You are on page 1of 19

Learning 

Objectives Parametric vs. Nonparametric Statistics
Parametric Statistics are statistical techniques based 
Recognize the advantages and disadvantages of  on assumptions about the population from which the 
nonparametric statistics. sample data are collected.
Understand how to use the runs test to test for  Assumption that data being analyzed are randomly
randomness. selected from a normally distributed population. 
Know when and how to use the Mann‐Whitney U y test, , Requires quantitative measurement that yield interval
the Wilcoxon matched‐pairs signed rank test, the  or ratio level data.
Kruskal‐Wallis test, and the Friedman test. Nonparametric Statistics are based on fewer 
assumptions about the population and the 
parameters. 
Sometimes  called “distribution‐free” statistics.
A variety of nonparametric statistics are available for
use with nominal or ordinal data.

Advantages of Nonparametric Techniques Disadvantages of Nonparametric Statistics

Sometimes there is no parametric alternative to the  Nonparametric tests can be wasteful of data if 
use of nonparametric statistics. parametric tests are available for use with the data.
Certain nonparametric test can be used to analyze  Nonparametric tests are usually not as widely 
nominal data. available and well know as parametric tests.
Certain nonparametric test can be used to analyze  For large samples, the calculations for many 
g p , y
ordinal data. nonparametric statistics can be tedious.
The computations on nonparametric statistics are 
usually less complicated than those for parametric 
statistics, particularly for small samples.
Probability statements obtained from most 
nonparametric tests are exact probabilities.

Page 1
Learning Objectives Parametric vs. Nonparametric Statistics
Parametric Statistics are statistical techniques based 
Recognize the advantages and disadvantages of  on assumptions about the population from which the 
nonparametric statistics. sample data are collected.
Understand how to use the runs test to test for  Assumption that data being analyzed are randomly
randomness. selected from a normally distributed population. 
Know when and how to use the Mann‐Whitney U y test, , Requires quantitative measurement that yield interval
the Wilcoxon matched‐pairs signed rank test, the  or ratio level data.
Kruskal‐Wallis test, and the Friedman test. Nonparametric Statistics are based on fewer 
assumptions about the population and the 
parameters. 
Sometimes  called “distribution‐free” statistics.
A variety of nonparametric statistics are available for
use with nominal or ordinal data.

Advantages of Nonparametric Techniques Disadvantages of Nonparametric Statistics

Sometimes there is no parametric alternative to the  Nonparametric tests can be wasteful of data if 
use of nonparametric statistics. parametric tests are available for use with the data.
Certain nonparametric test can be used to analyze  Nonparametric tests are usually not as widely 
nominal data. available and well know as parametric tests.
Certain nonparametric test can be used to analyze  For large samples, the calculations for many 
g p , y
ordinal data. nonparametric statistics can be tedious.
The computations on nonparametric statistics are 
usually less complicated than those for parametric 
statistics, particularly for small samples.
Probability statements obtained from most 
nonparametric tests are exact probabilities.

Page 1
Mann‐Whitney U Test Mann‐Whitney U Test:  
Sample Size Consideration
Mann‐Whitney U test ‐ a nonparametric counterpart  Size of sample one:  n1
of the t test used to compare the means of two  Size of sample two:  n2
independent populations. If both n1 and n2 are  10, the small sample procedure 
Nonparametric counterpart of the t test for  is appropriate.
independent samples If either n1 or n
If either n or n2 is greater than 10, the large sample 
is greater than 10 the large sample
Does not require normally distributed populations procedure is appropriate.
May be applied to ordinal data
Assumptions
Independent Samples
At Least Ordinal Data

