SME11e PPT ch11std

Slides by:
Andrew Stephenson
Georgia Gwinnett College
© 2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in
a license distributed with a certain product or service or otherwise on a password-protected website for classroom use.
Chapter 11
11.2
Introduction to Hypothesis Testing
• Concepts of Hypothesis Testing

• Testing the Population Mean When the Population Standard
Deviation Is Known
• Calculating the Probability of a Type II Error
• The Road Ahead
Introduction
11.3
• In chap. 10, we introduced estimation(추정) and showed how it is
used.
• Hypothesis testing(가설검정) is the second form of statistical
inference.
• The purpose of hypothesis testing is to determine whether
enough statistical evidences exists to enable us to conclude
that a belief or hypothesis about a parameter is supported by
the data.
• Hypothesis testing has greater applicability in business and
economics, as well as many other fields.
• This chapter will lay the foundation upon which the rest of the
book is based. It will help you to understand an inferential
statistics.
Statistical Inference
11.4
• In many ways the formal procedure for hypothesis testing is similar
to the scientific method.
• The scientist observes nature, formulates a theory, and then
tests this theory against observation.
• In our context scientist poses a hypothesis concerning one or
more population parameters that they equal specified values.
She then samples the population and compares her observations
with the hypothesis.
• If the observations disagree with the hypothesis, the scientist
rejects it.
• If not, the scientist concludes either that the hypothesis is true
or that the sample did not detect the difference between the
real and hypothesized values of the population parameter.
11-1 Concepts of Hypothesis Testing
11.5
• To understand the concepts we’ll start with an example of
nonstatistical hypothesis testing.
• When a person is accused of a crime, he or she faces a trial(재판).
• When the prosecution(검사) presents its case, a jury(배심원) must
make a decision on the basis of the evidence presented. In
fact, the jury conducts a test of hypothesis.
• There are two hypotheses that are tested:
• The null hypothesis (귀무가설) 𝐻0 : The defendant is innocent
(피고는 무죄이다).
• The alternative hypothesis (대립가설) or research hypothesis

(연구가설) 𝐻1 : The defendant is guilty (피고는 유죄이다).
* 𝐻0 : 𝐻 − 𝑛𝑜𝑢𝑔ℎ𝑡
11.6
• The jury does not know which hypothesis is true. The members
must make a decision on the basis of evidence presented by
both the prosecution and the defense.
• In the language of statistics convicting the defendant* is
equivalent to rejecting the null hypothesis in favor of the
alternative hypothesis.**
• That is, the jury is saying that there is enough evidence to
conclude that the defendant is guilty (i.e., there is enough
evidence to support the alternative hypothesis).
* 피고가 유죄라고 선언하는 것

** 대립가설을 찬성하여 귀무가설을 기각하는 것
11.7
• If the jury acquits the defendant*, it is stating that not rejecting the
null hypothesis**, which means that the jury decided that there
was not enough evidence to conclude that the defendant was
guilty.
• Notice that we do not say that the defendant is innocent, only

say that there is not enough evidence to support the alternative
hypothesis. That is why we never say that we accept the null
hypothesis***.
• Our justice system does not allow this decision****.
* 피고가 무죄라고 선고하는 것; ** 귀무가설을 기각하지 못함

*** 귀무가설을 채택한다고 표현하지 않음; ****재판제도는 이러한 결정을 허용하지 않음
11.8
• There are two possible errors:
• A Type I error (제1종 오류) occurs when we reject a true null
hypothesis. That is, a Type I error occurs when the jury
convicts an innocent person.
• A Type II error (제2종 오류) occurs when we don’t reject a false
null hypothesis. That occurs when a guilty defendant is
acquitted*.
* 유죄인 피고가 무죄를 선고받을 때 발생
11.9
• The probability of a Type I error is denoted as α (Greek letter

alpha) which is also called the significance level(유의수준).
• The probability of a Type II error is denoted by β (Greek letter
beta).
• The two error probabilities 𝛼 𝑎𝑛𝑑 𝛽 are inversely related:

decreasing one increases the other.
11.10
Table 11.1 Terminology of Hypothesis Testing
Decision 𝑯𝟎 is true (=Defendant 𝑯𝟎 is false (= Defendant

is innocent) is guilty)
Reject 𝑯𝟎 Type 𝑰 𝑬𝒓𝒓𝒐𝒓 Correct decision
(Convict Defendant) 𝑷 𝑻𝒚𝒑𝒆 𝑰 𝑬𝒓𝒓𝒐𝒓 = 𝜶
Do not reject 𝑯𝟎 Correct decision Type 𝑰𝑰 Error

(Acquit defendant) 𝑷 𝑻𝒚𝒑𝒆 𝑰𝑰 𝑬𝒓𝒓𝒐𝒓 = 𝜷
11.11
• In our justice system, Type I errors are regarded as more serious.
We try to avoid convicting innocent people. We are more willing
to acquit guilty people.
• As a consequence, the system is set up so that the probability
of a Type I error is small.
• This is arranged by placing the burden of proof (입증책임) on the
prosecution(the prosecution must prove guilt.) and by having
judges instruct the jury to find the defendant guilty only if
there is “evidence beyond a reasonable doubt(합리적인 의심을
넘어서는 증거가 있을 때만).” (= “sufficient evidence”)
11.12
• The consequence of this arrangement is that the probability of
acquitting guilty people is relatively large.
• Oliver Wendell Holmes*, a United States Supreme Court justice,

once phrased the relationship between the probabilities of
Type I and Type II errors in the following way: “Better to
acquit 100 guilty men than convict one innocent one.”
• In Justice Holmes’s opinion, the probability of a Type I error
should be 1/100 of the probability of a Type II error.
* Holmes, Oliver Wendell,Jr.(1841-1935): US Supreme Court associate justice

(연방대법원 판사) 1902-32; the son of physician and essayist Oliver Wendell Holmes.
He became well known for his strong, articulate , and often dissenting opinions
(반대의견).
11.13
• The critical concepts in hypothesis testing follow.
1. There are two hypotheses, the null and the
alternative(research) hypotheses.
2. The procedure begins with the assumption that the null
hypothesis is true.
3. The goal is to determine whether there is enough evidence
to infer that the alternative hypothesis is true.
4. There are two possible decisions:
• Conclude that there is enough evidence to support the
alternative hypothesis.
• Conclude that there is not enough evidence to support
the alternative hypothesis.
11.14
5. Two possible errors can be made.
Type I error: Reject a true null hypothesis
Type II error: Do not reject a false null hypothesis.
𝑃 Type I error = 𝛼
𝑃 Type II error = 𝛽
• Let’s extend these concepts to statistical hypothesis testing.
11.15
• In statistics we frequently test hypotheses about parameters.
• The hypotheses we test are generated by questions that
managers need to answer.
11.16
• Consider Example 10.1 (mean demand for computers during
assembly lead time) again.
• Rather than estimate the mean demand, our operations
manager wants to know whether the mean is different from 350
units, which may be the point at which the current inventory
policy needs to be altered.
• We can rephrase this request into a test of the hypothesis:
𝐻0 : 𝜇 = 350
• Thus, our alternative(research) hypothesis becomes:
𝐻1 : 𝜇 ≠ 350
⇒ Is there enough evidence to conclude that 𝜇 is not
equal to 350?
11.17
• The testing procedure begins with the assumption that the null
hypothesis is true.
• Thus, until we have further statistical evidence, we will
assume:
𝐻0 : 𝜇 = 350 (assumed to be TRUE)
• The goal of the process is to determine whether there is enough
evidence to infer that the alternative hypothesis is true.
• That is, is there sufficient statistical information to determine if
this statement is true?
𝐻1 : 𝜇 ≠ 350
• When we state the hypotheses, we list the null first followed by

