Exercises – Two sample t-test

Cruise Ship Ratings. Condé Nast Traveler conducts an annual survey in which
readers rate their favourite cruise ship. All ships are rated on a 100-point scale,
with higher values indicating better service. A sample of 37 ships that carry fewer
than 500 passengers resulted in an average rating of 85.36, and a sample of 44
ships that carry 500 or more passengers provided an average rating of 81.40.
Assume that the population standard deviation is 4.55 for ships that carry fewer
than 500 passengers and 3.97 for ships that carry 500 or more passengers.
1. What is the point estimate of the difference between the population mean
rating for ships that carry fewer than 500 passengers and the population
mean rating for ships that carry 500 or more passengers?
2. At 95% confidence, what is the margin of error?
3. What is a 95% confidence interval estimate of the difference between the
population mean ratings for the two sizes of ships?

Increases in Customer Satisfaction. Will improving customer service result in

higher stock prices for the companies providing the better service? “When a
company’s satisfaction score has improved over the prior year’s results and is
above the national average (75.7), studies show its shares have a good chance of
outperforming the broad stock market in the long run.” The following satisfaction
scores of three companies for the 4th quarters of two previous years were obtained
from the American Customer Satisfaction Index. Assume that the scores are based
on a poll of 60 customers from each company. Because the polling has been done
for several years, the standard deviation can be assumed to equal 6 points in each
case.

Company Year 1 Year 2

Rite Aid 73 76
Expedia 75 77
J.C. Penney 77 78

➢ For Rite Aid, is the increase in the satisfaction score from year 1 to year
2 statistically significant? Use alpha = 0.05, What can you conclude?
 ➢ Can you conclude that the year 2 score for Rite Aid is above the national
average of 75.7? Use alpha = 0.05.
➢ For Expedia, is the increase from year 1 to year 2 statistically
significant? Use alpha = 0.05.
➢ When conducting a hypothesis test with the values given for the
standard deviation, sample size, and α, how large must the increase
from year 1 to year 2 be for it to be statistically significant?
➢ Use the result of part (d) to state whether the increase for J.C. Penney
from year 1 to year 2 is statistically significant.

Miles Driven Per Day. The U.S. Department of Transportation provides the
number of miles that residents of the 75 largest metropolitan areas travel per day
in a car. Suppose that for a simple random sample of 50 Buffalo residents the
mean is 22.5 miles a day and the standard deviation is 8.4 miles a day, and for an
independent simple random sample of 40 Boston residents the mean is 18.6 miles
a day and the standard deviation is 7.4 miles a day.
➢ What is the point estimate of the difference between the mean number
of miles that Buffalo residents travel per day and the mean number of
miles that Boston residents travel per day?
➢ What is the 95% confidence interval for the difference between the
two population means?

Comparing Financial Consultant Ratings. Periodically, Merrill Lynch

customers are asked to evaluate Merrill Lynch financial consultants and services.
Higher ratings on the client satisfaction survey indicate better service, with 7 the
maximum service rating. Independent samples of service ratings for two financial
consultants are summarized here. Consultant A has 10 years of experience,
whereas consultant B has 1 year of experience. Use alpha=0.5 and test to see
whether the consultant with more experience has the higher population mean
service rating.
 ➢ State the null and alternative hypotheses.
➢ Compute the value of the test statistic.
➢ What is the p-value?
➢ What is your conclusion?

Television Commercials and Product Purchase Potential. A market research

firm used a sample of individuals to rate the purchase potential of a particular
product before and after the individuals saw a new television commercial about
the product. The purchase potential ratings were based on a 0 to 10 scale, with
higher values indicating a higher purchase potential. The null hypothesis stated
that the mean rating “after” would be less than or equal to the mean rating
“before.” Rejection of this hypothesis would show that the commercial improved
the mean purchase potential rating. Use α = 0.05 and the following data to test
the hypothesis and comment on the value of the commercial.
Purchase Rating
Individual After Before
1 6 5
2 6 4
3 7 7
4 4 3
5 3 5
6 9 8
7 7 5
8 6 6
Credit Card Expenditures. Bank of America’s Consumer Spending Survey
collected data on annual credit card charges in seven different categories of
expenditures: transportation, groceries, dining out, household expenses, home
furnishings, apparel, and entertainment. Using data from a sample of 42 credit
card accounts, assume that each account was used to identify the annual credit
card charges for groceries (population 1) and the annual credit card charges for
dining out (population 2). Using the difference data, the sample mean difference
was 𝑑̅ = $850, and the sample standard deviation was Sd= $1123.
➢ Formulate the null and alternative hypotheses to test for no difference
between the population mean credit card charges for groceries and the
population mean credit card charges for dining out.
➢ Use a .05 level of significance. Can you conclude that the population means
differ? What is the p-value?
➢ Which category, groceries or dining out, has a higher population mean
annual credit card charge? What is the point estimate of the difference
between the population means? What is the 95% confidence interval
estimate of the difference between the population means?