Mann‐Whitney U Test: Small Sample  Mann‐Whitney U Test: Small Sample 
Example‐Demonstration Example‐Demonstration
H0:  µ1 = µ2  = .05 Compensation Rank Group
Ha:  µ1 π µ2 Drug A Drug B 18.75 1 H
19.80 2 H
20.10 26.19 If the final p‐value < .05, reject H0. 20.10 3 H
20.75 4 H
19.80 23.88
21.64 5 E
22 36
22.36 25 50
25.50 21 90
21.90 6 H
18.75 21.64 W1 = 1 + 2 + 3 + 4 + 6 + 7 + 8 = 31 22.36 7 H
22.96 8 H
21.90 24.85 23.45 9 E
22.96 25.30 W2 = 5 + 9 + 10 + 11 + 12 + 13 + 14 + 15 = 89 23.88 10 E
24.12 11 E
20.75 24.12 24.85 12 E
23.45 25.30 13 E
25.50 14 E
26.19 15 E

Page 2
Mann‐Whitney U Test Mann‐Whitney U Test:  
Sample Size Consideration
Mann‐Whitney U test ‐ a nonparametric counterpart  Size of sample one:  n1
of the t test used to compare the means of two  Size of sample two:  n2
independent populations. If both n1 and n2 are  10, the small sample procedure 
Nonparametric counterpart of the t test for  is appropriate.
independent samples If either n1 or n
If either n or n2 is greater than 10, the large sample 
is greater than 10 the large sample
Does not require normally distributed populations procedure is appropriate.
May be applied to ordinal data
Assumptions
Independent Samples
At Least Ordinal Data

Mann‐Whitney U Test: Small Sample  Mann‐Whitney U Test: Small Sample 
Example‐Demonstration Example‐Demonstration
H0:  µ1 = µ2  = .05 Compensation Rank Group
Ha:  µ1 π µ2 Drug A Drug B 18.75 1 H
19.80 2 H
20.10 26.19 If the final p‐value < .05, reject H0. 20.10 3 H
20.75 4 H
19.80 23.88
21.64 5 E
22 36
22.36 25 50
25.50 21 90
21.90 6 H
18.75 21.64 W1 = 1 + 2 + 3 + 4 + 6 + 7 + 8 = 31 22.36 7 H
22.96 8 H
21.90 24.85 23.45 9 E
22.96 25.30 W2 = 5 + 9 + 10 + 11 + 12 + 13 + 14 + 15 = 89 23.88 10 E
24.12 11 E
20.75 24.12 24.85 12 E
23.45 25.30 13 E
25.50 14 E
26.19 15 E

Page 2
Mann‐Whitney U Test: Small Mann‐Whitney U Test:  
Sample Example Formulas for Large Sample Case
n (n  1)  Since U2 < U1, U = 3.
U n n
1 1 2

2
1 1
W1
(7)(8) p‐value  = .0011*2  n1 (n1  1)  
n n
 (7)(8)   31 U  n1n 2  W 1
1 2

2 (for a two‐tailed test) = .022 2
U 2

 53 < .05, reject H0. where : n1  number in group 1 n  n n  n 


1
 U

1 2 1

12
2

n (n  1) n  number in group 2

U n n
2 1 2
 2 2
2
W 2 2
Z 
U  U

(8)(9)
W 1
 sum or the ranks of  U

 (7)(8) n1 n2   89 values in group 1


2
3

Example – Mann Whitney U for large  Ranks of Income from Combined


samples Groups of PBS and Non‐PBS Viewers
H0: µ1 = µ2 PBS Non‐PBS Data value  Rank Group Data value Rank Group
Ha: µ1 π µ2 24,500 41,000 16,000 1 Non‐PBS 39,500 15 Non‐PBS
39,400 32,500 21,000 2 Non‐PBS 40,500 16 Non‐PBS
36,800 33,000 21,500 3 Non‐PBS 41,000 17 Non‐PBS
44,300 21,000 24,500 4 PBS 43,000 18 PBS
57 960
57,960 40 500
40,500 27,600 5 Non‐PBS
Non PBS 43,500 19.5 PBS
n1 = 14 27,800 6 Non‐PBS 43,500 19.5 Non‐PBS
  .05 32,000 32,400
32,000 7 PBS 51,900 21 Non‐PBS
61,000 16,000
If Z   1.96 or Z  1.96 , reject H o 34,000 21,500 32,400 8 Non‐PBS 53,000 22 PBS
n2 = 13
43,500 39,500 32,500 9 Non‐PBS 55,000 23 PBS
33,000 10 Non‐PBS 57,960 24 PBS
55,000 27,600
34,000 11 PBS 61,000 25 PBS
39,000 43,500
36,800 12 PBS 61,400 26 PBS
62,500 51,900
39,000 13 PBS 62,500 27 PBS
61,400 27,800
39,400 14 PBS
53,000

