Chapter 9: Introduction to Hypothesis Testing
Chapter 9: Introduction to Hypothesis Testing
Chapter 9: Introduction to Hypothesis Testing
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
422 CHAPTER 9 • INTRODUCTION TO HYPOTHESIS TESTING<br />
A random sample of 50 accounts revealed an average<br />
additional payment of $114 with a sample standard<br />
deviation of $50.<br />
a. <strong>Testing</strong> at a significance level of 0.10, do the<br />
sample data refute the direc<strong>to</strong>r’s claim?<br />
b. Determine the largest sample mean (with the<br />
same sample size and standard deviation) that<br />
would fail <strong>to</strong> refute the direc<strong>to</strong>r’s claim.<br />
9-80. Bowman Electronics sells electronic components<br />
for car stereos. It claims that the average life of a<br />
component exceeds 4,000 hours. To test this claim,<br />
it has selected a random sample of n 12 of the<br />
components and traced the life between installation<br />
and failure. The following data were obtained:<br />
1,973 4,838 3,805 4,494 4,738 5,249<br />
4,459 4,098 4,722 5,894 3,322 4,800<br />
a. State the appropriate null and alternative<br />
hypotheses.<br />
b. Assuming that the test is <strong>to</strong> be conducted using<br />
a 0.05 level of significance, what conclusion<br />
should be reached based on these sample data?<br />
Be sure <strong>to</strong> examine the required normality<br />
assumption.<br />
9-81. The college of business at a state university has a<br />
computer literacy requirement for all graduates:<br />
Students must show proficiency with a computer<br />
spreadsheet software package and with a wordprocessing<br />
software package. To assess whether<br />
students are computer literate, a test is given at the<br />
end of each semester. The test is designed so that at<br />
least 70% of all students who have taken a special<br />
microcomputer course will pass. Suppose that, in a<br />
random sample of 100 students who have recently<br />
finished the microcomputer course, 63 pass the<br />
proficiency test.<br />
a. Using a significance level of 0.05, what conclusions<br />
should the administra<strong>to</strong>rs make regarding<br />
the difficulty of the test?<br />
b. Describe a Type II error in the context of this<br />
problem.<br />
9-82. In a March 2002 article, Golf Digest reported on<br />
a survey conducted involving 300 golfers in which<br />
the respondents were asked their views about the<br />
impact of new technologies on the game of golf.<br />
Before the study, a group of United States<br />
Golf Association officials believed that less than<br />
50% of golfers believed that professional golfers<br />
should have different equipment rules than amateurs.<br />
The survey conducted by the Golf Digest<br />
found 67% did not favor different equipment<br />
rules.<br />
a. If the claim made by the USGA is <strong>to</strong> be<br />
tested, what should the null and alternative<br />
hypotheses be?<br />
b. Based on the sample data, and an alpha level<br />
equal <strong>to</strong> 0.05, use the p-value approach <strong>to</strong> conduct<br />
the hypothesis test.<br />
9-83. USA Today reports (Darryl Haralson, “It’s all about<br />
Overs<strong>to</strong>ck.com,” July 25, 2005) on an ad for<br />
Overs<strong>to</strong>ck.com which sells discounted merchandize<br />
on its Web site. To evaluate the effectiveness of the<br />
ads, Harris Interactive conducted a nationwide poll<br />
of 883 adults. Of the 883 adults, 168 thought the<br />
ads were very effective. This was compared <strong>to</strong> the<br />
Harris Ad Track average of 21%.<br />
a. Determine if the sample size is large enough for<br />
the test in part b <strong>to</strong> warrant approximating the<br />
sample proportion’s distribution with a normal<br />
distribution.<br />
b. Does the Harris poll provide evidence <strong>to</strong><br />
contend that the proportion of adults who find<br />
Overs<strong>to</strong>ck.com’s ads <strong>to</strong> be very effective is<br />
smaller than the Harris Ad Track average?<br />
Use a significance level of 0.05.<br />
9-84. ACNielsen is a New York-based leading global<br />
provider of marketing research information services,<br />
analytical systems and <strong>to</strong>ols, and professional<br />
client service. A recent issue of its magazine<br />
(Todd Hale, “Winning Retail Strategies Start with<br />
High Value Consumers,” Consumer Insight, Spring<br />
2005) addressed, in part, consumers attitudes <strong>to</strong><br />
self-checkout lines. Of the 17,346 EDLP (every<br />
day low price) shoppers, only 3,470 indicated an<br />
interest in this service. If Wal-Mart’s CEO, Lee<br />
Scott, had decided not <strong>to</strong> install self-checkout lines<br />
unless consumer interest was more than 17.5%,<br />
would he order the installation?<br />
a. Determine if the sample size for the test indicated<br />
is large enough <strong>to</strong> warrant approximating the<br />
sample proportion’s distribution with a normal<br />
distribution.<br />
b. Use a significance level of 0.05 and the p-value<br />
approach <strong>to</strong> answer the question put forward<br />
above.<br />
9-85. The Arrow Tire and Rubber Company plans <strong>to</strong><br />
warranty its new mountain bike tire for 12 months.<br />
However, before it does this, the company wants <strong>to</strong><br />
be sure that the mean lifetime of the tires is at least<br />
18 months under normal operations. It will put the<br />
warranty in place unless the sample data strongly<br />
suggest that the mean lifetime of the tires is less<br />
than 18 months. The company plans <strong>to</strong> test this statistically<br />
using a random sample of tires. The test<br />
will be conducted using an alpha level of 0.03.<br />
a. If the population mean is actually 16.5 months,<br />
determine the probability the hypothesis test<br />
will lead <strong>to</strong> incorrectly accepting the null<br />
hypothesis. Assume that the population standard<br />
deviation is known <strong>to</strong> be 2.4 months and the<br />
sample size is 60.