CARLOS HILADO MEMORIAL STATE COLLEGE

COLLEGE OF ARTS AND SCIENCES

MASTERS OF PUBLIC
ADMINISTRATION

Second Semester 2019 - 2020

Chapter 4. Non-Parametric Test
According to LaMorte (2017), the nonparametric tests are sometimes called distribution-free tests because they
are based on fewer assumptions (e.g., they do not assume that the outcome is approximately normally distributed).
Parametric tests involve specific probability distributions (e.g., the normal distribution) and the tests involve
estimation of the key parameters of that distribution (e.g., the mean or difference in means) from the sample data.
In addition, the cost of fewer assumptions is that nonparametric tests are generally less powerful than their
parametric counterparts (i.e., when the alternative is true, they may be less likely to reject H 0). There are some
situations when it is clear that the outcome does not follow a normal distribution. These include situations:
 when the outcome is an ordinal variable or a rank,
 when there are definite outliers or
 when the outcome has clear limits of detection.

Nonparametric test and its Parametric counterpart

Wilcoxon Singed Rank

The Wilcoxon signed-rank test is the nonparametric test equivalent to the dependent t-test. As the Wilcoxon
signed-rank test does not assume normality in the data, it can be used when this assumption has been violated and
the use of the dependent t-test is inappropriate. It is used to compare two sets of scores that come from the same
participants (Leard Statistics)

Example:
The table below shows the hours of relief provided by two analgesic drugs in 12 patients suffering from arthritis.
Is there any evidence that one drug provides longer relief than the other?
Patients Drug 1 Drug 2
1 2.0 3.5
2 3.6 5.7
3 2.6 2.9
4 2.6 2.4
5 7.3 9.9
6 3.4 3.3
7 14.9 16.7
8 6.6 6.0
9 2.3 3.8
10 2.0 4.0
11 6.8 9.1
12 8.5 20.9
 By examining the final Test Statistics table, we can discover whether these changes, changing drug 1 to drug 2,
led overall to a statistically significant difference in relief Scores. We are looking for the "Asymp. Sig. (2-tailed)" value,
which in this case is 0.013. This is the p-value for the test. We report the Wilcoxon signed-ranks test, there was
significant difference between drug 1 and drug to [W = 7.00, p = 0.012] at 0.05 level of significance.

ays. There are 10 employees in the department. The following table shows the average daily output for 30 days before the introduction of

Employee 1 2 3 4 5 6 7 8 9 10
Before music 90 80 92 85 81 85 72 85 70 88
During music 99 85 98 83 88 99 80 91 80 94

Mann-Whitney U-test
The Mann-Whitney U test is used to compare differences between two independent groups when the
dependent variable is either ordinal or continuous, but not normally distributed. For example, you could use the Mann-
Whitney U test to understand whether attitudes towards pay discrimination, where attitudes are measured on an
ordinal scale, differ based on gender (i.e., your dependent variable would be "attitudes towards pay discrimination" and
your independent variable would be "gender", which has two groups: "male" and "female") (Leard Statistics).
Alternately, you could use the Mann-Whitney U test to understand whether salaries, measured on a continuous
scale, differed based on educational level (i.e., your dependent variable would be "salary" and your independent
variable would be "educational level", which has two groups: "high school" and "university"). The Mann-Whitney U test
is often considered the nonparametric alternative to the independent t-test although this is not always the case.

Example: A researcher gives a random sample of 15 college men and an independent random sample of 20 college
women a test to measure their knowledge of ecological issues. The table of scores shows, if there is significant
difference in the level of the participants knowledge on ecological issues when grouped according to sex?

Men’s Scores Women’s Scores

18.50 17.00 12.40 25.00 19.10 15.00 18.00
14.00 16.00 15.20 23.00 18.75 21.00 18.25
20.00 12.50 12.50 16.20 21.10 18.50 24.00
19.00 12.00 19.25 19.75 17.50 17.25 18.30
19.50 10.00 11.00 20.00 17.75 16.30 19.20
 This table shows us the actual significance value of the test. Specifically, the Test Statistics table provides the
test statistic, U statistic, as well as the asymptotic significance (2-tailed) p-value. From this data, it can be concluded that
level of knowledge of women was statistically significantly higher than the men group. Therefore, there was significant
difference in the level of the participants knowledge on ecological issues when grouped according to sex (U (33) = 234,
p = 0.005) at 0.05 level of significance.

roduct by Method A. Subjects in the second group learn about it by Method B. At the end of the experiment each subject is given a test to

Method A 50 59 60 71 80 81 80 78 72 77 73
Method B 52 54 58 78 65 69 61 60 72 60 59

Kruskal-Wallis H-test
The Kruskal‐Wallis (Kruskal & Wallis, 1952) is a nonparametric statistical test that assesses the differences
among three or more independently sampled groups on a single, non‐normally distributed continuous variable. Non‐
normally distributed data (e.g., ordinal or rank data) are suitable for the Kruskal‐Wallis test.
In contrast, the one‐way analysis of variance (ANOVA), which is a parametric test, may be used for a normally
distributed continuous variable. The Kruskal‐Wallis test is an extension of the two‐group Mann‐Whitney U (Wilcoxon
rank) test. Thus, the Kruskal‐Wallis is a more generalized form of the Mann‐Whitney U test and is the nonparametric
version of the one‐way ANOVA.
Example:

The Department of Agriculture which to know whether they can conclude that three types of fertilizer have
different effects on the mean yield in bushels per acre of a certain grain. Each of the three types of fertilizer is applied to
four one-acre plots of ground. These plots are as alike with respect to relevant variables as possible. The plots are all
treated alike during the growing season. The table (next page) shows the yields of the 30 plots. Is there significant
difference in the mean yield of plots when grouped according to types of fertilizers?

Fertilizer A Fertilizer B Fertilizer C

45 42 53
40 44 56
41 43 54
46 47 55
45 45 53
50 47 54
47 48 56
41 50 55
43 51 52
44 49 53

A Kruskal-Wallis H test showed that there was a statistically significant difference in mean yields between the
different fertilizers, [χ2(2) = 20.635, p < 0.001], with a mean rank yields of 8.35 for fertilizer A, 12.65 for fertilizer B and
25.50 for fertilizer C.
 There was significant difference in the mean yield of plots when grouped according to types of fertilizers [χ2(2)
= 20.635, p < 0.001] at 0.05 level of significance.

r management potential. Each interviewer gives the trainee a score from 0 to 100. The Personnel Manager wishes to know if the 3 intervie

Interviewer
Trainee 1 2 3
A 71 94 70
B 84 68 71
C 60 72 54
D 78 66 57
E 65 85 78
F 53 47 45
G 72 75 53

Is there significant differece on the interviers scores?

References:

Kruskal & Wallis, (1952), Kruskal-Wallis H-test,

https://onlinelibrary.wiley.com/doi/abs/10.1002/9780470479216.corpsy0491, Retrieved: May 9, 2020, Saturday
LaMorte, W (2017), When to Use a Nonparametric Test, http://sphweb.bumc.bu.edu/otlt/MPH-
Modules/BS/BS704_Nonparametric/BS704_Nonparametric2.html, Retrieved: May 9, 2020, Saturday

Leard Statistics, Mann-Whitney U-test, https://statistics.laerd.com/spss-tutorials/mann-whitney-u-test-using-spss-

statistics.php, Retrieved May 9, 2020, Saturday.

Leard Statistics, Wilconxon Singed Rank, https://statistics.laerd.com/spss-tutorials/wilcoxon-signed-rank-test-using-spss-

statistics.php, Retrieved May 9, 2020, Saturday.