a difference between a parameter and a statistic or states that there is a

difference between two parameters.

A decision rule to resolve which of these two opposing hypotheses is

more likely to be true will be followed. The null hypothesis will be rejected in
favor of the alternative hypothesis if the sample evidence strongly suggests
that it is false. In the same manner, if we favor the null hypothesis, we reject
the alternative hypothesis.

Two types of tests were considered for hypothesis testing. A type of test
used for directional hypothesis is known as a one-tailed test, whereas the
one used for a non-directional hypothesis is known as a two-tailed test. A
one-tailed test can only be right-tailed or left-tailed, which leans in the
direction of the inequality of the alternative hypothesis.

To state the hypotheses, we must translate the words into mathematical

symbols. The basic symbols used are as follows:

Equal to = Greater than >

Not equal to ≠ Less than <

The null and alternative hypotheses are stated together, and the null
hypothesis contains the equals sign, as shown (where µo represents a
statistic).

Two-tailed test Right-tailed test Left-tailed test

Ho: µ = µo Ho: µ = µo Ho: µ = µo
Ha: µ ≠ µo Ha: µ > µo Ha: µ < µo

Three different scenarios will be used to show how to identify the

parameter to be tested given a real-life problem and to illustrate how
hypotheses should be stated.

SCENARIO 1

A Virologist is interested in finding out whether a new COVID-19

vaccine will have some severe cases of side effects. He is particularly
concerned with the body temperature of the patients who were injected with
the vaccine. Will the body temperature increase, decrease, or remain
unchanged after a patient is injected with a vaccine? The Virologist knows
that the mean body temperature for the population under study is 37 °C.

Parameter: µ = 37 °C

3
 Hypothesis:

Ho: µ = 37 °C and Ha: µ ≠ 37 °C.

The null hypothesis stipulates that the mean will not be different, and
the alternative hypothesis asserts that it will be different. This test is called a
two-tailed test for the possible side effects of the vaccine could be to raise or
lower the body temperature.

SCENARIO 2

There is a recently passed law which makes child car safety seat
compulsory. It is reported that the average cost of a child car safety seat is
greater than ₱1 000. A customer wanted to know if the report is really true.

Parameter: µ = ₱1 000

Hypothesis:

Ho: µ = ₱1 000 and Ha: µ > ₱1 000.

In this situation, the customer is interested only in finding out if the

reported cost of a child car safety seat is really greater than ₱1 000, so her
alternative hypothesis is that the mean is greater than ₱1 000. The null
hypothesis is that the mean is equal to ₱1 000. This test is called right-tailed
(one-tailed) since the interest is whether the cost is really greater.

SCENARIO 3

An actuarial scientist wanted to find out if there is a significant decrease

in the number of people violating the COVID-19 protocols in Davao City last
December 2020. The reported average number of violators per month is 1 756
individuals.

Parameter: µ = 1 756

Hypothesis:

Ho: µ =1 756 and Ha: µ < 1 756

This test is a left-tailed test since the actuarial scientist is interested

only in finding out if there is a significant decrease of COVID-19 violators in
Davao City.

4
Level of Significance

When one has formulated the hypothesis, the next step is to make a
research design. The researcher decides on the right statistical test and
selects a fitting level of significance. For example, in Scenario 1, a sample of
patients who will be given the vaccine will be chosen. After letting an
appropriate period for the vaccine to be absorbed, the researcher will take
each patient’s temperature.

Let us remember that when samples of size n are carefully chosen from
a population, the means of these samples will differ from the population mean,
and the spread of these sample means will be nearly normal when the sample
size is 30 or more. This means that even if the null hypothesis is accepted,
the mean of the temperature of the sample of patients will not, in most cases,
be exactly equal to the population mean of 37 °C.

The decision on whether to accept or reject the hypothesis is based on

statistics, and not on hunches or opinions. That is, the difference must be
significant. Therefore, if we accept the null hypothesis, even if the sample
mean differs from the population mean, the difference is not significant at all.
Thus, we say that they are statistically equal. This is where the ideas of
statistical tests and level of significance are used.

The level of significance, usually designated by the alpha (a) symbol

pertains to the degree of confidence we require in order to reject the null
hypothesis in favor of the alternative hypothesis. It is the highest probability
of committing a Type I error. The significance testing that we currently use is
a combination of the Ronald Fisher’s idea of utilizing the probability value p
as an index of the weight of evidence against a null hypothesis, and Jerzy
Neyman and Egron Pearson’s notion of testing a null hypothesis against an
alternative hypothesis.

Fisher proposed that 95% is a useful threshold of confidence for only

when we are 95% positive that a result is accurate should we accept it as true.
In other words, if there is only a 5% chance (a = .05) of committing an error,
then we say that it is a statistically significant finding.

Rejection Region

In testing a hypothesis, the researcher decides the level of significance

to be used. The gravity of the type I error will be the basis for the level to be
used. Selecting a critical value from a table, given the appropriate test, will
then follow.

