ELEMENTS OF

HYPOTHESIS
TESTING
WHAT IS
HYPOTHESIS?
A hypothesis is basically a
statement about the target
population. This is formulated as a
result of years of observation and
researches.
2
STATISTICAL
HYPOTHESIS

It is an assertion or
conjecture
concerning one or
more populations.
 TWO FORMS OF
HYPOTHESIS
Null hypothesis, 𝐻0 , is the hypothesis to be
tested and it represents what the investigation
doubts to be true.

Alternative hypothesis, 𝐻1 , is the operational

statement of the theory that the experimenter or
researcher believes to be true and whishes to
prove.
• Moreover, the alternative hypothesis
is a statement that contradicts or
disagrees with the null hypothesis. If
the null hypothesis is rejected as a
result of sample evidence, then
alternative hypothesis is the
conclusion.
• If there is no sufficient evidence to
reject the null hypothesis, it is
retained, but not accepted. In this
case, the null hypothesis necessarily
true, but it just cannot be rejected
from the current evidence from the
sample data.
5
 EXAMPLE
A computer shop owner receives a delivery of
a big number of computer sets from a
salesman. The delivery is regarded as a
satisfactory if the percentage of defective
computer sets does not exceed 5% (i.e., 𝑃 ≤
5%); otherwise, it is regarded as unsatisfactory.
In as much as it is not feasible to inspect all the
computer sets that time, the percentage of
defective sets is an unknown parameter.
Hence, sampling inspection is done to decide
whether the delivery is satisfactory (accept) or
unsatisfactory (reject). 6
DIRECTIONAL AND
NONDIRECTIONAL TESTS
A one-tailed (or A two-tailed (or nondirectional)
directional) test occurs test of a hypothesis is a test
when researchers do have where the alternative
an a priori expectation hypothesis does not specify a
about the sample value directional difference for the
they expect to observe. parameter of interest.

What the researchers The researchers would simply

wanted to do is to just test want to see if the sample mean
whether a mean is larger is different from the
or smaller than the (hypothesized) population
(hypothesized) population mean in either direction.
mean.
7
 TYPES OF ERRORS
A type I error is incurred A type II error is incurred
if a researcher rejects a when the researcher accepts
null hypothesis when in the null hypothesis when in
fact it is true. The fact it is false. The
probability of probability of committing a
committing a type I error type II error is denoted by
is denoted by the Greek the Greek letter beta.
letter alpha.

8
EXAMPLE
• In a court of law, a type I error is committed when
an innocent person is convicted while a type II
error is committed when a guilty person is
declared not guilty and set free.
• It should be noted that 𝛼 and 𝛽 are inversely
related for a fixed sample size, n. To reduce both,
there is a need to increase the sample size.
• In practice, researchers traditionally choose the
cut-off (critical region), so that 𝛼 is controlled
(fixed) at 0.05, and then, let 𝛽 be whatever it turns
out to be.
9
 TYPES OF ERRORS INCURRED BY
RESEARCHERS
• The choice of alpha usually Null Hypothesis
depends on the consequences
associated with making a type I Decision True False
error.
• The smaller the value of alpha, Reject 𝐻0 Type I Correct
the smaller will be the rejection Accept 𝐻0 error Decision
region.
Correct Type II
• The rejection region depends
on the form of the alternative decision error
hypothesis.
10
CONSEQUENCES OF INCURRING A
TYPE I ERROR
𝜶 −Value Consequences of a Type I error

0.01 or similar Very serious

0.05 Moderately serious
0.10 Not too serious

If the test leads to the rejection of the null hypothesis, the researcher
can then conclude that there is sufficient evidence supporting the
alternative hypothesis at alpha level of significance. If the test leads to
the acceptance of the null hypothesis, we hold judgment.
11
 STEPS IN HYPOTHESIS TESTING
Step 1: Formulate the null and the alternative
hypotheses.

The null hypothesis is usually that the observations are

the result purely of chance or the hypotheses of no
effect.

