Hypothesis Testing

Learning Objectives: To explain why there is a need to test the hypothesis and to identify the
test statistics to be used in testing the significance of difference between means.

Performance Standard:
· able to formulate null and alternative hypotheses,
· identify the appropriate test statistics to be used, and
· be able to perform simple tests of hypotheses.

Topics:
• Test of Significance
• Hypothesis: Null and Alternative
• Hypothesis Testing
• Level of Significance, Errors & Rejection/Acceptance Region
• Hypothesis Testing Approaches: Critical Value, p-Value
• The 5-step Solution
• Z-test and t-test

Test of Significance
Once sample data has been gathered, statistical inference allows researchers to assess
evidence in favor of some claims about the population from which the sample has been drawn.
The methods of inference used to support or reject claims based on sample data are known as
test of significance. In English, “significant” means important, while in Statistic “significant”
means probably true. When Statisticians say the result is “highly significant” they mean it is
probably true. They do not necessarily mean it is highly important. (Blay, n.d.)

According to BYJU’s (Tests of Significance, 2019), in statistics, it is important to know if

the result of an experiment is significant enough or not. In order to measure the significance, there
are some predefined tests which could be applied. These tests are called the tests of
significance or simply the significance tests.
This statistical testing is subjected to some degree of error. For some experiments, the
researcher is required to define the probability of sampling error in advance. In any test which
does not consider the entire population, the sampling error does exist. The testing of significance
is very important in statistical research.
The significance level is the level at which it can be accepted if a given event is statistically
significant. This is also termed as p-value. It is observed that the bigger samples are less prone
to chance, thus the sample size plays a vital role in measuring the statistical significance. One
should use only representative and random samples for significance testing.
In short, the significance is the probability that a relationship exists. Significance tests tell
us about the probability that if a relationship we found is due to random chance or not and to
which level. This indicates the error that would be made by us if the found relationship is assumed
to exist.

In Descriptive Statistics, we have learned that when Group 1 for instance got an average
score of 90 while Group 2 has an average score of 85, we readily conclude that Group 1
performed better than Group 2. It is because Group 1 has a higher mean score than Group 2. But
it is a different thing in Inferential Statistics, we need to search for more evidence for saying so
beyond what numbers tell us. We need to find out the difference “5” (90-85) is big enough for us
to conclude that Group 1 really performed better than Group 2 by TESTING THE SIGNIFICANCE
OF DIFFERENCE between these two mean scores using some test statistic which we are about
to discuss. If the difference is found to be significant, it is only then that we can make a conclusion
that Group 1 performed better than Group 2.

Testing the significance of difference can also be done between two portions. Forexample,
if survey says 40% of respondents are in favor of divorce while 47% are not in favor, with 13%
undecided, we cannot just conclude that the difference between the proportion of favorand not in
favor is significant unless we do testing. But we cannot just test this difference. We need to write
it in the form of HYPOTHESIS, thus the process is called “HYPOTHESIS TESTING”.

Hypothesis tests are not limited to means and proportions. It is also conducted in
regression and correlation analysis to determine if the regression relationship and the correlation
coefficient (r) are statistically significant. A goodness of fit test refers to hypothesis test in which
the null hypothesis is that the population has a specific probability distribution, such as a normal
probability distribution. Nonparametric statistical methods involve a variety of hypothesis-testing
procedures.

Hypotheses: Null and Alternative

Hypothesis is an educated guess about the population parameter. It is an assumption
about the population parameter. Statistical Hypotheses is a guess or prediction made by the
researcher regarding the possible outcome of the study. Hypothesis testing is the process of
making an inference or generalization on population parameters based on the results of the study
on samples.

Null Hypothesis (Ho) is always hoped to be rejected. It always contains the “=” sign (≤ or
≥). Whilethe Alternative Hypothesis (Ha) challenges Ho. It never contains the “=” sign. Uses “< or
> or ≠. Itgenerally represents the idea which the researcher wants to prove.

Every test of significance begins with a Null Hypothesis. It represents a theory that has
been put forward, either because it is believed to be true or because it is to be used as a basis
for argument but has not been proved. The Alternative Hypothesis is a statement of what a
statistical hypothesis test is set up to establish.

The logic of traditional hypothesis testing requires that we set up two competing
statements in the form of null hypothesis and alternative hypothesis. These hypotheses are
mutually exclusive and exhaustive. You will have no problem in the null hypothesis since it always
uses “=” sign. Your little trouble is the alternative, since you have three choices, “< or > or ≠. But
do not worry there are glaring clues.
Hypothesis Testing
Hypothesis testing is a procedure for deciding if the null hypothesis should be rejected in
favor of an alternative hypothesis or will not be rejected. There are types of hypothesis tests that
are determined by the alternative hypothesis. An alternative hypothesis may be one-sided
(one-tail right or left directional) or two-sided (two-tail non directional) depending upon the
problem. A one-tail left/right test is used if Ha claims that a parameter is smaller/greater than the
value given by the null hypothesis. A two-tail test claims that a parameter is simply not equal to
the value given by the null hypothesis, thus the direction does not matter. In short, If Ha uses >
or <, then the hypothesis test is One-Tail right or left directional respectively, and if Ha uses ≠, the
hypothesis test is Two-Tail Non-directional.