Page 3
Mann‐Whitney U Test: Small Mann‐Whitney U Test:  
Sample Example Formulas for Large Sample Case
n (n  1)  Since U2 < U1, U = 3.
U n n
1 1 2

2
1 1
W1
(7)(8) p‐value  = .0011*2  n1 (n1  1)  
n n
 (7)(8)   31 U  n1n 2  W 1
1 2

2 (for a two‐tailed test) = .022 2
U 2

 53 < .05, reject H0. where : n1  number in group 1 n  n n  n 


1
 U

1 2 1

12
2

n (n  1) n  number in group 2

U n n
2 1 2
 2 2
2
W 2 2
Z 
U  U

(8)(9)
W 1
 sum or the ranks of  U

 (7)(8) n1 n2   89 values in group 1


2
3

Example – Mann Whitney U for large  Ranks of Income from Combined


samples Groups of PBS and Non‐PBS Viewers
H0: µ1 = µ2 PBS Non‐PBS Data value  Rank Group Data value Rank Group
Ha: µ1 π µ2 24,500 41,000 16,000 1 Non‐PBS 39,500 15 Non‐PBS
39,400 32,500 21,000 2 Non‐PBS 40,500 16 Non‐PBS
36,800 33,000 21,500 3 Non‐PBS 41,000 17 Non‐PBS
44,300 21,000 24,500 4 PBS 43,000 18 PBS
57 960
57,960 40 500
40,500 27,600 5 Non‐PBS
Non PBS 43,500 19.5 PBS
n1 = 14 27,800 6 Non‐PBS 43,500 19.5 Non‐PBS
  .05 32,000 32,400
32,000 7 PBS 51,900 21 Non‐PBS
61,000 16,000
If Z   1.96 or Z  1.96 , reject H o 34,000 21,500 32,400 8 Non‐PBS 53,000 22 PBS
n2 = 13
43,500 39,500 32,500 9 Non‐PBS 55,000 23 PBS
33,000 10 Non‐PBS 57,960 24 PBS
55,000 27,600
34,000 11 PBS 61,000 25 PBS
39,000 43,500
36,800 12 PBS 61,400 26 PBS
62,500 51,900
39,000 13 PBS 62,500 27 PBS
61,400 27,800
39,400 14 PBS
53,000

Page 3
PBS and Non‐PBS : Calculation of U PBS and Non‐PBS : Conclusion

 n n U  
W  4  7  11  12  13  14  18  19.5  22  23  24  25  26  27  1 2
Z  U


1 U 2
 2455
. 14 13 U
 41.5  91
n n  1  
2
20.6
U  n1 n2 
1 1
W  91
2 1
  2.40
1415 n  n n  n 
1
 1413   2455
.  U

1 2 1

12
2

2
 415
. 14 13 28

 20.6
12
Z Cal
 2 .40   1.96 , reject H o

Wilcoxon Matched‐Pairs Signed Rank Test Wilcoxon Matched‐Pairs Signed Rank Test

Mann-Whitney U test is a nonparametric alternative Differences of the scores of the two matched samples


to the t test for two independent samples. If the two Differences are ranked, ignoring the sign
samples are related, the U test is not applicable.
Handle related data
Ranks are given the sign of the difference
Serves as a nonparametric alternative to the t test for two Positive ranks are summed
related samples Negative ranks are summed
Negative ranks are summed
A nonparametric alternative to the t test for related
samples T is the smaller sum of ranks
Before and After studies
Studies in which measures are taken on the same
person or object under different conditions
Studies or twins or other relatives

