Hypothesis Testing

-is basically a statement about the target

population. Hypotheses are formulated
because of years’ experience, observations, and
researchers.
Three general statements to state the null and alternative hypothesis:

1. The null hypothesis suggests that there are no signi7icant differences

between compared entities. If the null hypothesis involves numerical
value of parameters, then the null hypothesis would be the equality
of the compared parameters.
2. The alternative hypothesis is the negation of the null hypothesis. It
simply suggests that there is a signi7icant difference between
compared objects or values.
3. The null and alternative hypotheses are constructed in a way that
when we reject the null hypothesis, there is no other possibility other
than to accept the alternative hypothesis. With this, the hypothesis
proposed by researchers which is usually in favor of the alternative
hypothesis has been supported.
Example:

Scenario, a defendant is brought to a court room for trial.

(This scenario similar to any case where hypothesis testing
is conducted). Since the defendant should be proven guilty
beyond reasonable doubt, we may put the part of
“hypothesis testing” for the judge to test with the purpose
to detect guilt.

𝐻! : 𝑡ℎ𝑒 𝑑𝑒𝑓𝑒𝑛𝑑𝑎𝑛𝑡 𝑖𝑠 𝑖𝑛𝑛𝑜𝑐𝑒𝑛𝑡

𝐻" ∶ 𝑡ℎ𝑒 𝑑𝑒𝑓𝑒𝑛𝑑𝑎𝑛𝑡 𝑖𝑠 𝑔𝑢𝑖𝑙𝑡𝑦
As the trial proceed, several cases of decision could be made.
1.The defendant is innocent but was convicted.
2.The defendant is innocent and was acquitted.
3.The defendant is guilty and was convicted.
4.The defendant is guilty but was acquitted.
With the above decisions, it is evident that some are incorrect. This leads
us to possible errors committed during hypothesis testing. Similarly, we
may have the possible situations in hypothesis testing.
1. We have rejected the null hypothesis when it is true.
2. We have accepted the null hypothesis when it is true.
3. We have rejected the null hypothesis when it is false.
4. We have accepted the null hypothesis when it is false.
Types of Error

Type I Error: Rejection of the null hypothesis

when in fact it is true.

Type II Error: Acceptance of the null

hypothesis when in fact it is false.
The decisions
from a hypothesis Table 7.1
testing could be Decision Ho is True Ho is false
summarized as Reject Incorrect Correct
decision decision
the table below (Type I Error)
indicates. Accept Correct Incorrect
decision decision
(Type II Error)
Level of Significance (𝛼 − 𝑙𝑒𝑣𝑒𝑙)
Is the probability of rejecting the null hypothesis when it is true. With a
higher 𝛼 − 𝑙𝑒𝑣𝑒𝑙, it will provide a greater probability of committing type I
error. Want we want is to put the 𝛼 − 𝑙𝑒𝑣𝑒𝑙 to the minimum as we can.

For instance, if we may have 𝛼 − 0.01, this mean that we are 99% conFident
in our decision. Comparing to the case of choosing 𝛼 − 0.05 which only gives
95% accuracy the latter alpha level gives us more conFidence in making a
correct decision. There are common alpha levels used in research, 0.01, 0.05
and 0.1. We may say that the 𝛼 − 𝑙𝑒𝑣𝑒𝑙 gives the certainty of a decision made.
The Two Forms of Alternative hypothesis

1. Non-Directional Alternative Hypothesis. States that there is a

significant di;erence between compared units but does not
imply any direction of the di;erence nor gives the source and
magnitude of the di;erence.
2. Directional Hypothesis. States the direction of the di;erence.
As like for comparing two objects A and B, we may say that
A<B or A>B.
The critical Regions
Critical regions are the regions where null hypothesis is rejected. The
three consecutive figures below show the critical and acceptance region
in the z-test distribution at 0.05 level of significance.
To Perform hypothesis testing, the following steps are usually followed:

1. Formulate and state the null hypothesis and alternative hypothesis.

These hypotheses should be in their appropriate forms.
2. Decide on the level of significance.
3. Determine the type of test statistics, critical values (tabular values) and
critical regions or acceptance regions.
4. Compute for the test statistics.
5. Make your decision whether to accept your null hypotheses or reject it
based on the computed test statistics referring to the critical values and
critical regions or acceptance region. If the absolute value of the
computed value of the test statistic is greater than the absolute value of
the critical value, then reject and accept Ha. Otherwise, accept Ho and
reject Ha.
Some Parametric Test