The alternative hypothesis is that there is a real effect;

or that the observations are the result of this real effect
plus chance variations. 12
 STEPS IN HYPOTHESIS TESTING
Example
In symbols, 𝐻0 : 𝜃 = 𝜃0 where
𝜃 = a parameter of the population
𝜃0 = hypothesized (expected value) of the parameter
𝜃 versus an appropriate alternative hypothesis
from one of the following:
𝐻1 ∶ 𝜃 ≠ 𝜃0 (two-tailed test or non-directional test)
𝜃 > 𝜃0 (one-tailed or directional test; right-tail)
𝜃 < 𝜃0 (one-tailed or directional test; left-tail)
13
 STEPS IN HYPOTHESIS TESTING
Example
𝐻0 : Smokers and nonsmokers have the same
prevalence of oral cancer.
𝐻1 ∶ The prevalence of oral cancer is not the same
for smokers and nonsmokers. (two-tailed test)

The prevalence of oral cancer is higher

among smokers. (one-tailed test)
14
 STEPS IN HYPOTHESIS TESTING
Step 2: Set the level of significance (𝜶).
• The level of significance refers to the probability level
that is considered to warrant support of the hypothesis
being tested.
• In social and educational researches, the level of
significance is usually set at 0.05.
• Setting alpha at this point means that if the probability
of occurrence of the sample under the null hypothesis
is less than or equal to this, then the sample does not
support the null hypothesis and thus can be rejected.15
 STEPS IN HYPOTHESIS TESTING
Step 3: Select the appropriate test statistic.
• There is a specific test statistic that is appropriate
for each kind of statistical hypothesis and each
test statistic has a probability distribution.
• Test of hypothesis also involves the application
of statistical tests. The choice of the statistical
test depends on the following:
a. Level of measurement (nominal, ordinal,
interval, or ratio) 16
 STEPS IN HYPOTHESIS TESTING
b. Objectives/Purpose of the study – whether the
researcher would like to compare means, or whether to
compare the sample to some population value, or
compare two groups, or more than two groups.
c. Design of the study – whether case control, cohort,
cross sectional, and so on.
d. Whether the samples are related or independent –
in independent samples, the probability of selection of
samples in one group is not affected by the selection in
the other group/s. 17
 STEPS IN HYPOTHESIS TESTING
e. Assumptions on the distribution of the test

Parametric tests can be used if the assumptions about

the parameters in the population like normality,
independence, and homogeneity hold true.
Nonparametric test can be used when assumptions for
the use of parametric tests are questionable in the data.
Nonparametric tests have fewer and less stringent
assumptions about the population parameter. Also called
“distribution-free” tests, applicable to nominal and ordinal
18
 STEPS IN HYPOTHESIS TESTING
Step 4: Establish the critical (rejection) region.
Determine the value (and region) that the test statistic
must attain (or fall into) to be declared statistically
significant, that is, to be able to reject the null hypothesis.
• All absolute values of the test statistics greater than or
equal to these values have very low probability of
occurrence under the null hypothesis.
• If the test statistics computed for the sample data fall in
this region, then there is a basis for rejecting the null
hypothesis. 19
 STEPS IN HYPOTHESIS TESTING
Step 5: Compute the value of the test statistic from the
sample data.
Perform the required calculations in the test statistic (from
step 3) using the sample data on hand and compare the
computed value of the test statistic with the critical values
(from step 4) to determine if it falls within the critical
(rejection) region.

20
 STEPS IN HYPOTHESIS TESTING
Step 6: State your conclusion.
Reject the null hypothesis (conclude 𝐻1 ) if the test
statistic has a value that falls into the critical
(rejection); otherwise, do not reject the null
hypothesis.

21
 TESTING HYPOTHESIS ABOUT A
SINGLE MEAN USING LARGE SAMPLES
• When a researcher wishes to test whether an
acceptable or established mean value for a certain
product is still true then a test of a single population
mean can be used to accomplish the objective.

