1

In-class exercises for Unit Three

note: Use the p-value approach to hypothesis testing as the default approach in any hypothesis
testing scenario encountered below. Also, state all conclusions in the context of the problem (using
words, not symbols).

1. A company claims that its Model Z dishwasher uses only 3 gallons of water per normal cycle. An
independent agency tested a random sample of 16 Model Z dishwashers, for which the mean water
usage in a normal cycle was 3.3 gallons. Assume that the water usage levels of Model Z
dishwashers in a normal cycle are approximately normally distributed with a standard deviation
of .2 gallons. (a) Does the sample data provide evidence that the mean water usage of Model Z
dishwashers in a normal cycle is greater than 3 gallons? (b) Would the null hypothesis be rejected
at the .01 significance level (i.e., at α = .01)?

2. A company that manufactures ceramic vases periodically monitors its production processes. For its
Simplicity ceramic vases, a mean shatterstrength of at least 100 psi is desired. For a random sample
of 15 Simplicity ceramic vases produced over the past hour, the mean shatterstrength was 98 psi.
Assume that the current process is producing vases with approximately normally distributed
shatterstrengths having a standard deviation of 7 psi. (a) Does this sample data provide evidence
that its current process is producing Simplicity ceramic vases with a mean shatterstrength below
100 psi? (b) Would the null hypothesis be rejected at the .05 significance level?

3. A wholesaler sells “80-ounce” jars of olive oil to restaurant chains. As part of a quality control check,
an employee of the wholesaler randomly selected 20 jars of olive oil filled within the past hour and
determined that their mean content of olive oil was 80.2 ounces with a standard deviation of .7 ounces.
Assume approximately normally distributed content levels with a standard deviation of .6 ounces. (a)
Does this sample provide evidence that the mean content of the jars of olive oil being filled with the
current process is not 80.0 ounces?
We conclude that H0 may be true: ie the mean content of jars IS NOT 80 ounces.
(b) Would the null hypothesis be rejected at α = .10?

p-value ≤  ?
.1362 ≤ .10? NO, DO NOT reject H0. At the .10 significance level the mean jar content IS 80 ounces.
 2

4. The manager of a fast food establishment has set as a standard that the mean service time for drive-
through customers be at most 120 seconds. For a random sample of 51 drive-through customers
over the past month, the mean service time was 125 seconds with a standard deviation of 21
seconds. (a) Does the sample data provide evidence that the mean service time for drive-through
customers is greater than 120 seconds (i.e., that the standard is not being met)? (b) Would the null
hypothesis be rejected at α = .01?

a) Conclusion: There is STRONG evidence that Ha is true; ie, that the mean drive-thru time is
GREATER than 120 seconds.

b) p-value ≤ ?

.048 ≤ .01? ,NO

At the .01 significance level DO NOT reject Ho; ie the mean drive-thru time may be at most 120
seconds.

5. A production manager would want to switch to a new procedure for assembling leather sofas if it
would result in a mean assembly time per sofa of less than 20 minutes. Assume that the assembly
times using the new procedure would be approximately normally distributed. The new procedure is
put in place on a trial basis. For 25 sofas made with the new procedure, the mean assembly time is
19.6 minutes with a standard deviation of 1.2 minutes. (Treat those 25 sofas as a random sample of
sofas that could be made with the new procedure.) (a) Does the trial data suggest the mean
assembly time with the new procedures would be less than 20 minutes? (b) Would the null
hypothesis be rejected at the .05 significance level?

a) There is SOME evidence that Ha is true; that the mean assembly time IS less than 20 minutes.
b) p-value: .054290 <= .05? NO,
c) Do we reject Ho? No, DO NOT REJECT Ho; At alpha = .05 the new assembly time MAY NOT be
less than 20 minutes.

IN T.I. CALCULATOR : STAT -> TESTS – 2: T-TEST, PLUG IN STATS AND CHOOSE THE
APPROPRIATE HYPOTHESIS AND THEN HIT CALCULATE.
 3
6. A Honda auto dealer bases its labor charge for a 36000-mile maintenance check-up on a Honda
Accord on the assumption that the mean time to perform such a check-up is 40 minutes. For a
random sample of 23 Honda Accords receiving that check-up last quarter, the mean time to perform
the check-up was 38 minutes with a standard deviation of 5 minutes. Assume that the time to
perform a 36000-mile maintenance check-up on a Honda Accord is approximately normally
distributed. (a) Does the sample data suggest the mean time to perform such a check-up is not 40
minutes? (b) Would the null hypothesis be rejected at the .01 significance level?

