Você está na página 1de 6

MAT 102 - Introduction to Statistics Chapters 9 & 10 – Hypothesis Testing

Chapter 9 – Introduction to Hypothesis Testing


9.2 – Hypothesis Testing
 Hypothesis testing is an example of inferential statistics
 We use sample information to draw conclusions about the population from which we sample
 We can never be 100% certain about the conclusions we draw about a population based upon one
sample since a different sample selected from the same population could possibly lead one to draw
a different conclusion about the same population. It is necessary to use probability to help explain
how reliable our conclusions are.
 We always assume that the sample data was selected randomly from the population

Null and Alternative Hypotheses


We begin a hypothesis test by stating the claims of the statistical test.
There are two hypotheses that are formulated:

Null Hypothesis (denoted by Ho) is the statement being tested which states a value for a population
parameter and is initially assumed to be true. It is formulated for the sole purpose of trying to reject it by
seeking evidence against it.

Alternative Hypothesis (denoted by Ha) is a statement about the value of a population parameter which
is different from the value stated in the null hypothesis. This hypothesis is suspected to be true and is
supported by the rejection of the null hypothesis.

Example 9.2 pg. 459 – State the null and alternative hypotheses for the following research
A national poll states that 60% of the U.S. population believe that the Space Shuttle program should be
continued. A researcher who decides to test this statement would like to show that more than 60% of the
U.S. population believe the Space Shuttle program should be continued.

Example 9.3 pg. 459 – State the null and alternative hypotheses for the following research
A college newspaper claims that full-time college students work an average of 20 hours a week. A
marketing professor who believes this claim is too high decides to conduct a study to test the newspaper’s
claim.

Example – State the null and alternative hypotheses for the following research
An automatic opening device for parachutes has a stated mean release time of 10 seconds. A local
parachute club believes the release time is not 10 seconds.

A directional alternative hypothesis considers only one specified direction of difference away from the
value stated in the null hypothesis and uses word equivalent to “less than or greater than or too low or
has increased”

A nondirectional alternative hypothesis considers both directions away from the value stated in the null
hypothesis and uses word equivalent to “is not or not equal to or different than”

1
MAT 102 - Introduction to Statistics Chapters 9 & 10 – Hypothesis Testing

Hypothesis Testing Procedure


1. Formulate the hypotheses
2. Determine the model to test the null hypothesis
3. Formulate a decision rule
4. Analyze the sample data
5. State the conclusion

There are two types of errors a researcher can make when performing a hypothesis test:

Type I error: made when we reject the null hypothesis when it is true
Type II error: made when we fail to reject the null hypothesis when it is false

Possible Outcomes to a Statistical Hypothesis Test


Reality:
Conclusion: H 0 is True H 0 is False

Reject H 0 Type I Error Correct Decision

Fail to Reject H 0 Correct Decision Type II Error

9.3 – The Development of a Decision Rule


The objective of a hypothesis test is to use sample data to decide if one should reject or fail to reject the
null hypothesis. Researchers use a predetermined decision rule to decide whether the sample data
supports or refutes the null hypothesis. To decide whether to reject Ho or fail to reject Ho, the researcher
compares the statistical result to the value(s) of the decision rule, referred to as the critical value(s).

The level of significance (denoted by the Greek letter alpha, α) is the probability of a statistical result
falling beyond the critical value, assuming that Ho is true. It is also the probability of committing a type I
error.
 The level of significance,  , is usually given to you in the form of 1% or 5%.

The statistical result computed from the sample that we are going to compare to the critical value is called
the test statistic. The test statistic is computed from the sample information and is used to decide
whether to reject or fail to reject the null hypothesis.
 It is the z or t score calculated for the sample mean.