𝑥−𝜇
𝑧= z-test for a single mean
𝜎/ 𝑛

22
 TESTING HYPOTHESIS ABOUT A
SINGLE MEAN USING LARGE SAMPLES
• In a recent survey of nurses in region I, it was found out
that the average monthly net income of nurses is Php8
048.25. Suppose a researcher wants to test this figure
by taking a random sample of 158 nurses in Region I to
determine whether the monthly net income has
changed. Suppose further that the average net monthly
income of the 158 nurses is Php9 568.40 and the
population standard deviation was found out to be
Php1 563.42 23
• In a recent survey of nurses in region I, it was found out that the
average monthly net income of nurses is Php8 048.25. Suppose a
researcher wants to test this figure by taking a random sample of
158 nurses in Region I to determine whether the monthly net
income has changed. Suppose further that the average net monthly
income of the 158 nurses is Php9 568.40 and the population
standard deviation was found out to be Php1 563.42
Solution:
Step 1: 𝐻0 : The average monthly net income of nurses
is still Php8 048.25 (𝜇 = 𝑃ℎ𝑝8 048.25)
𝐻1 : The average monthly net income is no
longer Php8 048.25 (𝜇 ≠ 𝑃ℎ𝑝8 048.25)
Step 2: 𝛼 = 0.05 two-tailed test (nondirectional)
Step 3: z-test for a single mean 24
• In a recent survey of nurses in region I, it was found out that the
average monthly net income of nurses is Php8 048.25. Suppose a
researcher wants to test this figure by taking a random sample of
158 nurses in Region I to determine whether the monthly net
income has changed. Suppose further that the average net monthly
income of the 158 nurses is Php9 568.40 and the population
standard deviation was found out to be Php1 563.42
Step 4: Reject 𝐻0 if the computed z-value is greater than
1.96 or less than -1.96.
Step 5: 𝑧 = 12.22 (reject 𝐻0 )
Step 6: Statistically, the researcher has enough
evidence to reject the null hypothesis.
Conclusion: The average monthly net income of nurses
is no longer Php8 048.25. 25
Testing Hypothesis About a Single Mean Using
Small Samples: 𝝈 Unknown
• Suppose a researcher is testing a
hypothesis about a single population
mean, however, due to some reasons, he
was able to gather a small random
sample 𝑛 < 30 of data.
ҧ
𝑥−𝜇
𝑡 = 𝑠 where 𝑑𝑓 = 𝑛 − 1
𝑛