11.18
• Now suppose that the current inventory policy is based on an
analysis that the actual mean demand during lead time is 350.
• After a vigorous advertising campaign, the manager
suspects that there has been an increase in demand and
thus an increase in mean demand during lead time.
• To test whether there is an evidence of increase, the
manager would specify the alternative hypothesis as
𝐻1 : 𝜇 > 350
• Because the manager knew that the mean was (and
maybe still is) 350, the null hypothesis would state
𝐻0 : 𝜇 = 350 (assumed to be TRUE)
11.19
• Further suppose that the manager does not know the actual mean
demand during lead time, but the current inventory policy is
based on the assumption that the mean is less than or equal to
350.
• If the advertising campaign increases the mean to a
quantity larger than 350, the hypotheses become
𝐻0 : 𝜇 ≤ 350
𝐻1 : 𝜇 > 350
• Notice that in both illustrations (𝐻0 : 𝜇 = 350 and 𝐻0 : 𝜇 ≤ 350 ) the

alternative hypothesis is designed to determine whether there is
enough evidence to conclude that the mean is greater than 350.
11.20
• If there is enough evidence to conclude that the alternative
hypothesis is true when we assume that the mean is equal to 350,
we would certainly draw the same conclusion when we assume
that the mean value is less than 350.
• As a result, the null hypothesis will always state that the parameter
equals the value specified in the alternative hypothesis. (Use the
equal sign in the 𝐻0 .)
• Suppose that the manager now wanted to determine whether

there has been a decrease in the mean demand during lead
time. We express the null and alternative hypotheses as
𝐻0 : 𝜇 = 350
𝐻1 : 𝜇 < 350
11.21
• The hypotheses are often set up to reflect a manager’s decision
problem wherein the null hypothesis represents the status quo
(현재상황).
• If there is evidence of an increase or decrease in the value of

the parameter, a new course of action will be taken.
• Examples include deciding to produce a new product,
switching to a better drug to treat an illness, or sentencing a
defendant to prison.
11.22
• Once the null and alternative hypotheses are stated, the next step
is to randomly sample the population and calculate a test statistic
(검정통계량) (in this example, the sample mean).
• The test statistic is the criterion on which we base our decision

about the hypothesis.
• The test statistic is based on the best estimator (최량 추정량) of the
parameter. (see chap. 10)
• If the test statistic’s value is inconsistent with the null hypothesis we
reject the null hypothesis and infer that the alternative hypothesis
is true.
11.23
• For example, if we’re trying to decide whether the mean is not
equal to 350, a large value of 𝑋ത (say, 600) would provide
enough evidence.
• If 𝑋ത is close to 350 (say, 355) we would say that this does not
provide a great deal of evidence to infer that the population
mean is greater than 350.
• In the absence of sufficient evidence, we do not reject the
null hypothesis in favor of the alternative. (In then absence of
sufficient evidence of guilt, a jury finds the defendant not
guilty.)
• Sufficient evidence = “evidence beyond a reasonable doubt”
11.24
• Two possible errors can be made in any test:
• A Type I error occurs when we reject a true null hypothesis
and
• A Type II error occurs when we don’t reject a false null
hypothesis.
• There are probabilities associated with each type of error:
𝑃 Type I error = 𝛼
𝑃 Type II error = 𝛽
• 𝛼 is called the significance level (유의수준).
• (1 − 𝛼) is called the confidence level (신뢰수준)
11-2 Testing the Population Mean with
Known 𝜎 11.25
• Example 11.1 Department Store’s New Billing System

• Data Xm 11-01
Example 11.1 Department Store’s New Billing
System 11.26
• The manager of a department store is thinking about establishing a new
billing system for the store's credit customers.
• She determines that the new system will be cost-effective only if the
mean monthly account is more than $170. A random sample of 400
monthly accounts is drawn, for which the sample mean is $178 (Data
Xm 11-01).
• The manager knows that the accounts are approximately normally

distributed with a standard deviation of $65. Can the manager
conclude from this that the new system will be cost-effective?
Example 11.1 New Billing System
11.27
IDENTIFY
• The system will be cost effective if the mean account balance for all
customers is greater than $170.
• We express this belief as our research hypothesis, that is:

𝐻1 : 𝜇 > 170 (Install new system)
• Thus, our null hypothesis becomes:

𝐻0 : 𝜇 = 170 (Do not install new system)
Example 11.1 New Billing System
11.28
IDENTIFY
• What we want to show:

𝐻0 : 𝜇 = 170 (we’ll assume this is true)
𝐻1 : 𝜇 > 170
• We know:
𝑛 = 400, 𝑋ത = 178, 𝑎𝑛𝑑 𝜎 = 65
• What to do next?!
• Since the sample mean 𝑋ത is the best estimator of the
population mean, 𝑋ത is the test statistic.
• To conduct this test, we ask and answer the following
question: Is a sample mean of 178 sufficiently greater than
170 to allow us to confidently infer that the population
mean is greater than 170?
Example 11.1 New Billing System(새로운 청구 제도)
COMPUTE 11.29
• To test our hypotheses, we can use two different approaches:

• The rejection region approach (기각역 방법) (typically used
when computing statistics manually), and
• The p-value approach (p-값 방법) (which is generally used with a
computer and statistical software).
• We will explore both in turn…
11.2a Rejection region (기각역)
COMPUTE 11.30
• Definition 11.1 (Rejection Region)
• The rejection region is a range of values such that if the test
statistic falls into that range, we decide to reject the null
hypothesis in favor of the alternative hypothesis. (대립가설을
선호하여 귀무가설을 기각)
• Suppose we define the value of the sample mean that is just