The critical value determines the critical and non-critical regions. The
symbol for critical value is C.V. Critical values had to be computed by
remarkably brilliant people like Fisher. His tables for particular probability
values (.05, .02 and .01) led to a trend that state test statistics as being
significant in today’s well-known p < .05 and p < .01.

5
 The critical or rejection region is the range of values of the test
statistic that signifies that there is a significant difference and that the null
hypothesis should be rejected. In contrast, the non-critical or acceptance
region is the range of values of the test statistic that implies that the
difference was possibly coincidental and that the null hypothesis should be
accepted.

A one-tailed test is a directional test in which the rejection region is

situated either on the left or right tail of a normal curve. In connection to this,
the critical value can be on the right or on the left side of the mean for a one-
tailed test. Its position depends on the sign of the inequality sign of the
alternative hypothesis.

For illustration, in Scenario 2, where a customer wanted to know if the

reported average cost of a child car safety seat is indeed greater than ₱1 000,
the alternative hypothesis is Ha: µ > ₱1 000. Since the inequality sign is >, the
null hypothesis will only be rejected when the sample mean is significantly
greater than ₱1 000. Hence, the critical value must be on the right side of the
mean. Therefore, this test is called a right-tailed test.

Shown below is the illustration of the acceptance and rejection regions

of the left-tailed and right-tailed tests based on the statement of the
alternative hypothesis.

Left-tailed Test
(From scenario 3)

Rejection Region Acceptance Region

–1.65
Ha: µ < ₱1 756
a = 0.05

Right-tailed Test
(From scenario 2)

Acceptance Region Rejection Region

1.65
Ha: µ > ₱1 000
a = 0.05

6
 A two-tailed test is a non-directional test in which the rejection region
is situated on both tails of a normal curve. The illustration of the acceptance
and rejection regions is shown below.

Two-tailed Test
(From scenario 1)

Rejection Region Rejection Region

Acceptance Region

– 2.58 2.58
Ha: µ ≠ 37 °C
a = 0.01

Type I and Type II Errors

The hypothesis testing procedure has four possible outcomes since the
null hypothesis (Ho) may be true or false, and the decision to reject or accept
it is based only on the statistics taken from samples. As a result, there are
two possibilities for a correct decision and two possibilities for an incorrect
decision, as illustrated below.
True Ho False Ho

Accepted Ho Correct
Type II Error
Decision
False Negative
True Positive

Correct
Type I Error
Rejected Ho Decision
False Positive
True Negative

If a true Ho is rejected, then it will result in a type I error. If we apply

a Fisher’s criterion, then the probability of this error is 5% (a-level). It implies
that out of 100 samples, we could assume that in 5 cases we could obtain a
test statistic large enough to make us believe that there is a real effect in the
population even if there isn’t. On the other hand, if a false Ho is accepted, then
a type II error occurs. Cohen (1992) suggested that the utmost acceptable
probability of a Type II error would be 20% (β-level). That would imply that if

7
we selected 100 samples from a population in which a change exists, we would
miss detecting that change in 20 of those samples.
In Scenario 1, a significant effect brought by the COVID-19 vaccine on
the body temperature of the patients might coincidentally exists in the sample
but not in the entire population. Herein, the virologist will reject the Ho when
it is really true, thus causing a type I error.
Contrariwise, the COVID-19 vaccine may not perhaps change the body
temperature of the subjects in the sample, but when it is given to the overall
population, it may result in a significant increase or decrease in the body
temperature of takers. The virologist, on the source of the data obtained from
the sample, will accept the Ho, thus causing a type II error.
In Scenario 2, the cost of a child car safety seat may not be greater
than ₱1 000, but it might be in the sample. In this case, a true Ho would be
rejected, resulting in a type I error. On the other hand, the cost of the car
safety seat might not be greater than ₱1 000 in the sample, but it is, in reality.
The customer, on the basis of evidence obtained from the sample, would
accept the null hypothesis, thus causing a type II error.

Let Us Practice

DIRECTION: Identify the parameter to be tested and state the null and
alternative hypotheses for each scenario.

1. An Obstetrician-Gynecologist believes that certain maternal milk will

increase the birth weight of babies. The average birth weight of a Filipino
newborn is 3 kg.
2. Mr. Velasco hypothesizes that the production cost of printing modules can
be decreased by using a Risograph instead of a Photocopier. The mean cost
per 500 sheets is ₱350.00.
3. A Grade-11 class adviser observed that the ages of her incoming students
vary from 15 to 21 years. The average age of incoming senior high school
students is 17 years.
4. A report states that the average monthly income of public high school
teachers is ₱26 000. A random sample of salaries of 100 public high school
teachers showed an average monthly salary of ₱27, 500.
5. A new Internet Service Provider (ISP) offers a cheaper alternative for an
internet connection. The average cost for the same speed as other existing
ISP is ₱1 699.00.