SlideShare a Scribd company logo
1 of 63
Download to read offline
Kruskal-Wallis test
Friedman test
Spearmanā€™s Correlation
Dr. S. A. Rizwan, M.D.,
Public Health Specialist,
Saudi Board of Preventive Medicine,
Riyadh, Kingdom of Saudi Arabia.
With thanks to Dr. Tajali Nazir Shora
Checklist of four questions
ā€¢ Q1. What scales of measurement has been used?
ā€¢ Q2. Which hypothesis has been tested?
ā€¢ Q3. If the hypothesis of difference has been
tested, are the samples independent or
dependent?
ā€¢ Q4. How many sets of measures are involved?
2
Kruskal-Wallis test
3
KW test
ā€¢ The Kruskal-Wallis test is a nonparametric
test that can be used to determine whether
three or more independent samples were
selected from populations having the same
distribution.
H0: There is no difference in the group medians
Ha: There is a difference in the group medians
4
KW test
ā€¢ This test is appropriate for use under the
following circumstances:
ā€“ we have three or more conditions that you want to
compare;
ā€“ each condition is performed by a different group of
participants;
ā€“ the data do not meet the requirements for a
parametric test.
5
Requirements
ā€¢ The data:
ā€“ not normally distributed;
ā€“ if the variances for the different conditions are
markedly different;
ā€“ if the data are measurements on an ordinal scale.
ā€¢ If the data meet the requirements for a
parametric test, it is better to use a one-way
independent-measures Analysis of Variance
(ANOVA)
6
Given three or more independent samples, the test statistic H
for the Kruskal-Wallis test is
Where: N represents the number of participants,
Tc is the rank total for each group
nc is the number of participants in each group
Reject the null hypothesis when H is greater than the critical number.
(always use a right tail test.)
KW test
7
Procedure
1
ā€¢ Combine the observations of the various groups
2
ā€¢ Arrange them in order of magnitude from lowest to highest
3
ā€¢ Assign ranks to each of the observations and replace them in each of the
groups
4
ā€¢ Original ratio data has therefore been converted into ordinal or ranked
data
5
ā€¢ Ranks are summed in each group and the test statistic, H is computed
6
ā€¢ Ranks assigned to observations in each of the groups are added
separately to give rank sums
8
Example
ā€¢ Does physical exercise alleviate depression?
ā€“ We find some depressed people and check that they
are all equivalently depressed to begin with. Then we
allocate each person randomly to one of three
groups:
ā€¢ no exercise;
ā€¢ 20 minutes of jogging per day; or
ā€¢ 60 minutes of jogging per day.
ā€“ At the end of a month, we ask each participant to
rate how depressed they now feel, on a Likert scale
that runs from 1 ("totally miserable") through to 100
(ecstatically happy"). 9
Appropriate test?
ā€¢ We have three separate groups of participants.
ā€¢ In each group each participant gives us a single
score on a rating scale.
(Ratings are examples of an ordinal scale of
measurement, and so the data are not suitable
for a parametric test)
10
KW test
The Kruskal-Wallis test will tell us if the
differences between the groups are so large that
they are unlikely to have occurred by chance.
11
KW test
12
Step 1: Rank all of the scores
13
14
Step 2: Find "Tc", the total of the ranks for each group. Just add
together all of the ranks for each group in turn.
ā€¢ Tc1 (the rank total for the "no exercise" group)
is 76.5.
ā€¢ Tc2 (the rank total for the "20 minutes" group)
is 79.5.
ā€¢ Tc3 (the rank total for the "60 minutes" group)
is 144.
15
Step 3: Find "H".
16
KW test
17
18
Step 4: Calculate the degrees of freedom
ā€¢ The degrees of freedom is the number of
groups minus one.
Degrees of Freedom = Number of Groups - 1
ā€¢ Here we have three groups, and so we have 2
d.f.
19
Step 5: Assess the significance of H
ā€¢ Assessing the significance of H depends on
ā€“ the number of participants and
ā€“ the number of groups.
20
KW test
ā€¢ Three groups, with ā‰¤ 5 participants in each
group, then use the special table for small
sample sizes.
ā€¢ For more than five participants per group, treat
H as Chi-Square.
21
22
23
KW test
ā€¢ H is statistically significant if it is equal to or
larger than the critical value of Chi-Square for
particular df
ā€¢ Here, we have 8 participants per group, and so
we treat H as Chi-Square.
ā€¢ H is 7.27, with 2 df
24
KW test
ā€¢ Look along the row that corresponds to number
of degrees of freedom.
ā€¢ We look along the row for 2 df
ā€¢ We compare our obtained value of H to each of
the critical values in that row of the table,
starting on the left hand side and stopping once
our value of H is no longer equal to or larger
than the critical value.
25
KW test
ā€¢ So here, we start by comparing our H of 7.27 to
5.99. With 2 degrees of freedom, a value of Chi-
Square as large as 5.99 is likely to occur by
chance only 5 times in a hundred: i.e. it has a p
of .05.
ā€¢ Our obtained value of 7.27 is even larger than
this, and so this tells us that our value of H is
even less likely to occur by chance.
ā€¢ Our H will occur by chance with a probability
of less than 0.05. 26
KW test
ā€¢ Move on, and compare our H to the next value
in the row, 9.21. 9.21 will occur by chance one
time in a hundred, i.e. with a p of .01. However,
our H of 7.27 is less than 9.21, not bigger than
it.
ā€¢ This tells us that our value of H is not so large
that it is likely to occur with a probability of
0.01.
27
KW test
ā€¢ The likelihood of obtaining a value of H as
large as the one we've found, purely by chance,
is somewhere between 0.05 and 0.01 - i.e. pretty
unlikely, and so we would conclude that there is
a difference of some kind between our three
groups.
28
KW test
ā€¢ Note that the Kruskal-Wallis test merely tells you that
the groups differ in some way: you need to inspect the
group means or medians
ā€¢ However in this particular case, the interpretation
seems fairly straightforward: exercise does seem to
reduce self-reported depression, but only in the case of
participants who are doing 60 minutes.
ā€¢ There seems to be no difference between those
participants who took 20 minutes of exercise per day,
and those who did not exercise at all.
29
KW test
ā€¢ We could write this up as follows:
ā€¢ "A Kruskal-Wallis test revealed that there was a
significant effect of exercise on depression
levels (H (2) = 7.27, p < .05).
ā€¢ Inspection of the group means suggests that
compared to the "no exercise" control
condition, depression was significantly reduced
by 60 minutes of daily exercise, but not by 20
minutes of exercise".
30
KW test
ā€¢ The post hoc test for a significant KW is called
Dunnā€™s test with Bonferroni correction for
multiple testing
31
Friedman Test
32
Friedman Test
ā€¢ This is similar to the Wilcoxon singed rank test,
except that we can use it with three or more
conditions.
ā€¢ Each subject participates in all of the different
conditions of the experiment.
33
FT
ā€¢ Does background music affect the
performance of Laboratory workers?
ā€¢ We take a group of five workers, and measure
