Module 1.

Introduction to Nonparametric Statistics 1

1
Introduction to Nonparametric
Statistics
We are often confronted with a lot of information every day. We classify and
process them according to our own standards and inclinations. In this section, we will try
classify and process those information according to a more systematic and scientific way.
We will recall some of our knowledge regarding the “what- about” and “how- about” of
information or data processing to which we termed as Statistics. But this time, we will
focus only on Nonparametric Statistics.

SELF-CHECK!!!

After finishing this module, you are expected to achieve the following
objectives
1. Describe Nonparametric Statistics
2. Compare Parametric and Nonparametric Statistics
3. Identify the Preconceived Advantages and Disadvantages of
Nonparametric Statistical Tests
4. Differentiate –value from Tabular Value
5. Recall the Steps in Hypothesis Testing
Module 1. Introduction to Nonparametric Statistics 2

LEARNING RESOURCES
For further readings, you may refer to the following books;

Bee Ching U. Ong Kian Koc and Vivien M. Talisayon (1998). Seminar in Advance
Statistics Applied to Education. Quezon City: UP Open University. (Modules 1-8)
Hair, J., Anderson, R., Tatham, R. and Black, W.(1998).Multivariate Data
Analysis with Readings. New Jersey: Prentice Hall, Inc.
Redecilla Glenn (2020). Parametric and Nonparametric Statistics for Statistician
Or view Nonparametric Statistics with Illustrations by: Gabino P. Petilos, Ph.D at
http://www.causeweb.org

EXPLORE!!!!
We will now try to build up our understanding about

Nonparametric Statistics. Let us start learning the basic concepts

relevant to our course of study.

NONPARAMETRIC STATISTICS

We cannot indicate a direct definition of Nonparametric Statistics, however

we can associate this to several relevant tests. We can attest that we are

engaged in Nonparametric Statistics if we are analysing categorical data, or if we

are assessing the equivalence of the whole population or distribution. Some of

the experts had coined Nonparametric Statistics with statistical tests. Petilos

(2017), described Nonparametric Statistics in terms of Nonparametric Statistical

Tests, as tests that are distribution free.

Module 1. Introduction to Nonparametric Statistics 3

Nonparametric Statistical Tests

The Nonparametric Statistical Tests or Nonparametric Tests is based on a

model that specifies only very general condition about the respondents where no

assumptions about the exact shape of the population distribution need be made.

It is distribution free. Observations are independent with underlying continuity but

with fewer and weaker assumptions. This test involves less computational work

and hence is easier and quicker to apply. It is often described as test that does

not require a normal distribution.

Uses of Nonparametric Tests

Ong Kian Koc, et. al (1998) had mentioned several conditions when we

can use Nonparametric Tests. It includes the following:

1. It is appropriate for categorical data.

2. It is used to assess the equivalence of the whole population.

3. It may be used for interval – ratio data where the distribution producing the

data is either unspecified or specified except for an infinite number of

unknown parameter.

4. It is good for small sample size.

PARAMETRIC AND NONPARAMETRIC STATISTICS

Let us try to have some sort of comparison between Parametric and

Nonparametric Statistics. I will present the comparison made by Ong Kian Koc,

et. al (1998) between Parametric and Nonparametric Statistics indicated on their

book entitled “Seminar in Advance Statistics Applied to Education”.

Module 1. Introduction to Nonparametric Statistics 4

Parametric Statistics Nonparametric Statistics

1. The distribution is normal. 1. The distribution is free.

2. The samples are randomly chosen. 2. Data in the samples are independent

of one another.

3. The level of measurement is 3. The level of measurement is

interval – ratio. categorical (nominal or ordinal)

4. Statistical procedures are used to 4. Statistical procedures are used to

assess the equivalence of assess the equivalence of population

population parameters. or complete distribution.

5. Can be used for larger sample size. 5. Can be used for smaller sample

size.

It is evident from the table that there is a great difference between

Parametric and Nonparametric Statistics. As what we can observe, Parametric

Statistics requires stricter norm as compared with Nonparametric Statistics. I will

leave the tasks for you to analyse their similarities.

PRECONCIEVED ADVANTAGES AND DISADVANTAGES OF

NONPARAMETRIC STATISTICS

I do not have all the knowledge in this universe, so I have to ask for help

from experts in this field. I would like to mention the list of advantages of

Nonparametric Statistics indicated by Petilos (2017) and here are the following:
Module 1. Introduction to Nonparametric Statistics 5

Preconceived Advantages of Nonparametric Statistics

1. They can be used for small sample size.

2. They make fewer assumptions about the data and hence may be more

relevant to the situation mentioned in the research.

3. They are available to analyse data where numerical scores have the

strength of ranks.

4. They are available to treat data that are simple classificatory or

categorical, i.e., measured in a nominal scale.

5. There are suitable Nonparametric tests that can handle data or

observations obtained from several different populations.

6. They are easier to learn than parametric tests and the results can be

interpreted directly.

Preconceived Disadvantages of Nonparametric Statistics

1. If all conditions for parametric tests are met and the nonparametric tests

are less powerful, then the use of nonparametric tests is wasteful.

