SlideShare a Scribd company logo
1 of 4
Download to read offline
IOSR Journal of Mathematics (IOSR-JM)
e-ISSN: 2278-5728.Volume 5, Issue 6 (Mar. - Apr. 2013), PP 20-23
www.iosrjournals.org
www.iosrjournals.org 20 | Page
Post Component Analysis of Categorical data
U. Sangeetha1
, M. Subbiah2
, M.R. Srinivasan3
1(
Department of Management Studies, SSN College of Engineering, Chennai, India)
2
(Department of Mathematics, L. N. Government College, Ponneri, India)
3
(Department of Statistics, University of Madras, Chennai, India)
Abstract: Categorical data are often stratified into two dimensional I × J tables in order to test the
independence of attributes. Agresti (1999) has discussed testing methods with partitioning property of chi
square to extract components that describe certain aspects of the overall association in a table. In this work an
attempt has been made to study the pattern in which sub-tables exhibit a sign of reverse association when
compared to a significant association of attributes with regard to the original I × J table. Simulation studies
and subsequent results of vote counting method indicate that 2 × 2 tables have the reversal component
association when compared to higher order sub-tables; interestingly, in more than 90% cases. The
computationally extensive and exhaustive procedure provide a better tool to understand the association between
the categories that focus on the strongest differences among all comparisons, and could be practically and other
aspects in experimental studies such as clinical trials.
Keywords: Contingency tables, Chi-Square test, Component Analysis.
I. INTRODUCTION
Categorical data analysis has found a rapid development since 1960’s which might be due to the
methodological sophistication of the social and biomedical sciences. It is not uncommon in social sciences for
measuring attitudes and opinions on various issues and demographic characteristics such as gender race and
social class. Categorical scales have been used extensively in biomedical sciences to measure many factors, such
as severity of injury, degree of recovery from surgery, and stage of a disease etc. This progress has been found
as an outcome of stimulus ties between social scientists and / or biomedical scientists and Statisticians.
However, one of the most influential contributions by Karl Pearson and Yule on association between categorical
variables has also been used extensively in many research works and by the practitioners in this large data
driven era. The data for this analysis are displayed in a table with I rows and J columns and many software
perform this analysis to provide a test of significance using a chi-square analysis testing the null hypothesis that
there is no association between the two variables.
If the chi-square test results in a p-value of 0.05or smaller, the results are deemed significant, the null
hypothesis is rejected. In such cases when there are only two groups being compared (a 2 × 2 table) it is easy to
observe the association between them; however when there are three or more groups, there has been a need to
identify the significant/non-significant association between pairs of groups. An extensive literature such as
Berkson, (1938), Norton, (1945), Simpson, (1951), Cochran, (1954), Mantel and Haenszel, (1959), Goodman,
(1964), Goodman, (1969), Blyth, 1972, Agresti, (1990), Agresti (1999), Sergio and Paulette (2011) provides a
method based on portioning of chi-square it has laid a practical limitation to the users, in that partitioning the
given I × J table satisfying the basic assumptions of the procedure.
This paper has identified a need based analysis between the practice and the theory and has provided a
procedure based on the post component analysis which is similar to the multiple comparisons in ANOVA for
continuous response variables. A primitive approach and subsequent analysis issues have been discussed with
limited size of the underlying classification table; this paper includes a simulation study to understand the
decisions on the Reversal Association Pattern (RAP) in a I × J table. Also, this attempt has laid a foundation to
visualize this post component analysis under various model settings and with large contingency tables.
II. METHOD
A bivariate relationship has been defined by the joint distribution of the two associated variables. If X
and Y denote two categorical response variables, X having I levels and Y having J levels then the IJ possible
combinations of classification could be displayed in a contingency table, where the cells contains frequency
counts of outcomes. Such a contingency table is referred to as an I × J table.
In a two-way contingency tables, the null hypothesis of statistical independence has been tested using
Pearson chi-square statistic (Agresti 1990)
Post Component Analysis of Categorical data
www.iosrjournals.org 21 | Page
χ2
=ΣΣ
nij-mij
mij
2
where nij and mij are observed and expected frequencies of (i, j) cell ( i=1,2,…,I; j=1,2,..,J).
Interestingly mij used in χ2
depend on row and column marginal totals, but not on the order in which the rows
and columns are listed. Chi square does not change under permutations of rows or columns and a treats both
classifications as nominal scales.
To understand the notion that the components represent certain aspects of the associations, the present
attempt has relaxed the assumptions and include all possible a x b sub-tables. (2 ≤ a ≤ I and 2 ≤ b ≤ J). Chi
square statistic have been based on this exhaustive and extensive list of sub-tables [(2I
– I – 1) (2J
– J – 1)-1] to
portray the association pattern compared to the original table. In particular interest lies on the I × J tables which
have significant over all association. By fixing size of the tables, the cell counts are simulated using resampling
technique by considering arbitrary yet meaningful limits for lower and upper limits and suitable probability
limits.
