.12 9-60

 Chapter 9
LO9-8: Perform a hypothesis test for a proportion and find
the p-value (continued, 4).

Steps in Testing a Proportion (continued)

• Step 3: Collect sample data and calculate the test statistic.

If H0 is true, then the test statistic should be near 0 because the
sample mean should be near μ0. The value of the test statistic is
given next.

• Step 4: Since the test statistic lies in the left-tail rejection

region, we reject the null hypothesis H0: π ≥ .13.

Figure 9.12 9-61

 Chapter 9
LO9-8: Perform a hypothesis test for a proportion and find
the p-value (continued, 5).
Calculating the p-Value
For our test statistic zcalc = −1.975, the p-value (.02413) can be obtained from
Excel’s cumulative standard normal =NORM.S.DIST(−1.975). Alternatively, if
we round the test statistic to two decimals, we can use the cumulative normal
table in Appendix C-2.

Using the p-value, we reject H0 at α = .05, but the decision would be very close
if we had used α = .025.

9-62
 Chapter 9
LO9-8: Perform a hypothesis test for a proportion and find
the p-value (continued, 6).

The Effect of α

Would the decision be the same if we had used a different level of

significance? While the test statistic zcalc = −1.975 is the same regardless of
our choice of α, our choice of α does affect the decision. Refer to Table 9.6.

9-63
 Chapter 9
LO9-8: Perform a hypothesis test for a proportion and find
the p-value (continued, 7).

The Effect of α (continued)

• Which level of significance is the “right” one? They all are.
• It depends on how much Type I error we are willing to allow.
• Before concluding that α = .01 is “better” than the others because it
allows less Type I error, you should remember that smaller Type I
error leads to increased Type II error.
• In this case, Type I error would imply that there has been a change
in return rates when in reality nothing has changed, while Type II
error implies that the software had no effect on the return rate, when
in reality the software did decrease the return rate.

9-64
 Chapter 9
LO9-9: Check whether normality may be assumed in testing
a proportion.

Small Samples and Non-Normality

Our rule is to assume normality if nπ0 ≥ 10 and n(1 − π0) ≥ 10.

When this is satisfied, the test statistic will be a z-score.

When this is not satisfied, we will have to use the exact

binomial test. We can use Excel to perform the exact binomial
test.

9-65
 Chapter 9
9.7 Power Curves and OC Curves
(Optional)
LO9-10: Interpret a power curve or OC curve (optional).

Please refer to the text for the content in this section.

Figure 9.19 9-66

 Chapter 9
9.8 Tests for One Variance (Optional)

LO9-11: Perform a hypothesis test for a variance (optional).

Hypothesis Test for a Single Population Variance.

Null Hypothesis: H0: σ2 = σ02

Alternative Hypothesis:

H1: σ2 > σ02 H1: σ2 < σ02 H1: σ2 ≠ σ02

Test Statistic: χ2 (calculated) = ((n − 1)s2) ⁄ σ02

Figure 9.19 9-67

 Chapter 9
LO9-11: Perform a hypothesis test for a variance
(optional) (continued).
Hypothesis Test for a Single Population Variance
(continued).

Decision Rule:

1. Right-tail test: Reject H0 if χ2 (calculated) > χ2

(upper).
2. Left-tail test: Reject H0 if χ2 (calculated) < χ2
(lower).
3. Two-tail test: Reject H0 if χ2 (calculated) < χ2
(lower) or if Reject H0 if χ2 (calculated) > χ2
(upper).

Figure 9.19 9-68

 Chapter 9
LO9-11: Perform a hypothesis test for a variance
(optional) (continued, 2).

Example: Attachment Times

Historical statistics show that the standard deviation

of attachment times for an instrument panel in an
automotive assembly line is σ = 7 seconds.
Observations on 20 randomly chosen attachment
times are shown in Table 9.11 (see the next slide). At
α = .05, does the variance in attachment times differ
from the historical variance (σ2 = 72 = 49)?

Figure 9.19 9-69

 Chapter 9
LO9-11: Perform a hypothesis test for a variance
(optional) (continued, 3).

Example: Attachment Times (continued)

Figure 9.19 9-70

 Chapter 9
LO9-11: Perform a hypothesis test for a variance
(optional) (continued, 4).

Example: Attachment Times (continued)

The sample mean is ̅ = 129.400 with a standard deviation

s = 7.44382. We ignore the sample mean since it is irrelevant to
this test. For a two-tailed test, the hypotheses are:

H0: σ2 = 49 H0: σ2 ≠ 49

For a test of one variance, assuming a normal population, the test

statistic follows the chi-square distribution with degrees of freedom
equal to d.f. = n – 1 = 20 – 1 = 19.

Figure 9.19 9-71

 Chapter 9
LO9-11: Perform a hypothesis test for a variance
(optional) (continued, 5).

Example: Attachment Times (continued)

Denoting the hypothesized variance as σ02, the test statistic is:

Figure 9.19 9-72

 Chapter 9
LO9-11: Perform a hypothesis test for a variance
(optional) (continued, 6).

Example: Attachment Times (continued)

For a two-tailed test, the decision rule based on the upper and
lower critical values of chi-square is

Reject H0 if χ2calc < χ2lower or if χ2calc > χ2upper.

Otherwise, do not reject H0.

We can use the Excel function =CHISQ.INV to get the critical
values. View the next slide.

Figure 9.19 9-73

 Chapter 9
LO9-11: Perform a hypothesis test for a variance
(optional) (continued, 7).

Example: Attachment Times (continued)

χ2lower = CHISQ.INV(α ⁄ 2, d.f.) = CHISQ.INV(0.025, 19) = 8.907

χ2upper = CHISQ.INV(1 − α ⁄ 2, d.f.) = CHISQ.INV(0.975) = 32.854

The value of the test statistic is:

The decision rule is illustrated in Figure 9.22 on the next slide.

Figure 9.19 9-74

 Chapter 9
LO9-11: Perform a hypothesis test for a variance
(optional) (continued, 8).

Example: Attachment Times (continued)

Because the test statistic is within the middle range, we conclude that the
population variance does not differ significantly from 49; that is, the assembly
process variance is unchanged.
Figure 9.19 9-75
 Chapter 9
LO9-11: Perform a hypothesis test for a variance
(optional) (continued, 9).
When to Use Tests for One Variance
• In general, we would be interested in a test of variances when it is not the
center of the distribution, but rather the variability of the process, that
matters.
• More variation implies a more erratic data-generating process.
• For example, variance tests are important in manufacturing processes
because increased variation around the mean can be a sign of wear and
tear on equipment or other problems that would require attention.

Caution

The chi-square test for a variance is not robust to non-normality of the

population. If normality cannot be assumed (e.g., if the data set has outliers or
severe skewness), you might need to use a bootstrap method (see
LearningStats Unit 08) to test the hypothesis, using specialized software. In
such a situation, it is best to consult a statistician.
Figure 9.19 9-76