large enough to reject the null hypothesis as 𝑥ҧ𝐿 .
• The rejection region is 𝑥ҧ > 𝑥ҧ𝐿 .
COMPUTE 11.31
• It seems reasonable to reject the null hypothesis in favor of the
alternative if the value of the sample mean is large relative to 170.
• If the calculated sample were 500, we would reject 𝐻0 .
• If values of 𝑥ҧ were 171, we would not reject it.
• Because Type I error is defined as rejecting a true null hypothesis,
𝛼 = 𝑃 Type 𝐼 error
= 𝑃(reject 𝐻0 given that 𝐻0 is true)
= 𝑃(Value of test statistic is in RR when 𝐻0 is true)
= 𝑃(𝑥ҧ > 𝑥ҧ𝐿 ).
COMPUTE 11.32
• Since the sampling distribution of 𝑥ҧ is normal or approximately
normal (Section 9-1), with mean 𝜇 and standard deviation 𝜎/ 𝑛,
we can standardize 𝑥ҧ and obtain the following probability:
ҧ
𝑥−𝜇 𝑥ҧ 𝐿 −𝜇 𝑥ҧ 𝐿 −𝜇
𝑃 𝑥ҧ > 𝑥ҧ𝐿 =𝑃 > =𝑃 𝑍> =𝛼
𝜎/ 𝑛 𝜎/ 𝑛 𝜎/ 𝑛
• From Section 8-2, we defined 𝑧𝛼 to be the value of a standard

normal random variable such that
𝑃 𝑍 > 𝑧𝛼 = 𝛼
𝑥ҧ 𝐿 −𝜇
∴ = 𝑧𝛼
𝜎/ 𝑛
• We can calculate this 𝑥ҧ𝐿 based on any level of significance (𝛼)

we want…
COMPUTE 11.33
• At a 5% significance level (i.e. 𝛼 = 0.05), we get

𝑥ҧ 𝐿 −𝜇
= 𝑧𝛼 and 𝑧𝛼 = 𝑧.05 = 1.645
𝜎/ 𝑛
𝑥ҧ 𝐿 −170 65
→ = 1.645 → 𝑥ҧ𝐿 =170 + (1.645)( ) = 175.35
65/ 400 20
• Since our sample mean(178) is greater than the critical value we

calculated(175.35), (in other words, it is in the rejection region) we
reject the null hypothesis in favor of 𝐻1 (𝜇 > 170) and that it is cost
effective to install the new billing system.
11-2b Standardized Test Statistic
• An easier method is to use the standardized test statistic: 11.34
ҧ
𝑥−𝜇
𝑧=
𝜎/ 𝑛
and compare its result to 𝑧𝛼 : (rejection region: z > 𝑧𝛼 )

ҧ
𝑥−𝜇 178−170
𝑧= = = 2.46
𝜎/ 𝑛 65/ 400
Since 𝑧 = 2.46 > 1.645(= 𝑧.05 ), we reject 𝐻0 in favor of 𝐻1 …

• Because statistical software packages employ it, the
standardized test statistic will be used throughout this book.
(standardized test statistic = test statistic)*
* 표준화 검정 통계량 = 검정 통계량
11-2b Standardized Test Statistic
• When a null hypothesis is rejected, the test is said to be 11.35
statistically significant at whatever significance level the test was
conducted!!
• The test was significant at the 5% significance level. (5%
유의수준에서 통계적으로 유의하다.)
Example 11.1 The Big Picture Again
11.36
Fig. 11.2
Sampling Distribution of 𝑋ത 𝑥ҧ 𝐿 = 175.34
for Example 11.1 ҧ 178
𝑥=
Z
Fig. 11.3
Sampling Distribution of Z for
Example 11.1
.05
H0: 𝜇 = 170
H1: 𝜇 > 170
Reject H0 in favor of
Z.05=1.645 z = 2.46
11-2c p-Value of a Test (검정의 p값)
11.37
• Definition 11.2 (The p-value of a test)
• The p-value of a test is the probability of observing a test statistic at
least as extreme as the one computed given that the null hypothesis
is true.
(검정의 p값은 귀무가설이 옳다는 가정하에서 계산되는 검정 통계량의 값보다 더 큰
검정 통계량의 값이 관측되는 확률이다.)
• In the case of our department store example, what is the probability of

observing a sample mean at least as extreme as the one already
observed (i.e. 𝑥=
ҧ 178), given that the null hypothesis (𝐻0 : 𝜇 = 170) is
true?
ҧ
𝑥−𝜇 178−170
𝑃 𝑥ҧ > 178 = 𝑃 > = 𝑃 𝑍 > 2.46 = .0069
𝜎/ 𝑛 65/ 400
p-value
p-Value of a Test
11.38
Fig. 11.4 p-Value for Example 11.1
p-value =.0069
p-value = P(Z > 2.46)
z =2.46
11-2d Interpreting the p-value
11.39
• Remember that the sampling distribution allows us to make
probability statements about a sample statistic assuming
knowledge of the population parameter.
• Thus, the probability of observing a sample mean at least as
large as 178 from a population whose mean is 170 is .0069,
which is very small.
• In other words, we have just observed an unlikely event, an
event so unlikely that we seriously doubt the assumption that
began the process—that the null hypothesis is true.
• Consequently, we have reason to reject the null hypothesis

and support the alternative.
11.40
• Do not interpret that the p-value is the probability that the null
hypothesis is true.
• You cannot make a probability statement about a
parameter. It is not a random variable.
• The p-value of a test provides valuable information because it is

a measure of the amount of statistical evidence that supports
• To understand this interpretation fully, refer to Table 11.2 where
we list several values of 𝑥,ҧ their z-statistics, and p-values for
Example 11.1.
11.41
Table 11.2 Test Statistics and p-values for Example 11.1
Test Statistic
ഥ−𝝁
𝒙 ഥ − 𝟏𝟕𝟎
𝒙
𝒛= =
Sample mean 𝒙
ഥ 𝝈/ 𝒏 𝟔𝟓/ 𝟒𝟎𝟎 𝒑-value
170 0 .5000
172 0.62 .2676
174 1.23 .1093
176 1.85 .0322
178 2.46 .0069
180 3.08 .0010
Figure 11.5 p-Values for Example 11.1
11.42
z = 0.62; 𝒙
ഥ = 𝟏𝟕𝟐
p-value=P(Z> .62) =.2676
z =1.856; 𝒙
ഥ = 𝟏𝟕𝟔
p-value=P(Z> 1.85) =.0322

11.43
• Notice that the closer 𝑥ҧ is to the hypothesized mean, 170, the
larger the p-value is. The farther 𝑥ҧ is above 170, the smaller the p-
value is.
• The smaller the p-value, the more statistical evidence exists to
support the alternative hypothesis.
11.44
• This raises the question: How small does the p-value have to be to
infer that the alternative hypothesis is true?
• In general, the answer depends on a number of factors,
including the cost of making Type I and Type II errors. In
Example 11.1, a Type I error would occur if the manager
adopts the new billing system when it is not cost-effective.
• If the cost of this error is high, we attempt to minimize its
probability.
• In the rejection region method, we do so by setting the
significance level quite low—say 1%.
• Using p-value method, we would insist that the p-value be
quite small, providing sufficient evidence to infer that the
mean monthly account is greater than $170 before
proceeding with the new billing system.
11-2e Describing the p-value (p-값의 해석)
11.45
• p-values can be translated using the following descriptive terms.
• If the p-value is less than 1%, we say that there is overwhelming
evidence to infer that the alternative hypothesis is true. We
also say that the test is highly significant*.
• If the p-value is between 1% and 5%, there is strong evidence
to infer that the alternative hypothesis is true. The result is
deemed to be significant**.
• If the p-value is between 5% and 10%, there is weak evidence
that supports the alternative hypothesis. We say that the result
is weakly significant***.
• If the p-value exceeds 10%, there is little to no evidence that
supports the alternative hypothesis.
* overwhelming(압도적인); highly significant(매우 유의하다)