each worker's productivity (in terms of the
number of Stool Microscopies per hour) thrice
ā€“ once while the worker is listening to ā€œEasy-
listening musicā€™,
ā€“ once while the worker is listening to ā€˜Marching
Band music,ā€™ and
ā€“ once while the same worker is working in ā€˜Silence.ā€™34
Step 1:
Rank each subject's scores individually
Worker No Music Easy Listening Marching Band
Raw
Score
Ranked
Score
Raw
Score
Ranked
Score
Raw
Score
Ranked
Score
1 4 1 5 2 6 3
2 2 1 7 2.5 7 2.5
3 6 1.5 6 1.5 8 3
4 3 1 7 3 5 2
5 3 1 8 2 9 3
35
Step 2:
Find the rank total for each condition
Worker No Music Easy Listening Marching Band
Raw
Score
Ranked
Score
Raw
Score
Ranked
Score
Raw
Score
Ranked
Score
1 4 1 5 2 6 3
2 2 1 7 2.5 7 2.5
3 6 1.5 6 1.5 8 3
4 3 1 7 3 5 2
5 3 1 8 2 9 3
Total 5.5 11 13.5
36
Step 3:
Work out Ļ‡r2
ā€¢ Where :
ā€“ C is the number of conditions,
ā€“ N is the number of subjects
ā€“ Ī£Tc2 is the sum of the squared rank totals for
each condition.
37
FT
ā€¢ To get Ī£Tc2:
ā€¢ Take each rank total and square it.
Worker No Music Easy Listening Marching Band
Raw
Score
Ranked
Score
Raw
Score
Ranked
Score
Raw
Score
Ranked
Score
1 4 1 5 2 6 3
2 2 1 7 2.5 7 2.5
3 6 1.5 6 1.5 8 3
4 3 1 7 3 5 2
5 3 1 8 2 9 3
Total 5.5 11 13.5
Square of
Rank Total
30.25 121 182.25
38
FT
39
Step 4: Calculate degrees of
freedom
ā€¢ The degrees of freedom are given by the
number of conditions minus one.
C - 1 = 3 - 1 = 2.
40
Step 5:
Assess the statistical significance of Xr2
ā€¢ It depends on the number of subjects and the
number of groups.
ā€“ more than 9 subjects, we can use a Chi-Square table.
ā€“ less than nine subjects, special tables of critical
values are available.
ā€¢ Compare obtained Xr2 value to the critical value
of Chi-Square for df
41
Number of
Subjects
Number of
Conditions
42
FT
ā€¢ If your obtained value of Xr2 is bigger than the
critical Chi-Square value, the difference between
conditions is statistically significant.
ā€¢ As with the Kruskal- Wallis test, Friedman's test
only tells you that some kind of difference
exists;
ā€¢ Inspection of the median score for each
condition will usually be enough
ā€¢ Post hoc test for FT is called Dunnā€™s test
43
Spearmanā€™s Correlation
44
Types of correlation
ā€¢ Pearsonā€™s
ā€¢ Spearmanā€™s
ā€¢ Hoeffdingā€™s D
ā€¢ Distance correlation
ā€¢ Mutual Information and the Maximal
Information Coefficient
45
Correlation coefficient
ā€¢ A correlation coefficient is a succinct (single-
number) measure of the strength of association
between two variables.
ā€¢ There are various types of correlation
coefficient for different purposes.
ā€¢ Commonly used correlation coeff. are
ā€“ Pearson's r
ā€“ Spearman's rho
46
Difference between Pearsonā€™s r
and Spearmanā€™s rho
ā€¢ Spearmanā€™s rho makes fewer assumptions about the
nature of the data on which the correlation is to be
performed
ā€¢ The data need only be measured on an ordinal scale
ā€¢ Pearsonā€™s ā€œrā€ is a measure of the strength of the
linear relationship between two variables
ā€¢ Spearmanā€™s rho is a measure of the strength of the
monotonic relationship between them.
47
Difference between Pearsonā€™s r
and Spearmanā€™s rho
ā€¢ If a monotonic relationship exists, it simply means that
one of the variables increases (or decreases) by some
amount when the value of the other variable changes.
ā€¢ A linear relationship is thus a special case of a
monotonic relationship.
ā€¢ Thus, if there is a monotonic but non-linear
relationship between two variables, itā€™s better to use
Spearmanā€™s rho because Pearsonā€™s ā€œrā€ will tend to
underestimate the strength of the relationship.
48
Difference between Pearsonā€™s r
and Spearmanā€™s rho
ā€¢ Spearmanā€™s rho wonā€™t be ā€œfooledā€ by the fact
that the relationship isnā€™t a linear one.
49
Monotonic and non-monotonic
50
Linear relationships
Monotonic and non-monotonic
ā€¢ Monotonic variables
increase (or decrease)
in the same direction,
but not always at the
same rate.
ā€¢ Linear variables
increase (or decrease)
in the same direction
at the same rate.
51
How to decide a correlation test?
ā€¢ Before running a test, you should make a scatter plot first to
view the overall pattern of your data.
ā€¢ The strength and direction of a monotonic relationship between
two variables can be measured by the Spearman Rank-Order
Correlation.
ā€¢ If your variables are monotonic and linear, a more appropriate
test might be Pearsonā€™s Correlation as long as the assumptions
for Pearsonā€™s are met. for example, if your data is highly skewed,
has high kurtosis, or is heteroscedastic, you cannot use Pearsonā€™s.
You can, however, still use Spearmanā€™s, which is a non-
parametric test.
52
Example for Spearmanā€™s rho
ā€¢ Vitamin treatment enhances the memory of
people who take it
ā€¢ If there is a correlation between the number of
vitamin treatments that a subject has had, and
their performance on some memory test. We
would have two scores for each subject: number
of vitamin treatments, and that personā€™s
memory-test score.
53
Spearmanā€™s rho
54
Spearmanā€™s rho
55
Problems In Interpreting Correlations
ā€¢ Correlation does not imply causality
ā€¢ Factors affecting the size of the correlation
ā€“ The smaller the sample, the more likely it is to show
sampling variation, and the more variable the
correlation can be.
ā€“ Thus, when a correlation coefficient is calculated, it
does not represent the one-and-only correlation
between those two variables for that population.
ā€“ Different samples might produce higher or lower
correlations.
56
Sample and population
correlations
Here are the limits within which 80% of sample rā€™s will
fall, when the true correlation (i.e., in the population) is
zero
57
Requirements for Pearsonā€™s
correlation
Linearity of the relationship
Homoscedasticity
Effect of discontinuous distributions
58
Linearity of the relationship
ā€¢ Pearsonā€™s r is a measure of the linear relationship between two
variables and Spearmanā€™s rho is a measure of the monotonic
relationship between them.
ā€¢ Two variables may have a very strong relationship to each other,
but of a non-linear or non-monotonic kind (for example they
might have a curvilinear relationship).
ā€¢ If so, the correlation coefficient will underestimate the strength
of the relationship between the two variables.
ā€¢ This is why you should always make a scatterplot of the data, so
that you can look at the trend before calculating a correlation.
59
Homoscedasticity
ā€¢ Homoscedasticity (equal variability): For
Pearsonsā€™ r to be valid, scores should have a
reasonably constant amount of variability at all
points in their distribution
60
Effect of discontinuous
distributions
61
Take home messages
ā€¢ KW test NP equivalent of One Way ANOVA, >2
independent groups
ā€¢ FT test NP equivalent of Repeated Measures ANOVA, >2
paired groups
ā€¢ Spearmanā€™s rank order correlation NP equivalent of
Pearsonā€™s correlation for monotonic relationships
62
Thank you
Kindly email your queries to sarizwan1986@outlook.com
63

