OIS 2340
–
Business Statistics
Estimating Population Parameters
1
Hypothesis Tests
Goal:
Find statistical evidence to support or refute a claim.
The Hypotheses
1.
Null Hypothesis
a.
Includes the “status quo” or the assumed state of affairs.
b.
Includes the equality in the claim. (i.e. > or < or =)
c.
Denoted as H
0
d.
Is either “rejected” or “not rejected” based on the statistical analysis.
2.
Alternative Hypothesis
a.
Encompasses the rest of the sample space not included in the null.
b.
Does not include equality.
(i.e. > or <)
c.
Denoted as H
A
d.
Often called a “research hypothesis” or an “assumption” that we are
challenging.
Types of Hypothesis Tests
1.
One-Tail Test
a.
Upper Tail Test
Example:
Suppose someone claims that the average GPA of a student at
the University of Utah is above 3.0.
H
0
:
H
A
:
This is an
upper tail test
because we will reject the null hypothesis if
enough of our sample data falls ABOVE 3.0, or in the upper tail of the
normal curve.

OIS 2340
–
Business Statistics
Estimating Population Parameters
2
b.
Lower Tail Test
Example:
Suppose someone claims that fewer than 40% of the
population favors candidate A for election.
H
0
:
H
A
:
This is a
lower tail test
because we will reject the null hypothesis if
enough of our sample data falls BELOW 0.40, or in the lower tail of the
normal curve.
2.
Two-Tail Test (similar to confidence intervals)
Example:
We want to test whether a particular fuel injection unit will provide 24
miles per gallon.
H
0
:
H
A
:
This is a
two tail test
because we will reject the null hypothesis if enough
of our sample data falls ABOVE or BELOW 24, or in either tail of the
normal curve.

OIS 2340
–
Business Statistics
Estimating Population Parameters
3
Formulating Hypotheses and Determining Test Type
Formulating Hypothesis and Determining Test Type
It is claimed that the mean salary of college-educated women in Utah is greater than the
national average for all college-educated women, which is known to be $45,000. A random
sample of 100 college-educated women in Utah is taken. Their average salary is determined to
be $46,500 with a standard deviation of $5,200.

#### You've reached the end of your free preview.

Want to read all 11 pages?

- Fall '11
- Statistics, Statistical hypothesis testing, Type I and type II errors