** strong(강한); significant(유의하다)
*** weak (약한); weakly significant (약하게 유의하다)
11-2e Describing the p-value
11.46
Figure 11.6 Describing p-values
Overwhelming Evidence
(Highly Significant)
Strong Evidence Weak Evidence No Evidence
(Strongly Significant)(Weakly Significant) (Not Statistically
Significant)
0 0.05 1.0
0.01 0.10
p=.0069
• We observe a p-value of .0069, hence there is overwhelming
evidence to support 𝐻1 : 𝜇 > 170.
11-2f The p-value and Rejection Region
Method 11.47
• Compare the p-value with the selected value of the significance

level:
• If the p-value is less than α, we judge the p-value to be small
enough to reject the null hypothesis.
• If the p-value is greater than α, we do not reject the null
hypothesis.
• Since p-value = .0069 < α = .05, we reject H0 in favor of H1.
11-2g Solving Manually and Using Excel
11.48
COMPUTE: Excel Workbook
• Instructions
• Type or import the data into one column. (Open Xm11-01.) In
any empty cell, calculate the sample mean
(=AVERAGE(A1:A401).)
• Open the Test Statistics in Excel Workbooks folder and click the
z-Test Mean tab. In Cell B3, type or copy the value of the
sample mean. In cells B4-B7, type the value of 𝜎(65), the value
of n(400), the value of 𝜇 under the null hypothesis(170), and the
value of 𝛼 .05 , respectively.
• The spreadsheet reports the value of the test statistic, 𝑧 = 2.46.
The p-value of the test is .0069. Excel reports this probability as
𝑃 𝑍 ≤ 𝑧 𝑜𝑛𝑒 𝑡𝑎𝑖𝑙 . (see appendix)
Example 11.1
11.49
z-Test of a Mean
Sample mean 178 z Stat 2.46

Population standard
deviation 65 P(Z<=z) one-tail 0.0069
Sample size 400 z Critical one-tail 1.6449
Hypothesized mean 170 P(Z<=z) two-tail 0.0138
Alpha 0.05 z Critical two-tail 1.9600
11-2h Interpreting the Results of a Test
11.50
• In Example 11.1, we rejected the null hypothesis. Does this prove
that the alternative hypothesis is rue?
• The answer is no; because our conclusion is based on sample
data(and not on the entire population), we can never prove
anything by using statistical inference.
• Consequently, we summarize the test by stating that there is
enough statistical evidence to infer that the null hypothesis is
false and that the alternative hypothesis is true.
Conclusions of a Test of Hypothesis
11.51
• If we reject the null hypothesis, we conclude that there is
enough evidence to infer that the alternative hypothesis is true.
• If we do not reject the null hypothesis, we conclude that there

is not enough statistical evidence to infer that the alternative
hypothesis is true.
• Remember: The alternative hypothesis is the more important

one. It represents what we are investigating, which is why it is
also called the research hypothesis.
Chapter-Opening Example: SSA Envelope
Plan 11.52
• Federal Express (FedEx) sends invoices to customers requesting

payment within 30 days.
• The bill lists an address and customers are expected to use their
own envelopes to return their payments.
• Currently the mean and standard deviation of the amount of
time taken to pay bills are 24 days and 6 days, respectively.
• The chief financial officer (CFO) believes that including a stamped
self-addressed (SSA) envelope would decrease the amount of
time.
SSA Envelope Plan
11.53
• She calculates that the improved cash flow from a 2-day
decrease in the payment period would pay for the costs of the
envelopes and stamps.
• Any further decrease in the payment period would generate
a profit.
• To test her belief she randomly selects 220 customers and

includes a stamped self-addressed envelope with their
invoices.
• The numbers of days until payment is received were recorded.

Can the CFO conclude that the plan will be profitable?
SSA Envelope Plan
11.54
IDENTIFY
• The objective of the study is to draw a conclusion about the

mean payment period. Thus, the parameter to be tested is the
population mean 𝜇.
• We want to know whether there is enough statistical evidence to

show that the population mean is less than 22 days. Thus, the
alternative hypothesis is
𝐻1 : 𝜇 < 22
• The null hypothesis is
𝐻0 : 𝜇 = 22
SSA Envelope Plan
11.55
IDENTIFY
• The test statistic is

ҧ
𝑥−𝜇
𝑧=
𝜎/ 𝑛
• We wish to reject the null hypothesis in favor of the alternative

only if the sample mean and hence the value of the test
statistic is small enough.
• As a result we locate the rejection region in the left tail of the

sampling distribution.
• We set the significance level at 10%.
SSA Envelope Plan
COMPUTE: Manually
11.56
• Rejection region: 𝑧 < −𝑧𝛼 = −𝑧.10 = −1.28
• From the data in Xm11-00 we compute
σ220
𝑖=1 𝑥𝑖 4,759 ҧ
𝑥−𝜇 21.63−22
𝑥ҧ = = = 21.63 and 𝑧 = = = −.91,
220 220 𝜎/ 𝑛 6/ 220
• Because the value of test statistic , 𝑧 = −.91, is not less than -1.28,
we do not reject the null hypothesis and we do not conclude that
the alternative hypothesis is true.
• There is insufficient evidence to infer that the mean is less than
22 days.
SSA Envelope Plan
COMPUTE: Manually
11.57
• We can determine the p-value of the test as follows:

p-value = 𝑃 𝑍 < −.91 =.1814
• In this type of one-tail (left-tail) test of hypothesis, we calculate the

p-value as 𝑃 𝑍 < 𝑧 , where z is the actual value of the test statistic.
• Figure 11.7 depicts the sampling distribution, rejection region, and
p-value.
SSA Envelope Plan
11.58
z-Test of a Mean
Sample mean 21.63 z Stat -0.91

Population standard deviation 6 P(Z<=z) one-tail 0.1802
Sample size 220 z Critical one-tail 1.2816
Hypothesized mean 22 P(Z<=z) two-tail 0.3604
Alpha 0.1 z Critical two-tail 1.6449
SSA Envelope Plan
11.59
Figure 11.7 Sampling Distribution for SSA Envelope
p-value=.1814
-1.28 -.91
Rejection region
SSA Envelope Plan
INTERPRET 11.60
• Conclusion:
• The value of the test statistic is -.91, and its p-value is .1814: this
figure does not allow us to reject the null hypothesis.
• We say that there is not enough evidence to infer that the
mean payment period is less than 22 days. In the absence of
evidence, we cannot infer that the plan would be profitable.
• A Type I error occurs when we conclude that the plan works
when it actually does not. The cost of this mistake Is not high.
• A Type II error occurs when we don’t adopt the SSA envelope
plan when it would reduce costs. The cost of this mistake can
be high.
• As a consequence, we would like to minimize the probability
of a Type II error. Thus, we chose a large value for the
probability of a Type I error; we set 𝛼 = .10.
11-2i One– and Two–Tail Tests (단측 검정과 양측검정)
11.61
• The statistical tests conducted in department store
example(Example 11.1) is called one tail tests, because the
rejection region is located in only one tail of the sampling
distribution:
• More correctly, this was an example of a right tail test.