More Related Content

What's hot

Four steps to hypothesis testing
Four steps to hypothesis testingFour steps to hypothesis testing
Four steps to hypothesis testing
Hasnain Baber
Ā 
Friedman test Stat
Friedman test Stat Friedman test Stat
Friedman test Stat
Kate Malda
Ā 

What's hot (20)

Mann Whitney U Test | Statistics
Mann Whitney U Test | StatisticsMann Whitney U Test | Statistics
Mann Whitney U Test | Statistics
Ā 
Non parametric tests
Non parametric testsNon parametric tests
Non parametric tests
Ā 
NON-PARAMETRIC TESTS by Prajakta Sawant
NON-PARAMETRIC TESTS by Prajakta SawantNON-PARAMETRIC TESTS by Prajakta Sawant
NON-PARAMETRIC TESTS by Prajakta Sawant
Ā 
DIstinguish between Parametric vs nonparametric test
 DIstinguish between Parametric vs nonparametric test DIstinguish between Parametric vs nonparametric test
DIstinguish between Parametric vs nonparametric test
Ā 
One way anova final ppt.
One way anova final ppt.One way anova final ppt.
One way anova final ppt.
Ā 
Sign test
Sign testSign test
Sign test
Ā 
Four steps to hypothesis testing
Four steps to hypothesis testingFour steps to hypothesis testing
Four steps to hypothesis testing
Ā 
wilcoxon signed rank test
wilcoxon signed rank testwilcoxon signed rank test
wilcoxon signed rank test
Ā 
non parametric statistics
non parametric statisticsnon parametric statistics
non parametric statistics
Ā 
The mann whitney u test
The mann whitney u testThe mann whitney u test
The mann whitney u test
Ā 
Parametric tests
Parametric  testsParametric  tests
Parametric tests
Ā 
Advance Statistics - Wilcoxon Signed Rank Test
Advance Statistics - Wilcoxon Signed Rank TestAdvance Statistics - Wilcoxon Signed Rank Test
Advance Statistics - Wilcoxon Signed Rank Test
Ā 
Student t-test
Student t-testStudent t-test
Student t-test
Ā 
Friedman test Stat
Friedman test Stat Friedman test Stat
Friedman test Stat
Ā 
Regression ppt
Regression pptRegression ppt
Regression ppt
Ā 
Sample and sample size
Sample and sample sizeSample and sample size
Sample and sample size
Ā 
Sample size determination
Sample size determinationSample size determination
Sample size determination
Ā 
The t test
The t testThe t test
The t test
Ā 
Point Estimation
Point Estimation Point Estimation
Point Estimation
Ā 
Mann Whitney U test
Mann Whitney U testMann Whitney U test
Mann Whitney U test
Ā 

Similar to Kruskal Wallis test, Friedman test, Spearman Correlation

Research method ch08 statistical methods 2 anova
Research method ch08 statistical methods 2 anovaResearch method ch08 statistical methods 2 anova
Research method ch08 statistical methods 2 anova
naranbatn
Ā 
7- Quantitative Research- Part 3.pdf
7- Quantitative Research- Part 3.pdf7- Quantitative Research- Part 3.pdf
7- Quantitative Research- Part 3.pdf
ezaldeen2013
Ā 
allnonparametrictest-210427031923.pptx
allnonparametrictest-210427031923.pptxallnonparametrictest-210427031923.pptx
allnonparametrictest-210427031923.pptx
SoujanyaLk1
Ā 

Similar to Kruskal Wallis test, Friedman test, Spearman Correlation (20)