Page 4
PBS and Non‐PBS : Calculation of U PBS and Non‐PBS : Conclusion

 n n U  
W  4  7  11  12  13  14  18  19.5  22  23  24  25  26  27  1 2
Z  U


1 U 2
 2455
. 14 13 U
 41.5  91
n n  1  
2
20.6
U  n1 n2 
1 1
W  91
2 1
  2.40
1415 n  n n  n 
1
 1413   2455
.  U

1 2 1

12
2

2
 415
. 14 13 28

 20.6
12
Z Cal
 2 .40   1.96 , reject H o

Wilcoxon Matched‐Pairs Signed Rank Test Wilcoxon Matched‐Pairs Signed Rank Test

Mann-Whitney U test is a nonparametric alternative Differences of the scores of the two matched samples


to the t test for two independent samples. If the two Differences are ranked, ignoring the sign
samples are related, the U test is not applicable.
Handle related data
Ranks are given the sign of the difference
Serves as a nonparametric alternative to the t test for two Positive ranks are summed
related samples Negative ranks are summed
Negative ranks are summed
A nonparametric alternative to the t test for related
samples T is the smaller sum of ranks
Before and After studies
Studies in which measures are taken on the same
person or object under different conditions
Studies or twins or other relatives

Page 4
Wilcoxon Matched‐Pairs Signed Rank Test:   Wilcoxon Matched‐Pairs Signed Rank Test: 
Sample Size Consideration Small Sample Example
n is the number of matched pairs Consider the survey by American Demographics that  
If n > 15, T is approximately normally distributed, estimated the average annual household spending
and a Z test is used. on healthcare. The U.S. metropolitan average was 
If n  15, a special “small sample” procedure is  $1,800.  
followed. Suppose six families in Pittsburgh, Pennsylvania, are 
pp g , y ,
The paired data are randomly selected. matched demographically with six families in 
The underlying distributions are symmetric. Oakland,  California, and their amounts of household 
spending on healthcare for last year are obtained.

Wilcoxon Matched‐Pairs Signed Rank Test:   Wilcoxon Matched‐Pairs Signed Rank Test:  
Small Sample Example Small Sample Example
Family 
H0:  Md = 0 Pair Sample A Sample B d Rank
Ha:  Md  0 Pair Sample B 1 1,950 1,760 190 +4
Sample A
1 1,950 1,760 2 1,840 1,870 ‐30 ‐1
n = 6
2 1,840 1,870 3 2,015 1,810 205 +5
3 2,015 1,810 4 1 580
1,580 1 660
1,660 ‐80
80 ‐2
2
 =0.05
4 1,580 1,660 5 1,790 1,340 450 +6
6 1,925 1,765 160 +3
5 1,790 1,340
If Tobserved  1, reject H0. 6 1,925 1,765
T = minimum(T+, T-) T = 3 > Tcrit = 1, do not reject H0.
T+ = 4 + 5 + 6 + 3= 18
T- = 1 + 2 = 3
T=3

Page 5
Wilcoxon Matched‐Pairs Signed Rank Test:   Wilcoxon Matched‐Pairs Signed Rank Test: 
Sample Size Consideration Small Sample Example
n is the number of matched pairs Consider the survey by American Demographics that  
If n > 15, T is approximately normally distributed, estimated the average annual household spending
and a Z test is used. on healthcare. The U.S. metropolitan average was 
If n  15, a special “small sample” procedure is  $1,800.  
followed. Suppose six families in Pittsburgh, Pennsylvania, are 
pp g , y ,
The paired data are randomly selected. matched demographically with six families in 
The underlying distributions are symmetric. Oakland,  California, and their amounts of household 
spending on healthcare for last year are obtained.