• The p-value is also computed by finding the area in one tail of
the sampling distribution.
11-2i One– and Two–Tail Testing
11.62
• The SSA Envelope example is a left tail test because the
rejection region was located in the left tail of the sampling
distribution.
Right-Tail Testing
11.63
Left-Tail Testing
11.64
Two–Tail Testing
11.65
• Two tail testing is used when we want to test a research hypothesis
that a parameter is not equal (≠) to some value
Example 11.2 Comparison of AT&T and
Its Competitor 11.66
• In recent years, a number of companies have been

formed to compete with AT&T in long-distance calls. All
advertise that their rates are lower than AT&T's, and as a
result their bills will be lower.
• AT&T has responded by arguing that for the average

consumer there will be no difference in billing.
• Suppose that a statistics practitioner working for AT&T

determines that the mean and standard deviation of
monthly long-distance bills for all its residential customers
are $17.09 and $3.87, respectively.
Example 11.2 Comparison of AT&T
and Its Competitor 11.67
• He then takes a random sample of 100 customers and

recalculates their last month's bill using the rates quoted by a
leading competitor.
• Assuming that the standard deviation of this population is the

same as for AT&T, can we conclude at the 5% significance level
that there is a difference between AT&T's bills and those of the
leading competitor?
Example 11.2
11.68
IDENTIFY
• In this problem, we want to know whether the mean monthly

long-distance bill is different from $17.09.
• Consequently, we set up the alternative hypothesis to express
this condition:
𝐻1 : 𝜇 ≠ 17.09
• The null hypothesis specifies that the mean is equal to the value
specified under the alternative hypothesis. Hence
𝐻0 : 𝜇 = 17.09
Example 11.2
11.69
COMPUTE
• The rejection region is set up so we can reject the null hypothesis
when the test statistic is large or when it is small.
stat is “small” stat is “large”

• That is, we set up a two-tail rejection region. The total area in the
rejection region must sum to α, so we divide this probability by 2.
Example 11.2
11.70
COMPUTE
• At a 5% significance level (i.e. α = .05), we have

α/2 = .025. Thus, z.025 = 1.96 and our rejection region is:
z < –1.96 -or- z > 1.96
z
-z.025 0 +z.025
Example 11.2
COMPUTE 11.71
• From the data (Xm11-02), we calculate 𝑋ത = 17.55
ҧ
𝑥−𝜇
• Using our standardized test statistic, 𝑧 = , we find that:
𝜎/ 𝑛
ҧ
𝑥−𝜇 17.55−17.09
𝑧= = = 1.19
𝜎/ 𝑛 3.87/ 100
• Since z = 1.19 is neither greater than 1.96, nor less than –1.96, we
cannot reject the null hypothesis in favor of 𝐻1 . That is “there is
insufficient evidence to infer that there is a difference between the
bills of AT&T and the competitor.”
Two-Tail Test p-value
11.72
COMPUTE
• We can also calculate the p-value of the test. Because it is a

two-tail test, we determine the p-value by finding the area in
both tails; that is,
𝑝 − 𝑣𝑎𝑙𝑢𝑒 = 2𝑃(𝑍 > 𝑧 )
where z is the actual value of the test statistic and |z| is its
absolute value.
• For Example 11.2 we find

𝑝 − 𝑣𝑎𝑙𝑢𝑒 = 2𝑃 𝑍 > 1.19 = 2 .1170 = .2340
Example 11.2
11.73
Figure 11.8 Sampling Distribution for Example 11.2
• There is not enough evidence to infer that the mean long-
distance bill is different from AT&T’s mean of $17.09.
• Figure 11.8 depicts the sampling distribution for this example.
𝒑 − 𝒗𝒂𝒍𝒖𝒆
=. 𝟏𝟏𝟕𝟎
𝟐
-1.96 1.19 1.96

Rejection region Rejection region
11-2j When Do We Conduct One- and
Two-Tail Tests? 11.74
• A two-tail test is conducted whenever the alternative hypothesis
specifies that the mean is not equal to the value stated in the null
hypothesis—that is, when the hypotheses assume the following
form:
𝐻0 : 𝜇 = 𝜇0 vs. 𝐻1 : 𝜇 ≠ 𝜇0
11-2j When Do We Conduct One- and
Two-Tail Tests? 11.75
• We conduct a one-tail test that focuses on the right tail of the
sampling distribution whenever we want to know whether there is
enough evidence to infer that the mean is greater than the
quantity specified by the null hypothesis:
𝐻0 : 𝜇 = 𝜇0 vs. 𝐻1 : 𝜇 > 𝜇0
• The left tail test focuses on the left tail of the sampling distribution
to determine whether there is enough evidence to infer that the
mean is less than the value of the mean stated in the null
hypothesis:
𝐻0 : 𝜇 = 𝜇0 vs. 𝐻1 : 𝜇 < 𝜇0
Summary of One- and Two-Tail Tests…
11.76
One-Tail Test Two-Tail Test One-Tail Test
(left tail) (right tail)
𝐻0 : 𝜇 = 𝜇0 𝐻0 : 𝜇 = 𝜇0 𝐻0 : 𝜇 = 𝜇0
𝐻1 : 𝜇 < 𝜇0 𝐻1 : 𝜇 ≠ 𝜇0 𝐻1 : 𝜇 > 𝜇0
11-2k Testing Hypothesis and Confidence
Interval Estimators 11.77
• The test statistic and the confidence interval estimator are both
derived from the sampling distribution.
• So we can use the confidence interval estimator to test
hypothesis.
• To illustrate, consider Example 11.2. The 95% confidence
interval estimate of the population mean is
𝜎 3.87
𝑥ҧ ± 𝑧𝛼/2 = 17.55 ± 1.96 = 17.55 ± .76
𝑛 100
LCL=16.79 and UCL=18.31
• We estimate that 𝜇 lies between 16.79 and 18.31. Because this
interval includes 17.09, we cannot conclude that there is
sufficient evidence to infer that the population mean differs
from 17.09.
• In Example 11.1, the 95% confidence interval estimate is

𝜎 65
𝑥ҧ ± 𝑧𝛼/2 = 178 ± 1.96 = 178 ± 6.37
𝑛 400
LCL=171.67 and UCL=184.37
• The interval estimate excludes 170, allowing us to conclude
that the population mean account is not equal to $170.
• The confidence interval estimator can be used to conduct tests

of hypotheses. This process is equivalent to the rejection region
approach.
• Instead of finding the critical values of the rejection region
and determining whether the test statistic falls into the region,
we compute the interval estimate and determine whether
the hypothesized value of the mean falls into the interval.
• Using the interval estimator to test hypotheses has the
advantage of simplicity.
• We don’t need the formula for the test statistic
• We need only the interval estimator
• There are two serious drawbacks:

• First, when conducting a one-tail test, our conclusion may not
answer the original question.
• Second, the confidence interval estimator does not yield a p-
value, which we have argued is the better way to draw
inferences about a parameter.
• Using the confidence interval estimator forces the decision
maker into making a reject-don’t reject decision rather
than providing information about how much statistical
evidence exists to be judged with other factors in the
decision process.
11-2l Developing an Understanding of
Statistical Concepts 1 11.81
• As is the case with the confidence interval estimator, the test of

hypothesis is based on the sampling distribution of the sample
statistic.
• The result of a test of hypothesis is a probability statement
about the sample statistic.
• We assume that the population mean is specified by the null
hypothesis.
• We then compute the test statistic and determine how likely it
is to observe this large (or small) a value when the null
hypothesis is true.
• If the probability is small, we conclude that the assumption
that the null hypothesis is true is unfounded and we reject it.
11-2m Developing an Understanding of
Statistical Concepts 2 11.82
• When we (or the computer) calculate the value of the test statistic
ҧ
𝑥−𝜇
𝑧= ,
𝜎/ 𝑛
we’re also measuring the difference between the sample statistic
𝑥ҧ and the hypothesized value of the parameter 𝜇 in terms of the
standard error 𝜎Τ 𝑛.
• In Example 11.2, the value of the test statistic was 𝑧 = 1.19. This
means that the sample mean was 1.19 standard error above the
hypothesized value of 𝜇.
• The standard normal probability table told us that this value is
not considered unlikely. As a result, we did not reject the null
hypothesis.
• The unit of measurement of the difference is the standard error.
11-3 Probability of a Type II Error (β)
11.83
• To properly interpret the results of a test of hypothesis, you must
be able to specify an appropriate significance level or to judge
the p-value of a test.
• However, you must also understand the relationship between a
Type I and Type II errors.
• In this section, we describe how the probability of a Type II error is
computed and interpreted.
11.84
• Recall Example 11.1…
𝐻0 : 𝜇 = 170
𝐻1 : 𝜇 > 170
, where we conducted the test using the sample mean as the
test statistic and we computed the rejection region (with 𝛼 = .05) as
𝑥ҧ > 175.34
• At a significance level of 5% we rejected H0 in favor of H1 since
our sample mean (178) was greater than the critical value of 𝑋ത
(175.34).
11.85
• A Type II error occurs when a false null hypothesis is not rejected.
• In example 11.1, this means that if 𝑋ത is less than 175.34 (our
critical value) we will not reject our null hypothesis, which
means that we will not install the new billing system.
• Thus, we can see that:
𝛽 = 𝑃(𝑋ത < 175.34 𝑔𝑖𝑣𝑒𝑛 𝑡ℎ𝑎𝑡 𝑡ℎ𝑒 𝑛𝑢𝑙𝑙 ℎ𝑦𝑝𝑜𝑡ℎ𝑒𝑠𝑖𝑠 𝑖𝑠 𝑓𝑎𝑙𝑠𝑒)
• The condition that the null hypothesis is false tells us only that
the mean is not equal to 170.
• If we want to compute 𝛽, we need to specify a value for 𝜇.
Example 11.1 (revisited)
11.86
• For example, suppose that if the mean account balance is $180
the new billing system will be so profitable that the manager
would hate to lose the opportunity to install it.
• She would like to determine the probability of not installing the
new system when it would produce large cost savings.
• Calculate the probability of not installing the new system
when 𝜇 is equal to 180:
𝛽 = 𝑃(𝑋ത < 175.34, 𝑔𝑖𝑣𝑒𝑛 𝑡ℎ𝑎𝑡 𝜇 = 180), thus
ത
𝑋−𝜇 175.34−180
𝛽=𝑃 < = 𝑃 𝑍 < −1.43 = .0764∎
𝜎/ 𝑛 65/ 400
• This tells us that when the mean account is actually $180, the
probability of incorrectly not rejecting the null hypothesis is .0764.
• Fig. 11.9 graphically depicts how the calculation was performed.
Example 11.1 (revisited)
11.87
Figure 11.9 Calculating 𝛽 𝑓𝑜𝑟 𝜇 = 180, 𝛼 = .05, 𝑎𝑛𝑑 𝑛 = 400
Our original hypothesis…
our new assumption…
𝛽 = 𝑃(𝑥ҧ < 175.34, 𝑔𝑖𝑣𝑒𝑛 𝑡ℎ𝑎𝑡 𝜇 = 180)

11-3a Effects on β of Changing α
11.88
• Decreasing the significance level 𝛼, increases the value of 𝛽 and
vice versa.
• Change 𝛼 to .01 in Example 11.1.
• Stage 1: Rejection region
𝑧 > 𝑧𝛼 = 𝑧.01 = 2.33
ҧ
𝑥−𝜇 ҧ
𝑥−170
𝑧= = > 2.33
𝜎/ 𝑛 65/ 400
• Solving for 𝑥,
ഥ we find the rejection region in terms of
unstandardized test statistic
𝑥ҧ > 177.57
Effects on β of Changing α
11.89
• Stage 2: Probability of a Type II error
•
𝛽 = 𝑃 𝑥ҧ < 177.57 𝜇 = 180
ҧ
𝑥−𝜇 177.57−180
=𝑃 <
𝜎/ 𝑛 65/ 400
= 𝑃 𝑧 < −.75 = .2266∎
11-3a Effects on β of Changing α
11.90
• Decreasing the significance level α, increases the value of β and vice versa.
• Consider this diagram again. Shifting the critical value line to the right (to
decrease α) will mean a larger area under the lower curve for β… (and vice
versa)
Fig. 11.9 Fig. 11.10
11-3b Judging the Test (검정의 판단)
11.91
• A statistical test of hypothesis is effectively defined by the significance level
(α) and the sample size (n), both of which are selected by the statistics
practitioner.
• Therefore, if the probability of a Type II error (β) is judged to be too large, we

can reduce it by
• Increasing α, and/or increasing the sample size, n.
11-3b Judging the Test
11.92
• For example, suppose we increased n from a sample size of 400
account balances to 1,000 in Example 11.1.
Stage 1: Rejection region