Impact of Pinterest on Purchase Decisions. Forbes reports that women trust

recommendations from Pinterest more than recommendations from any other
social network platform (Forbes.com). But does trust in Pinterest differ by
gender? The following sample data show the number of women and men who
stated in a recent sample that they trust recommendations made on Pinterest.

Women Men
Sample 150 170
Trust Recommendations Made on 117 102
Pinterest

➢ What is the point estimate of the proportion of women who trust

recommendations made on Pinterest?
➢ What is the point estimate of the proportion of men who trust
recommendations made on Pinterest?
 ➢ Provide a 95% confidence interval estimate of the difference between
the proportion of women and men who trust recommendations made
on Pinterest.

Voter Turnout. Minnesota had the highest turnout rate of any state for the 2016
presidential election (United States Election Project website). Political analysts
wonder if turnout in rural Minnesota was higher than turnout in the urban areas
of the state. A sample shows that 663 of 884 registered voters from rural
Minnesota voted in the 2016 presidential election, while 414 out of 575 registered
voters from urban Minnesota voted.
➢ Formulate the null and alternative hypotheses that can be used to test
whether registered voters in rural Minnesota were more likely than
registered voters in urban Minnesota to vote in the 2016 presidential
election.
➢ What is the proportion of sampled registered voters in rural Minnesota
that voted in the 2016 presidential election?
➢ What is the proportion of sampled registered voters in urban
Minnesota that voted in the 2016 presidential election?
➢ At α=0.05, test the political analysts’ hypothesis. What is the p-value,
and what conclusion do you draw from your results?

Comparing Risk of Mutual Funds. Investors commonly use the standard

deviation of the monthly percentage return for a mutual fund as a measure of the
risk for the fund; in such cases, a fund that has a larger standard deviation is
considered more risky than a fund with a lower standard deviation. The standard
deviation for the American Century Equity Growth fund and the standard
deviation for the Fidelity Growth Discovery fund were recently reported to be
15.0% and 18.9%, respectively. Assume that each of these standard deviations is
based on a sample of 60 months of returns. Do the sample results support the
conclusion that the Fidelity fund has a larger population variance than the
American Century fund? Which fund is more risky?
Repair Costs as Automobiles Age. In its 2016 Auto Reliability
Survey, Consumer Reports asked subscribers to report their maintenance and
repair costs. Most individuals are aware of the fact that the average annual repair
cost for an automobile depends on the age of the automobile. A researcher is
interested in finding out whether the variance of the annual repair costs also
increases with the age of the automobile. A sample of 26 automobiles 4 years old
showed a sample standard deviation for annual repair costs of $170 and a sample
of 25 automobiles 2 years old showed a sample standard deviation for annual
repair costs of $100.
➢ State the null and alternative versions of the research hypothesis that the
variance in annual repair costs is larger for the older automobiles.
➢ At a .01 level of significance, what is your conclusion? What is the p-
value? Discuss the reasonableness of your findings.

