MANB 1123 ASSIGNMENT 1 (PART 2): HYPOTHESIS TEST

RULES:
1. This assignment should be conducted in pair.
2. Submit the answer in softcopy format via e-learning link provided
3. Please be make sure to run the normality test before proceed with statistical test

Question 1:
The Center on Budget and Policy Priorities reported that average out-of-pocket medical expenses for
prescription drugs for privately insured adults with incomes over 200% of the poverty level was $173 in
2002. Suppose an investigation was conducted in 2009 to determine whether the increased availability of
generic drugs, Internet prescription drug purchases, and cost controls have reduced out-of-pocket drug
expenses. The investigation randomly sampled 196 privately insured adults with incomes over 200% of
the poverty level, and the respondents’ 2009 out-of-pocket medical expenses for prescription drugs were
recorded. These data are in the file Drug Expenses. Based on the sample data, can it be concluded that
2009 out-of-pocket prescription drug expenses are lower than the 2002 average reported by the Center
on Budget and Policy Priorities? Use a level of significance of 0.01 to conduct the hypothesis test.

ANSWER:
Normality test

Data is normally distributed thus use t-test on single mean:

Hypothesis Statement
Hnull: 2009 average out-of-pocket prescription drug expenses ≥ $173
Halt: 2009 average out-of-pocket prescription drug expenses < $173
Summary:
• Reject the null hypothesis
• The mean difference is lower $5.48

Question 2:
Hono Golf is a manufacturer of golf products in Taiwan and China. One of the golf accessories it produces
at its plant in Tainan Hsing, Taiwan, is plastic golf tees. The injector molder produces golf tees that are
designed to have an average height of 66 mm. To determine if this specification is met, random samples
are taken from the production floor. One sample is contained in the file labeled THeight.
a. Determine if the process is not producing the tees to specification. Use a significance level of 0.05.
ANSWER:
Normality test

Data is normally distributed thus use t-test on single mean

 Hypothesis Statement
Hnull: Average height of injector molder produces golf tees = 66mm
Halt: Average height of injector molder produces golf tees ≠ 66mm

Summary:
• Reject the null hypothesis thus the average height of injector molder produces golf tees
≠ 66mm. Therefore, the specification is not being met
• The mean difference is 0.31 higher than the hypothesized mean of 66mm

b. If the hypothesis test determines the specification is not being met, the production process will
be shut down while causes and remedies are determined. At times this occurs even though the
process is functioning to specification. What type of statistical error would this be?

This is a Type 1 error because we are supposing to accept the hnull (specification met) but we
reject the hnull. Thus, we will change the significance level to 0.01.

Question 3:
Joseph-Armand Bombardier in the 1930s founded the company that is now known as Seadoo in Canada.
His initial invention of the snowmobile in 1937 led the way to what is now a 7,600-employee, worldwide
company specializing in both snow and water sports vehicles. The company stresses high quality in both
manufacturing and dealer service. Suppose that the company standard for customer satisfaction is 95%
“highly satisfied.” Company managers recently completed a customer survey of 700 customers from
around the world. The responses to the question “What are your overall level of satisfaction with
Seadoo?” are provided in the file called Seadoo.
a. State the appropriate null and alternative hypotheses to be tested.
Hypothesis Statement
Hnull: The categories for satisfaction rating occur with equal probabilities
Halt: The categories for satisfaction rating occur with not equal probabilities
b. Using an alpha level of 0.05, conduct the hypothesis test and discuss the conclusions.

Summary:
• Reject the null hypothesis thus the categories for satisfaction rating occur with not equal
probabilities
• See the difference in the chart below

Summary:
• From the chart and the frequency table it shows that almost 91% of the customer is
highly satisfied with Seedoo.
Question 4:
Cell phones are becoming an integral part of our daily lives. Commissioned by Motorola, a new behavioral
study took researchers to nine cities worldwide from New York to London. Using a combination of
personal interviews, field studies, and observation, the study identified a variety of behaviors that
demonstrate the dramatic impact cell phones are having on the way people interact. The study found cell
phones give people a newfound personal power, enabling unprecedented mobility and allowing them to
conduct their business on the go. Interesting enough, gender differences can be found in phone use.
Women see their cell phone as a means of expression and social communication, whereas males tend to
use it as an interactive toy. A cell phone industry spokesman stated that half of all cell phones in use are
registered to females.
a. State the appropriate null and alternative hypotheses for testing the industry claim.
ANSWER:
Hypothesis Statement
Hnull: The proportion of female registered cell phone is = 50% (claim)
Halt: The proportion of female registered cell phone is ≠ 50%

b. Based on a random sample of cell phone owners shown in the data file called Cell Phone Survey,
test the null hypothesis. (Use α = 0.05.)
ANSWER:

c. The study also wants to test which group gender spent money on their mobile phone. Conduct
the statistical test at significance level 0.05. Discuss your finding.