𝑧 > 𝑧𝛼 = 𝑧.05 = 1.645
ҧ
𝑥−𝜇 ҧ
𝑥−170
𝑧= = > 1.645
𝜎Τ 𝑛 65Τ 1,000
𝑥ҧ > 173.38
11-3b Judging the Test
11.93
Stage 2: Probability of a Type II error
𝛽 = 𝑃 𝑥ҧ < 173.38 𝜇 = 180
ҧ
𝑥−𝜇 173.38−180
=𝑃 <
𝜎Τ 𝑛 65Τ1,000
= 𝑃 𝑧 < −3.22
= 0 (𝑎𝑝𝑝𝑟𝑜𝑥𝑖𝑚𝑎𝑡𝑒𝑙𝑦)
Fig. 11.9 and 11.11 Compare β at n=400 and n=1,000
… 𝜎Τ 𝑛
11.94
n=400 175.35
• The sampling distribution of

the mean is narrower because
the standard error of the
mean becomes smaller as n
increases.
173.38
n=1,000
• By increasing the sample size,

we reduce the probability of a
Type II error.
11-3c Developing an Understanding of
Statistical Concepts 11.95
• The calculation of the probability of a Type II error for n = 400

and for n = 1,000 illustrates a concept whose importance cannot
be overstated.
• By increasing the sample size we reduce the probability of a

Type II error. By reducing the probability of a Type II error we
make this type of error less frequently.
• And hence, we make better decisions in the long run. This finding
lies at the heart of applied statistical analysis and reinforces the
book's first sentence, "Statistics is a way to get information from
data."
11-3c Developing an Understanding of
Statistical Concepts 11.96
• Throughout this book we introduce a variety of applications in
finance, marketing, operations management, human resources
management, and economics.
• In all such applications the statistics practitioner must make a
decision, which involves converting data into information. The
more information, the better the decision.
• Without such information, decisions must be based on
guesswork, instinct, and luck. A famous statistician, W. Edwards
Deming* said it best: "Without data you're just another person
with an opinion."
* W. Edwards Deming(Oct. 14, 1900-Dec. 20, 1993): an American engineer, statistician,

professor…, management consultant; He developed the theory of Total Quality
Management (TQM) that focuses on quality and continuous improvement. He also
introduced the Plan-Do-Study-Act(PDSA) cycle for problem solving.
11-3d Power of a Test (검정력)
11.97
• Another way of expressing how well a test performs is to report its
power: the probability of its leading us to reject the null hypothesis
when it is false. Thus, the power of a test is 1 − 𝛽.
• When more than one test can be performed in a given situation,

we would naturally prefer to use the test that is correct more
frequently.
• If (given the same alternative hypothesis, sample size, and

significance level) one test has a higher power than a second
test, the first test is said to be more powerful. (첫번째 검정이 보다
강력하다고 말한다.)
11-3e Operating Characteristic Curve
(검사특성곡선)
11.98
• To compute the probability of a Type II error, we must specify the
significance level(𝛼), the sample size (𝑛), and an alternative value
of the population mean.
• One way to keep track of all these components is to draw the
operating characteristic(OC)curve, which plots the value of 𝛽
versus the values of 𝜇.
• We used the Excel function NORMDIST to compute the
probability of a Type II error in Example 11.1 for 𝜇 =
170, 171, … , 184, with n=400.
• Figure 11.12 depicts this curve*.
* 20221201 OC curve chap.11
11.99
• Excel Function Instruction
• With 𝜎 = 65 and 𝑛 = 400, the standard error of the mean is
𝜎 65
𝜎𝑥ҧ = = = 3.25
𝑛 400
• To calculate the probability of a Type II error in Example 11.1,

we open Excel and in any empty cell type
=NORMDIST(175.35, [𝜇], 3.25, True)
• For example, to compute 𝛽 when 𝜇 =180, we type
=NORMDIST(175.35, 180, 3.25, True)
11.100
• Notice as the alternative value of 𝜇 increases the value of 𝛽
decreases.
• In other words, it becomes easier to distinguish between 𝜇 =170
and other values of 𝜇 when 𝜇 is farther from 170.
• When 𝜇=170, 𝛽 = 1 − 𝛼.
11.101
• The OC curve can also be useful in selecting a sample size.
• Figure 11.13 shows the OC curve for Example 11.1 with 𝑛 =
100, 400, 1,000, 𝑎𝑛𝑑 2,000.
• An examination of this chart sheds some light on the effect
increasing the sample size has on how well the test performs at
different values of 𝜇.
• Smaller sample sizes will work well to distinguish between 170
and values of larger than 180.
• However, to distinguish between 170 and smaller values of
requires larger sample sizes.
11.102
11-3f Determining the Alternative Hypotheses to
Define Type I and Type II Errors 11.103
• We’ve already discussed how the alternative hypotheses is
determined?
• In Example 11.1, we wanted to know whether there was
sufficient statistical evidence to infer that the new billing system
would be cost-effective—that is, whether the mean monthly
account is greater than $170.
• In general, you will find that the question can be posed in two
ways.
• In Example 11.1, we asked whether there was evidence to
conclude that the new system would be cost-effective.
• Another way of investigating the issue is to determine whether
there is sufficient evidence to infer that the new system would
not be cost-effective.
• In a criminal trial, the burden of proof falls on the prosecution to
prove that the defendant is guilty.
• In other countries with less emphasis on individual rights, the
defendant is required to prove his or her innocence.
• This is so because we consider the conviction of an innocent
defendant to be the greater error. (committing a Type I error is a
more serious one.)
• Thus the test is set up with the null and alternative hypotheses as
described in Section 11-1.
• In a statistical test where we are responsible for both asking and
answering a question, we must ask the question so that we
directly control the error that is more costly.
• As you have already seen, we control the probability of a Type
I error by specifying its value(the significance level, 𝛼).
• In Example 11.1, there are two possible errors:
(1) conclude that the billing system is cost-effective when it isn’t;
(2) conclude that the system is not cost-effective when it is.
• If the manager concludes that the billing plan is cost-effective,

the company will install the new system. If, in reality, the system
is not cost-effective, the company will incur a loss.
• If the manager concludes that the billing plan is not going to
be cost-effective, the company will not install the system.
However, if the system is actually cost-effective, the company
will lose the potential gain from installing it.
• Which cost is greater?
• Suppose we believe that the cost of installing a system that is not
cost-effective is higher than the potential loss of not installing an
effective system.
• The error we wish to avoid is the erroneous conclusion that the
system is cost-effective. We define this as a Type I error. The
burden of proof is placed on the system to deliver sufficient
statistical evidence that the mean account is greater than
$170. The null and alternative hypotheses are as formulated
previously.
𝐻0 : 𝜇 = 170 vs 𝐻1 : 𝜇 > 170
• If we believe that the potential loss of not installing the new system
when it would be cost-effective is larger cost, we would place the
burden of proof on the manager to infer that the mean monthly
account is less than $170.
• Consequently, the hypotheses would be
𝐻0 : 𝜇 = 170 vs 𝐻1 : 𝜇 < 170
• This discussion emphasizes the need in practice to examine the