a) There is SOME evidence that Ha is true: the mean Accord check-up time IS NOT 40 minutes.

b) .068 <= .01? NO, DO NOT REJECT THE Ho; the mean check-up time MAY be equal to 40
minutes at the .01 significance level.
7. A medical records transcription company has developed a new coding system intended to reduce
the percentage of transactions recorded incorrectly, which under the current coding system is 4%.
For a random sample of 1100 transactions recorded using the new coding system, 19 were recorded
incorrectly. (a) Does this sample data suggest that the new coding system would be superior to the
current coding system? (b) Would the null hypothesis be rejected at α = .01?

a) Conclusion: We have EXTREMELY STRONG evidence that the new coding system would be
superior to the current coding system.
b) .0001 <= .01?, YES, reject Ho at significance level of .01; ie the new coding system would be
superior to the current coding system at the .01 significance level.
 4

8. A legislator is considering proposing a bill to pay tuition and room & board expenses for
undergraduates who major in math or science education and commit to teach in their major area in
the public school system for at least 5 years. A survey of 920 randomly selected college-bound
high school seniors was conducted; 103 of the seniors surveyed indicated they would be highly
likely to accept such an offer. (a) Does this sample data provide evidence that more than 10% of
college-bound high school seniors would indicate a high likelihood of accepting such an offer? (b)
Would the null hypothesis be rejected at α = .10?

a) There is evidence that Ho may be true, ie. that NO MORE THAN 10% of college-bound high school
seniors would indicate a high likelihood of accepting such an offer.
b) .11 <= .10? NO, DO NOT reject Ho ie., NO MORE THAN 10% of college-bound high school
seniors would indicate a high likelihood of accepting such an offer at the .10 significance level.

9. A national real estate association estimated that 15% of first home buyers in 2010 had (at the time
of purchase) a family income below $40,000. To test that claim, a survey was conducted. Of a
random sample of 810 first home buyers in 2010, 127 had a family income below $40,000. (a)
Does the sample data suggest the percentage is other than 15%? (b) Would the null hypothesis be
rejected at α = .05?

a) There may be evidence that Ho is true; ie, the percentage of first home buyers with incomes below
$40,000 may be equal to 15%.
b) .58 <= .05?, NO, DO NOT reject Ho, ie the percentage of first home buyers with incomes below
$40,000 may be equal to 15% at the .05 significance level.

10. A polygraph (lie detector) is an instrument used to determine if an individual is telling the truth.
These tests are considered to be 95% reliable. In other words, if an individual lies, there is a 0.95
probability that the test will detect a lie. Let there also be a 0.005 probability that the test
erroneously detects a lie even when the individual is actually telling the truth. Consider the null
hypothesis, “the individual is telling the truth,” to answer the following questions.
 5
a) What is the probability of a Type I error (alpha = as given value)?
Type I error: you reject a true H0
Answer:
b) What is the probability of a Type II error (Beta = 1- bigger value)?
Type II error: you accepting a false H0
Answer:
c) What are the consequences of Type I and Type II errors?
Ho: Individual is telling the truth
Ha: Individual is lying
Type 1 error (rejecting a true Ho)
Consequence:
Type 2 error (accepting a false Ho)
Consequence:
d) What is wrong with the statement, “I can prove that the individual is telling the truth on
the basis of the polygraph result”? There is a possibility of a Type 1 or 2 error.

11. The screening process for detecting a rare disease is not perfect. Researchers have developed a
blood test that is considered fairly reliable. It gives a positive reaction in 98% of the people who
have that disease. However, it erroneously gives a positive reaction in 3% of the people who do not
have the disease. Consider the null hypothesis “the individual does not have the disease” to answer
the following questions.

a) What is the probability of a Type I error?

b) What is the probability of a Type II error?
c) What are the consequences of Type I and Type II errors?
Ho:
Ha:
Type I: rejecting a true Null
Consequence:
Type II: accepting a false Null.
Consequence:
d) What is wrong with the statement, “The blood test result has proved the individual is free
of disease”?

