Vous êtes sur la page 1sur 9

1

IE - 2333
SWN
CONCEPTS OF HYPOTHESIS TESTING
As the previous section indicated, we can use sample data
to estimate the value of an known parameter. Another way of
drawing inferences about a population is hypothesis testing.
The objective of this from of statistical inference is to
determine whether or not the sample data support some
belief or hypothesis about the population.
STRUCTURE OF HYPOTHESIS TESTING.
The tests of hypotheses that the present in this section are
called parametric test, because they test the value of a
population parameter.
2
IE - 2333
SWN
These tests consist of the following four components :
1. Null hypothesis
2. Alternative hypothesis
3. Test statistic
4. Rejection region

3
IE - 2333
SWN
STRUCTURE OF
HYPOTHESIS TESTING
TEST STATISTIC
The test statistic is the sample
statistic upon which we base our
decision to either reject or not reject
the null hypothesis
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
ALTERNATIVE HYPOTHESIS/*
The alternative hypothesis denoted H
1
or H
A
;
answers the question by specifying that the
parameter is one of the following
1. Greater than the value shown in the null
hypothesis.
2. Less than the value shown in the null hypothesis
3. Different from the value shown in the null
hypothesis

NULL HYPOTHESIS
The null hypothesis, which is denoted
H
o
, must specify that the parameter is
equal to a single value
* = RESEARCH HYPOTHESIS
4
IE - 2333
SWN
RESULTS OF A TEST OF THE NULL
HYPOTHESIS
DECISION H
o
is true H
o
is false

Reject H
o
Type I error
P(type I error) = o

Correct Decision
Do not
reject H
o

Correct Decision
Type II error
P(type II error) = |
o is also known as the level of statistical significance or the significance level of the test
STATE OF NATURE
Note : When hypothesis testing is viewed as a problem in
decision making, two alternative actions can be taken :
accept H
o
or reject H
o
. The two alternatives, truth or falsity
of hypothesis H
o
, are viewed as states of nature or states
of the world
5
IE - 2333
SWN
Example :
Suppose we can tolerate a size of type I error up to 0,06 when
testing the null hypothesis : Ho : = 10 versus H1 : > 10 for
the assembly time problem. Assume the distribution of the time
required to assemble a unit is normal with standard
deviation o=1,4 minutes. Say we observe the assembly time of
25 randomly selected units and chose the sample mean
as the test statistic. In particular, we want to compare the
following three critical regions:
X
: Reject 10, 65
o
Test A H if X > : Reject 10, 45
o
Test B H if X >
: Reject 10, 25
o
Test C H if X >
to determine which one satisfies the size of type I error that can
be tolerated and which has the smallest | among the three.
SOLUTION : discuss in class-room

6
IE - 2333
SWN
EXERCISES
1. Suppose you wish to test the hypothesis H
o
: u = 5
against the alternative H
1
: u = 8
by means of a single observed value of a random variable
with probability density function
( )
1
exp ; 0
( ; )
0 ;
x
x
f x
otherwise
u u
u

>

If the maximum size of type I error that can be tolerated is


0,15, which of the following tests is best for choosing
between the two hypotheses? a. Reject H
o
if X > 9
b. Reject H
o
if X > 10
c. Reject H
o
if X > 11
7
IE - 2333
SWN
2. Suppose a manufacturer of memory chips observes that
the probability of chip failure is p = 0,05. A new procedure
is introduced to improve the design of chips. To test this
new procedure and tested.
Let random variable X denote the number of these 200
chips that fail. We set the test rule that we would accept
the new procedure, if X s 5.
Let H
o
: p = 0,05, versus
H
1
: p < 0,05
Find the probability of a type I error
8
IE - 2333
SWN
3. Let be the mean of a random sample of size n = 36
from NOR (, 9). Our decision rule is to reject H
o
: = 50
and to accept H
1
: > 50 if
Determine the OC () curve and evaluate it at = 50,0 ;
50,5; 51,0; 51,5. What is the significance level of the test?
4. Consider a NOR (, o
2
=40) distribution. To test H
o
: = 32
against H
1
: > 32, we reject H
o
if the sample mean
Find the sample size n and the constant c such that
OC (=32) = 0,90 and OC(=35) = 0,15

X
50, 8 X >
. X c >
9
IE - 2333
SWN
5. Let Y have binomial distribution with parameter n and p. In
a test of H
o
: p = 0,25, against H
1
: p < 0,25 we reject H
o
if
Y s c. Find n and c if OC (p=0,25)=0,90 and OC
(p=0,20)=0,05.
state your assumptions

Vous aimerez peut-être aussi