Вы находитесь на странице: 1из 17

1) A study was recently conducted to estimate the mean cholesterol for adult males over the age of 55 years.

The
following random sample data were observed:

245 304 135 202 300


196 210 188 256 390

Given this information, what is the point estimate for the population mean?
A) About 73.35
B) ±102
C) About 242.6
D) Can't be determined without knowing the confidence level.
Answer: C

2) The following data represent a random sample of bank balances for a population of checking account customers
at a large eastern bank. Based on these data, what is the 95 percent confidence interval estimate for the true
population mean?

$2,300 $756 $325 $1,457 $208 $2,345


$1,560 $124 $356 $3,179 $457 $789
$120 $2,760 $998 $508 $210 $789

A) Approximately $1,069 ± $484.41


B) About $839.40 to $1,298.60
C) Approximately $1,069 ± 2.1098
D) None of the above
Answer: A

3) Assuming the population of interest is approximately normally distributed, construct a 95% confidence interval
estimate for the population mean given the following values:

A) (16.73, 20.07)
B) (13.22, 23.58)
C) (15.86, 20.94)
D) (14.20, 22.60)
Answer: C

4) Allante Pizza delivers pizzas throughout its local market area at no charge to the customer. However, customers
often tip the driver. The owner is interested in estimating the mean tip income per delivery. To do this, she has
selected a simple random sample of 12 deliveries and has recorded the tips that were received by the drivers. These
data are:

$2.25 $2.50 $2.25 $2.00 $2.00 $1.50


$0.00 $2.00 $1.50 $2.00 $3.00 $1.50

Based on these sample data, what is the best point estimate to use as an estimate of the true mean tip per delivery?
A) 1.875
B) 1.811
C) 1.50
D) 1.312
Answer: A

5) One of the major oil products companies conducted a study recently to estimate the mean gallons of gasoline
purchased by customers per visit to a gasoline station. To do this, a random sample of customers was selected with
the following data being recorded that show the gallons of gasoline purchased.

8.7 22.4 9.5 13.3 18.9


8-1
Copyright © 2014 Pearson Education, Inc.
22 14.4 35.7 19 24.9
5.7 15.7 8.9 22.5 15.9

Based on these sample data, construct and interpret a 95 percent confidence interval estimate for the population
mean.
Answer: Since the population standard deviation is unknown, and since the size of the sample is fairly small, the
format for the confidence interval estimate is:
±t

If we assume that the population distribution is fairly symmetric or normally distributed, the t-distribution can be
used. The first thing we need to do is compute both and s from the sample data. First for the sample mean, we get:

= = = 17.1667

Next, we compute the sample standard deviation using:

s= = 7.77

Next, for 95 percent confidence and degrees of freedom equal to n-1 = 14, the critical value for the t-distribution is
2.1448. Then the confidence interval estimate is:
±t = 17.667 ± 2.1448

17.1667 ± 4.3029

Thus, the 95 percent confidence interval estimate is 12.8638 - 21.4696. We interpret this as follows: "Based on a
random sample of 15 customers selected from the population of interest, with 95 percent confidence we believe that
the true mean purchase per visit to the gas station is between 12.8638 gallons and 21.4696 gallons."

6) A cell phone manufacturer claims that its phone will last for more than 8 hours of continuous talk time when the
battery is fully charged. To test this claim a sample of n = 18 phones were tested. The results showed a sample mean
of 8.2 hours and a sample standard deviation of 0.4 hour. Conduct the hypothesis test using a 0.5 level of
significance and determine whether or not the company's claim is supported.
Answer: Since the claim "more than 8 hours" is a strict inequality, this needs to be the alternate hypothesis so that
that "equal to 8 hours" is included in the null:
H0 : μ ≤ 8
HA : μ > 8
So this will be a one-tailed test with the rejection region in the upper tail. In addition, the t-distribution must be used
since a sample standard deviation is all we have available. There are 18 - 1 = 17 degrees of freedom. So the critical
value for α = 0.05 and 17 d.f. is t = 1.7396 and the decision rule is:
Reject the null if t > 1.7396

The formula for the test statistic is:

t=

Then the test statistic is:

t= = = 2.12

Since the test statistic t = 2.12 is greater than the critical value (1.7396) we should reject the null hypothesis and
conclude that the mean is significantly larger than 8 hours, which means the manufacturer's claim is supported.