12. An economist wanted to compare the starting salaries of recently graduated (with a baccalaureate
degree) accounting and marketing majors employed in Georgia. For a random sample of 50
recently graduated marketing majors, the mean starting annual salary was $40 K with a standard
deviation of $3 K; for an independent random sample of 45 recently graduated accounting majors,
the mean starting annual salary was $55 K with a standard deviation of $8 K. (a) Does the sample
data provide evidence that the mean starting salary of recently graduated marketing majors is less
than the mean starting salary of recently graduated accounting majors? (b) Would the null
hypothesis be rejected at the .05 significance level?
 6
13. As part of a study to assess the benefits of exercise to senior citizens, a health insurance company
determined the number of medications taken by independent random samples of senior citizens
with low and high exercise levels, respectively. The sample data is summarized below, and is
consistent with the assumption of approximately normally distributed medication totals for each
category of seniors.
Number of medications
Sample Sample size
mean variance
Seniors with low exercise levels 15 5.6 2.5
Seniors with high exercise levels 19 3.1 .8
(a) Does the sample data provide evidence that the mean number of medications taken by senior
citizens with low exercise levels is greater than the mean number taken by senior citizens with high
exercise levels? (b) Would the null hypothesis be rejected at the .01 significance level?

14. The purchasing manager of a large chain of retail department stores initiated a study to compare
the breaking strengths of two varieties of suspension rods suitable for hanging clothes, one made by
supplier A and the other made by supplier B. For a sample of 24 supplier A rods, the mean
breaking strength was 620 pounds and the standard deviation in the breaking strengths was 25
pounds. For an independent sample of 16 supplier B rods, the mean breaking strength was 570
pounds and the standard deviation in the breaking strengths was 30 pounds. Assume approximately
normally distributed breaking strengths for each supplier’s rods. (a) Does the sample data provide
evidence that the mean breaking strengths of the two varieties of rods are different? (b) Would the
null hypothesis be rejected at α = .05?

15. A stock market investor wished to assess whether companies, on average, experienced a change in
their P/E (price to earnings) ratio between year 1 and year 2. The investor selected a random
sample of 35 companies from the Handbook of Common Stocks and determined for each company
the difference between its year 1 and year 2 P/E, obtaining a mean difference of -4.98 and a
standard deviation in the differences of 18.75 [adapted from Black (2012)]. (a) Does the sample
data provide evidence that, on average, companies experienced a change in their P/E ratio between
year 1 and year 2? (b) Would the null hypothesis be rejected at α = .05?

16. A random sample of 650 beer drinkers was cross-classified by gender and beer preference (light
versus dark versus regular) as follows:
light regular dark
male 80 100 170
female 175 80 45
(a) Does this sample data provide evidence that, for beer drinkers, gender and beer preference are
related? (b) Would the null hypothesis be rejected at α = .05?

17. The table below shows the results of cross-classifying a random sample of 436 auto assembly line
employees by union membership (yes or no) and job satisfaction level (low, moderate, or high).
 7

Job Satisfaction Level

Union Low Moderate High

p
membershi
Yes 28 76 112 216

No 65 71 84 220
93 147 196 436
(a) Does this sample data provide evidence that, for auto assembly line employees, union membership
and job satisfaction level are related? (b) Would the null hypothesis be rejected at α = .01?

18. Below are X = # hours worked per week last term, and Y = the GPA last term, of a sample of 8
juniors signed up for 12 credit hours last term at a large public university. It can be readily
x=25, s ≈11.95, y=3.1, and s ≈.42
x y
determined that .
x 30 10 40 30 20 40 20 10
y 2.8 3.7 2. 3.3 3.5 3.2 2.8 3.1
4
(a) construct a scatter diagram of GPA versus hours worked per week; (b) determine the covariance
of/between hours worked per week and GPA; and (c) determine and interpret the correlation
coefficient of/between hours worked per week and GPA.

For problems 19-25 below, rely on the specified online data sources, and incorporate into your
answers as needed the appropriate units (e.g., billions of dollars, %).

19. (www.census.gov) What was the estimated median household income in the Atlanta-Sandy
Springs-Marietta MSA (metropolitan statistical area) in 2012?
20. (www.census.gov) As of 2012, what estimated percentage of the Georgia population 25 years
and older had a bachelor’s degree or higher?
21. (www.worldbank.org) In 2010, how many passenger cars per 1000 people were there:
(a) in China?
(b) in the U.S.?
22. (www.bea.gov) In the year 2012, what were the total U.S. personal consumption expenditures
(in 2011 dollars) on:
(a) motor vehicles and parts?
(b) health care?
23. (www.stlouisfed.org) What is the secondary market interest rate for a 1-Year (U.S.) Treasury
bill purchased on March 7, 2014?
24. (www.sec.gov) What was the (unaudited) net income of American Eagle Outfitters Inc for:
(a) the thirteen weeks ending October 30, 2010?
(b) the thirteen weeks sending October 31, 2009?
 8
25. (www.bls.gov) For May, 2014, in the Atlanta-Sandy Springs-Marietta MSA, what are the
estimates of:
(a) the size of the civilian labor force (16 years and over)?
(b) the unemployment rate?