Kruskal wallis Friedman Spearman Dr Tajali Shora
Kruskal wallis Friedman Spearman Dr Tajali ShoraKruskal wallis Friedman Spearman Dr Tajali Shora
Kruskal wallis Friedman Spearman Dr Tajali Shora
Ā 
Workshop on Data Analysis and Result Interpretation in Social Science Researc...
Workshop on Data Analysis and Result Interpretation in Social Science Researc...Workshop on Data Analysis and Result Interpretation in Social Science Researc...
Workshop on Data Analysis and Result Interpretation in Social Science Researc...
Ā 
Non parametric-tests
Non parametric-testsNon parametric-tests
Non parametric-tests
Ā 
Research method ch08 statistical methods 2 anova
Research method ch08 statistical methods 2 anovaResearch method ch08 statistical methods 2 anova
Research method ch08 statistical methods 2 anova
Ā 
Statistics using SPSS
Statistics using SPSSStatistics using SPSS
Statistics using SPSS
Ā 
non parametric test.pptx
non parametric test.pptxnon parametric test.pptx
non parametric test.pptx
Ā 
chi_square test.pptx
chi_square test.pptxchi_square test.pptx
chi_square test.pptx
Ā 
univariate and bivariate analysis in spss
univariate and bivariate analysis in spss univariate and bivariate analysis in spss
univariate and bivariate analysis in spss
Ā 
Non parametric test
Non parametric testNon parametric test
Non parametric test
Ā 
Summary of statistical tools used in spss
Summary of statistical tools used in spssSummary of statistical tools used in spss
Summary of statistical tools used in spss
Ā 
Parametric tests
Parametric testsParametric tests
Parametric tests
Ā 
T test^jsample size^j ethics
T test^jsample size^j ethicsT test^jsample size^j ethics
T test^jsample size^j ethics
Ā 
Biomedical statistics
Biomedical statisticsBiomedical statistics
Biomedical statistics
Ā 
UNIT 5.pptx
UNIT 5.pptxUNIT 5.pptx
UNIT 5.pptx
Ā 
Statistics for Anaesthesiologists
Statistics for AnaesthesiologistsStatistics for Anaesthesiologists
Statistics for Anaesthesiologists
Ā 
7- Quantitative Research- Part 3.pdf
7- Quantitative Research- Part 3.pdf7- Quantitative Research- Part 3.pdf
7- Quantitative Research- Part 3.pdf
Ā 
Chi square test final
Chi square test finalChi square test final
Chi square test final
Ā 
Some statistical concepts relevant to proteomics data analysis
Some statistical concepts relevant to proteomics data analysisSome statistical concepts relevant to proteomics data analysis
Some statistical concepts relevant to proteomics data analysis
Ā 
allnonparametrictest-210427031923.pptx
allnonparametrictest-210427031923.pptxallnonparametrictest-210427031923.pptx
allnonparametrictest-210427031923.pptx
Ā 
All non parametric test
All non parametric testAll non parametric test
All non parametric test
Ā 

More from Rizwan S A

More from Rizwan S A (20)

Introduction to scoping reviews
Introduction to scoping reviewsIntroduction to scoping reviews
Introduction to scoping reviews
Ā 
Sources of demographic data 2019
Sources of demographic data 2019Sources of demographic data 2019
Sources of demographic data 2019
Ā 
Effect sizes in meta-analysis
Effect sizes in meta-analysisEffect sizes in meta-analysis
Effect sizes in meta-analysis
Ā 
Presenting the results of meta-analysis
Presenting the results of meta-analysisPresenting the results of meta-analysis
Presenting the results of meta-analysis
Ā 
Heterogeneity in meta-analysis
Heterogeneity in meta-analysisHeterogeneity in meta-analysis
Heterogeneity in meta-analysis
Ā 
Overview of the systematic review process
Overview of the systematic review processOverview of the systematic review process
Overview of the systematic review process
Ā 
Biases in meta-analysis
Biases in meta-analysisBiases in meta-analysis
Biases in meta-analysis
Ā 
Moderator analysis in meta-analysis
Moderator analysis in meta-analysisModerator analysis in meta-analysis
Moderator analysis in meta-analysis
Ā 
Fixed-effect and random-effects models in meta-analysis
Fixed-effect and random-effects models in meta-analysisFixed-effect and random-effects models in meta-analysis
Fixed-effect and random-effects models in meta-analysis
Ā 
Inverse variance method of meta-analysis and Cochran's Q
Inverse variance method of meta-analysis and Cochran's QInverse variance method of meta-analysis and Cochran's Q
Inverse variance method of meta-analysis and Cochran's Q
Ā 
Data extraction/coding and database structure in meta-analysis
Data extraction/coding and database structure in meta-analysisData extraction/coding and database structure in meta-analysis
Data extraction/coding and database structure in meta-analysis
Ā 
Introduction & rationale for meta-analysis
Introduction & rationale for meta-analysisIntroduction & rationale for meta-analysis
Introduction & rationale for meta-analysis
Ā 
Types of correlation coefficients
Types of correlation coefficientsTypes of correlation coefficients
Types of correlation coefficients
Ā 
Checking for normality (Normal distribution)
Checking for normality (Normal distribution)Checking for normality (Normal distribution)
Checking for normality (Normal distribution)
Ā 
Analysis of small datasets
Analysis of small datasetsAnalysis of small datasets
Analysis of small datasets
Ā 
A introduction to non-parametric tests
A introduction to non-parametric testsA introduction to non-parametric tests
A introduction to non-parametric tests
Ā 
Kolmogorov Smirnov good-of-fit test
Kolmogorov Smirnov good-of-fit testKolmogorov Smirnov good-of-fit test
Kolmogorov Smirnov good-of-fit test
Ā 
Mantel Haenszel methods in epidemiology (Stratification)
Mantel Haenszel methods in epidemiology (Stratification) Mantel Haenszel methods in epidemiology (Stratification)
Mantel Haenszel methods in epidemiology (Stratification)
Ā 
Use of checklists in critical appraisal of health literature
Use of checklists in critical appraisal of health literatureUse of checklists in critical appraisal of health literature
Use of checklists in critical appraisal of health literature
Ā 
Critical Appraisal of health literature
Critical Appraisal of health literatureCritical Appraisal of health literature
Critical Appraisal of health literature
Ā 