26
Example: A certain brand of laundry soap is
advertised to have a net weight of 500
grams. If the net weights of a random
sample of 10 boxes are 495, 503, 507, 498,
490, 505, 510, 502, 493, and 506 grams, can
it be concluded that the average net weight
of the boxes is less than the advertised
amount? Use 𝛼 = 0.01.
Step 1: 𝑯𝟎 : 𝝁 = 𝟓𝟎𝟎 𝒈𝒓𝒂𝒎𝒔
𝑯𝟏 : 𝝁 < 𝟓𝟎𝟎 𝒈𝒓𝒂𝒎𝒔
Step 2: 𝛼 = 0.01 ; one-tailed test 27
Example: A certain brand of laundry soap is advertised to have
a net weight of 500 grams. If the net weights of a random
sample of 10 boxes are 495, 503, 507, 498, 490, 505, 510, 502,
493, and 506 grams, can it be concluded that the average net
weight of the boxes is less than the advertised amount? Use 𝛼 =
0.01.
Step 3: t-test
ҧ
𝑥−𝜇
𝑡 = 𝑠 where 𝑑𝑓 = 𝑛 − 1
𝑛
Step 4: Reject 𝑯𝟎 if the computed t-value at
nine degrees of freedom is less than 2.81.
28
Example: A certain brand of laundry soap is advertised to have
a net weight of 500 grams. If the net weights of a random
sample of 10 boxes are 495, 503, 507, 498, 490, 505, 510, 502,
493, and 506 grams, can it be concluded that the average net
weight of the boxes is less than the advertised amount? Use 𝛼 =
0.01.
Step 5:
ҧ
𝑥−𝜇
𝑡= 𝑠 where 𝑑𝑓 = 𝑛 − 1
𝑛
𝟓𝟎𝟎. 𝟗 − 𝟓𝟎𝟎
𝒕=
𝟔. 𝟔𝟏/ 𝟏𝟎
𝒕 = 𝟎. 𝟒𝟑𝟎𝟔
29
Example: A certain brand of laundry soap is advertised to have
a net weight of 500 grams. If the net weights of a random
sample of 10 boxes are 495, 503, 507, 498, 490, 505, 510, 502,
493, and 506 grams, can it be concluded that the average net
weight of the boxes is less than the advertised amount? Use 𝛼 =
0.01.
Step 6: Since the computed t-value of
0.4306 is lower than the critical t-value of
2.821 at 0.01 level of significance with 9
degrees of freedom, then there is no
sufficient evidence to reject the null
hypothesis.
30
 TESTING A HYPOTHESIS ABOUT A
PROPORTION The z-test of population
proportion is given as
A proportion, P, is a value between zero
𝒑−𝑷
and one (0 ≤ 𝑃 ≤ 1) that expresses the 𝒛=
part of the whole that has a given 𝒑𝒒
𝒏
characteristic. Comparatively, means are
computed by averaging measurements Where:
while proportions are calculated by 𝑝 is the sample
proportion
counting the number of items in the
𝑃 is the population
population that have a characteristic and proportion
then dividing the number by the total. 𝑞 is equal to 1 − 𝑝
31
Example: The sales manager at Video City
estimates that 60% of its video rentals are
returned on the same day they are
borrowed. A random sample of 115 rental
returns showed that 95 of them were in fact
returned on the same day. Can we conclude
at 95% confidence that the proportion is not
equal to 60%?
Step 1: 𝑯𝟎 : 𝑷 = 𝟎. 𝟔𝟎
𝑯𝟏 : 𝑷 ≠ 𝟎. 𝟔𝟎
Step 2: 𝛼 = 0.05 ; two-tailed test 32
Example: The sales manager at Video City estimates that 60% of its video
rentals are returned on the same day they are borrowed. A random
sample of 115 rental returns showed that 95 of them were in fact returned
on the same day. Can we conclude at 95% confidence that the proportion
is not equal to 60%?
Step 3: test statistics = z-test of P
Step 4: Reject 𝑯𝟎 if the computed z-value is
greater than 1.96 or lesser than -1.96.
𝒑−𝑷
Step 5: 𝒛 = 𝒑𝒒
𝒏
𝟎. 𝟖𝟑 − 𝟎. 𝟔𝟎
𝒛= = 𝟓. 𝟎𝟒
𝟎. 𝟔 𝟎. 𝟒
𝟏𝟏𝟓 33
Example: The sales manager at Video City estimates that 60%
of its video rentals are returned on the same day they are
borrowed. A random sample of 115 rental returns showed that
95 of them were in fact returned on the same day. Can we
conclude at 95% confidence that the proportion is not equal to
60%?
Step 5: Reject 𝐻0 since the computed z-value
5.04 exceeds the critical value of 1.96.
Conclusion:
The estimated proportion of video rentals
returned on the same day is not equal to
60%.
34
TESTING A HYPOTHESIS ABOUT A
VARIANCE
In situations where a researcher wants to
test a hypothesis of population variance of
which normalcy of the population is
assumed, the formula below is used:
𝑛−1 𝑠 𝑠
𝑋2 = where: 𝑑𝑓 = 𝑛 − 1
𝜎2

