Chapter 9: Introduction to Hypothesis Testing
Chapter 9: Introduction to Hypothesis Testing
Chapter 9: Introduction to Hypothesis Testing
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
412 CHAPTER 9 • INTRODUCTION TO HYPOTHESIS TESTING<br />
error. To find beta for this test procedure under these conditions, the engineers can use the<br />
following steps:<br />
Step 1 Specify the null and alternative hypotheses.<br />
The null and alternative hypotheses are<br />
H 0<br />
: 2.25 (status quo)<br />
H A<br />
: 2.25<br />
Step 2 Specify the significance level.<br />
The two-tailed hypothesis test will be conducted using α 0.05.<br />
Step 3 Determine the critical values, z (/2)L<br />
and z (/2)U<br />
, from the standard<br />
normal distribution.<br />
The critical value from the standard normal is z (α/2)L<br />
and<br />
z (α/2)U<br />
z 0.025<br />
1.96.<br />
Step 4 Calculate the critical values.<br />
x<br />
<br />
.<br />
,<br />
z0. 025<br />
225 . 196 .<br />
0 005 → xL2. 2478;<br />
x U<br />
2.<br />
2522<br />
n<br />
20<br />
LU<br />
Thus, the null hypothesis will be rejected if x 2.2478 or x 2.2522<br />
Step 5 Specify the stipulated value of .<br />
The stipulated value of is 2.255.<br />
Step 6 Compute the z-values based on the stipulated population mean.<br />
The z-values based on the stipulated population mean is<br />
xL<br />
z 2. 24782.<br />
255<br />
xU<br />
<br />
644<br />
. and z 2. 25222.<br />
255<br />
<br />
250<br />
.<br />
0.<br />
005<br />
0.<br />
005<br />
n 20<br />
n 20<br />
Step 7 Determine beta and reach a conclusion.<br />
Beta is the probability from the standard normal distribution between<br />
z 6.44 and z 2.50. From the standard normal table, we get<br />
(0.5000 0.5000) (0.5000 0.4938) 0.0062<br />
Thus, beta 0.0062. There is a very small chance (only 0.0062) that<br />
this hypothesis test will fail <strong>to</strong> detect that the mean diameter has shifted<br />
<strong>to</strong> 2.255 inches from desired mean of 2.25 inches. This low beta will<br />
give the engineers confidence that their test can detect problems when<br />
they occur.<br />
As shown in Section 9.2, many business applications will involve hypotheses tests<br />
about population proportions rather than population means. Example 9-12 illustrates the<br />
steps needed <strong>to</strong> compute the beta for a hypothesis test involving proportions.