Recently uploaded

Jaipur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Jaipur NošŸ’°...
Jaipur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Jaipur NošŸ’°...Jaipur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Jaipur NošŸ’°...
Jaipur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Jaipur NošŸ’°...
Sheetaleventcompany
Ā 
šŸ‘‰ Amritsar Call Girls šŸ‘‰šŸ“ž 8725944379 šŸ‘‰šŸ“ž JustšŸ“² Call Ruhi Call Girl Near Me Amri...
šŸ‘‰ Amritsar Call Girls šŸ‘‰šŸ“ž 8725944379 šŸ‘‰šŸ“ž JustšŸ“² Call Ruhi Call Girl Near Me Amri...šŸ‘‰ Amritsar Call Girls šŸ‘‰šŸ“ž 8725944379 šŸ‘‰šŸ“ž JustšŸ“² Call Ruhi Call Girl Near Me Amri...
šŸ‘‰ Amritsar Call Girls šŸ‘‰šŸ“ž 8725944379 šŸ‘‰šŸ“ž JustšŸ“² Call Ruhi Call Girl Near Me Amri...
Sheetaleventcompany
Ā 
Pune Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Pune NošŸ’°Adva...
Pune Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Pune NošŸ’°Adva...Pune Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Pune NošŸ’°Adva...
Pune Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Pune NošŸ’°Adva...
Sheetaleventcompany
Ā 
Electrocardiogram (ECG) physiological basis .pdf
Electrocardiogram (ECG) physiological basis .pdfElectrocardiogram (ECG) physiological basis .pdf
Electrocardiogram (ECG) physiological basis .pdf
MedicoseAcademics
Ā 
Nagpur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Nagpur NošŸ’°...
Nagpur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Nagpur NošŸ’°...Nagpur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Nagpur NošŸ’°...
Nagpur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Nagpur NošŸ’°...
Sheetaleventcompany
Ā 
šŸ’šChandigarh Call Girls šŸ’ÆRiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh NošŸ’°Advance...
šŸ’šChandigarh Call Girls šŸ’ÆRiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh NošŸ’°Advance...šŸ’šChandigarh Call Girls šŸ’ÆRiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh NošŸ’°Advance...
šŸ’šChandigarh Call Girls šŸ’ÆRiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh NošŸ’°Advance...
Sheetaleventcompany
Ā 
Cara Menggugurkan Kandungan Dengan Cepat Selesai Dalam 24 Jam Secara Alami Bu...
Cara Menggugurkan Kandungan Dengan Cepat Selesai Dalam 24 Jam Secara Alami Bu...Cara Menggugurkan Kandungan Dengan Cepat Selesai Dalam 24 Jam Secara Alami Bu...
Cara Menggugurkan Kandungan Dengan Cepat Selesai Dalam 24 Jam Secara Alami Bu...
Cara Menggugurkan Kandungan 087776558899
Ā 
Premium Call Girls Dehradun {8854095900} ā¤ļøVVIP ANJU Call Girls in Dehradun U...
Premium Call Girls Dehradun {8854095900} ā¤ļøVVIP ANJU Call Girls in Dehradun U...Premium Call Girls Dehradun {8854095900} ā¤ļøVVIP ANJU Call Girls in Dehradun U...
Premium Call Girls Dehradun {8854095900} ā¤ļøVVIP ANJU Call Girls in Dehradun U...
Sheetaleventcompany
Ā 
šŸ’šChandigarh Call Girls Service šŸ’ÆPiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh No...
šŸ’šChandigarh Call Girls Service šŸ’ÆPiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh No...šŸ’šChandigarh Call Girls Service šŸ’ÆPiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh No...
šŸ’šChandigarh Call Girls Service šŸ’ÆPiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh No...
Sheetaleventcompany
Ā 
šŸ‘‰Chandigarh Call Girl ServicešŸ“²Niamh 8868886958 šŸ“²Book 24hours NowšŸ“²šŸ‘‰Sexy Call G...
šŸ‘‰Chandigarh Call Girl ServicešŸ“²Niamh 8868886958 šŸ“²Book 24hours NowšŸ“²šŸ‘‰Sexy Call G...šŸ‘‰Chandigarh Call Girl ServicešŸ“²Niamh 8868886958 šŸ“²Book 24hours NowšŸ“²šŸ‘‰Sexy Call G...
šŸ‘‰Chandigarh Call Girl ServicešŸ“²Niamh 8868886958 šŸ“²Book 24hours NowšŸ“²šŸ‘‰Sexy Call G...
Sheetaleventcompany
Ā 

Recently uploaded (20)