Wilcoxon Matched‐Pairs Signed Rank Test:   Wilcoxon Matched‐Pairs Signed Rank Test:  
Small Sample Example Small Sample Example
Family 
H0:  Md = 0 Pair Sample A Sample B d Rank
Ha:  Md  0 Pair Sample B 1 1,950 1,760 190 +4
Sample A
1 1,950 1,760 2 1,840 1,870 ‐30 ‐1
n = 6
2 1,840 1,870 3 2,015 1,810 205 +5
3 2,015 1,810 4 1 580
1,580 1 660
1,660 ‐80
80 ‐2
2
 =0.05
4 1,580 1,660 5 1,790 1,340 450 +6
6 1,925 1,765 160 +3
5 1,790 1,340
If Tobserved  1, reject H0. 6 1,925 1,765
T = minimum(T+, T-) T = 3 > Tcrit = 1, do not reject H0.
T+ = 4 + 5 + 6 + 3= 18
T- = 1 + 2 = 3
T=3

Page 5
Wilcoxon Matched‐Pairs Signed Rank Test:  Wilcoxon Matched‐Pairs Signed Rank Test: 
Large Sample Formulas Large Sample Formulas
For large samples, the T statistic is approximately normally 
distributed and a z score can be used as the test statistic. 

Wilcoxon Matched‐Pairs Signed Rank Test:  Example
Large Sample Formulas 
n n  1 H0: Md = 0   .05
 T
 Ha: Md  0
4 If Z   1.96 or Z  1.96 , reject H o
nn  12n  1
T  24
City 1979 2011
1 20.3
20 3 22.8
22 8 25
d Rank
‐2.5 ‐8
8
City 1979 2011
10 20.3
20 3 20.9
20 9
d Rank
‐0.6
06 ‐1
1
T  T 2 19.5 12.7 6.8 17 11 19.2 22.6 ‐3.4 ‐11.5
Z 3 18.6 14.1 4.5 13 12 19.5 16.9 2.6 9
 T
4 20.9 16.1
5 19.9 25.2
4.8
‐5.3
15
‐16
13 18.7 20.6
14 17.7 18.5
‐1.9 ‐6.5
‐0.8 ‐2
where : n = number of pairs 6 18.6 20.2 ‐1.6 ‐4 15 21.6 23.4 ‐1.8 ‐5
T = total ranks for either + or - differences, which ever is less 7 19.6 14.9 4.7 14 16 22.4 21.3 1.1 3
8 23.2 21.3 1.9 6.5 17 20.8 17.4 3.4 11.5
9 21.8 18.7 3.1 10

Page 6
Wilcoxon Matched‐Pairs Signed Rank Test:  Wilcoxon Matched‐Pairs Signed Rank Test: 
Large Sample Formulas Large Sample Formulas
For large samples, the T statistic is approximately normally 
distributed and a z score can be used as the test statistic. 

Wilcoxon Matched‐Pairs Signed Rank Test:  Example
Large Sample Formulas 
n n  1 H0: Md = 0   .05
 T
 Ha: Md  0
4 If Z   1.96 or Z  1.96 , reject H o
nn  12n  1
T  24
City 1979 2011
1 20.3
20 3 22.8
22 8 25
d Rank
‐2.5 ‐8
8
City 1979 2011
10 20.3
20 3 20.9
20 9
d Rank
‐0.6
06 ‐1
1
T  T 2 19.5 12.7 6.8 17 11 19.2 22.6 ‐3.4 ‐11.5
Z 3 18.6 14.1 4.5 13 12 19.5 16.9 2.6 9
 T
4 20.9 16.1
5 19.9 25.2
4.8
‐5.3
15
‐16
13 18.7 20.6
14 17.7 18.5
‐1.9 ‐6.5
‐0.8 ‐2
where : n = number of pairs 6 18.6 20.2 ‐1.6 ‐4 15 21.6 23.4 ‐1.8 ‐5
T = total ranks for either + or - differences, which ever is less 7 19.6 14.9 4.7 14 16 22.4 21.3 1.1 3
8 23.2 21.3 1.9 6.5 17 20.8 17.4 3.4 11.5
9 21.8 18.7 3.1 10

Page 6
T Calculation Conclusion

T  minimum(T  ,T )  n n  1 1718


 T

4

4
 765
.
T 
 17  13  15  14  6.5  10  9  3  11.5
n n  1 2n  1 1718 35
 99  T

24

24
 211
.