8-2
Copyright © 2014 Pearson Education, Inc.
7) If a real estate market is strong there will be a close relationship between the asking price for homes and the
selling price. Suppose that one analyst believes that the mean difference between asking price and selling price for
homes in a particular market area is less than $2,000. To test this using an alpha level equal to 0.05, a random
sample of n = 15 homes that have sold recently was selected. The sample differences between asking price and
selling price are in the following table:

$2,053 $1,693 $1,854 $1,747 $869


$1,396 $2,473 $1,931 $2,303 $1,502
$1,038 $2,755 $2,084 $1,664 $2,104

Based on these sample data, what is the critical value expressed in dollars?
Answer: The analyst claims that the mean difference will be less than $2,000. Given this research hypothesis, we
can form the following null and alternative hypotheses to test statistically:
Ho : μ ≥ $2,000
Ha : μ < $2,000
The test is a one-tail, lower-tail test with 0.05 area in the rejection region. Since we do not know the population
standard deviation and the sample size is relatively small, if we assume that the population is normally distributed,
the t-distribution can be used. The critical t-value for degrees of freedom 15-1 = 14 and 0.05 in one tail is -1.7613.
We find the critical value as follows:
μ–t = 2000 - 1.7613

We compute the sample standard deviation, s, from the data as follows:

s= = $506.59.

Then the critical value is 2000 - 1.7613 = $1,769.62

Thus, if < $1,769.62 reject the null hypothesis; otherwise do not reject. We find as follows:

= = = $1,831.067.

Since $1,831.067 ≥ $1,769.62, we do not reject the null hypothesis.

8) A company makes a device that can be fitted to automobile engines to improve the mileage. The company claims
that if the device is installed, owners will observe a mean increase of more than 3.0 mpg. Assuming that the
population standard deviation of increase is known to be 0.75 mpg, and a sample of size 64 cars is selected with an
= 3.25 mpg, use the p-value approach to test the null hypothesis using a significance level of 0.05.
Answer: The research hypothesis is based on the claim made by the company that mean mileage will increase by
more than 3.0 mpg. Thus, we set up the following null and alternative hypotheses:
H0 : μ ≤ 3.0 mpg
Ha : μ > 3.0 mpg.
The next step in determining the p-value is to compute the test statistic. Since we assume that the population
standard deviation is known, the test statistic will be a z-value computed as follows:

z= = = 2.67

We then go to the standard normal table to find the probability associated with z = 2.68. This value is 0.4962. The p-
value is found by taking 0.5000 - 0.4962 = 0.0038. Since 0.0038 < α = 0.05, we reject the null hypothesis. The
difference between 3.25 mpg and 3.0 is too much to attribute to sampling error alone. This conclusion means that
the manufacturer's claim is supported by the sample data.

9) A company makes a device that can be fitted to automobile engines to improve the mileage. The company claims
that if the device is installed, owners will observe a mean increase of more than 3.0 mpg. Assuming that the
8-3
Copyright © 2014 Pearson Education, Inc.
population standard deviation of increase is known to be 0.75 mpg, and a sample of size 64 cars is selected, what is
the probability of "accepting" the null hypothesis if the true population mean is 3.10 mpg increase? Assume that the
test will be performed using a 0.05 level of significance.
Answer: The objective here is to compute beta, the probability of a Type II error. We begin by formulating the null
and alternative hypotheses:
H0 : μ ≤ 3.0 mpg
Ha : μ > 3.0 mpg
Next, we need to determine the critical value as follows:
μ+z = 3.0 + 1.645 = 3.1542.

The next step in solving for beta is to determine the probability of finding a sample mean of 3.1542 or less given a
true population mean of 3.10 mpg. We do this by finding the test statistic z-value associated with 3.1542 as follows:

z= = = .5781 ≈ .58.

From the standard normal table, the probability for a z-value equal to 0.58 is 0.2190. To find beta, we add 0.5000 to
0.2190, giving 0.7190. Thus, if the true mean is 3.10 mpg, there is a 71.9 percent chance that the hypothesis test will
not detect that and, instead, will lead to not rejecting the null hypothesis that the mean is 3.0 or less.

10) A small city is considering breaking away from the county school system and starting its own city school
system. City leaders believe that more than 60 percent of residents support the idea. A poll of n = 215 residents is
taken and 134 people say they support starting a city school district. Using a 0.10 level of significance, conduct a
hypothesis test to determine whether this poll supports the belief of city leaders.
Answer: Since "more than 60 percent" is a strict inequality, this must go in the alternate hypothesis, so the
hypotheses are:
H0 : p ≤ 0.6
HA : p > 0.6
Since this a proportion, the normal distribution is used, and the test statistic for α = 0.10 is z = 1.28 and the decision
rule is:
Reject the null if Z > 1.28.
The test statistic formula is

z=

And the sample proportion is p = 134/215 = 0.6233.