ANSWER:
Question 5:
Suppose a professional job-placement firm that monitors salaries in professional fields is interested in
determining if the fluctuating price of oil and the outsourcing of computer-related jobs have had an effect
on the starting salaries of chemical and electrical engineering graduates. Specifically, the job placement
firm would like to know if the 2007 average starting salary for chemical engineering majors is higher than
the 2007 average starting salary for electrical engineering majors. To conduct its test, the job placement
firm has selected a random sample of 124 electrical engineering majors and 110 chemical engineering
majors who graduated and received jobs in 2007. Each graduate was asked to report his or her starting
salary. The results of the survey are contained in the file Starting Salaries. Conduct a hypothesis test to
determine whether any differences between mean starting salary for 2007 graduates in chemical
engineering with mean starting salary for 2007 graduates in electrical engineering. Conduct the test at the
0.05 level of significance. Be sure to test the normality assumption.

ANSWER:
Normality test on the differences
 Data is normally distributed thus use paired t-test
Hypothesis Statement
Hnull: There is No difference between mean starting salary for 2007 graduates in chemical engineering
with mean starting salary for 2007 graduates in electrical engineering. (diff in μave = 0)
Halt: There is a difference between mean starting salary for 2007 graduates in chemical engineering with
mean starting salary for 2007 graduates in electrical engineering. (diff in μave ≠ 0)

Summary:
• Reject the null hypothesis thus there is a difference between mean starting salary for 2007
graduates in chemical engineering with mean starting salary for 2007 graduates in electrical
engineering.
• The mean difference is $2904.61

Question 6:
A treadmill manufacturer has developed a new machine with softer tread and better fans than its current
model. The manufacturer believes these new features will enable runners to run for longer times than
they can on its current machines. To determine whether the desired result is achieved, the manufacturer
randomly sampled 35 runners. Each runner was measured for one week on the current machine and for
one week on the new machine. The weekly total number of minutes for each runner on the two types of
machines was collected. The results are contained in the file Treadmill. At the 0.02 level of significance,
can the treadmill manufacturer conclude that the new machine has the desired result?
 ANSWER:
Normality test on the differences

Data is normally distributed thus use paired t-test

Hypothesis Statement
Hnull: There is No difference in weekly running time between the new machine with the current. (diff in
μaverunningtime = 0)
Halt: There is a difference in weekly running time between the new machine with the current. (diff in
μaverunningtime ≠ 0)

Summary:
• Accept the null hypothesis thus there is no difference in weekly running time between the new
machine with the current.
Question 7:
A business statistics instructor at State University has been experimenting with her testing procedure. This
term, she has taken the approach of giving two tests over each section of material. The first test is a
problem-oriented exam, in which students have to set up and solve applications problems. The exam is
worth 50 points. The second test, given a day later, is a multiple-choice test, covering the concepts
introduced in the section of the text covered by the exam. This exam is also worth 50 points. In one class
of 15 students, the observed test scores over the first section of material in the course are contained in
the file State University.
a. If the instructor is unwilling to make the assumptions for the paired-sample t-test, what should
she conclude based on these data about the distribution of scores for the two tests if she tests at
a significance level of 0.05?

ANSWER:
Using Wicoxon signed rank test to test the median of differences between test 1 and test 2

b. In the context of this problem, define a Type II error.

To avoid Type II error by increasing the sample size.

Question 8:
USA Today notes (Mary Beth Marklein, “College Gender Gap Widens: 57% Are Women”) that there are
more men than women ages 18–24 in the United States—15 million versus 14.2 million. The male/female
ratio in colleges today is 42.6/57.4. However, there is a discrepancy in the percentage of males’ dependent
on their parents’ income. The file entitled Diversity contains the gender of undergrads (18–24) whose
parents’ income is in two categories: (1) low income—less than $30,000, and (2) upper income—$70,000
or more.
a. Determine if the sample sizes are large enough so that the sampling distribution of the difference
between the sample proportions of male undergraduates in the two income categories can be
approximated by a normal distribution.
ANSWER:

The sample size is large enough with size, n = 147. However, we cannot determine the variable
approximated by normal distribution because the variables are nominal variable.

b. Perform a test of hypothesis to determine that the proportion of male undergraduates in the
upper income category is more than 1% greater than that of the low-income category. Use a
significance level of 0.01.
ANSWER:

Upper Income Binomial Test Low Income Binomial Test