Hypotheses are always stated in terms of population parameters, such as “μ” for mean
and “P” for proportion.

Title of study Problem/Objective Null Hypothesis Alternative Hypothesis

1. An Evaluation of To determine if online Ho: μ = 80 Ha: μ > 80

the Effectiveness learning has The average GPA of The average GPA of
Online Learning significantly increased students in WIX students in WIX
the average GPA College is equal to 80 College is greater than
(Grade Point average) or Online learning has 80 or Online learning
of students in WIX not significantly has significantly
College from the increased the increased the average
known GPA which is average GPA of GPA of students in
80. The GPA of 200 students in WIX WIX College.
randomly selected College
students was found to
be 83.

2. The Effects of To determine if there is Ho: μ after = μ before Ha: μ after ≠ μ before
New a significant difference There is no significant There is a significant
Compensation on the job satisfaction difference between difference between the
Package on the level of employees the level of job level of job satisfaction
Job Satisfaction of before and after the satisfaction of of employees in Banko
Banko Metro new compensation employees in Banko Metro before and after
Employees package is introduced. Metro before and the new compensation
after the new package is introduced.
compensation
package is
introduced.

3. An Evaluation of To determine the Ho: P = 0.10 Ha; P < 10%

Hotel King & proportion of The proportion of The proportion of
Queen Customer dissatisfied hotel dissatisfied hotel dissatisfied hotel
Relations and guests is significantly guests is equal to guests is lower than
Services lower than 10% 10% 10%
 In study number 1, the hypothesis on means is one-tail right directional. Ha uses >,
because the researcher is interested in knowing if Online learning has significantly increased the
average of GPA of WIX students.

While the study number 2 the hypothesis on means is two-tail non-directional. Ha uses ≠,
because the manager is just interested to know if there is a significant difference on the job
satisfaction level of employees before and after the new compensation package is introduced.
Whether it has increased or decreased will be seen after the test has been done.

On the other hand, the study number 2 the hypothesis on proportion is one-tail left-
directional. Ha uses < because the human resource researcher wants to find out if the number of
dissatisfied customers is significantly lower than 10%.

Once the test has been carried out the decision is always given the Null Hypothesis. We
either “Reject Ho in favor of Ha” or “Do not Reject Ho”; we never say, “reject Ha”, or even
“accept Ha”. If we decide “not to reject”, this does not necessarily mean that the null hypothesis
is true, it only suggests that there is not sufficient evidence against Ho in favor of Ha. Rejecting
the null hypothesis then, suggests that the alternative hypothesis may be true.

Level of Significance, Errors and Rejection/Acceptance Region

Below is taken from the discussion of level of significance from the website of Statistics by
Jim (Frost, n.d.) The significance level, also denoted as alpha or α, is a measure of the strength of
the evidence that must be present in your sample before you will reject the null hypothesis and
conclude that the effect is statistically significant. The researcher determines the significance level
before conducting the experiment.

The significance level is the probability of rejecting the null hypothesis when it is true. For
example, a significance level of 0.05 indicates a 5% risk of concluding that a difference exists
when there is no actual difference. Lower significance levels indicate that you require stronger
evidence before you will reject the null hypothesis.

Use significance levels during hypothesis testing to help you determine which hypothesis
the data support. Compare your p-value to your significance level. If the p-value is less than your
significance level, you can reject the null hypothesis and conclude that the effect is statistically
significant. In other words, the evidence in your sample is strong enough to be able to reject the
null hypothesis at the population level.

Another interpretation of the significance level α, based in decision theory, is that it

corresponds to the value for which one chooses to reject or not to reject the null hypothesis Ho.
In decision theory, is known as a Type 1 error. The probability of a Type error I is equal to the
significance level alpha, and the probability of Type II is equal to β(1- α). The following illustrationwill
help you comprehend the meaning of these hypotheses testing.
Source: https://microbenotes.com/type-i-and-type-ii-error/

So, it is very clear that the Level of Significance is the probability of committing Type I
Error, this is the basis to Reject or Not to Reject Ho and in short it is the area of the Rejection
Region.

Usually, researchers use either the 0,05 level or sometimes called 5% level, written as α
=0.05 or the 0.01 level or 1% level, written as α =0.01, although the choice of levels is largely
subjective. The lower the significance level, the more the data must diverge from the null
hypothesis to be significant. Therefore, the 0.01 level is more conservative than the 0.05 level.
The next illustrations show most of the concepts that we defined and discussed previously. These
include: Types of hypothesis Test (One-Tail/Two Tail Test), Level of Significance “α“,Rejection
and Acceptance Regions.