So the test statistic is z = = = 0.6976

This does not fall in the rejection region since 0.6976 < 1.28, so we do not reject the null. This indicates the
proportion of residents who support a city school district is not significantly greater than 0.6.

11) The Gordon Beverage Company bottles soft drinks using an automatic filling machine. When the process is
running properly, the mean fill is 12 ounces per can. The machine has a known standard deviation of 0.20 ounces.
Each day, the company selects a random sample of 36 cans and measures the volume in each can. They then test to
determine whether the filling process is working properly. The test is conducted using a 0.05 significance level.
What is the critical value in ounces?
Answer: The company will take action if it determines that the fill process is putting too little or too much in the
cans on average. Therefore, the research hypothesis is that the population mean is not equal to 12 ounces. Given this,
the null and alternative hypotheses are:
H0 : μ = 12 ounces
Ha : μ ≠ 12 ounces.
We note that the hypothesis test is two-tailed since we can reject it if the sample mean gets too large or too small.
8-4
Copyright © 2014 Pearson Education, Inc.
Thus, there are actually two critical values, one on each side of the distribution. We find the critical values as
follows:
μ±z

Because this is a two-tailed test, we must split the alpha into two parts of 0.025 each. The z-value from the standard
normal distribution for 0.5000 - 0.025 = 0.475 is 1.96. Given this, the critical values are determined as:
12.0 ± 1.96 or 12.0 ± .0653

Thus, the decision rule is:


If 11.9347 ≤ ≤ 12.0653 do not reject the null hypothesis, otherwise reject.

12) The Gordon Beverage Company bottles soft drinks using an automatic filling machine. When the process is
running properly, the mean fill is 12 ounces per can. The machine has a known standard deviation of 0.20 ounces.
Each day, the company selects a random sample of 36 cans and measures the volume in each can. They then test to
determine whether the filling process is working properly. The test is conducted using a 0.05 significance level.
Using the test statistic approach, what conclusion should the company reach if the sample mean is 12.02 ounces?
What type of statistical error may have been committed?
Answer: The company will take action if it determines that the fill process is putting too little or too much in the
cans on average. Therefore, the research hypothesis is that the population mean is not equal to 12 ounces. Given this,
the null and alternative hypotheses are:
H0 : μ = 12 ounces
Ha : μ ≠ 12 ounces.
We note that the hypothesis test is two-tailed since we can reject if the sample mean gets too large or too small.
The critical z-value from the standard normal table for a two-tailed test with alpha = 0.05 is 1.96, the decision rule is
reject the null if z < -1.96 or z > 1.96. The test statistic is computed as:

z= = = .60

Since 0.60 < 1.96, we do not reject the null hypothesis based on these sample data. So we can conclude that the
filling process does not need adjusting.
Thus, the statistical error that could have been committed is a Type II error since the null hypothesis was not
rejected. If a Type II error was committed this would mean that the filling machine actually needs adjusting, but we
failed to detect it.

8-5
Copyright © 2014 Pearson Education, Inc.
13) A national car rental chain believes that more than 80 percent of its customers are satisfied with the check-in
process that the company is using. To test this, a random sample of n = 200 customers are surveyed. These sample
results show 168 that say they were satisfied. If the test is to be conducted using a .05 level of significance, what is
the critical value?
Answer: This is a hypothesis about a single population proportion. The appropriate null and alternative hypotheses
are:
H0 : p ≤ .80
Ha : p > .80

The critical value for this one-tailed test is found as: p + z

The z-value from the standard normal table for a 0.05 level of significance and one-tailed test is 1.645. Then the
critical value is: .80 + 1.645 = .8465. Thus, if the sample proportion exceeds 0.8465, we reject the

null hypothesis, since = = = .84

Since 0.84 is just slightly less than 0.8465, we do not reject the null hypothesis.

14) A contract with a parts supplier calls for no more than .04 defects in the large shipment of parts. To test whether
the shipment meets the contract, the receiving company has selected a random sample of n = 100 parts and found 6
defects. If the hypothesis test is to be conducted using a significance level equal to 0.05, what is the test statistic and
what conclusion should the company reach based on the sample data?
Answer: This is a hypothesis about a single population proportion. The appropriate null and alternative hypotheses
are:
H0 : p ≤ .04
Ha : p > .04
Note that the research hypothesis is that the shipment contains more than the suggested proportion of defects. The
test statistic when testing about proportions is a z-value computed as:

z= .

