Statistics

MODULE 24: Steps in Hypothesis Testing

LEARNING OUTCOMES

At the end of the module, you are expected to exhibit the following competencies:
1. Identify the steps in hypothesis testing.
2. Illustrate level of significance and corresponding rejection region.
3. Calculate the probabilities of committing an error in a test of hypothesis.

IMPORTANT CONCEPTS

A test of hypothesis is a series of steps that starts with the formulation of the null and the alternative hypotheses
and ends with stating the conclusion. Each step has several components to consider. It has parallelism with court
proceedings which could be used as a motivational activity.

• As in a court proceeding, the first step is to state the accusation or the statement of what will be evaluated as
true or false. Parallel to hypothesis testing, we first formulate the hypotheses to be tested. Remember that we
do not know the true state of nature of the hypothesis, that is, whether the hypothesis is true or false. Like in a
court trial, we do not know whether the accused is guilty or not.

• As discussed in the previous lesson, there are two types of hypotheses to state: the null and the alternative. In a
court proceeding, we can state the null hypothesis as Ho: The accused is not guilty. While the alternative is
stated as Ha: The accused is guilty.

• The second step is to state the decision rule that we will follow in making a decision on whether to reject or fail
to reject (accept) the null hypothesis. In a court proceeding, it is a guideline that the court uses to evaluate the
quantity and quality of evidences to be presented. And based on this guideline, the court decides whether to
reject or accept the hypothesis that the accused is not guilty. To be able to specify the decision rule in a
hypothesis testing procedure, there is a need to specify the components of the rule. These components include
the following:

1. We specify a level of significance, which is usually denoted as α in doing the test of hypothesis. It is the
same α that we encounter in the discussion of the (1-α)% confidence interval estimate. A level of
significance is the probability of rejecting a true null hypothesis or committing a Type I error in the test of
hypothesis. Since it is a probability of committing an error, it is usually a small value and it is between 0
and 1.
2. We identify the test statistic to use in the decision rule. Usually, the test statistic is a standardized
expression of the point estimator of the parameter identified in the hypothesis. Also, the distribution of
the test statistic is also needed to be specified.
3. Part of the decision rule is the specification of the rejection region. The rejection region is that part of the
distribution of the test statistic where we reject the null hypothesis.

An example of a decision rule is stated as follows:

“At a given α = 0.05, we reject Ho if the computed test statistic (denoted as tc) is greater than a tabular value of
the t distribution with n-1 degrees of freedom. Otherwise, we fail to reject Ho.”
Module 24 Page 1 of 4
Steps in Hypothesis Testing

In this decision rule, the level of significance is set at α equal to 0.05 and the test statistic is denoted by tc which is
assumed to follow the Student’s t-distribution with n-1 degrees of freedom. The rejection region is the area to
the right of the tabular value obtained from the Student’s t-distribution with n-1 degrees of freedom. Such
rejection region is illustrated in the following figure.

• The third step is then to compute the value of the test statistic using a random sample of observations gathered
or collected for the purpose of the test of hypothesis. In a court proceeding, this is the time that the gathered
evidences are presented.

• With the computed value of the test statistic, the next step is to use the decision rule to make a decision
whether to reject or fail to reject (accept) the null hypothesis. As in a court proceeding, the jury or the court will
decide whether the accused is guilty or not based on the evidences presented.

• Lastly, as a consequence of the decision, conclusions are made which are in relation to the purpose of the test
of hypothesis. In a court proceeding, this is the time when the court gives its verdict on the accused. In both
scenarios, this last step is the most awaited part of the procedure.

At this point, we can summarize the steps as follows:

1. Formulate the null and alternative hypotheses.
2. Identify the test statistic to use. With the given level of significance and the distribution of the test statistics,
state the decision rule and specify the rejection region.
3. Using a simple random sample of observation, compute for the value of the test statistic.
4. Make a decision whether to reject or fail to reject (accept) Ho.
5. State the conclusion

Note that in drawing the conclusions based on the test of hypothesis, we are not 100% sure of our decisions and
also with our conclusions. Like in a court proceeding when the court declares the accused is guilty, the decision is
not made with certainty. In other words, the court is not 100% sure that the accused is guilty. There is still that
possibility that the accused is not guilty and the court is making an error with its decision.

Recall that there are two types of errors that one can commit in decision making and these are Type I and Type II
error. Type I error is rejecting a hypothesis when in fact the hypothesis is true. Thus, the court commits a Type I
error when it declares that the accused is guilty when in fact the accused is not guilty of the crime. In other
words, the court had convicted an innocent person. On the other hand, Type II error is committed when a false
hypothesis is accepted, that is, the court freed a person guilty of the crime.