Acceptance Rejection
Rejection
Region 95% Region
Region
Area = 0.025 Area = 0.025

0.025 - Rejection Region, 2 Tail non-directional, Ha uses not equal sign/symbol

An α =0.05 means, “The probability of being right is 95%, and probability of being wrong is 5%.
 Rejection
Region Acceptance
Area = 0.01 Region 99%

0.01 - Rejection Region, One-Tail Left Directional , Ha uses < symbol

An α =0.01 means, “The probability of being right is 99%, and probability of being wrong is 1%.

Hypothesis Testing Approaches

1. The Critical Value Approach - One way of deciding whether to reject Ho is by comparing
the value of the test statistics with the critical value. The critical value that test statistics (Z
or T) must exceed for the null hypothesis to be rejected. We reject Ho if the absolute value
of the computed Z or T ≥ the absolute value of the critical value.
Please watch this video to understand the z-test https://youtu.be/oEW8Hd_xy1k

2. The P-Value Approach - We compare p-value with the level of significance or alpha, thus
making our simpler. The p-value is often called the observed level of significance for the test.
Again, p-value refers to probability or the expected value that the phenomenon is likely to
occur. If the hypothesis is to be tested at α = 0.05, then the area of the rejection is 0.05. In this
approach, comparison is made between “α” and p-value. We reject Ho if the p-value ≤ α
(0.05), otherwise Ho cannot be rejected. The decision rule is in this process is “Reject Ho if
the p-value is less than or equal to α.
Please watch this video to fully understand this concept
https://www.youtube.com/watch?v=-FtlH4svqx4

3. The 5-Step Solution - These five steps will guide us in the hypothesis testing process.
a. Critical Value Approach
1. Ho:
Ha:
2. α = ; critical value = ; 1-T/2-T
3. Test Statistic Reject Ho if the [computed-value] ≥ [critical-value]
4. Decision: Reject/Do not reject Ho because………
5. Conclusion:
 b. p-value vs α
1. Ho:
Ha:
2. α = ; critical value = ; 1-T/2-T
3. Decision rule: Reject Ho if p-value ≤ α
4. Decision: Reject/Do not reject Ho because………
5. Conclusion:

Please watch this video to supplement your understand about critical value approach and
p-value approach part 1: https://youtu.be/4mDemwLHzbo and part 2:
https://youtu.be/JkLCPOZcc4A

Testing the Significance of Difference Between/Among Means

The three known statistical hypothesis tests for means are T-test, Z-test and F-Test or
ANOVA. Please see illustration below

Statistical Hypotheses Test

Testing the significance of Difference Between/Among Means

Z – test Normal
n ≥ 30 means Z- distribution

t-test Student
n < 30 & n ≥ 30 2 means T- distribution

F-test 2 or more F
(ANOVA) means distribution

Z-test is a statistical hypothesis test that follows a normal distribution, the T- test follows a
Student’s T-distribution while the F-Test or ANOVA follows an F-distribution.

Z-test & T-test are both basically the same. They compare between two means to suggest
whether one or both samples come from the same population. Z-tests are often applied to
enlarged samples n>30 if the distribution is normal. It is used if these conditions are met;
otherwise, other statistical tests like T-test are applied in substitute. Z-test is used to test if there
is a significant difference between one sample mean and population mean and for two sample
means. Please watch this video to fully understand this concept
https://www.youtube.com/watch?v=5ABpqVSx33I
 T-test is also referred to as “Student T-test”. It is the most used in Statistical Data Analysis
procedure for hypothesis testing since it is straightforward and easy to use. Additionally, it is
flexible and adaptable to a broad range of circumstances. T-test is best applied, if you have a
limited sample size n < 30 if the variables are approximately normally distributed and the variation
of scores in the two groups is not reliably different. It is however noted that if ‘n'' becomes larger,
the T-distribution becomes close to Z- distribution. Thus, T-test can be used not only if n<30 but
also if n is large or n ≥ 30 and if the populations’ standard deviation (σ) is not known.

The t-distribution is like the Z-distribution. They are both symmetrical about the mean.
Both are bell shaped, but the t-distribution is more variable since t values depend on the
fluctuation of the mean and standard deviation, whereas the Z values depend only on the
fluctuation of the mean from sample to sample. The two distributions differ in standard deviation,
Z has a standard deviation of 1 while t has a standard deviation which is always greater than 1.

The divisor n-1 in the formula for variance and standard deviation is called the degrees of
freedom (df). Degrees of freedom is the number of variables which are free to vary. If the mean
and standard deviation are computed from samples of size n, the values t is said to belong to a t
distribution with df = n-1. Therefore, we have a different t curve for each possible sample size,
such that a curve becomes more and more like the standard normal curve as n becomes larger
or as n approaches infinity.

References:

Blay, B. E. (n.d.). Elementary Statistics.

Frost, J. (n.d.). Significance level - Statistics By Jim. Retrieved October 22, 2020, from

http://statisticsbyjim.com/glossary/significance-level/

Lorenzana, E. (2020). Statistical Methods.

Tests of Significance. (2019, October 10). BYJU’S. https://byjus.com/maths/tests-of-

significance/