Chapter 8: Hypothesis Testing P a g e | 113

CHAPTER 8
HYPOTHESIS TESTING

Introduction

A statistical test of hypothesis is a fairly clear-cut procedure that enables an

experimenter to either reject or accept null hypothesis with measure risks 𝛼 𝑎𝑛𝑑 𝛽.
The experimenter can control the risk of falsely rejecting the null hypothesis by
selecting an appropriate value of 𝛼. On the other hand, the value of 𝛽depends on the
sample size and the values of the parameter under the test that are of practical
importance to the experimenter. When this information is not available, an
experimenter may decide to select an affordable sample size, in the hope that the
sample will contain sufficient information to reject the null hypothesis. The chance
that this decision is in error is given by 𝛼, whose value has been set in advance. If
the sample does not provide sufficient evidence to reject 𝐻0 , the experimenter may
wish to state the results of the test as “The data do not support the rejection of 𝐻0 ”
rather than accepting the 𝐻0 without knowing the chance of error 𝛽.

A hypothesis is a conjectural statement of the relation between two or more

variables. Hypotheses are always declarative sentence form, and they relate either
generally or specifically.

According to Deveza (1996), the following are the criteria of “good” hypotheses and
hypotheses statements:

1. Hypotheses are statements about the relations between variables.

2. Hypotheses carry clear implications for testing the stated relations.

These two criteria mean that hypothesis statements contain two or more variables
that are measurable or potentially measurable and that they specify how the
variables are related.

A research hypothesis is a predictive statement that represents a very specific

proposed answer to the problem statement. Hypotheses should be stated in an
empirically testable form, and should indicate the specific nature of the connection
between or among variables that was initially identified in the problem statement.

Statistics and Probability by SMAGalman

 Chapter 8: Hypothesis Testing P a g e | 114

Types of Hypothesis

1. Null Hypothesis (𝐻𝑜 ) is primarily used for statistical purposes. It is the hypothesis
to be tested and it represents what the investigation doubts to be true.

2. Alternative Hypothesis (𝐻𝑎 ) states the nature of the connection between or

among variables that we expect on theory and/or prior research. In other words, it
is the operational statement of the theory that the experimenter or researcher
believes to be true and wishes to prove.

The alternative hypothesis is a statement that contradicts or disagrees with the null
hypothesis. If the null hypothesis is rejected, the conclusion is reflected by the
alternative hypothesis. If there is no enough statistical evidence to reject the null
hypothesis, it is retained, but not accepted. In this case, the null hypothesis is not
necessarily true, but it just cannot be rejected from the current evidence from the
sample data.

Example 1

If we are going to test the correlation of Physics and Mathematics subjects, we

formulate the null and alternative hypotheses as follows:

𝐻0 : There is no significant relationship between students’ academic

performance in Physics and Mathematics subjects

or : Students’ academic performance in Physics and Mathematics are not

correlated.

𝐻𝑎 : There is a significant relationship between students’ academic performance

in Physics and Mathematics subjects

or : Students’ academic performance in Physics and Mathematics are correlated.

Example 2

A computer shop receives a delivery of a big number of laptops from a salesman.

The delivery is regarded as satisfactory if the percentage of defective laptops does
not exceed 3%(i.e., 𝑃 ≤ 3%); otherwise, it is regarded as unsatisfactory. The
percentage of defective sets is an unknown parameter hence, the sampling
inspection is done to decide whether the delivery is to be accepted or rejected. The
following portray the hypotheses of this problem.

𝐻0 : The delivery is satisfactory; i.e., 𝑃 ≤ 0.03

𝐻𝑎 : The delivery is satisfactory; i.e., 𝑃 > 0.03

Statistics and Probability by SMAGalman

 Chapter 8: Hypothesis Testing P a g e | 115

Directional and Nondirectional Tests

A test of hypothesis is categorized into two:

a) One-tailed test is also known as the directional test. It occurs when researchers
do not have a priori expectation about the ample value they expect to observe. This
is usually done if the researcher just want to test whether a mean is larger or
smaller the hypothesized population mean.

Figure 8.1
Region of Rejection and Acceptance in One-tailed Test

b) Two-tailed test is also known as the nondirectional test of a hypothesis. It is a

test where alternative hypothesis does not specify a directional difference for the
parameter of interest. It is appropriate when researchers do not have a priori
expectation regarding the value they expect to observe in the sample. The
researchers would simply want to see if the sample mean is different from the
hypothesized population in either direction.

Figure 8.2
Region of Rejection and Acceptance in Two-tailed Test

Statistics and Probability by SMAGalman

 Chapter 8: Hypothesis Testing P a g e | 116

Level of Confidence

In hypothesis testing, the conclusion reached is not based on certainty, but

on a probability estimate. This probability estimate or risk of committing an error in
the conclusion is determined by the researcher. This is known as the alpha level.

If we set the alpha level at 5% (𝛼 = 0.05), then we are taking a 5% risk of

being wrong and 95% risk of being right. That is, we have a probability error of 5%
that our conclusion is wrong.