Variance in Fund Amounts: Merrill Lynch versus Morgan Stanley. Barron’s

has collected data on the top 1000 financial advisers. Merrill Lynch and Morgan
Stanley have many of their advisers on this list. A sample of 16 of the Merrill
Lynch advisers and 10 of the Morgan Stanley advisers showed that the advisers
managed many very large accounts with a large variance in the total amount of
funds managed. The standard deviation of the amount managed by the Merrill
Lynch advisers was s1= $587. The standard deviation of the amount managed by
the Morgan Stanley advisers was s2= $489. Conduct a hypothesis test at alpha
equal to 0.05 to determine if there is a significant difference in the population
variances for the amounts managed by the two companies. What is your
conclusion about the variability in the amount of funds managed by advisers from
the two firms?
Solutions

Cruise Ship Ratings.

a. 1 = population mean for smaller cruise ships

2 = population mean for larger cruise ships
x1 − x2 = 85.36 – 81.40 = 3.96

 12  22
z.025 +
n1 n2
b.
(4.55) 2 (3.97) 2
1.96 + = 1.88
37 44

c. 3.96 ± 1.88 (2.08 to 5.84)

Increases in Customer Satisfaction.

a. This is an upper-tail hypothesis test.

H0: 1 − 2  0

Ha: 1 − 2  0

( x1 − x2 ) (76 − 73)
z= = = 2.74
 2
 2
62 62
1
+ 2
+
n1 n2 60 60

p-value = area in upper tail at z = 2.74

p-value = 1.0000 - .9969 = .0031

Because .0031  α = .05, we reject the null hypothesis. The difference is

significant. We can conclude that customer service has improved for Rite Aid.

b. This is another upper-tail test, but it only involves one population.

H0:   75.7
Ha:   75.7

( x1 − 75.7 ) (76 − 75.7)

z= = = .39
 2
62
n1 60
 p-value = area in upper tail at z = .39

p-value = 1.0000 – .6517 = .3483

Because .3483 > α = .05, we cannot reject the null hypothesis. The
difference is not statistically significant.

c. This is an upper-tail test similar to the one in part a.

H0: 1 − 2  0
Ha: 1 − 2  0

( x1 − x2 ) (77 − 75)
z= = = 1.83
 2
 2
62 62
1
+ 2
+
n1 n2 60 60

p-value = area in upper tail at z = 1.83

p-value = 1.0000 – .9664 = .0336

Because .0336  α = .05, we reject the null hypothesis. The difference is

significant. We can conclude that customer service has improved for Expedia.

d. We will reject the null hypothesis of “no increase” if the p-value ≤ .05. For an
upper-tail hypothesis test, the p-value is the area in the upper tail at the value of
the test statistic. A value of z = 1.645 provides an upper-tail area of .05. So, we
must solve the following equation for x1 − x2 .

( x1 − x2 )
z= = 1.645
62 62
+
60 60

62 62
x1 − x2 = 1.645 + = 1.80
60 60

This tells us that as long as the year 2 score for a company exceeds the year 1
score by 1.80 or more the difference will be statistically significant.

e. The increase from year 1 to year 2 for J.C. Penney is not statistically significant
because it is less than 1.80. We cannot conclude that customer service has
improved for J.C. Penney.
Miles Driven Per Day.

a. x1 − x2 = 22.5 – 18.6 = 3.9

2 2
 s12 s22   8.42 7.42 
 +   + 
df =  n1 n2 
=  50 40 
= 87.1
2 2 2 2
1  s12  1  s22  1  8.42  1  7.42 
  +     +  
n1 − 1  n1  n2 − 1  n2  49  50  39  40 
b.

Use df = 87, t.025 = 1.988

8.42 7.42
3.9  1.988 +
50 40
3.9 + 3.3 (.6 to 7.2)

Comparing Financial Consultant Ratings.

a. H0: 1 − 2  0
Ha: 1 − 2  0

( x1 − x2 ) − D0 (6.82 − 6.25) − 0
t= = = 1.99
2 2
s s .642 .752
1
+ 2
+
n1 n2 16 10
b.
2 2
 s12 s22   .642 .752 
 +   + 
df =  n1 n2 
=  16 10 
= 16.9
c. 2 2 2 2
1  s12  1  s22  1  .642  1  .752 
  +     +  
n1 − 1  n1  n2 − 1  n2  15  16  9  10 

Use df = 16

Using t table, p-value is between .025 and .05.

Exact p-value corresponding to t = 1.99 is .0320.

d. p-value  .05, reject H0. The consultant with more experience has a higher
population mean rating.

Television Commercials and Product Purchase Potential.

Difference = rating after – rating before

H0: μd ≤ 0
 Ha: μd > 0

d = .625 and sd = 1.30

d − d .625 − 0
t= = = 1.36
sd / n 1.30 / 8
df = n – 1 = 7

Using t table, p-value is between .10 and .20.