For this situation, the calculated z-value is z = = 1.02.

We compare this value to the z-value from the standard normal distribution table for 0.45 which is z = 1.645. The
decision rule is: If z calculated > 1.645, reject the null hypothesis, otherwise do not reject. Since 1.02 < 1.645, we
cannot reject the null hypothesis based on the sample data. Thus, there is insufficient evidence to conclude that the
proportion of defects in the shipment exceeds 0.04.

15) A major U.S. oil company has developed two blends of gasoline. Managers are interested in estimating the
difference in mean gasoline mileage that will be obtained from using the two blends. As part of their study, they
have decided to run a test using the Chevrolet Impala automobile with automatic transmissions. They selected a
random sample of 100 Impalas using Blend 1 and another 100 Impalas using Blend 2. Each car was first emptied of
all the gasoline in its tank and then filled with the designated blend of the new gasoline. The car was then driven 200
miles on a specified route involving both city and highway roads. The cars were then filled and the actual miles per
gallon were recorded. The following summary data were recorded:

Blend 1 Blend 2
Sample Size 100 100
Sample Mean 23.4 mpg 25.7 mpg
Sample St. Dev. 4.0 mpg 4.2 mpg

Based on these sample data, compute and interpret the 95 percent confidence interval estimate for the difference in
8-6
Copyright © 2014 Pearson Education, Inc.
mean mpg for the two blends.
Answer: The general format for a confidence interval estimate is:

Point estimate ± critical value (standard error).

In this case, the point estimate is 1 - 2 = 23.4 - 25.7 = -2.3 mpg.


This value forms the center of the interval estimate. To complete the confidence interval, we need to compute the
standard error. Since the sample sizes in this example are large, we compute the standard error using:

sc = = = 0.58

The critical value for a 95 percent confidence interval based on large samples is z = 1.96 from the standard normal
table. Therefore, the 95 percent confidence interval estimate for the difference in population means is:

1- 2±z . Substituting we get

(23.4 - 25.7) ± 1.96 which gives -2.3 ± 1.96(.58) or -2.3 ± 1.1368.

Thus we get: -3.4368 mpg to -1.1632 mpg.


We interpret this to mean that based on the sample information, with 95 percent confidence, Impalas using Blend 1
will get between 1.1632 and 3.4368 mpg less than Impalas using Blend 2.

16) A major U.S. oil company has developed two blends of gasoline. Managers are interested in determining
whether a difference in mean gasoline mileage will be obtained from using the two blends. As part of their study,
they have decided to run a test using the Chevrolet Impala automobile with automatic transmissions. They selected a
random sample of 100 Impalas using Blend 1 and another 100 Impalas using Blend 2. Each car was first emptied of
all the gasoline in its tank and then filled with the designated blend of the new gasoline. The car was then driven 200
miles on a specified route involving both city and highway roads. The cars were then filled and the actual miles per
gallon were recorded. The following summary data were recorded:

Blend 1 Blend 2
Sample Size 100 100
Sample Mean 23.4 mpg 25.7 mpg
Sample St. Dev. 4.0 mpg 4.2 mpg

Based on the sample data, using a 0.05 level of significance, what conclusion should the company reach about
whether the population mean mpg is the same or different for the two blends? Use the test statistic approach to test
the null hypothesis.
Answer: The hypothesis test involves a test for the difference between population means and is based on large
samples from each population. The appropriate null and alternative hypotheses are:
H0 : μ1 = μ2
Ha : μ 1 ≠ μ 2
This will be a two-tailed test since we are interested in testing whether there is a difference between the two
population means with respect to miles per gallon and neither blend of gasoline is predicted to be superior to the
other.

The critical value for a two-tailed test with a significance level of 0.05 is found in the Standard Normal table to be z
= ±1.96. The test statistic in this case is computed using:

8-7
Copyright © 2014 Pearson Education, Inc.
z= = = = -3.9655

Since z = -3.9655 < -1.96, we reject the null hypothesis and conclude that the population means are not equal. Based
on the samples, we infer that Blend 2 will provide higher mean mileage.