Furthermore, the level of significance determines the critical value of the test
statistic. This value serves as the reference point for rejecting or not rejecting a null
𝛼
hypothesis. In a two-tailed test, the level of significance is given by . For instance, if
2
the indicated level of significance is 5% and the test is a two-tailed test, the level of
significance would be determined by 2.5%.

Types of Errors

Whichever hypothesis is correct. But there is a chance that sampling error

will cause a researcher to reject the hypothesis. We may reject the null hypothesis
that is true or we may accept a null hypothesis that is false. These two types of
mistakes in hypothesis testing are called type I and type II error, respectively.

Steps in Hypothesis Testing

The logical steps that should be followed in order to arrive at an objective

and meaningful result are the following:

1. State the null hypothesis (𝐻0 ).

2. Determine the measurement level of the variables under study. The level of
measurement of the variable determines to a large extent the statistical test the
hypothesis. Parametric tests are generally used for interval and ratio data of
measurement while nonparametric tests are usually applied to nominal and ordinal
types.

3. Choose the statistical test appropriate to test the hypothesis.

4. Specify a significance level and the sample size.

5. Compute the value of a statistical test.

6. Determine the significance of the computed value.

7. Interpret and discuss the result.

Statistics and Probability by SMAGalman

 Chapter 8: Hypothesis Testing P a g e | 117

8.1 z-test

This section aims to:

1. determine the conditions when to use the z-test; and

2. test hypothesis using z-test.

This parametric test is used when the given data is population mean and
population standard deviation and 𝑛 > 30. The formula is given by

𝑥̅ − 𝜇
𝑧= 𝜎
√𝑛

Where: 𝑥̅ − 𝑠𝑎𝑚𝑝𝑙𝑒 𝑚𝑒𝑎𝑛

𝜇 − 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑚𝑒𝑎𝑛

𝜎 − 𝑝𝑜𝑝𝑢𝑙𝑎𝑡𝑖𝑜𝑛 𝑠𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝑑𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛

𝑛 − 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑒𝑠

Use z test when:

• Your sample size should be greater than 30. Otherwise, use a t test.
• Data points should be independent from each other.
• Your data should be normally distributed. However, for large sample sizes
(over 30) this doesn’t always matter.
• Your data should be randomly selected from a population, where each item has
an equal chance of being selected.
• Sample sizes should be equal if at all possible.

Example

In a recent survey of Information Technologists in Nueva Ecija, it was found out that
the average monthly net income of IT experts is P12, 542. Suppose a researcher
wants to test this figure by taking a random sample of 165 IT experts in the province
to determine whether the monthly net income has changed. Suppose further that
the average net monthly income of 165 IT experts is P13,861 and the population
standard deviation was found out to be P1,459. Test the significance at 5% level.

Statistics and Probability by SMAGalman

 Chapter 8: Hypothesis Testing P a g e | 118

Solution:

Following the steps in hypothesis testing,

Step 1. 𝐻0 : The average monthly net income of IT experts is still P12, 542

𝜇 = 𝑃12, 542

𝐻𝑎 : The average monthly net IT experts is no longer P12, 542

𝜇 ≠ 𝑃12,542

Step 2. 𝛼 = 5%, two-tailed test (nondirectional)

Step 3. z-test for a single mean

𝑥̅ − 𝜇
𝑧= 𝜎
√𝑛

Step 4. Reject 𝐻0 if the computed z-value is greater than 1.96 or less than -1.96

Step 5. Given : 𝑥̅ = 13,861

𝜇 = 12,542

𝑛 = 165

𝜎 = 1,459

𝑥̅ − 𝜇
𝑧= 𝜎
√𝑛
13,861−12,542
𝑧= 1,459
√165

= 11.61

Step 6. Reject 𝐻0 since 𝑧𝑐 > 1.96.

Step 7. Interpretation: The average monthly income of IT experts is no longer

P12,542.

Statistics and Probability by SMAGalman

 Chapter 8: Hypothesis Testing P a g e | 119

Name: ___________________________________________________________ Date: _____________________

Section: _________________________________________________________ Score: ____________________

Exercises 8.1
z-test

1. In the population, the average IQ is 100 with a standard deviation of 15. A team of
scientists wants to test a new medication to see if it has either a positive or negative
effect on intelligence, or no effect at all. A sample of 30 participants who have taken
the medication has a mean of 140. Did the medication affect intelligence, using alpha
= 0.05?

2. The mean Verbal SAT score for the population of first students at NEUST is 520.
The standard deviation of scores in this population is 95. An investigator believes
that the mean Verbal SAT of first year psychology majors is significantly different
from the mean score of the population. The mean of a sample of 36 first year
psychology majors is 548. Please test the investigator's prediction using an alpha
level of .05.

Statistics and Probability by SMAGalman

 Chapter 8: Hypothesis Testing P a g e | 120

3. The mean number of close friends for the population of people living in the
Philippines. is 5.7. The standard deviation of scores in this population is 1.3. An
investigator predicts that the mean number of close friends for introverts will be
significantly different from the mean of the population. The mean number of close
friends for a sample of 26 introverts is 6.5. Do these data support the investigator's
prediction? Use an alpha level of .05.

Statistics and Probability by SMAGalman