Hypotheses are constructed depending on the

nature of the problems and the underlying
assumptions about the population. Different
parametric tests are used depending on the
assumptions about the parameters.
Z-Test
When a population is normal, the distribution of its sample means regardless of its
size is also normal. On the other hand, if the population is skewed, the sampling
distribution of the sample means is approximately normal when sample sizes are
sufficiently large. For both of these cases if the population is normal or at least
approximately normal with sufficiently large sizes, testing the hypothesis should be
known to the researcher.
Table 1.2
The table below shows the Critical Values of z at varying level of significance
diVerent critical values of z in (𝛼 )
hypothesis testing at 0.10, 0.05, Level of One-tailed Two-tailed
0.025 and 0.01 levels of significance (𝛼 ) test Test
significance. Note that z-test is 0.10 Z= 1.28 Z=1.64
just one of the test statistics 0.05 Z=1.64 Z=1.96
that can be used in hypothesis 0.025 Z=1.96 Z=2.33
testing. 0.01 Z=2.33 Z=2.58
Using z-test for single samples
One of the common parameters that are
estimated for a population from any
sample is its mean. Now consider a
x̄ !"K
hypothesis testing about the mean of a
population with its variance known.
Z= L
Given the hypothesized mean 𝜇! for the M
population together with the
corresponding sample mean x̄ , the critical
region can be determined using the test
statistics.
Using z-test for two samples

We may not restrict the case of

having to compare two samples
coming from two di6erent (̄ x; (̄ x< )
populations. Having the same z= =; =<
assumptions as with single samples, )
the two samples should also be of >; ><
large sizes and are normal or at least
approximately normal.
Example#1:

A random sample of 100 suspected leptospirosis patients

in Metro Manila last year shows that the symptoms
exhibit within 72 hours. Does this project that the
symptoms would exhibit earlier than the common 80
hours with a standard deviation of 16 hours? Test the
hypothesis at 𝛼 =0.05 level of significance.
Example#1:
Following the steps in hypothesis testing we have the
following results.
1. The null hypothesis, Η# : 𝜇 = 80 ℎ𝑜𝑢𝑟𝑠 The
alternative hypothesis Η$: < 80 ℎ𝑜𝑢𝑟𝑠 (𝑜𝑛𝑒 𝑡𝑎𝑖𝑙𝑒𝑑)
2. 𝛼 =0.05 level of signiFicance was speciFied
A random sample of 100 suspected
leptospirosis patients in Metro Manila 3. Since n=100 we can use z-test. Using table 1.2,
last year shows that the symptoms Z%&'('%)* = 1.64
exhibit within 72 hours. Does this project
that the symptoms would exhibit earlier 4. Z computed value
than the common 80 hours with a x̄ +, 72+-.
standard deviation of 16 hours? Test the Z= " ! = $% = -5
hypothesis at 𝛼 =0.05 level of # $&&
significance. 5. Decision: Since the absolute value of z-value = 5 is
greater than the absolute value of the critical value of
z=1.64 then we reject the null hypothesis and accept
alternative. This means that symptoms of leptospirosis
exhibit in less than 80 hours.
Example #2:

A program organizer wishes to determine which among two popular

local bands A and B is more favored by teenagers in Pangasinan.
Two random samples of 250 and 300 teenagers were gathered in
two concert areas in Pangasinan. Band A performed in the 1st group
and band B performed 2nd group. After their performance, the
audience evaluated the two bands from 1-10with 10 as the highest.
Band A scored 8.73 with a population standard deviation of 1.15 and
band B scored 8.85 with a population standard deviation of 0.80.
Does this mean that band B was more favored? Let alpha level be at
0.05.
Example #2:
Following the steps in hypothesis testing we have the
A program organizer wishes to following results.
determine which among two 1. Η# : 𝜇/ = 𝜇0 , Η$: 𝜇/ < 𝜇0 (𝑜𝑛𝑒 𝑡𝑎𝑖𝑙𝑒𝑑)
popular local bands A and B is
more favored by teenagers in 2. 𝛼 =0.05 level of signiFicance was speciFied
Pangasinan. Two random samples
of 250 and 300 teenagers were 3. We are going to use z-test for two samples since the
gathered in two concert areas in two populations are both of large size. Based on
Pangasinan. Band A performed in
the 1st group and band B performed Table 1.2, the critical value of z=1.64
2nd group. After their performance, 4. Z computed value (̄ x "̄ x )
the audience evaluated the two ! " (%.'("%.%))
bands from 1-10with 10 as the z= #! #"
=
highest. Band A scored 8.73 with a $ !.!&" '.("
population standard deviation of $! $" $
"&' )''
1.15 and band B scored 8.85 with a
population standard deviation of
0.80. Does this mean that band B 5. Decision: Since the absolute value of z-value = 1.39 is
was more favored? Let alpha level less than the absolute value of the critical value of
be at 0.05.
z=1.64 then we accept Ho. Hence, we cannot say then
that band B is more favored than A.