Exact p-value corresponding to t = 1.36 is .1080.

Do not reject H0; we cannot conclude that seeing the commercial improves
the mean potential to purchase.

∑ 𝑑 −114.06
𝑑̅ = 𝑖𝑛 𝑖 = 25 = −4.56

Credit Card Expenditures.

a. μ1 = population mean grocery expenditures

μ2 = population mean dining-out expenditures
H0: d = 0
Ha: d  0

d − d 850 − 0
t= = = 4.91
sd / n 1123 / 42
b.
df = n – 1 = 41
p-value  0

Conclude that there is a difference between the annual population mean

expenditures for groceries and for dining out.

c. Groceries has the higher mean annual expenditure by an estimated $850.

sd
d  t.025
n

1123
850  2.020
42

850  350 (500 to 1200)

Impact of Pinterest on Purchase Decisions.

a.The point estimate of the proportion of women who trust recommendations made on
Pinterest is p1 = 117/150 = .78.

b. The point estimate of the proportion of men who trust recommendations made on
Pinterest is p2 = 102/170 = .60.

c. p1 − p2 = .78 – .60 = .18

p1 (1 − p1 ) p2 (1 − p2 )
.18  z.025 +
n1 n2
.78(.22) .60(.40)
.18  1.96 +
150 170
.18 ± .0991 (.0809 to .2791)

The 95% confidence interval estimate of the difference between the proportion of
women and men who trust recommendations made on Pinterest is .18 ± .0991 or
(.0809 to .2791).

Voter Turnout.

Let p1 = the population proportion of voters in rural Minnesota voted in the 2016 election
p2 = the population proportion of voters in urban Minnesota voted in the 2016 election

a. H0: p1  p2
Ha: p1  p2

b. p1 = 663/884 = .75 75% of voters in rural Minnesota voted in the 2016

election.

c. p2 = 414/575 = .72 72% of voters in urban Minnesota voted in the 2016

election.

n1 p1 + n2 p2 663 + 414
d. p= = = .7382
n1 + n2 884 + 575
p1 − p2 .75 − .72
z= = = 1.27
1 1   1 1 
p (1 − p )  +  .7382 (1 − .7382 )  + 
 n1 n2   884 575 

Upper-tail p-value is the area to the right of the test statistic.

Using normal table with z = 1.27: p-value = 1 – .8980 = .1020.

p-value > α ; do not reject H 0 .

We cannot conclude that voters from rural Minnesota voted more frequently
 than voters from urban Minnesota in the 2016 presidential election.

Comparing Risk of Mutual Funds.

For this type of hypothesis test, we place the larger variance in the numerator. So the Fidelity
variance is given the subscript of 1.

H 0 : 12   22

H a : 12   22

s12 18.92
F= = = 1.59
s22 15.02

Degrees of freedom in the numerator and denominator are both 59.

Using the F table, p–value is between .05 and .025.

Exact p–value corresponding to F = 1.59 is .0387.

p–value  .05; reject H0. We conclude that the Fidelity fund has a greater
variance than the American Century fund.

Repair Costs as Automobiles Age.

a. Population 1 is four-year-old automobiles.

H 0 : 12   22

H a : 12   22

s12 1702
F= = = 2.89
s22 1002
b.

Degrees of freedom 25 and 24.

Using F table, p–value is less than .01.

Exact p–value corresponding to F = 2.89 is .0057.

p–value  .01; reject H0. Conclude that 4 year old automobiles have a larger
variance in annual repair costs compared to two-year-old automobiles. This is
expected because older automobiles are more likely to have some more expensive
repairs that lead to greater variance in annual repair costs.
Variance in Fund Amounts: Merrill Lynch versus Morgan Stanley.

We place the larger sample variance in the numerator. So, the Merrill Lynch variance is given
the subscript of 1.

H 0 : 12 =  22

H a : 12   22

s12 587 2
F= = = 1.44
s22 4892

Degrees of freedom 15 and 9.

Using F table, area in tail is greater than .10.

Two–tail p–value is greater than .20 .

Exact p–value corresponding to F = 1.44 is .5906.

p–value > .10; do not reject H0. We cannot conclude there is a statistically
significant difference between the variances for the two companies.