LESSON

13
NON-PARAMETRIC TESTS

CONTENTS
13.0 Aims and Objectives
13.1 Introduction
13.2 “U” Tests
13.3 Cox and Stuart Test
13.4 Kruskal-Wallis Test
13.5 Kalmogorov-Smirnov Test
13.6 Run-Test for Randomness
13.7 Sign-Test
13.8 Let us Sum up
13.9 Lesson-end Activity
13.10 Keywords
13.11 Questions for Discussion
13.12 Suggested Readings

13.0 AIMS AND OBJECTIVES

This lesson is intended to discuss non-parametric tests for hypothesis testing. After
studying this lesson you will be able to:
(i) determine whether two independent samples have been drawn from the same
population.
(ii) analyse Kruskal-Wallis test for more than two population.
(iii) identify the pairs and decide whether the pairs have more or less similar
characteristic by using sign-test.

13.1 INTRODUCTION
As we have seen in the previous lesson that f-test may not be applicable in all cases of
analysing data of two related samples. In such cases for analysing the data, we may use
non-parametric statistical tests of two related samples. There are three most commonly
used methods for two related samples, which are discussed below.

13.2 "U" TESTS

(Rank Sum test)
This test is used to determine whether two independent samples have been drawn from
the same population. Suppose an experiment has obtained two sets of samples from two
populations and the study wishes to examine whether the two populations are identical.
Research Methods for Example: A computer company XYZ would like to choose the performance of
Management
programmers, working in 2 branches, located in different cities. The performance indices
of employees:
Branch – A Branch – B
84 76
68 77
78 64
49 62
45 53

To find out whether there is any difference in the performance indices of employees of
the two branches.

13.3 COX AND STUART TEST

This test is used to examine the presence of trends. A set of numbers is said to show
upward trend if the latter numbers in the sequence are greater than the former numbers.
And similarly, one can define a downward trend. How to examine whether a trend is
noticeable in a sequence? Example: Suppose a marketer wants to examine whether its
sales are showing a trend or just fluctuating randomly. Suppose the company has gathered
the monthly sales figures during the past one year month-wise:

Month 1 2 3 4 5 6 7 8 9 10 11 12
Sales 200 250 280 300 320 278 349 268 240 318 220 380
From the given data, analyse the sales trend.

13.4 KRUSKAL-WALLIS TEST

We can use the Mann Whitney test; when two populations are involved, the Kruskal-
Wallis test is used, when more than two populations are involved. This test will enable us
to know whether independent samples have been drawn from the same population or
from different populations having the same distribution. This test is an extension of "Mann
Whitney test".
This is a type of Rank Sum test. This test is used to find out whether two or more
independent samples are drawn from an identical population. This test is also called the
H Test. Mann Whitney test is used when only two populations are involved and Kruskal-
Wallis test is used when more than two populations are involved.
Example: In an assembling unit, three different workers do assembly work in shifts.
The data is tabulated as follows:
Shift No. Worker-1 Worker-2 Worker-3
1 25 28 29
2 31 28 30
3 35 29 27
4 33 28 36
5 35 32 31
6 31 32 34
Check whether there is any difference in the production quantum of the three workers:
Illustration: (Kruskal-Wallis Test, H-Test)
Let us assume that there are three categories of workers involved in a building construction.
The wages depends on the skills possessed by them and their availability. The wages of
three categories, namely painter carpenter and plumber are as follows:
Item Sample 1 Sample 2 Sample 3
Daily wages Daily wages (Carpenter Daily wages
(Painter Rs.) Rs.) (Plumber Rs.)
1 64 72 51
2 66 74 52
3 72 75 54
144 4 74 78 56
5 80
 Wage-Painter Wage-Carpenter Rs./day Wage-Plumber Non-parametric Tests
Item Rs./day Rs./day
Rs Rank Rs Rank Rs Rank
1 64 5 72 7.5 51 1
2 66 6 74 9.5 52 2
3 72 7.5 75 11 54 3
4 74 9.5 78 12 56 4
5 80 13
Total 276 R1=28 379 R2=53 213 R3=10

Use H-test and state whether the three populations are same or different.
H0 - The wages of the three occupation are the same.
H1 - The wages of the three occupations is not the same.
n1= 4 n2=5 n3=4
n = n1+ n2 n3 = 4+5+4 = 13
R1=28 R2= 53 R3 = 10