35
Example: A local manufacturing company
produces lead pencils. One batch of these lead
pencils is tagged to be 5.5 inches long. A quality
controller inspects 16 of the lead and found out
that on the average, the lead pencils are about
5.5 inches long. However, the measurements
vary. Supposing the variance is 1.12, can we
reject the hypothesis that the variance is 1.02 at
0.10 level of significance?
Step 1: 𝑯𝟎 : 𝝈 = 𝟏. 𝟎𝟐
𝟐

𝑯𝟏 : 𝝈𝟐 ≠ 𝟏. 𝟎𝟐
36
Example: A local manufacturing company produces lead pencils. One
batch of these lead pencils is tagged to be 5.5 inches long. A quality
controller inspects 16 of the lead and found out that on the average, the
lead pencils are about 5.5 inches long. However, the measurements vary.
Supposing the variance is 1.12, can we reject the hypothesis that the
variance is 1.02 at 0.10 level of significance?
Step 2: 𝜶 = 𝟎. 𝟏𝟎; two-tailed test
Step 3: 𝑿 = test for a variance
𝟐

Step 4: Reject 𝑯𝟎 if the computed 𝑿𝟐 -value

exceeds 24.9956 or if it is less than 7.26094
𝑑𝑓 = 𝑛 − 1 → 16 − 1 = 15
2
𝑋 (1−0.05),15 → 7.26094
2
𝑋 0.05,15 → 24.9958 37
Example: A local manufacturing company produces lead pencils. One
batch of these lead pencils is tagged to be 5.5 inches long. A quality
controller inspects 16 of the lead and found out that on the average, the
lead pencils are about 5.5 inches long. However, the measurements vary.
Supposing the variance is 1.12, can we reject the hypothesis that the
variance is 1.02 at 0.10 level of significance?
𝒏−𝟏 𝒔𝟐
Step 5: 𝑿𝟐 =
𝝈𝟐
𝟐
(𝟏𝟔 − 𝟏)(𝟏. 𝟐𝟓𝟒𝟒)
𝑿 = = 𝟏𝟖. 𝟒𝟒𝟕
𝟏. 𝟎𝟐
Step 6: Since the computed 𝑿 of 18.447 is 𝟐

within the nonrejection region, the variance

in length is 1.02.
38
TWO SAMPLE TESTS OF SIGNIFICANCE
CONCERNING MEANS
• Tests that are included are: test between
proportion (z-test) and test for independent
and dependent samples (t-test).
• There are two types of commonly used test
of significant difference between two
samples, the independent t-test and the
correlated t-test.
39
TWO SAMPLE TESTS OF SIGNIFICANCE
CONCERNING MEANS
• These techniques are used for a small sample,
that is, when 𝑛 < 30. These tests help
researchers or individuals to decide the nature
of interrelatedness or differences between two
samples being studied.
• The independent t-test is used when the two-
sample means are taken from a separate groups
of respondents or populations.
40
TWO SAMPLE TESTS OF SIGNIFICANCE
CONCERNING MEANS
𝑋ത1 − 𝑋ത2
𝑡=
𝑆1 2 𝑆2 2
−
𝑛1 𝑛2
Where: 𝑋ത1 = Mean of the first group
𝑋ത2 = Mean of the second group
𝑆1 2 = Variance of the first group
𝑆2 2 = variance of the second group
𝑛1 = number of cases in the first group
𝑛2 = number of cases in the second group