T  T
T 
 8  16  4  1  11.5  6.5  2  5
Z 
54  765
.
 107

.
211.
 54 T

T  minimum(99,54)
 54 1.96  Z Cal   1.07  1.96 , do not reject H o

Kruskal‐Wallis Test Kruskal‐Wallis K Statistic

Kruskal‐Wallis Test ‐ A nonparametric alternative
12  C T j 
2
to one‐way analysis of variance
May used to analyze ordinal data
K 
n  n  1  j 1 n j 
 3  n  1
 
No assumed population shape
where : c = number of groups
Assumes that the C groups are independent
Assumes that the C groups are independent
n = total number of items
Assumes random selection of individual items
T j
 total of ranks in a group
n j = number of items in a group
K  χ , with df = c -1
2

Page 7
T Calculation Conclusion

T  minimum(T  ,T )  n n  1 1718


 T

4

4
 765
.
T 
 17  13  15  14  6.5  10  9  3  11.5
n n  1 2n  1 1718 35
 99  T

24

24
 211
.

T  T
T 
 8  16  4  1  11.5  6.5  2  5
Z 
54  765
.
 107

.
211.
 54 T

T  minimum(99,54)
 54 1.96  Z Cal   1.07  1.96 , do not reject H o

Kruskal‐Wallis Test Kruskal‐Wallis K Statistic

Kruskal‐Wallis Test ‐ A nonparametric alternative
12  C T j 
2
to one‐way analysis of variance
May used to analyze ordinal data
K 
n  n  1  j 1 n j 
 3  n  1
 
No assumed population shape
where : c = number of groups
Assumes that the C groups are independent
Assumes that the C groups are independent
n = total number of items
Assumes random selection of individual items
T j
 total of ranks in a group
n j = number of items in a group
K  χ , with df = c -1
2

Page 7
Number of Patients per Day per Physician Number of Patients per Day per Physician
in Three Organizational Categories in Three Organizational Categories 
Suppose a researcher wants to determine whether the  Ho:  The three populations are identical
number of physicians in an office produces significant   Ha:  At least one of the three populations is different
differences in the number of office patients seen by each   Three or 
physician per day.    0.05 Two  More 
Sh t k
She takes a random sample of physicians from practices 
d l f h i i f ti df  C  1  3  1  2 P
Partners P
Partners HMO
in which (1) there are only two  partners, (2) there are  13 24 26

2
three or more partners, or (3) the office is a health  .05, 2
 5.991 15 16 22
20 19 31
maintenance organization (HMO). If K  5.991, reject Ho. 18 22 27
23 25 28
14 33
17

Patients per Day Data: Kruskal‐Wallis  Patients per Day Data: Kruskal‐Wallis 
Preliminary Calculations Calculations and Conclusion
Three or 
12  C T j 
Two  More  2

Partners
Patients
Partners HMO
Rank Patients Rank Patients Rank
K 
n n  1  j 1 n j 
 3 n  1
13 1 24 12 26 14
15 3 16 4 22 9.5
12  29 2 52.52 89.52 
20 8 19 7 31 17       318  1
18 6 22 95
9.5 27 15 1818  1  5 7 6 
23 11 25 13 28 16
12
14 2 33 18  1,897   318  1
17 5 1818  1
T1 = 29 T2 = 52.5 T3 = 89.5
 9.56
n1 = 5 n2 = 7 n3 = 6
n = n1 + n2 + n3 = 5 + 7 + 6 = 18

2

.05, 2
 5.991
K  9.56  5.991, reject Ho.