17) A major U.S. oil company has developed two blends of gasoline. Managers are interested in determining
whether a difference in mean gasoline mileage will be obtained from using the two blends. As part of their study,
they have decided to run a test using the Chevrolet Impala automobile with automatic transmissions. They selected a
random sample of 100 Impalas using Blend 1 and another 100 Impalas using Blend 2. Each car was first emptied of
all the gasoline in its tank and then filled with the designated blend of the new gasoline. The car was then driven 200
miles on a specified route involving both city and highway roads. The cars were then filled and the actual miles per
gallon were recorded. The following summary data were recorded:

Blend 1 Blend 2
Sample Size 100 100
Sample Mean 23.4 mpg 25.7 mpg
Sample St. Dev. 4.0 mpg 4.2 mpg

Based on the sample data, using a 0.05 level of significance, what conclusion should the company reach about
whether the population mean mpg is the same or different for the two blends? Use the p-value approach to test the
null hypothesis.
Answer: The hypothesis test involves a test for the difference between population means and is based on large
samples from each population. The appropriate null and alternative hypotheses The appropriate null and alternative
hypotheses are:
H0 : μ1 = μ2
Ha : μ 1 ≠ μ 2
This will be a two-tailed test since we are interested in testing whether there is a difference between the two
population means with respect to miles per gallon and neither blend of gasoline is predicted to be superior to the
other.

To obtain the p-value, we first compute the test statistic. The test statistic in this case is computed using:

z= = = = -3.9655.

From the standard normal table, the probability associated with a z = -3.9655 value is essentially 0.5000. This is the
area between z = 3.9655 and the mean of the standard normal distribution. We want the area in the tail of the
distribution. Thus the p-value for this test is 0.5000 - 0.5000 ≈ 0.0000. This value can be compared with α/2 = 0.025.
Since 0.0000 < 0.025, we reject the null hypothesis and conclude that the population means are not equal.

18) A PC company uses two suppliers for rechargeable batteries for its notebook computers. Two factors are
important quality features of the batteries: mean use time and variation. It is desirable that the mean use time be high
and the variability be low. Recently, the PC maker conducted a test on batteries from the two suppliers. In the test, 9
randomly selected batteries from Supplier 1 were tested and 12 randomly selected batteries from Supplier 2 were
tested. The following results were observed:

Supplier 1 Supplier 2
n1 = 9 n2 = 12
1 = 67.25 min. 2 = 72.4 min.

8-8
Copyright © 2014 Pearson Education, Inc.
S1= 11.2 min S2 = 9.9 min

Based on these sample results, can the PC maker conclude that a difference exists between the two batteries with
respect to the population mean use time? Test using a 0.10 level of significance.
Answer: This test calls for us to test whether the population means are equal. Since the sample sizes are small and
the population variances are unknown, the t-distribution should be employed if we assume that the populations are
normally distributed and the populations have equal variances.

We can test the following null and alternative hypotheses:


H0 : μ1 = μ2
Ha: μ1 ≠ μ2

The critical value for a t-test for a two-tailed test is found from the t-distribution with degrees of freedom equal to n1
+ n2 - 2 = 9 + 12 - 2 = 19. Given a 0.10 significance level, the critical value from the t-distribution with degrees of
freedom equal to 19 is equal to ±1.7291.

The test statistic is computed using:

t= where Sp is the pooled standard deviation and is computed as:

sp = = = 10.4671

Then the test statistic is:

t= = = -1.1158

Since t = -1.1158 > -1.7291, we do not reject the null hypothesis. Thus, based on the sample data, we do not have
evidence to indicate that the two batteries differ in mean use time.

20) A real estate agent believes that home with swimming pools take longer to sell than home without swimming
pools. A random sample of each type of recently sold homes was taken where the number of days on the market is
recorded. Results are:

Homes with pool Homes w/o pool


Sample mean 77 days 65 days
Sample Standard deviation 8.4 days 7.2 days
Sample size 19 23

Assuming that the populations are normally distributed and the variances are equal, conduct the appropriate
hypothesis test to determine if the real estate agent is correct. Use the 0.05 level of significance.
Answer: The agent believes that homes with pools have a higher average number of days on the market. Let μ1 =the
mean days for homes with pools and μ2 = the mean days for homes without pools. Then the hypotheses are:
H0 : μ 1 - μ 2 ≤ 0
Ha : μ 1 - u 2 > 0

Since we have sample standard deviations we need to use the t-distribution with (19+23-2) = 40 degrees of freedom.
This is a one-tailed test with the rejection region of size 0.05 in the upper tail. So the critical value is t = 1.6839 and
we will reject the null if the test statistic exceeds 1.6839.

8-9
Copyright © 2014 Pearson Education, Inc.
The test statistic is computed using:

t= where Sp is the pooled standard deviation and is computed as:

sp = = = 7.763

Then the test statistic is:

t= = = 4.986

Since 4.986 > 1.6839 we reject the null hypothesis and can conclude that the alternate hypothesis is correct, which
says that homes with pools have a significantly higher number of average days on the market. So the real estate
agent is correct.