A one-tailed test (denoted by 1TT) is conducted when the alternative hypothesis is directional

A two-tailed test (denoted by 2TT) is conducted when the alternative hypothesis is nondirectional

2
MAT 102 - Introduction to Statistics Chapters 9 & 10 – Hypothesis Testing

Procedure to Formulate the Decision Rule:


1. Determine if the alternative hypothesis is directional or nondirectional and if it is a one tailed or
two tailed test.
2. Identify the level of significance
3. Construct the appropriate hypothesis test model
4. Find the critical z score(s) for the corresponding level of significance.
5. State the decision rule

3
MAT 102 - Introduction to Statistics Chapters 9 & 10 – Hypothesis Testing

Chapter 10 – Hypothesis Testing Involving a Population Mean


Step 1: Formulate the two hypotheses Ho and Ha.

Step 2: Determine the model to test the null hypothesis, Ho.

Under the assumption Ho is true, the expected results are:


a) The distribution of the test statistic is the hypothesis testing model and is either:
i. Approximately a normal distribution, if the standard deviation of the
population is known and the sample size is greater than 30,
or
ii. A t Distribution with degrees of freedom: df  n  1 if the standard deviation
of the population is unknown.

b) The mean of the sampling distribution of the mean is given by the formula:
x  

c) For a normal distribution: the standard error of the mean, denoted by  x is:

x 
n
or
For a t Distribution: the estimate of the standard error of the mean, written s x , is given by the
formula:
s
sx 
n

Step 3: Formulate the decision rule

a) Alternate hypothesis: directional or non-directional


b) Type of test: 1TT or 2TT
c) Significance level:  = 1% or  = 5%
d) Construct the hypothesis test model: find critical z score(s) zC , or zLC and zRC
OR find critical t score(s) from table and where df  n  1
e) State the decision rule: reject the null hypothesis if the test statistic is …

Step 4: Analyze the sample data


x  x
For a normal distribution the test statistic is: z  .
x
x  x
For a t distribution the test statistic is: t .
sx

Step 5: State the conclusion by comparing the test statistic to the critical score of the decision rule.
a) Reject H 0 and accept H a at  or
b) Fail to reject H 0 at 

4
MAT 102 - Introduction to Statistics Chapters 9 & 10 – Hypothesis Testing

Examples from textbook: Testing the hypothesis about a population mean, using information from a
sample that is drawn from a population

Example 10.7 pg. 510 - 511


An automatic opening device for parachutes has a stated mean release time of 10 seconds with a
population standard deviation of 3 seconds. A local parachute club decides to test this claim against the
alternative hypothesis that the release time is not 10 seconds. The club purchases 36 of these devices and
finds that the mean release time is 8.5 seconds. Does this sample result indicate that the opening device
does not have a mean release time of 10 seconds? Use  = 1%.

Example 10.8 pg. 511-512


A tire manufacturer advertises that its brand of redial tires has a mean life of 40,000 miles with a
population standard deviation of 1,500 miles. A consumer’s research team decides to investigate this
claim after receiving several complaints from people who believe this advertisement is false (i.e. the mean
life of 40,000 miles is too high). If the research team tests 100 of these radial tires and obtains a sample
mean tire life of 39,750, is the advertisement legitimate? Use  = 5%.

Example 10.13 pg. 517-518


An automobile manufacturer claims that their economy car averages 55 mpg for highway driving. An
engineer representing a leading automotive magazine believes this claim is too high. To test the
manufacturer’s claim, the engineer randomly selects sixteen of the economy cars and finds that they
averaged 53 mpg with s = 2.5 mpg. Does this sample result indicate that the manufacturer’s claim is too
high at a = 1%?

Example 10.14 pg. 518-520


A medical research team decides to study whether pressure exerted on a person’s upper arm will increase
bleeding time. The average bleeding time for a pricked finger where no pressure is applied to the upper
arm is 1.6 minutes. To test their claim the research team randomly selects 64 subjects. They find that the
average bleeding time is 1.9 minutes with s  0.80 minutes when 50 mm of pressure is applied to the
upper arm. Does this sample result support the claim that pressure applied to the upper arm will increase
bleeding time? (use  = 5%)

Review Example 10.15 pg. 520, and Example 10.16 pg. 521

5
MAT 102 - Introduction to Statistics Chapters 9 & 10 – Hypothesis Testing

Você também pode gostar