Scribd
Upload
7 views

7.-Other-Non-parametric-Tests

The document discusses non-parametric tests, specifically the Mann-Whitney U-test and the Kruskal-Wallis H-test, which are used to analyze differences between groups measured on an ordinal scale. It outlines the hypotheses, calculations, and assumptions required for each test, along with examples illustrating their application. The Mann-Whitney U-test compares two groups, while the Kruskal-Wallis H-test generalizes this approach for multiple groups.

Uploaded by

hamoudiguinal2
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
7 views

7.-Other-Non-parametric-Tests

The document discusses non-parametric tests, specifically the Mann-Whitney U-test and the Kruskal-Wallis H-test, which are used to analyze differences between groups measured on an ordinal scale. It outlines the hypotheses, calculations, and assumptions required for each test, along with examples illustrating their application. The Mann-Whitney U-test compares two groups, while the Kruskal-Wallis H-test generalizes this approach for multiple groups.

Uploaded by

hamoudiguinal2
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 11

Other Non-parametric Tests

Mann-Whitney U-test
When variables are measured in ordinal scale, the
Mann-Whitney U-test may be used to test whether
these two groups come from the same population or
whether or not there is a difference between the
samples. This is one of the most powerful of the
nonparametric test.
Let m be the number of items in the sample of
group x and n be the number of items/cases in the
sample from group y.
Supposed that the samples are from two populations, x and y, the
null hypothesis is that x and y must have the same direction. Then if
the null hypothesis is rejected, we may accept that the alternative
hypothesis that either x is larger than y or x is smaller than y.
So,
𝑁 = 𝑚 + 𝑛 = 𝑁1 + 𝑁2
𝑁(𝑁 + 1)
Σ𝑊𝑥 + Σ𝑊𝑦 =
2
Where
Wx = sum of ranks of group x with m cases
Wy = sum of ranks of group y with n cases
Then, Where
𝑁1 (𝑁1 + 1) U = is a sample statistics. The
𝑈1 = 𝑁1 𝑁2 + − Σ𝑊𝑥
2 larger value of U (either U1 or U2)
𝑁2 (𝑁2 + 1) is the one to be substituted in the
𝑈2 = 𝑁1 𝑁2 + − Σ𝑊𝑦
2 next formula.
Also,
𝑁1 𝑁2 𝜇𝑢 = mean of sampling
𝜇𝑢 =
2 distribution
𝑁1 𝑁2 𝑁1 + 𝑁2 + 1 𝜎𝑢 = standard deviation of the
𝜎𝑢 =
12 sampling distribution
𝑈 − 𝜇𝑢
𝑧= z = z-test under 2-tailed test
𝜎𝑢
Example: FEMALE STUDENTS MALE STUDENTS
Are female students more Score Rank Score Rank
anxious than male students in 48 1 47 2
45 3 44 4
learning statistics? Samples of 12 43 5 40 7
female students and 12 male 42 6 35 10
students were chosen and given a 39 8 32 12
36 9 31 13
questionnaire to rate the degree 33 11 30 14
of anxiety from “very anxious” to 28 16 29 15
“not anxious at all”. Summary of 23 19 25 17
21 20 24 18
scores for those 24 students are
15 22 20 21
presented in the following table: 14 23 12 24
Kruskal-Wallis H-test

It is a test for differences in central tendency in


design with one-between-subjects variable.
It is a generalization of the Mann-Whitney U-test
for k-samples.
It provides a non-parametric method in the
Analysis of Variance for one-way classification.
Suppose that we have k samples of sizes N1, N2, N3, …, Nk ,
with the total size of all samples taken together being given
by
N = N1 + N2 + N3 + … + Nk

Also, suppose that the data from all the samples are ranked
and the sum of ranks for the k samples are R1, R2, R3, …, Rk ,
respectively. Then, we define the H-test to be
𝑘 2
12 𝑅𝑗
𝐻= ෍ − 3(𝑁 + 1)
𝑁(𝑁 + 1) 𝑁𝑗
𝑗=1
Steps:
1.Convert the data into ranks (ignoring group
membership) and then find the sum of the ranks for
each group.
2.Record the sizes of each group and get the sum of all
of them.
3.Compute for H.
4.Finally, compute the significance using df = k-1 (the
table to be used is the chi-square table) with the
given level of significance of the problem.
Example:
A study measured the sorption (either absorption or adsorption) rates of
three different types of organic chemical solvents. These solvents are
used to clean industrial fabricated-metal parts and are potential
hazardous waste. Independent samples from each type of solvent were
tested, and their sorption rates were recorded as a mole percentage. Use
Kruskal-Wallis H-test at the 0.05 level of significance to determine if
the organic chemical solvents differ significantly in the sorption rate.
Aromatics Chloroalkanes Esters
1.06 0.95 1.58 1.12 0.29 0.43
0.79 0.65 1.45 0.06 0.51
0.82 0.57 0.44
SPSS Portion:
Assumptions for Mann-Whitney U-test
1.The dependent variable should be measured at the
ordinal, interval, or ratio levels.
2.Independent variables should consist of two
categorical, independent groups.
3.Every data or observations must be independent
among each other.
Assumptions for Kruskal-Wallis H-test
1.The dependent variable should be measured at the
ordinal, interval, or ratio levels.
2.Independent variables should consist of two or
more categorical, independent groups. Typically, it
is used when there are 3 or more categories.
3.Every data or observations must be independent
among each other.

You might also like