Chapter 8

Introduction to
Hypothesis Testing

1
The purpose of hypothesis testing is to determine
whether there is enough statistical evidence supporting
a certain belief (or claim) about a parameter.

Examples
• Is there statistical evidence that support the
hypothesis that more than p% of all potential
customers will purchase a new products?
• Is there statistical evidence that support the
hypothesis that a certain drug is effective?

2
8.1 Concepts of Hypothesis Testing
The Reject-Region method
Step 1: Two hypotheses are defined.

H0: The null hypothesis specifies our current belief

about the parameter we test. (m = 170, p = .4, etc.)
Must be a specific value.
H1: The alternative hypothesis specifies a range of
values for the parameter tested (m > 170; p  .4; etc.)
effected by the belief.
3
A person is assumed
to be innocent
before court trial Not guilty
H0 Needed in the test Do not reject H0

Reject H0

guilty
Claim
Claim HH000 HHAA
m=A
m=A m=A
m=A m≠A
m≠A 22-tail
-tail
–tail
or
or 2-side
2-sidetest
test
m>A
m>A m=A
m=A m>A
m>A 11–tail
–tail
or
or 1-side
1-sidetest
test
m≥A
m≥A m=A
m=A m<A 1 –tail
or 1-side test
m<A
m<A m=A
m=A m<A 1 –tail
or 1-side test
m≤A
m≤A m=A
m=A m>A 1 –tail
or 1-side test
8.2 Testing the Population Mean when the Population
Standard Deviation is Known
Example 1: The manager of a department determines
that new billing 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. Assume a standard deviation of
$65. Can the manager conclude from this that the new
system will be cost-effective?
Step 1: H0: µ = 170
H1: µ > 170
6
Step 2: Choose the significance level a.
H00 is really
true true H00 is really
false false
HAA is really
false false HAA isisreally
true true
true
Really innocent Really guilty
reject H00 Type I error OK
Verdict: Guilty a = Probability of committing the
type I error
Send an innocent person
to jail
Do not reject
accept H0 H00 OK Type II error
Verdict:
accept HNot
0 Guilty OK b=
TypeProbability
II error of committing the
b= Probability
type II error of committing the
type
Let aIIguilty
errorperson go free

The more severe the consequence of committing the

type-I error, the smaller/higher the value of a.
Example 1: a = 0.05
Step 3: Determine the sample size n and hence the
sampling distribution.
Example 1:n=400; N(0,1)

8
Step 4: Depending on the sampling distribution, the HA,
and the value of a, find the suitable critical value and
the reject region.
HA: μ > 170 (1-sided test)

aN(0,1)
=.05

Reject region: Z > 1.645

1.645
Critical value
Right-Tail Testing
Left-Tail Testing
Two–Tail Testing
Step 5: Collect data. x = 178
Calculate the standard test statistic .

Example 1:

HA: μ =170

x − m 178 − 170
z= = = 2.46
 / n 65 / 400
Step 6: If the test statistic is in the reject region, then
reject H0, otherwise do not reject H0.

Example 1: Reject H0 at a = .05

Reject H0: There is enough statistical evidence to

conclude that the alternative hypothesis is true.

Do not reject H0: There is not enough statistical

evidence to conclude that the alternative hypothesis is
true.
 A Left Hand Tail Test

H0: m = m 0
H1: m < m 0

Reject H0 if x falls here Critical

value

15
The SSA envelop plan example: The chief financial
officer in FedEx believes that including a stamped self-
addressed (SSA) envelop in the monthly invoice sent
to customers will decrease the amount of time it take
for customers to pay their monthly bills. Currently,
customers return their payments in 22 days on the
average, with a standard deviation of 6 days.
A random sample of 220 customers was selected and
SSA envelops were included with their invoice packs.
The mean time it took customers to pay their bill was
21.63
Can the CFO conclude that the plan will be successful
at 10% significance level?
16
Step 1: H0: m = 22; H1: m < 22
Step 2: a = 0.10
Step 3: n= 220
a = .10 Z
Step 4:
Step 5: −1.28

x − m 21.63 − 22
z= = = −.91
 / n 6 / 220

Step 6: Do not reject the null hypothesis 17

 A Two Tail Test

H0: m = m0
H1: m  m0

a /2 a /2
Reject H0 if x falls here Reject H0 if x falls here

Critical Critical
value value 18
Example 2: A statistician believes the monthly mean of
the long-distance bills for all AT&T residential customers
is $17.09.
A random sample of 100 customers of its leading
competitor yields x = 17.55
Assuming the standard deviation of the bills of the
competitors is 3.87, can we infer that there is a
difference between AT&T’s bills and the competitor’s
bills (on the average)?

19
Step 1: H0: m = 17.09; H1: m  17.09
Step 2: a = 0.05
Step 3: n= 100 a /2 = .025 a /2 = .025
Step 4:
Step 5: −1.96 1.96
Critical value Critical value

x − m 17.55 − 17.09
z= = = 1.19
 / n 3.87 / 100

Step 6: Do not reject the null hypothesis 20

 Pop dist:
Normal Z-dist
 known n  30
?
Pop dist:
n > 30
μ? Non-Normal Z-dist