Call Girls Bangalore - 450+ Call Girl Cash Payment šŸ’ÆCall Us šŸ” 6378878445 šŸ” šŸ’ƒ ...
Call Girls Bangalore - 450+ Call Girl Cash Payment šŸ’ÆCall Us šŸ” 6378878445 šŸ” šŸ’ƒ ...Call Girls Bangalore - 450+ Call Girl Cash Payment šŸ’ÆCall Us šŸ” 6378878445 šŸ” šŸ’ƒ ...
Call Girls Bangalore - 450+ Call Girl Cash Payment šŸ’ÆCall Us šŸ” 6378878445 šŸ” šŸ’ƒ ...
Ā 
Jaipur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Jaipur NošŸ’°...
Jaipur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Jaipur NošŸ’°...Jaipur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Jaipur NošŸ’°...
Jaipur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Jaipur NošŸ’°...
Ā 
ā¤ļøChandigarh Escorts Serviceā˜Žļø9814379184ā˜Žļø Call Girl service in Chandigarhā˜Žļø ...
ā¤ļøChandigarh Escorts Serviceā˜Žļø9814379184ā˜Žļø Call Girl service in Chandigarhā˜Žļø ...ā¤ļøChandigarh Escorts Serviceā˜Žļø9814379184ā˜Žļø Call Girl service in Chandigarhā˜Žļø ...
ā¤ļøChandigarh Escorts Serviceā˜Žļø9814379184ā˜Žļø Call Girl service in Chandigarhā˜Žļø ...
Ā 
šŸ‘‰ Amritsar Call Girls šŸ‘‰šŸ“ž 8725944379 šŸ‘‰šŸ“ž JustšŸ“² Call Ruhi Call Girl Near Me Amri...
šŸ‘‰ Amritsar Call Girls šŸ‘‰šŸ“ž 8725944379 šŸ‘‰šŸ“ž JustšŸ“² Call Ruhi Call Girl Near Me Amri...šŸ‘‰ Amritsar Call Girls šŸ‘‰šŸ“ž 8725944379 šŸ‘‰šŸ“ž JustšŸ“² Call Ruhi Call Girl Near Me Amri...
šŸ‘‰ Amritsar Call Girls šŸ‘‰šŸ“ž 8725944379 šŸ‘‰šŸ“ž JustšŸ“² Call Ruhi Call Girl Near Me Amri...
Ā 
Pune Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Pune NošŸ’°Adva...
Pune Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Pune NošŸ’°Adva...Pune Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Pune NošŸ’°Adva...
Pune Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Pune NošŸ’°Adva...
Ā 
Electrocardiogram (ECG) physiological basis .pdf
Electrocardiogram (ECG) physiological basis .pdfElectrocardiogram (ECG) physiological basis .pdf
Electrocardiogram (ECG) physiological basis .pdf
Ā 
Gastric Cancer: Š”linical Implementation of Artificial Intelligence, Synergeti...
Gastric Cancer: Š”linical Implementation of Artificial Intelligence, Synergeti...Gastric Cancer: Š”linical Implementation of Artificial Intelligence, Synergeti...
Gastric Cancer: Š”linical Implementation of Artificial Intelligence, Synergeti...
Ā 
Bandra East [ best call girls in Mumbai Get 50% Off On VIP Escorts Service 90...
Bandra East [ best call girls in Mumbai Get 50% Off On VIP Escorts Service 90...Bandra East [ best call girls in Mumbai Get 50% Off On VIP Escorts Service 90...
Bandra East [ best call girls in Mumbai Get 50% Off On VIP Escorts Service 90...
Ā 
tongue disease lecture Dr Assadawy legacy
tongue disease lecture Dr Assadawy legacytongue disease lecture Dr Assadawy legacy
tongue disease lecture Dr Assadawy legacy
Ā 
Nagpur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Nagpur NošŸ’°...
Nagpur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Nagpur NošŸ’°...Nagpur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Nagpur NošŸ’°...
Nagpur Call Girl Service šŸ“ž9xx000xx09šŸ“žJust Call DivyašŸ“² Call Girl In Nagpur NošŸ’°...
Ā 
Ahmedabad Call Girls Book Now 9630942363 Top Class Ahmedabad Escort Service A...
Ahmedabad Call Girls Book Now 9630942363 Top Class Ahmedabad Escort Service A...Ahmedabad Call Girls Book Now 9630942363 Top Class Ahmedabad Escort Service A...
Ahmedabad Call Girls Book Now 9630942363 Top Class Ahmedabad Escort Service A...
Ā 
šŸ’šChandigarh Call Girls šŸ’ÆRiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh NošŸ’°Advance...
šŸ’šChandigarh Call Girls šŸ’ÆRiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh NošŸ’°Advance...šŸ’šChandigarh Call Girls šŸ’ÆRiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh NošŸ’°Advance...
šŸ’šChandigarh Call Girls šŸ’ÆRiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh NošŸ’°Advance...
Ā 
Circulatory Shock, types and stages, compensatory mechanisms
Circulatory Shock, types and stages, compensatory mechanismsCirculatory Shock, types and stages, compensatory mechanisms
Circulatory Shock, types and stages, compensatory mechanisms
Ā 
Cara Menggugurkan Kandungan Dengan Cepat Selesai Dalam 24 Jam Secara Alami Bu...
Cara Menggugurkan Kandungan Dengan Cepat Selesai Dalam 24 Jam Secara Alami Bu...Cara Menggugurkan Kandungan Dengan Cepat Selesai Dalam 24 Jam Secara Alami Bu...
Cara Menggugurkan Kandungan Dengan Cepat Selesai Dalam 24 Jam Secara Alami Bu...
Ā 
Premium Call Girls Dehradun {8854095900} ā¤ļøVVIP ANJU Call Girls in Dehradun U...
Premium Call Girls Dehradun {8854095900} ā¤ļøVVIP ANJU Call Girls in Dehradun U...Premium Call Girls Dehradun {8854095900} ā¤ļøVVIP ANJU Call Girls in Dehradun U...
Premium Call Girls Dehradun {8854095900} ā¤ļøVVIP ANJU Call Girls in Dehradun U...
Ā 
šŸšŗLEELA JOSHI WhatsApp Number +91-9930245274 āœ” Unsatisfied Bhabhi Call Girls T...
šŸšŗLEELA JOSHI WhatsApp Number +91-9930245274 āœ” Unsatisfied Bhabhi Call Girls T...šŸšŗLEELA JOSHI WhatsApp Number +91-9930245274 āœ” Unsatisfied Bhabhi Call Girls T...
šŸšŗLEELA JOSHI WhatsApp Number +91-9930245274 āœ” Unsatisfied Bhabhi Call Girls T...
Ā 
ANATOMY AND PHYSIOLOGY OF REPRODUCTIVE SYSTEM.pptx
ANATOMY AND PHYSIOLOGY OF REPRODUCTIVE SYSTEM.pptxANATOMY AND PHYSIOLOGY OF REPRODUCTIVE SYSTEM.pptx
ANATOMY AND PHYSIOLOGY OF REPRODUCTIVE SYSTEM.pptx
Ā 
šŸ’šChandigarh Call Girls Service šŸ’ÆPiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh No...
šŸ’šChandigarh Call Girls Service šŸ’ÆPiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh No...šŸ’šChandigarh Call Girls Service šŸ’ÆPiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh No...
šŸ’šChandigarh Call Girls Service šŸ’ÆPiya šŸ“²šŸ”8868886958šŸ”Call Girls In Chandigarh No...
Ā 
šŸ‘‰Chandigarh Call Girl ServicešŸ“²Niamh 8868886958 šŸ“²Book 24hours NowšŸ“²šŸ‘‰Sexy Call G...
šŸ‘‰Chandigarh Call Girl ServicešŸ“²Niamh 8868886958 šŸ“²Book 24hours NowšŸ“²šŸ‘‰Sexy Call G...šŸ‘‰Chandigarh Call Girl ServicešŸ“²Niamh 8868886958 šŸ“²Book 24hours NowšŸ“²šŸ‘‰Sexy Call G...
šŸ‘‰Chandigarh Call Girl ServicešŸ“²Niamh 8868886958 šŸ“²Book 24hours NowšŸ“²šŸ‘‰Sexy Call G...
Ā 
šŸ’°Call Girl In Bangaloreā˜Žļø63788-78445šŸ’° Call Girl service in Bangaloreā˜ŽļøBangalo...
šŸ’°Call Girl In Bangaloreā˜Žļø63788-78445šŸ’° Call Girl service in Bangaloreā˜ŽļøBangalo...šŸ’°Call Girl In Bangaloreā˜Žļø63788-78445šŸ’° Call Girl service in Bangaloreā˜ŽļøBangalo...
šŸ’°Call Girl In Bangaloreā˜Žļø63788-78445šŸ’° Call Girl service in Bangaloreā˜ŽļøBangalo...
Ā 