Page 8
Number of Patients per Day per Physician Number of Patients per Day per Physician
in Three Organizational Categories in Three Organizational Categories 
Suppose a researcher wants to determine whether the  Ho:  The three populations are identical
number of physicians in an office produces significant   Ha:  At least one of the three populations is different
differences in the number of office patients seen by each   Three or 
physician per day.    0.05 Two  More 
Sh t k
She takes a random sample of physicians from practices 
d l f h i i f ti df  C  1  3  1  2 P
Partners P
Partners HMO
in which (1) there are only two  partners, (2) there are  13 24 26

2
three or more partners, or (3) the office is a health  .05, 2
 5.991 15 16 22
20 19 31
maintenance organization (HMO). If K  5.991, reject Ho. 18 22 27
23 25 28
14 33
17

Patients per Day Data: Kruskal‐Wallis  Patients per Day Data: Kruskal‐Wallis 
Preliminary Calculations Calculations and Conclusion
Three or 
12  C T j 
Two  More  2

Partners
Patients
Partners HMO
Rank Patients Rank Patients Rank
K 
n n  1  j 1 n j 
 3 n  1
13 1 24 12 26 14
15 3 16 4 22 9.5
12  29 2 52.52 89.52 
20 8 19 7 31 17       318  1
18 6 22 95
9.5 27 15 1818  1  5 7 6 
23 11 25 13 28 16
12
14 2 33 18  1,897   318  1
17 5 1818  1
T1 = 29 T2 = 52.5 T3 = 89.5
 9.56
n1 = 5 n2 = 7 n3 = 6
n = n1 + n2 + n3 = 5 + 7 + 6 = 18

2

.05, 2
 5.991
K  9.56  5.991, reject Ho.

Page 8
Friedman Test Friedman Test

Friedman Test ‐ A nonparametric alternative to the  C


12
randomized block design 
2

bc(c  1) j 1 R j
  3b(c  1)
2
r
Assumptions
The blocks are independent.
There is no interaction between blocks and treatments. where : c  number of treatment levels (columns)
Observations within each block can be ranked. b = number of blocks (rows)
Hypotheses R j= total ranks for a particular treatment level
Ho: The treatment populations are equal j = particular treatment level

 
2 2
Ha: At least one treatment population yields larger values  , with df = c - 1
r
than at least one other treatment population

Friedman Test: Friedman Test: 

Ho: The supplier populations are equal
Ha: At least one supplier population yields larger values
than at least one other supplier population   0.05
Supplier 1 Supplier 2 Supplier 3 Supplier 4
df  c  1  4  1  3


Monday 62 63 57 61 2
Tuesday 63 61 59 65  7.81473
Wednesday 61 62 56 63 .05,3


Thursday 62 60 57 64 2
Friday 64 63 58 66 If  7.81473, reject Ho.
r

Page 9
Friedman Test Friedman Test

Friedman Test ‐ A nonparametric alternative to the  C


12
randomized block design 
2

bc(c  1) j 1 R j
  3b(c  1)
2
r
Assumptions
The blocks are independent.
There is no interaction between blocks and treatments. where : c  number of treatment levels (columns)
Observations within each block can be ranked. b = number of blocks (rows)
Hypotheses R j= total ranks for a particular treatment level
Ho: The treatment populations are equal j = particular treatment level

 
2 2
Ha: At least one treatment population yields larger values  , with df = c - 1
r
than at least one other treatment population

Friedman Test: Friedman Test: 

Ho: The supplier populations are equal
Ha: At least one supplier population yields larger values
than at least one other supplier population   0.05
Supplier 1 Supplier 2 Supplier 3 Supplier 4
df  c  1  4  1  3


Monday 62 63 57 61 2
Tuesday 63 61 59 65  7.81473
Wednesday 61 62 56 63 .05,3


Thursday 62 60 57 64 2
Friday 64 63 58 66 If  7.81473, reject Ho.
r

Page 9
Friedman Test: Friedman Test:


12

Supplier 1 Supplier 2 Supplier 3 Supplier 4 2

bc(c  1) j 1 R j
  3b(c  1)
2
Monday 3 4 1 2 r
Tuesday 3 2 1 4
Wednesday 2 3 1 4 12
 (714)  3(5)(4  1)
Thursday 3 2 1 4
(5)(4)(4  1)
Friday 3 2 1 4
R j 14 13 5 18  10.68
2
Rj 196 169 25 324


2
4

 R  (196  169  25  324)  714


2 =10.68  7.81473, reject Ho.
j r
j 1

Page 10

You might also like