(
(
2
21 , 995 . 0
2
2
21 , 005 . 0
2
2
1 , 2 / 1
2
2
1 , 2 /
2
) 1 (
,
) 1 (
) 1 (
,
) 1 (
_ _
_ _
o o
s n s n
s n s n
n n
745)
( )
2
2 2
2
22
1
: 25.4
0.01
99% CI on population variances is:
n
n s
Sample s o
_
o
=
= =
=
 
 
12.88, 66.4
:
3.59, 8.15
C I o n i s t h e r e f o r e o
4 . 25 ) 1 22 (
,
4 . 25 ) 1 22 (
2
21 , 995 . 0
2
21 , 005 . 0 (
(
_ _
Single Population Tests of Hypothesis
(Moving on to Chapter 8)
In statistics, a Hypothesis is an
assumption or theory about one or more
populations.
A Hypothesis Test is a statistical
procedure that involves formulating a
hypothesis and using sample data to
decide on its validity.
The Null Hypothesis is a statement about
the population. It is referred to as H
0
.
The Alternative Hypothesis is a
statement about the population that is
the opposite to the null hypothesis. It is
referred to as H
a
.
4
A Test Statistic is a number that
captures the information in a sample
that is used to choose between the null
and alternative hypotheses.
The Rejection Region is the range of
values for a test statistic that will lead
you to reject the null hypothesis.
.
4

3
.
3

2
.
6

1
.
9

1
.
2

0
.
5
0
.
2
0
.
9
1
.
6
2
.
3
33
.
7
The 5Step Hypothesis Testing Procedure
Step 1: Formulate the null and alternative
hypotheses.
Step 2: Pick the value of o and find the
rejection region.
Step 3: Calculate the test statistic.
Step 4: Reject or fail to reject the null
hypothesis.
Step 5: Interpret the statistical decision in
terms of the problem.
Rejection Region for one type of TwoTail
Test on :
z cutoff z cutoff
The pvalue is defined to be the smallest
value of o for which you can reject H
0
. It
provides a measure of the strength of
the test conclusion.
A Type I Error occurs in hypothesis
testing when you reject the null
hypothesis and the null hypothesis is
actually true, i.e., you incorrectly reject a
null hypothesis due to sample error.
A Type II Error is made when you fail to
reject the null hypothesis and the null
hypothesis is false. This may occur if a
test procedure is too insensitive.
When evaluated quantitatively, the type
II error is defined in conditional terms,
i.e., the probability of failing to reject if
the true parameter value deviates from
the assumed parameter value by some
fixed amount.
5
The probability of making a Type I Error
is called o (alpha).
The probability of making a Type II Error
is called  (beta).
Summary of Testing Errors:
Decision: H
0
is True H
0
is false
Fail to Reject No error Type II error
Reject Type I error No error
The choice of o and  are not generally
independent. For example, choosing a
very small value of could make a
statistical test too insensitive to detect
an error in the assumed parameter value.
Situations Calling for Different and values:
Situation 1:  Design the test for Low , High
A manufacturer of paper towels tests large lots of
finished product for correct resin content because
incorrect resin content in paper towels can reduce wet
strength. Rejecting finished product greatly increases
production costs. Surveys show that most customers
for this product purchase based on price and don't
care much about wet strength.
Assume that the null hypothesis states that resin
content in a finished lot is at the correct level. The
alternative hypothesis states it is not at the correct
level.
Situations Calling for Different and values:
Situation 2:  Design the test for High , Low
A manufacturer of peanut products tests the finished
product for salmonella contamination. Contaminated
peanut products could become a serious public health
threat.
Assume that the null hypothesis states that salmonella
organisms in a given lot are at a safe level. The
alternative hypothesis states that salmonella
organisms are not at a safe level.
Two general rules for hypothesis
formulation in statistics are:
1. The equality component (=, , ) of
the test is placed in the null hypothesis.
2. Tests are usually designed to assure
that the null hypothesis is rejected only
in the case of strong evidence, i.e.,
beyond a reasonable doubt.
6
A TwoTail Test of the population mean
has the following null and alternative
hypotheses:
H
0
: =
0
(a specific number)
H
a
: =
0
05 . 0 = o
A LowerTail Test of the population
mean has the following null and
alternative hypotheses:
H
0
: >
0
(a specific number)
H
a
: <
0
:
.
:
0.05
o
a
H
vs
H
o
>
<
=
An UpperTail Test of the population
mean has the following null and
alternative hypotheses:
H
0
: s
0
(a specific number)
H
a
: >
0
:
:
o
a
H
H
s
>
7
For tests on population means with
normal input populations and known ,
the test statistic is given by:
Z = (Xbar
0
)/(/n) vs. Z
or Z
/2
In this test, the null hypothesis
assumes that =
0
Problem 82 (page 308)
Formulation of tests
82
a.) Correct
b.) Incorrect
c.) Incorrect
d.) Incorrect
e.) Incorrect
f.) Incorrect
g.) Correct
h.) Correct
Problem 811 (page 309)
Type I and II error concepts
811
( )
25 0.2 ( )
: 10 : 10 .)
.) 10.1032 9.8968
Prob 10.1032 9.8968
~ 10,
.01 is the type I error probability
o a
n kg known
H H a
b x x
x or x
x n
n
o
o
o
= =
= =
> s
> s
 

\ .
=
c.)
What is if 10.1 vs. 10?  =
( ) ( ) Prob 9.8968 10.1032 5.08 0.08
0.5319
9.8968 10.1 10.1032 10.1
5.08 0.08
.2 25 .2 25
X P Z  = s s = s s
~
= =

.

\

n
N x
o
, 1 . 10 ~
8
c.)
Similarly if 9.8 vs. 10? =
( ) 0078 . ) 58 . 7 42 . 2 ( 8 . 9 = < < = z P 
58 . 2 = c
d.)

.

\

n
N x
o
, 8 . 9 ~
e.)
f.)
0632 .
162 . 3
2 .
= =
n
o
10
1.96
.0632
X
=
Thus 10.1032 is replaced by X, where
for theupper limit, X =10.1239, for thelower limit, X=9.8761.
10.02
Since is neither 10.1239 or 9.8761, it is not
in the rejection region so it is still plausible that
=10
X
x
=
> s
g.)
10.1032 or 9.8968 iff Z 2.58 or 2.58 X X Z > s > s