Note: if there is still time, let learners discuss the consequences of committing the two types of error and
determine which of the two types of error—(1) convicting an innocent person or (2) freeing a person guilty of the
crime—has greater consequence.

Module 24 Page 2 of 4
Steps in Hypothesis Testing
However, unlike in a court proceeding, statistics can allow us to compute the probability of committing an error in
decision making. The probability of committing Type I error is defined earlier as the level of significance and it is
denoted by α. On the other hand, the probability of committing Type II error is usually denoted by b. Also, in
statistics the decisions on the test of hypothesis are made with the given probabilities of Type I and Type II error.
We have the assurance that the test procedures that we use in statistics were formulated with minimum
probabilities of committing an error.

The probability of committing an error is a conditional probability problem. It is the probability of making a
decision based on the uncertainty of the true state of nature of the hypothesis being tested. You may use the
numerical example provided in the previous lesson in this chapter to illustrate the computation of the
probabilities of committing Type I and Type II errors.

Numerical Example: In testing the null hypothesis “The average daily number of text messages that a Grade 11
student sends is equal to 100” against an alternative hypothesis stated as “The average daily number of text
messages that a Grade 11 student sends is greater than 100”. A random sample of 16 students were selected and
interviewed. The daily number of text messages she sends is obtained. The null hypothesis is said to be rejected if
the sample mean is at least 102, otherwise the null hypothesis will be accepted or we fail to reject Ho. It is
assumed that the number of text messages that a Grade 11 student sends in a day follows a normal distribution
with standard deviation equal to 5 text messages.

Computing for the probability of committing Type I error, we have

Thus, we say the probability of rejecting a true null hypothesis is 0.0058 or we say that on the average, we are
assured with 94.52% (1-0.0058 = 0.9452) confidence that we are making a correct decision in accepting a true null
hypothesis.

The alternative hypothesis is stated as “The average daily number of text messages that a Grade 11 student sends
is greater than 100.” If we assume that the true distribution of the number of text messages that a Grade 11
student sends in a day follows a normal distribution with a mean of 103 and a standard deviation equal to 5 text
messages, then the computed probability of Type II error is

In this case, the probability of accepting a false null hypothesis or accepting Ho given that the average number of
text messages that a Grade 11 student sends in a day is indeed 103 (greater than 100) is computed as 0.2119.

Module 24 Page 3 of 4
Steps in Hypothesis Testing

PRACTICE SKILLS

1. The Graduate Record Exam (GRE) is a standardized test required to be admitted to many graduate schools in
the United States. A high score in the GRE makes admission more likely. According to the Educational Testing
Service, the mean score for takers of GRE who do not have training courses is 555 with a standard deviation
of 139. Brain Philippines (BP) offers expensive GRE training courses, claiming their graduates score better
than those who have not taken any training courses. To test the company’s claim, a statistician randomly
selected 30 graduates of BP and asked their GRE scores.

a. Formulate the appropriate null and alternative hypotheses.

b. Identify situations when Type I and Type II errors are committed and state their possible consequences.
c. Suppose the decision rule is “Reject Ho if the mean score of the sampled BP graduates is greater than
590; otherwise, fail to reject Ho.” Compute for the level of significance for this test. Also, find the risk of
concluding that the BP graduates did not score better than 555 when in fact the mean score is 600.

2. Consider a manufacturing process that is known to produce bulbs that have life lengths with a standard
deviation of 75 days. A potential customer will purchase bulbs from the company that manufactures the
bulbs if she is convinced that the average life of the bulbs is 1550 days.

a. Formulate the appropriate null and alternative hypotheses.

b. Identify situations when Type I and Type II errors are committed and state their possible consequences.
c. Suppose the decision rule is “Reject Ho if a random sample of 50 bulbs has a life less than than 1532 days;
otherwise, fail to reject Ho.” Compute for the level of significance for this test. Also, find the risk of
concluding that the average is greater than 1550 days when in fact their mean score is 1500.

REFERENCES

Richardson, M, Using Dice to Introduce Sampling Distributions. STatistics Education Web (STEW). Retrieved from
http://www.amstat.org/education/stew/pdfs/UsingDicetoIntroduceSamplingDistributio ns.doc

De Veau, R. D., Velleman, P. F., and Bock, D. E. (2006). Intro Stats. Pearson Ed. Inc.

Workbooks in Statistics 1: 11th Edition. Institute of Statistics, UP Los Baños, College Laguna 4031

Probability and statistics: Module 24. (2013). Australian Mathematical Sciences Institute and Education Services
Australia. Retrieved from http://www.amsi.org.au/ESA_Senior_Years/PDF/InferenceProp4g.pdf

Module 24 Page 4 of 4