ÈR 2 ˘
12
H= Â Í 1 ˙ - 3(n +1)
n ( n +1) Í n1 ˙
Î ˚

12 È 282 532 102 ˘

H= ÂÍ + + ˙ - 3(3 +1) = 9.61
13 (13 +1) Í 4 5 4 ˙
Î ˚
At 5% level of significance, for d.f = (3-1)=2, the table value is 5.991. Computed value
9.61 is greater.
Conclusion: Reject the Null hypothesis that the three populations are different.
Application of SPSS
1. Open a new spread sheet.
2. Type the first group of numbers in the first column and the second group of numbers
in the second column.
3. Using the headings of the page and click "Analyse".
4. Click on "Non parametric Tests" and then click on "2" related sample test.
5. Add "Test type" click on "Kruskal".
6. Then shift "var0001" "and var0002" over to "Test pairs list".
7. Click on OK.
The output will appear on the screen.

13.5 KOLMOGOROV-SMIRNOV TEST

This is used for examining the efficacy of fit between observed samples and expected
frequency distribution of data when the variable is in the ordinal scale.
Example:
A manufacturer of cosmetics wants to test four different shades of the liquid foundation
compound - very light, light, medium and dark. The company has hired a market research
agency to determine whether any distinct preference exists towards either extremes. If
so, the company will manufacture only the preferred shade, otherwise, the company is
planning to market all shades. Suppose, out of a sample of hundred, 50 preferred "very
light shade" 30 liked light shade, 15 the medium shade, and 50 dark shade. Do you think
the results show any kind of preference?
Since the shade represents ordering (rank), this test can be used to find the preference. 145
Research Methods for
Management 13.6 RUN-TEST FOR RANDOMNESS
Consider the example of arrival of customers at a branch office of a telephone department
for payment of telephone bills after the due date. The senior officer of the telephone
department wants to verify whether the gender of arriving customer is random.
Example: Sequence of arriving customers is as shown below. M is Male F is Female.
MM FFF MMM FFF M F MMMM FF
No. of male = 10
No. of female = 9
No. of run = 8

13.7 SIGN-TEST
Sign-test is used with matched pairs. The test is used to identify the pairs and decide
whether the pair has more or less similar characteristics.
Example: Suppose, an experiment on the effect of brand name on quality perceptions is
to be conducted. 10 persons are selected and asked to taste and compare the two products
(beverage). One of them is identified as branded well known beverage, and the other is
a new beverage. In reality, the samples are identical. The respondents who tested were
asked to rate the two samples on an ordinal scale. Two hypotheses are set up as follows:
H0 - there is no difference between the perceived qualities of two beverages.
HA - there is a difference in the perceived qualities of two beverages.

Check Your Progress

Discuss the uses of non-parametric test in hypothesis testing.

13.8 LET US SUM UP

We have discussed various non-parametric tests for hypothesis testing. U-test, i.e., rank
sum test, is used to determine whether two independent sample have been drawn from
the same population. Kruskal-Wallis test, which is an extension of Mann Whitney test, is
used when more than two populations are involved. We have also discussed Sign-Test
for matched pairs.

13.9 LESSON-END ACTIVITY

A company has three categories of managers:
1. With professional qualifications but without work experience.
2. With professional qualifications accompanied by work experience.
3. Without professional qualifications but with work experience.
A study was conducted to measure the motivation level of each of the category of
managers. Formulate a hypothesis, suggesting testing procedures to show that there is
no relation between the category of managers and the level of motivation.

146
 Non-parametric Tests
13.10 KEYWORDS
Non-parametric test
Related Samples
Matching pair
Rank-sum Test
Rank correlation Test
Krustal-Wallis test
Rank sum
Multi variate analysis
Parametric test
Non parametric test
Type I error
Variable
Variance
Z-test
T-test
Mann Whitney
"U" test
Cox & Stuart test

13.11 QUESTIONS FOR DISCUSSION

1. What is cox and stuart test?
2. What is Rank sum Test?
3. Give an example of Kruskal-Wallis test.
4. Write a brief note on application of SPSS in non-parametric Test.
5. Explain run test for randomness

13.12 SUGGESTED READINGS

S. N. Murthy and U. Bhojanna, Business Research Methods, Excel Books, 2007.
Boyd, Westfall, and Stasch, "Marketing Research Text and Cases", All India Traveller
Bookseller, New Delhi.
Brown, F.E. "Marketing Research, a structure for decision making", Addison- Wesley
Publishing Company.
Kothari, C.R. "Research Methodology- Methods and Techniques", Wiley Eastern
Ltd.
Stockton and Clark, "Introduction to Business and Economic Statistics", D.B.
Taraporevala Sons and Co. Private Limited, Bombay.

147