21) The National Football League (NFL) is interested in testing to see whether there is a difference in the proportion
of male fans that prefer instant replay to review officials' calls and the proportion of female fans that prefer instant
replay. It is believed that males tend to favor the practice to a higher degree than do females. To test this, random
samples of male fans and female fans were selected and the following results were obtained:

Male Fans Female Fans


Sample size 300 100
Number Preferring 234 57

Using a significance level equal to 0.05, what conclusion should be reached based on the sample data?
Answer: The parameter of interest in this situation is the difference in population proportions. The research
hypothesis states that the proportion of males favoring instant replay exceeds the proportion of females who favor it.
Thus, we can formulate the following null and alternative hypotheses:
H0 : p 1 ≤ p 2
Hα: p1 > p2

This means that we are dealing with a one-tailed hypothesis test. The critical value for a one-tailed test for the
difference between population proportions with a 0.05 significance level is z = 1.645 from the standard normal
distribution. The test statistic is computed using:

z=

where = + = = .7275 and

1= = = .78 and 2= = = .57

Then, the test statistic is:

z= = = 4.0846

8-10
Copyright © 2014 Pearson Education, Inc.
Since 4.0846 > 1.645, we reject the null hypothesis and conclude that based on the sample data, the male fans do
favor instant replay by a higher proportion than do female fans.

22) A maker of toothpaste is interested in testing whether the proportion of adults (over age 18) who use its
toothpaste and have no cavities within a six-month period is any different from the proportion of children (18 and
under) who use the toothpaste and have no cavities within a six-month period. To test this, it has selected a sample
of adults and a sample of children randomly from the population of those customers who use their toothpaste. The
following results were observed.

Adults Children
Sample size 100 200
Number with 0 cavities 83 165

Based on these sample data and using a significance level of 0.05, what conclusion should be reached? Use the p-
value approach to conduct the test.
Answer: The parameter of interest in this situation is the difference in population proportions. Since we are
interested in whether there is a difference in the proportion of adults who get cavities versus children, we can
formulate the following null and alternative hypotheses:
H0 : p 1 = p 2
Hα : p 1 ≠ p 2
This means that we are dealing with a two-tailed hypothesis test. The critical value for a two-tailed test for the
difference between population proportions with a 0.05 significance level is z=1.96 from the standard normal
distribution. The test statistic is computed using:

z= where = + = = .8267 and

1= = = .83 and 2= = = .8250

Then, the test statistic is:

z= = = .1079

To find the p-value, we go to the standard normal table for z = 0.11 getting 0.0438. Subtracting 0.0438 from 0.5000
gives 0.4562, which is the p-value. This value should be compared to alpha/2 = 0.025. Since 0.4562 > 0.025, we do
not reject the null hypothesis and conclude that based on the sample data, there is no basis for concluding that the
proportion of adults with no cavities is any different from the proportion of children with no cavities.

23) A market research firm has come up with two coupon designs that could be inserted into the envelopes that go
out with the monthly statement to credit card customers. The coupons offer the customer an opportunity to become a
member of a travel club. The research firm is interested in estimating the difference in proportion of customers who
will join the club after receiving one or the other of the two coupons. To obtain this estimate, the research firm has
sent out coupon design 1 to a random sample of 100 customers. A second random sample of 100 customers received
coupon 2. For the first coupon, 11 customers joined the travel club, while 15 customers who received coupon 2
joined. Based upon this sample information, develop and interpret the desired 95 percent confidence interval
estimate.
Answer: The parameter of interest in this situation is the difference in population proportions. The general format
for any confidence interval estimate is:
Point estimate ± critical value (standard error).

We are given the following sample information:

8-11
Copyright © 2014 Pearson Education, Inc.
1= = = .11 and 2= + = = .15

In this case, the point estimate will be the difference between the two sample proportions which is:
1- 2 = .11 - .15 = -.04.

The standard error for the difference in proportions is:

= = 0.0475

The critical value for an estimate for the difference in population proportions is a z-value from the standard normal
table. In this case, given a 95 percent confidence level, the critical value is z=1.96.

The confidence interval estimate is found as:

1- 2±z

Substituting values, we get:


-.04 ± 1.96

-.04 ± 0.0931.

Thus, the 95 percent confidence interval is:


-0.133 to +0.0531.

This indicates that we estimate the difference in proportions of redeemed coupons to be between -0.1331 and
+0.0531. Thus, the range of possibilities for the difference in population proportions ranges from over 0.13 more for
coupon 2 versus coupon 1 all the way to over 0.05 more for coupon 1 versus coupon 2. Therefore, since the interval
crosses zero, we are unable to conclude that a difference exists between the two coupons with respect to the
proportions that will be redeemed.