Kruskal Wallis test, Friedman test, Spearman Correlation

  • 1. Kruskal-Wallis test Friedman test Spearmanā€™s Correlation Dr. S. A. Rizwan, M.D., Public Health Specialist, Saudi Board of Preventive Medicine, Riyadh, Kingdom of Saudi Arabia. With thanks to Dr. Tajali Nazir Shora
  • 2. Checklist of four questions ā€¢ Q1. What scales of measurement has been used? ā€¢ Q2. Which hypothesis has been tested? ā€¢ Q3. If the hypothesis of difference has been tested, are the samples independent or dependent? ā€¢ Q4. How many sets of measures are involved? 2
  • 4. KW test ā€¢ The Kruskal-Wallis test is a nonparametric test that can be used to determine whether three or more independent samples were selected from populations having the same distribution. H0: There is no difference in the group medians Ha: There is a difference in the group medians 4
  • 5. KW test ā€¢ This test is appropriate for use under the following circumstances: ā€“ we have three or more conditions that you want to compare; ā€“ each condition is performed by a different group of participants; ā€“ the data do not meet the requirements for a parametric test. 5
  • 6. Requirements ā€¢ The data: ā€“ not normally distributed; ā€“ if the variances for the different conditions are markedly different; ā€“ if the data are measurements on an ordinal scale. ā€¢ If the data meet the requirements for a parametric test, it is better to use a one-way independent-measures Analysis of Variance (ANOVA) 6
  • 7. Given three or more independent samples, the test statistic H for the Kruskal-Wallis test is Where: N represents the number of participants, Tc is the rank total for each group nc is the number of participants in each group Reject the null hypothesis when H is greater than the critical number. (always use a right tail test.) KW test 7
  • 8. Procedure 1 ā€¢ Combine the observations of the various groups 2 ā€¢ Arrange them in order of magnitude from lowest to highest 3 ā€¢ Assign ranks to each of the observations and replace them in each of the groups 4 ā€¢ Original ratio data has therefore been converted into ordinal or ranked data 5 ā€¢ Ranks are summed in each group and the test statistic, H is computed 6 ā€¢ Ranks assigned to observations in each of the groups are added separately to give rank sums 8
  • 9. Example ā€¢ Does physical exercise alleviate depression? ā€“ We find some depressed people and check that they are all equivalently depressed to begin with. Then we allocate each person randomly to one of three groups: ā€¢ no exercise; ā€¢ 20 minutes of jogging per day; or ā€¢ 60 minutes of jogging per day. ā€“ At the end of a month, we ask each participant to rate how depressed they now feel, on a Likert scale that runs from 1 ("totally miserable") through to 100 (ecstatically happy"). 9
  • 10. Appropriate test? ā€¢ We have three separate groups of participants. ā€¢ In each group each participant gives us a single score on a rating scale. (Ratings are examples of an ordinal scale of measurement, and so the data are not suitable for a parametric test) 10
  • 11. KW test The Kruskal-Wallis test will tell us if the differences between the groups are so large that they are unlikely to have occurred by chance. 11
  • 13. Step 1: Rank all of the scores 13
  • 14. 14
  • 15. Step 2: Find "Tc", the total of the ranks for each group. Just add together all of the ranks for each group in turn. ā€¢ Tc1 (the rank total for the "no exercise" group) is 76.5. ā€¢ Tc2 (the rank total for the "20 minutes" group) is 79.5. ā€¢ Tc3 (the rank total for the "60 minutes" group) is 144. 15
  • 16. Step 3: Find "H". 16
  • 18. 18
  • 19. Step 4: Calculate the degrees of freedom ā€¢ The degrees of freedom is the number of groups minus one. Degrees of Freedom = Number of Groups - 1 ā€¢ Here we have three groups, and so we have 2 d.f. 19
  • 20. Step 5: Assess the significance of H ā€¢ Assessing the significance of H depends on ā€“ the number of participants and ā€“ the number of groups. 20
  • 21. KW test ā€¢ Three groups, with ā‰¤ 5 participants in each group, then use the special table for small sample sizes. ā€¢ For more than five participants per group, treat H as Chi-Square. 21
  • 22. 22
  • 23. 23
  • 24. KW test ā€¢ H is statistically significant if it is equal to or larger than the critical value of Chi-Square for particular df ā€¢ Here, we have 8 participants per group, and so we treat H as Chi-Square. ā€¢ H is 7.27, with 2 df 24
  • 25. KW test ā€¢ Look along the row that corresponds to number of degrees of freedom. ā€¢ We look along the row for 2 df ā€¢ We compare our obtained value of H to each of the critical values in that row of the table, starting on the left hand side and stopping once our value of H is no longer equal to or larger than the critical value. 25
  • 26. KW test ā€¢ So here, we start by comparing our H of 7.27 to 5.99. With 2 degrees of freedom, a value of Chi- Square as large as 5.99 is likely to occur by chance only 5 times in a hundred: i.e. it has a p of .05. ā€¢ Our obtained value of 7.27 is even larger than this, and so this tells us that our value of H is even less likely to occur by chance. ā€¢ Our H will occur by chance with a probability of less than 0.05. 26
  • 27. KW test ā€¢ Move on, and compare our H to the next value in the row, 9.21. 9.21 will occur by chance one time in a hundred, i.e. with a p of .01. However, our H of 7.27 is less than 9.21, not bigger than it. ā€¢ This tells us that our value of H is not so large that it is likely to occur with a probability of 0.01. 27
  • 28. KW test ā€¢ The likelihood of obtaining a value of H as large as the one we've found, purely by chance, is somewhere between 0.05 and 0.01 - i.e. pretty unlikely, and so we would conclude that there is a difference of some kind between our three groups. 28
  • 29. KW test ā€¢ Note that the Kruskal-Wallis test merely tells you that the groups differ in some way: you need to inspect the group means or medians ā€¢ However in this particular case, the interpretation seems fairly straightforward: exercise does seem to reduce self-reported depression, but only in the case of participants who are doing 60 minutes. ā€¢ There seems to be no difference between those participants who took 20 minutes of exercise per day, and those who did not exercise at all. 29
  • 30. KW test ā€¢ We could write this up as follows: ā€¢ "A Kruskal-Wallis test revealed that there was a significant effect of exercise on depression levels (H (2) = 7.27, p < .05). ā€¢ Inspection of the group means suggests that compared to the "no exercise" control condition, depression was significantly reduced by 60 minutes of daily exercise, but not by 20 minutes of exercise". 30
  • 31. KW test ā€¢ The post hoc test for a significant KW is called Dunnā€™s test with Bonferroni correction for multiple testing 31
  • 33. Friedman Test ā€¢ This is similar to the Wilcoxon singed rank test, except that we can use it with three or more conditions. ā€¢ Each subject participates in all of the different conditions of the experiment. 33