costs of making both types of error before setting up the
hypotheses.
• [It is important for readers to understand that the questions posed in
exercises throughout this book have already take these costs into
consideration.]
11-4 The Road Ahead
11.109
• We had two principal goals to accomplish in chaps. 10 and 11.
• First, we wanted to present the concepts of estimation and
hypothesis testing.
• Second, we wanted to show how to produce confidence
interval estimates and conduct tests of hypotheses.
• The importance of both goals should not be underestimated.
• Sections 10-2 and 11-2 set the pattern for the ways in which
statistical techniques are applied.
• If you understand how to produce and use confidence interval
estimates and how to conduct and interpret hypothesis tests,
then you are well on your way to the ultimate goal of being
competent at analyzing, interpreting, and presenting data.
11-4 The Road Ahead
11.110
• In the chapters that follow, we plan to present about three dozen
different statistical techniques that can be (and frequently are)
employed by statistics practitioners.
• To calculate the value of test statistics or confidence interval
estimates requires nothing more than the ability to add,
subtract, multiply, divide and compute square roots.
• If you intend to use the computer, all you need to know are
the commands.
• The key to applying statistics is knowing which formula to
calculate or which set of commands to issue.
• The real challenge of the subject lies in being able to define
the problem and identify which statistical method is the most
appropriate one to use. (ICI approach: Identify-Compute-
Interpret)
11-4 The Road Ahead
11.111
• A number of factors determine which statistical method
should be used, but two are especially important: the type of
data and the purpose of the statistical inference.
• Type of data:
• Nominal data: categories such as marital status,
occupation and gender; count the number of times each
category is observed.
• Ordinal data: obtained from questions whose answers
represent a rating or a ranking system
• Interval data: real numbers, such as those representing
income, age, height, weight, and volume; computations of
means and variances is permissible.
11-4a Problem Objectives
11.112
• The second key factor in determining the statistical technique
is the purpose of doing the work. Every statistical method has
some specific objective. We address five such objectives in this
book.
1. Describe a population:
• Population’s incomes: mean and variance…
• The brand of computer: the proportion of the
population
2. Compare two populations
3. Compare two or more populations
4. Analyze the relationship between two variables
5. Analyze the relationship among two or more variables
11-4a Problem Objectives
11.113
Table 11.3 Guide to Statistical Inference Showing Where
Each Technique Is Introduced
Data Type
Problem Objective Nominal Ordinal Interval
Describe a population §12-3, 15-2 Not covered § 12-1, 12-2
Compare two populations §13-5, 15-2 §19-1,19-2 § 13-1, 13-3, 13-4,
19-1, 19-2
Compare two or more §15-2 §19-3 Chap.14, §19-3
populations
Analyze the relationship §15-2 §19-4 Chap.16
between two variables
Analyze the relationship Not covered Not covered Chaps.17,18
among two or more
variables
§: Section
11-4b Derivations
11.114
• Though this book is about statistical applications, it might be
helpful to have some understanding about the process that
produces the formulas.
• Factors such as the problem objective and the type of data
determine the parameter to be estimated and tested.
• For each parameter, statisticians have determined which statistic
to use. That statistic has a sampling distribution that can usually
be expressed as a formula.
11-4b Derivations
11.115
• For example, if the parameter of interest is the population mean
𝜇, then its best estimator is the sample mean 𝑋. ത (Each parameter
has a “best” estimator.)
• Assuming that the population standard deviation 𝜎 is known,
the sampling distribution of 𝑋ത is normal (or approximately so) with
mean 𝜇 and standard deviation 𝜎Τ 𝑛. The sampling distribution
can be described by the formula,
ത
𝑋−𝜇
𝑍=
𝜎Τ 𝑛
• This
formula also describes the test statistic for 𝜇 with known 𝜎.
the sampling distribution is often the formula for the test statistic.
• With a little algebra the confidence interval estimator can be
derived from the sampling distribution.
𝜎
𝑥ҧ ± 𝑧𝛼/2
𝑛
11.116
End of Chapter 11

SME11e PPT ch11std

Uploaded by

Copyright:

Available Formats

SME11e PPT ch11std

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SME11e PPT ch11std

Uploaded by

Copyright:

Available Formats

Slides by:

• Concepts of Hypothesis Testing

• The alternative hypothesis (대립가설) or research hypothesis

* 피고가 유죄라고 선언하는 것

• Notice that we do not say that the defendant is innocent, only

• Our justice system does not allow this decision****.

* 피고가 무죄라고 선고하는 것; ** 귀무가설을 기각하지 못함

* 유죄인 피고가 무죄를 선고받을 때 발생

• The probability of a Type I error is denoted as α (Greek letter

• The two error probabilities 𝛼 𝑎𝑛𝑑 𝛽 are inversely related:

Table 11.1 Terminology of Hypothesis Testing

Decision 𝑯𝟎 is true (=Defendant 𝑯𝟎 is false (= Defendant

Do not reject 𝑯𝟎 Correct decision Type 𝑰𝑰 Error

• Oliver Wendell Holmes*, a United States Supreme Court justice,

* Holmes, Oliver Wendell,Jr.(1841-1935): US Supreme Court associate justice

• Let’s extend these concepts to statistical hypothesis testing.

• When we state the hypotheses, we list the null first followed by

• Notice that in both illustrations (𝐻0 : 𝜇 = 350 and 𝐻0 : 𝜇 ≤ 350 ) the

• Suppose that the manager now wanted to determine whether

• If there is evidence of an increase or decrease in the value of

• The test statistic is the criterion on which we base our decision

• Example 11.1 Department Store’s New Billing System

• The manager knows that the accounts are approximately normally

• We express this belief as our research hypothesis, that is:

• Thus, our null hypothesis becomes:

• What we want to show:

• To test our hypotheses, we can use two different approaches:

• We will explore both in turn…

• Suppose we define the value of the sample mean that is just

• From Section 8-2, we defined 𝑧𝛼 to be the value of a standard

• We can calculate this 𝑥ҧ𝐿 based on any level of significance (𝛼)

• At a 5% significance level (i.e. 𝛼 = 0.05), we get

• Since our sample mean(178) is greater than the critical value we

and compare its result to 𝑧𝛼 : (rejection region: z > 𝑧𝛼 )

Since 𝑧 = 2.46 > 1.645(= 𝑧.05 ), we reject 𝐻0 in favor of 𝐻1 …

* 표준화 검정 통계량 = 검정 통계량

• In the case of our department store example, what is the probability of

Fig. 11.4 p-Value for Example 11.1

p-value = P(Z > 2.46)

• Consequently, we have reason to reject the null hypothesis

• The p-value of a test provides valuable information because it is

172 0.62 .2676

174 1.23 .1093

176 1.85 .0322

178 2.46 .0069

180 3.08 .0010

p-value=P(Z> .62) =.2676

p-value=P(Z> 1.85) =.0322

* overwhelming(압도적인); highly significant(매우 유의하다)

• Compare the p-value with the selected value of the significance

• Since p-value = .0069 < α = .05, we reject H0 in favor of H1.

Sample mean 178 z Stat 2.46

• If we do not reject the null hypothesis, we conclude that there

• Remember: The alternative hypothesis is the more important

• Federal Express (FedEx) sends invoices to customers requesting

• To test her belief she randomly selects 220 customers and

• The numbers of days until payment is received were recorded.

• The objective of the study is to draw a conclusion about the

• We want to know whether there is enough statistical evidence to