24) When estimating the difference between two population means, when should the normal distribution be used and
when should the t-distribution be used?
Answer: If the two standard deviations are known, then the normal distribution should be used. In most real
situations the standard deviations are not know and sample standard deviations must be used. In this case the t-
distribution should be used. However, as the number of degrees of freedom increases, the t-distribution approaches
the normal distribution. So if the sample sizes are large enough the normal distribution can be used because there
will be very little difference between the two. A sample size of n > 30 is often used a rule of thumb for a sufficiently
large sample.

25) There are two major companies that provide SAT test tutoring for high school students. At issue is whether
Company 1, which has been in business for the longer time, provides better results than Company 2, the newer
company. Specifically of interest is whether the mean increase in SAT scores for students who have already taken
the SAT test one time is higher for Company 1 than for Company 2. Two random samples of students are selected.
The following data reflect the number of points higher (or lower) that the students scored on the SAT test after
taking the tutoring. Prior to conducting the test, which compares the means, we should determine if the assumption
of equal variances is supported.

Conduct the appropriate hypothesis test to determine if the assumption of equal variances is supported using a 0.10
level of significance.

Company 1 Company 2
98 65
60 11
-14 30

8-12
Copyright © 2014 Pearson Education, Inc.
80 52
55 27
71 16
37 27
41 47
Answer: The following null and alternative hypotheses would be appropriate in this situation:
H0 : =

Ha : ≠

This will be a two-tailed test since we are testing to see whether a difference exists and are not predicting which
company will have more or less variability.

We first need to compute the sample variances for each sample using:

s2 =

The following variances are computed from the sample data:

= 1145.43 and = 345.7

In a two-tailed test of variances, the test statistic is an F-value that is formed by the ratio of the two sample variances
and by placing the larger sample variance in the numerator:

F= = = 3.3134

We then compare this value to a critical value from the F-distribution table. Since this is a two-tailed test with alpha
= .10, we use the table with 0.05 in the upper-tail of the F-distribution.
Two sets of degrees of freedom are used with the F-distribution. Across the top of the F-table we look for degrees of
freedom corresponding to n-1 where n is the sample size associated with the sample variance that was placed in the
numerator of the F-test statistic. In our case, n = 8, so the degrees of freedom is 8 - 1 = 7 for the numerator. The
degrees of freedom down the side in the F table is n-1 where n corresponds to the sample size for the sample
variance in the denominator of the F-test statistic. In our case that would be n = 8. Thus, the denominator degrees of
freedom are 8 - 1 = 7. Then the critical F from the .05 table with 7 and 7 degrees of freedom is 3.787. The decision
rule is:

If F calculated > 3.787, reject H0; otherwise do not reject.

Since our calculated F-test statistic is F = 3.3134 < 3.787, we do not reject the null hypothesis. Thus, based on the
sample information, we have no basis for believing that there is a difference in the two companies with respect to
population variance.

26) Frasier and Company manufactures four different products that it ships to customers throughout the United
States. Delivery times are not a driving factor in the decision as to which type of carrier to use (rail, plane, or truck)
to deliver the product. However, breakage cost is very expensive, and Frasier would like to select a mode of delivery
that reduces the amount of product breakage. To help it reach a decision, the managers have decided to examine the
dollar amount of breakage incurred by the three alternative modes of transportation under consideration. Because
each product's fragility is different, the executives conducting the study wish to control for differences due to type of
product. The company randomly assigns each product to each carrier and monitors the dollar breakage that occurs
over the course of 100 shipments. The dollar breakage per shipment (to the nearest dollar) is as follows:

8-13
Copyright © 2014 Pearson Education, Inc.
Rail Plane Truck
Product 1 $7,960 $8,053 $8,818
Product 2 $8,399 $7,764 $9,432
Product 3 $9,429 $9,196 $9,260
Product 4 $6,022 $5,821 $5,676

Was Frasier and Company correct in its decision to block for type of product? Conduct the appropriate hypothesis
test using a level of significance of 0.01.
A) Because F = 32.12 > Fα=0.01 = 9.78, reject the null hypothesis. Thus, based on these sample data we conclude
that blocking is effective.
B) Because F = 28.14 > Fα=0.01 = 7.63, reject the null hypothesis. Thus, based on these sample data we conclude
that blocking is effective.
C) Because F = 32.12 > Fα=0.01 = 9.78, do not reject the null hypothesis. Thus, based on these sample data we
conclude that blocking is not effective.
D) Because F = 28.14 > Fα=0.01 = 7.63, do not reject the null hypothesis. Thus, based on these sample data we
conclude that blocking is not effective.
Answer: A

