You are on page 1of 9

TUM School of Management

Production and Supply Chain Management

Prof Martin Grunow

Technische Universitt Mnchen

Designing a Test
The test procedure is designed in such a way that the probability
of the Type I error occurring is contained within a specified, small
value
For every test, we choose a test statistic that provides the
relationship between an estimator and the parameter about
which hypotheses are proposed. The distribution of the test
statistic will be known
Designing the test consists of identifying the Critical Region (CR):
the set of observed values of the test statistic that will lead to
rejection of the null hypothesis
2012 from "A First Course in Quality Engineering: Integrating Statistical
and Management Methods of Quality" by K.S. Krishnamoorthi. Reproduced
by permission of Taylor and Francis Group, LLC, a division of Informa plc.
Holly Ott

TUM School of Management

Production and Supply Chain Management
Prof Martin Grunow

Designing a Test

Holly Ott

TUM School of Management

Production and Supply Chain Management
Prof Martin Grunow

Test Concerning the Mean of a normal

population when is known
H0: = 0 (Hypothesize that the mean equals a number 0)
H1: > 0 (The alternate hypothesis is: if the mean is not equal
to 0 it must be greater than 0)
Test-Statistic:

! " 0
Z obs =
~ N(0,1)
! n

This test-statistic relates the sample average X-bar to the

population mean , and its distribution is known to be the N(0, 1)
distribution

2012 from "A First Course in Quality Engineering: Integrating Statistical

and Management Methods of Quality" by K.S. Krishnamoorthi. Reproduced
by permission of Taylor and Francis Group, LLC, a division of Informa plc.
Holly Ott

TUM School of Management

Production and Supply Chain Management
Prof Martin Grunow

Choosing the Critical Region

Z is called the critical value
of the statistic.
The values in the distribution
of the test statistic beyond Z
constitute the Critical Region
We draw the line at Z so that
the probability of Type I error
is limited to a small number,
.
2012 from "A First Course in Quality Engineering: Integrating Statistical
and Management Methods of Quality" by K.S. Krishnamoorthi. Reproduced
by permission of Taylor and Francis Group, LLC, a division of Informa plc.
Holly Ott

H0: = 0
H1: > 0
8

TUM School of Management

Production and Supply Chain Management
Prof Martin Grunow

Hypothesis Test for the Mean: Example

A supplier of cotton rope claims their new product has average
strength greater than 10 kg. A sample of 16 rope pieces gave an
average of 10.2 kg. If the standard deviation of the strength is known
to be 0.5 kg, test the hypothesis = 10 vs. > 10. Use = 0.01.

2012 from "A First Course in Quality Engineering: Integrating Statistical

and Management Methods of Quality" by K.S. Krishnamoorthi. Reproduced
by permission of Taylor and Francis Group, LLC, a division of Informa plc.
Holly Ott

TUM School of Management

Production and Supply Chain Management
Prof Martin Grunow

Hypothesis Test for the Mean: Example

For = 0.01, we
standard normal
tables that Z will
lie between 2.32
and 2.33.
The exact value is
Z = 2.326.

Holly Ott

10

TUM School of Management

Production and Supply Chain Management
Prof Martin Grunow

Hypothesis Test for the Mean: Example

The Zobs is not in the
critical region, and the null
hypothesis is not rejected.
The mean strength of the
population of cotton ropes
is not greater than 10 kg.
The suppliers claim is not
valid.

2012 from "A First Course in Quality Engineering: Integrating Statistical

and Management Methods of Quality" by K.S. Krishnamoorthi. Reproduced
by permission of Taylor and Francis Group, LLC, a division of Informa plc.
Holly Ott

11

TUM School of Management

Production and Supply Chain Management
Prof Martin Grunow

Believes mean is less than

the hypothesized value, H0

Believes mean is greater

than the hypothesized value,
H0

Believes the mean does not

equal the hypothesized
value, H0

The test statistic will be the same for all three cases; however the
location of the critical region will differ
2012 from "A First Course in Quality Engineering: Integrating Statistical
and Management Methods of Quality" by K.S. Krishnamoorthi. Reproduced
by permission of Taylor and Francis Group, LLC, a division of Informa plc.
Holly Ott

12

TUM School of Management

Production and Supply Chain Management
Prof Martin Grunow

Technische Universitt Mnchen

Practice
Now let's look at an another example of a
hypothesis test.
Please look at the next "Practice" module
before continuing with the lecture.

Holly Ott