  • 34. FT ā€¢ Does background music affect the performance of Laboratory workers? ā€¢ We take a group of five workers, and measure each worker's productivity (in terms of the number of Stool Microscopies per hour) thrice ā€“ once while the worker is listening to ā€œEasy- listening musicā€™, ā€“ once while the worker is listening to ā€˜Marching Band music,ā€™ and ā€“ once while the same worker is working in ā€˜Silence.ā€™34
  • 35. Step 1: Rank each subject's scores individually Worker No Music Easy Listening Marching Band Raw Score Ranked Score Raw Score Ranked Score Raw Score Ranked Score 1 4 1 5 2 6 3 2 2 1 7 2.5 7 2.5 3 6 1.5 6 1.5 8 3 4 3 1 7 3 5 2 5 3 1 8 2 9 3 35
  • 36. Step 2: Find the rank total for each condition Worker No Music Easy Listening Marching Band Raw Score Ranked Score Raw Score Ranked Score Raw Score Ranked Score 1 4 1 5 2 6 3 2 2 1 7 2.5 7 2.5 3 6 1.5 6 1.5 8 3 4 3 1 7 3 5 2 5 3 1 8 2 9 3 Total 5.5 11 13.5 36
  • 37. Step 3: Work out Ļ‡r2 ā€¢ Where : ā€“ C is the number of conditions, ā€“ N is the number of subjects ā€“ Ī£Tc2 is the sum of the squared rank totals for each condition. 37
  • 38. FT ā€¢ To get Ī£Tc2: ā€¢ Take each rank total and square it. Worker No Music Easy Listening Marching Band Raw Score Ranked Score Raw Score Ranked Score Raw Score Ranked Score 1 4 1 5 2 6 3 2 2 1 7 2.5 7 2.5 3 6 1.5 6 1.5 8 3 4 3 1 7 3 5 2 5 3 1 8 2 9 3 Total 5.5 11 13.5 Square of Rank Total 30.25 121 182.25 38
  • 39. FT 39
  • 40. Step 4: Calculate degrees of freedom ā€¢ The degrees of freedom are given by the number of conditions minus one. C - 1 = 3 - 1 = 2. 40
  • 41. Step 5: Assess the statistical significance of Xr2 ā€¢ It depends on the number of subjects and the number of groups. ā€“ more than 9 subjects, we can use a Chi-Square table. ā€“ less than nine subjects, special tables of critical values are available. ā€¢ Compare obtained Xr2 value to the critical value of Chi-Square for df 41
  • 43. FT ā€¢ If your obtained value of Xr2 is bigger than the critical Chi-Square value, the difference between conditions is statistically significant. ā€¢ As with the Kruskal- Wallis test, Friedman's test only tells you that some kind of difference exists; ā€¢ Inspection of the median score for each condition will usually be enough ā€¢ Post hoc test for FT is called Dunnā€™s test 43
  • 45. Types of correlation ā€¢ Pearsonā€™s ā€¢ Spearmanā€™s ā€¢ Hoeffdingā€™s D ā€¢ Distance correlation ā€¢ Mutual Information and the Maximal Information Coefficient 45
  • 46. Correlation coefficient ā€¢ A correlation coefficient is a succinct (single- number) measure of the strength of association between two variables. ā€¢ There are various types of correlation coefficient for different purposes. ā€¢ Commonly used correlation coeff. are ā€“ Pearson's r ā€“ Spearman's rho 46
  • 47. Difference between Pearsonā€™s r and Spearmanā€™s rho ā€¢ Spearmanā€™s rho makes fewer assumptions about the nature of the data on which the correlation is to be performed ā€¢ The data need only be measured on an ordinal scale ā€¢ Pearsonā€™s ā€œrā€ is a measure of the strength of the linear relationship between two variables ā€¢ Spearmanā€™s rho is a measure of the strength of the monotonic relationship between them. 47
  • 48. Difference between Pearsonā€™s r and Spearmanā€™s rho ā€¢ If a monotonic relationship exists, it simply means that one of the variables increases (or decreases) by some amount when the value of the other variable changes. ā€¢ A linear relationship is thus a special case of a monotonic relationship. ā€¢ Thus, if there is a monotonic but non-linear relationship between two variables, itā€™s better to use Spearmanā€™s rho because Pearsonā€™s ā€œrā€ will tend to underestimate the strength of the relationship. 48
  • 49. Difference between Pearsonā€™s r and Spearmanā€™s rho ā€¢ Spearmanā€™s rho wonā€™t be ā€œfooledā€ by the fact that the relationship isnā€™t a linear one. 49
  • 51. Monotonic and non-monotonic ā€¢ Monotonic variables increase (or decrease) in the same direction, but not always at the same rate. ā€¢ Linear variables increase (or decrease) in the same direction at the same rate. 51
  • 52. How to decide a correlation test? ā€¢ Before running a test, you should make a scatter plot first to view the overall pattern of your data. ā€¢ The strength and direction of a monotonic relationship between two variables can be measured by the Spearman Rank-Order Correlation. ā€¢ If your variables are monotonic and linear, a more appropriate test might be Pearsonā€™s Correlation as long as the assumptions for Pearsonā€™s are met. for example, if your data is highly skewed, has high kurtosis, or is heteroscedastic, you cannot use Pearsonā€™s. You can, however, still use Spearmanā€™s, which is a non- parametric test. 52
  • 53. Example for Spearmanā€™s rho ā€¢ Vitamin treatment enhances the memory of people who take it ā€¢ If there is a correlation between the number of vitamin treatments that a subject has had, and their performance on some memory test. We would have two scores for each subject: number of vitamin treatments, and that personā€™s memory-test score. 53
  • 56. Problems In Interpreting Correlations ā€¢ Correlation does not imply causality ā€¢ Factors affecting the size of the correlation ā€“ The smaller the sample, the more likely it is to show sampling variation, and the more variable the correlation can be. ā€“ Thus, when a correlation coefficient is calculated, it does not represent the one-and-only correlation between those two variables for that population. ā€“ Different samples might produce higher or lower correlations. 56
  • 57. Sample and population correlations Here are the limits within which 80% of sample rā€™s will fall, when the true correlation (i.e., in the population) is zero 57
  • 58. Requirements for Pearsonā€™s correlation Linearity of the relationship Homoscedasticity Effect of discontinuous distributions 58
  • 59. Linearity of the relationship ā€¢ Pearsonā€™s r is a measure of the linear relationship between two variables and Spearmanā€™s rho is a measure of the monotonic relationship between them. ā€¢ Two variables may have a very strong relationship to each other, but of a non-linear or non-monotonic kind (for example they might have a curvilinear relationship). ā€¢ If so, the correlation coefficient will underestimate the strength of the relationship between the two variables. ā€¢ This is why you should always make a scatterplot of the data, so that you can look at the trend before calculating a correlation. 59
  • 60. Homoscedasticity ā€¢ Homoscedasticity (equal variability): For Pearsonsā€™ r to be valid, scores should have a reasonably constant amount of variability at all points in their distribution 60
  • 62. Take home messages ā€¢ KW test NP equivalent of One Way ANOVA, >2 independent groups ā€¢ FT test NP equivalent of Repeated Measures ANOVA, >2 paired groups ā€¢ Spearmanā€™s rank order correlation NP equivalent of Pearsonā€™s correlation for monotonic relationships 62
  • 63. Thank you Kindly email your queries to sarizwan1986@outlook.com 63