2. The nonparametric tests are not systematic as compared to parametric

tests. In most cases, the different nonparametric tests are simply a

variation from a central theme.

3. The nonparametric statistical tests are designed for convenience. The

tables in nonparametric statistical tests are scattered widely and have

different formats.
Module 1. Introduction to Nonparametric Statistics 6

THE PHI VALUE AND TABULAR VALUE

Although I am not a statistician by profession, I feel that it is my duty to

clarify the meaning of phi – value ( – value) and tabular value. So, I will present

my discussion in the same way as how it was done by Petilos (2017).

Figure 1.

In interpreting the Statistical Package for Social Science Research (SPSS)

output, we emphasize the use of - value or significance level of the test which is

commonly used in presenting results of hypothesis testing. The – value is the

probability of obtaining a value of a test statistic as extreme or more

extreme than the one expected when the null hypothesis is true. In general,

when the – value associated to a test statistic is smaller than or equal to the

given level of significance , the null hypothesis is rejected. Take note that if the

– value is less than or equal to the level of significance, we should reject

the null hypothesis, ( ; reject ). This is opposite to what we do

when analysing data manually using tables or tabular value. When we are using

table or tabular value, when the absolute value of the computed test statistic

is greater than or equal to the corresponding tabular value, we should

Module 1. Introduction to Nonparametric Statistics 7

reject the null hypothesis. ( ; reject ). I will

provide the explanation for the equivalence of these two techniques in the next

paragraph.

In the figure, the area to the right of is equal to the level of significance

. The value of is called the critical or tabular value. When the data are

manually analysed, the value of the test statistic is computed and compared with

the critical value. When the computed value is greater than or equal to the

tabular value, the null hypothesis is rejected. This means that the computed

value will lie to the right of the critical value. The area to the right of the computed

value will be smaller than the given level of significance. This area is what we call

the – value or significance level of the test. Hence, when the – value is less

than or equal to the given level of significance, the null hypothesis is also

rejected. I hope that I did not provide confusion on your mind with this

explanation!

HYPOTHESIS TESTING

Perhaps you are familiar with this topic but, I will include it here to serve as

a review. I will not explain the steps in detailed because I assume that you

already know all of them.

The Six – Step model for Hypothesis Testing

1. Stating the null hypothesis ( ) and its alternative hypothesis ( ) or ( )

Module 1. Introduction to Nonparametric Statistics 8

In conducting a research with statistical tool, we usually express the

hypothesis in its null form initially and then express it in an alternative

hypothesis or .

2. Choosing a statistical test

The power of statistical treatment of data depends on the selection

of the statistical test.

3. Selecting the level of significance and the sample size

The level of significance or is the maximum probability of

rejecting a true null hypothesis (Type I error). Common values of are

0.05 and 0.01. Researches related to medicine are allowed only to use

0.01 since we are dealing with life.

4. Selecting the sampling distribution and establishing critical region

The sampling distribution is a theoretical distribution. It is a

probabilistic scale against which a particular sample outcome is

measured. If the sample size is infinitely large, a – distribution is used.

5. Collecting the data and computing the statistical test.

The selection of the statistical test to be used in analysing the data

depends on the level of measurement of the data.

6. Making a decision

This is the last step in hypothesis testing. The computed value of

the test statistic is compared with the critical region. Be careful in making

the decision for it may ruin your entire research undertakings.

Module 1. Introduction to Nonparametric Statistics 9

Since we have an allocated time for laboratory, I want you to do the following

activity. Make sure to submit your answers in MS Word format.

ACTIVITY

1. Surf the net and make an outline of the historical development of

Statistics. Make sure to include the important events and persons

involved. Include also a short biography for the persons involved in the

outline.

2. Discuss the different levels of measurement. Cite at least three examples

for each level.

3. Collect/search at least five sets of categorical data. Each set must have at

least 15 elements. Write a label for each set of data.

4. Surf the net and search for the different nonparametric tests.

DISCUSSION BOARD
We will now try to interact with each other in a two way

process at a least possible manner! I will post a question/s and try to answer it on

your own.

1. How can we determine if a particular research will use Nonparametric

tests or not?

2. Aside from the advantages and disadvantages of mentioned in this

section, can you still name some of them?

Module 1. Introduction to Nonparametric Statistics 10

POST COMPETENCY CHECKLISTS

Let us try to determine if you learned something from our topic. Answer the
following Self-Assessment Question.

SELF – ASSESSMENT QUESTION 1

1. In your own words, describe Nonparametric Statistics
2. Based from what you’ve learned, compare Parametric and
Nonparametric Statistics
3. What are the Preconceived Advantages and Disadvantages of
Nonparametric Statistical Tests
4. In your own words, differentiate –value from Tabular Value
5. Enumerate the Steps in Hypothesis Testing

ASSIGNMENT

Write the null and alternative hypothesis for the following statement. Indicate if

the hypothesis is directional (one – tailed) or non-directional (two tailed)

1. Boys are better than girls in manipulative skills.

2. There is a significant correlation between interest in science and the

career choice of high school graduate in science

3. Method A is more effective in developing higher order thinking skills of the

learners than Method B.

4. Girls are better than boys in taking notes.

5. Students from school A perform significantly better in spelling and spatial

activities than those from school B.