41
Example: In a study conducted to determine the
research skills of assistant and associate professors
in state universities and colleges, the following data
represent the mean scores of the professors:
Assistant Prof. 𝑋1 3.00 4.20 2.75 3.5 4.5 2.5 2.6
Associate Prof. 𝑋2 3.75 4.50 3.50 4.0 4.0 3.5 3.0
Step 1: 𝑯𝟎 : There is no significant difference
between the research skills of assistant and
associate professors.
𝑯𝟏 :There is a significant difference between
the research skills of assistant and associate
professors.
42
Step 2: 𝜶 = 𝟎. 𝟎𝟓; two-tailed test
Step 3: Reject 𝑯𝟎 if the computed value is greater than
1.782.
Step 4: Compute the value of the test statistic
Asst. Professor Assoc. Professor
( 𝑋1 ) ( 𝑋2 )
3.00 3.75 𝑋ത1 = 3.29 𝑋ത2 = 3.
4.20 4.50 𝑆1 2 = 0.64 𝑆2 2 = 0.23
2.75 3.50
3.50 4.00
4.50 4.00
2.50 3.50
2.60 3.00
෍ 𝑋1 = 23.05 ෍ 𝑋2 = 26.26 43
 𝑋ത1 − 𝑋ത2
𝑡=
2 2
𝑆1 𝑆2
−
𝑛1 𝑛2
3.29 − 3.75
𝑡=
0.64 0.23
+
7 7
0.46
𝑡=− = −1.3143
0.12
Note: Disregard the negative sign in the
interpretation of data.

44
The degree of freedom:
𝑑𝑓 = 𝑛1 + 𝑛2 − 2
𝑑𝑓 = 7 + 7 − 2
𝑑𝑓 = 14 − 2 = 12
Step 5: In as much as the computed t-value of
1.3143 is lower than the critical value, then there is no
sufficient evidence to reject the null hypothesis.
Step 6: Since the computed t-value is lower than the
critical t-value, it means that there is no significant
difference between the research skills of assistant
and associate professors. Thus, they manifest
comparable research skills.
45
Step 1: 𝑯𝟎 : There is no significant difference in the
mean post-test performance of the two groups of
students classified according to their learning
styles.
𝑯𝟏 :There is a significant difference in the
mean post-test performance of the two groups of
students classified according to their learning
styles.
Step 2: 𝜶 = 𝟎. 𝟎𝟓; two-tailed test
Step 3: Reject the null hypothesis if the computed
t-value is greater than the critical t-value of 1.734 at
𝛼 = 0.05 with 18 degrees of freedom. 46
Step 4: t-test for independent samples
Convergers Assimilators
𝑋ത1 = 26.26 𝑋ത2 = 20
2 2
𝑆1 = 8.67 𝑆2 = 8.92
𝑛1 = 10 𝑛2 = 10
𝑋ത1 − 𝑋ത2
𝑡=
𝑆1 2 𝑆2 2
−
𝑛1 𝑛2
26.26 − 20
𝑡=
8.92 8.97
+
10 10
𝑡 = 4.71
47
The degree of freedom:
𝑑𝑓 = 𝑛1 + 𝑛2 − 2
𝑑𝑓 = 10 + 10 − 2
𝑑𝑓 = 18
Step 5: Reject the null hypothesis since the
computed t-value of 4.71 is greater than 1.734.
Conclusion: There exists a significant difference
between the mean post-test performance of the
two groups of students classified according to their
learning styles after their exposures to simplified
instructional materials in Basic Statistics.
48
THE T-TEST FOR DEPENDENT
SAMPLES
• The t-test dependent samples is applied to
matched pairs or correlated samples.
• For example, in the research study on the degree of
seriousness of problems encountered by college
freshmen, data were taken before and after their
individual counseling sessions. From the 15-item
problem checklist, the corresponding degree of
seriousness of problems encountered by the
students before and after their individual sessions
comprises the data to be compared. This is referred
to as repeated measures.

49
THE T-TEST FOR DEPENDENT
SAMPLES
σ𝐷
𝑡=
𝑛 σ 𝐷2 − σ 𝐷 2
𝑛−1

Where:
𝐷 =difference
𝑛 =number of cases
50
Example: Consider the data below to determine if
there exists a significant difference in the degree of
seriousness of problems encountered by the
college freshmen.
Student Before After
1 3.00 4.00
2 3.25 3.50
3 3.00 3.50
4 2.50 3.60
5 2.75 3.45
6 2.50 2.75
7 2.75 4.75
8 3.75 4.00
9 3.50 3.51
10 3.60 4.50 51
THANK YOU