27) An insurance company has the business objective of reducing the amount of time it takes to approve applications
for life insurance. The approval process consists of underwriting, which includes a review of the application, a
medical information bureau check, possible requests for additional medical information and medical exams, and a
policy compilation stage in which the policy pages are generated and sent for delivery. The ability to deliver
approved policies to customers in a timely manner is critical to the profitability of this service. During a period of
one month, a random sample of 27 approved policies is selected, and the total processing time, in days, is collected.
These data, stored in insurance, are:
73 19 16 64 28 28 31 90 60 56 31 56 22 18 45 48
17 17 17 91 92 63 50 51 69 16 17
In the past, the mean processing time was 45 days. At the 0.05 level of significance, is there evidence that the mean
processing time has changed from 45 days?

28) A problem with a phone line that prevents a customer from receiving or making calls is upsetting to both
the customer and the telecommunications company. The file contains samples of 20 problems reported to two
different offices of a telecommunications company and the time to clear these problems (in minutes) from the
customers’ lines:
Central Office I Time to Clear problems (minutes)
1.48 1.75 0.78 2.85 0.52 1.60 4.15 3.97 1.48 3.10
1.02 0.53 0.93 1.60 0.80 1.05 6.32 3.93 5.45 0.97
Central Office II Time to Clear problems (minutes)
7.55 3.75 0.10 1.10 0.60 0.52 3.30 2.10 0.58 4.02
3.75 0.65 1.92 0.60 1.53 4.23 0.08 1.48 1.65 0.72

Assuming that the population variances from both offices are equal, is there evidence of a difference in the mean
waiting time between the two offices? (Use α=0.05).

29) Multiple myeloma, or blood plasma cancer, is characterized by increased blood vessel formulation
(angiogenesis) in the bone marrow that is a predictive factor in survival. One treatment approach used for multiple
myeloma is stem cell transplantation with the patient’s own stem cells. The data stored in and shown below,
represent the bone marrow microvessel density for patients who had a complete response to the stem cell transplant
(as measured by blood and urine tests). The measurements were taken immediately prior to the stem cell transplant
and at the time the complete response was determined.

Patient Before After


1 158 284
2 189 214
3 202 101

8-14
Copyright © 2014 Pearson Education, Inc.
4 353 227
5 416 290
6 426 176
7 441 290

At the 0.05 level of significance, is there evidence that the mean bone marrow microvessel density is higher before
the stem cell transplant than after the stem cell transplant?

30) Nine experts rated two brands of Colombian coffee in a taste-testing experiment. A rating on a 7-point
scale (1=extremely unpleasing, 7= extremely pleasing) is given for each of four characteristics: taste, aroma,
richness, and acidity. The following data stored in coffee contain the ratings accumulated over all four
characteristics:
Brand
Expert A B
C.C. 24 26
S.E. 27 27
E.G. 19 22
B.L. 24 27
C.M. 22 25
C.N. 26 27
G.N. 27 26
R.M. 25 27
P.V. 22 23
a. At the 0.05 level of significance, is there evidence of a difference in the mean ratings between the two
brands?
b. Construct and interpret a 95% confidence interval estimate of the difference in the mean ratings
between the two brands.

31) The effects of developer strength (Factor A) and development time (Factor B) on the density of photographic
plate film were being studied. Two strengths and two development times were used, and four replicates in each
of the four cells were evaluated. The results (with larger being best) are stored in and shown in the following
table:

Developer Strength Development Time (minutes)


10 14
1 0 1
1 5 4
1 2 3
1 4 2
2 4 6
2 7 7
2 6 8
2 5 7
At the 0.05 level of significance,
a. Is there an interaction between developer strength and development time?
b. Is there an effect due to developer strength?
c. Is there an effect due to development time?
d. What can you conclude about the effect of developer strength and development time on density?

32) You are working as a purchase manager for a company. The following information has been supplied to you by
two manufacturers of electric bulbs:

Company A Company B

8-15
Copyright © 2014 Pearson Education, Inc.
Mean Life (in hours) 1300 1248
Standard deviations (in hours) 82 93
Sample size 100 100
Which brand of bulbs are you going to purchase if you desire to take a risk of 5%?

8-16
Copyright © 2014 Pearson Education, Inc.
8-17
Copyright © 2014 Pearson Education, Inc.