A long sequence of tables of required size has been generated to study the nature of associations in sub-
tables compared to the original I × J table. Decisions are made based on vote counting procedures; by
calculating the proportions of sub-tables exhibiting RAP to the number of possible sub-tables. While in such
calculations it has been emphasized for classifying the sub-tables as basic 2 × 2 categories and rest of them as
higher order through the original I × J table has not been considered as a sub table ( as opposed to the notion of
subsets)
III. RESULTS
The RAP routine calculates the necessary values for the post component analysis and outputs a
standard table of results. This includes the number of sub-tables with RAP, corresponding row and column
number together with the conclusion of chi square test. The required proportions are calculated based on these
numbers and conclusions are drawn there in. It further provides the distribution of cell counts in the given I × J
table so as to understand the spread of the data that would be helpful in conjugating the new hypothesis pertain
to RAP. However, due to paucity of space, this current simulation study has been limited to size 3 ×4 and 4 x 4.
Figure 1 provides the box plot for the distribution of proportions of 2 × 2 sub-tables which have shown RAP; it
could be observed that the pattern has been consistently high varies over the range of 60 % to 80 % in all the
three cases considered and no extreme observation has been recorded. This pattern has a consistent behavior
over the designs of the study.
.
.
.
.
.
.
.
.
.
(a)
Post Component Analysis of Categorical data
www.iosrjournals.org 22 | Page
Figure 1: Box plot for the proportions of 2 x 2 tables which have RAP based on the simulation of three types of
contingency tables (a) 3 × 4 (1 – 80) (b) 3 × 4 (101 – 160) (c) 4 × 4 (0 – 40)
Further, the proportion of sub-tables that exhibit the characteristic of reversal association has been summarized
in Table 1. This includes the simulation parameters together with some summary measures to understand the
notion of RAP among the three contingency tables considered for the study. Numerical values for the mean as
well as minimum and maximum for the proportions pertain to 2 × 2 sub-tables show that the dominance in terms
of reversal associations. Also, it could be observed that the mean proportions of other higher order sub–tables
that have RAP, have been systematically less than that of corresponding 2 × 2 sub-tables. However, maximum
proportion does not differ as much as minimum proportions do as small as 0%, which indicates that in some
cases could have higher order sub-tables without RAP though there could be no sign of such pattern observed in
2 × 2 sub-tables
Table 1: Summary measures for the proportions of sub-tables which exhibit the Reversal Association Pattern
(RAP) based on the simulation study
Table
Size
Resample
Limits
Summary
Measure
Size of sub-tables with RAP Proportion of sub-
tables with RAP
Lower Upper 2 x 2 Other
3 x 4 101 160 Mean 0.7222 0.5091 0.5983
S E 0.0095 0.0142 0.0118
Minimum 0.3889 0.1200 0.2558
Maximum 0.8889 0.8800 0.8372
3 x 4 1 80 Mean 0.4592 0.1894 0.3023
S E 0.0095 0.0097 0.0094
1.0
.9
.8
.7
.6
.5
.4
.3
(c)
1.0
.9
.8
.7
.6
.5
.4
.3
(b)
Post Component Analysis of Categorical data
www.iosrjournals.org 23 | Page
Minimum 0.1667 0.0000 0.0930
Maximum 0.8889 0.6800 0.7674
4 x 4 0 40 Mean 0.6506 0.3589 0.4464
S E 0.0095 0.0128 0.0115
Minimum 0.3056 0.0476 0.1333
Maximum 0.9167 0.7500 0.8000
IV. CONCLUSION
In general a multi comparison procedure that has been applied in many statistical methods such as
ANOVA could share a common aim. If a significant main effect or interaction is found, then there could be a
significant difference amongst the levels of attributes somewhere. Yet still this significant effect has to be
isolated exactly where the differences lie. However, in the case of categorical data, an attempt has been to
understand the pattern in which such analysis could be performed; understanding the extensive possibilities of
multiple comparisons between different levels of categories, this study has attempted an initial investigation to
describe the pattern of reversal association compared to the original I × J table.
The study used a non - parametric simulation procedure to generate the required contingency tables that
has varied cell frequencies including zeros in few cells. It could be observed that in more than 90% cases 2 × 2
tables have the reversal component association when compared to higher order sub-tables and interestingly,
higher order tables that have this pattern have strong relationships with their 2 × 2 counterparts. Though, this
approach might lack some theoretical exactness but in terms of consistency and ease and breadth of application
this would enable to understand and draw general conclusions regarding the general purpose of any multi
comparison procedures. However, similar attempts could be made with higher order contingency tables from
possible parametric simulations over a range of values that would yield low or high cell frequencies, so that the
method of Post Component could be applied in many practical experimental studies such as clinical trials to
design a drug efficacy study and the associations between different levels of categories could be explicitly
analyzed to have cost effective yet pragmatic decisions.
References
[1] J.Berkson,. Some difficulties of interpretation encountered in the application of the chi-square test, Journal of the American
Statistical Association 33, 1938, 526-536.
[2] H.W,Norton, Calculation of chi-square for complex contingency tables, J. Amer. Statist. Assoc. 40, 1945, 251-258.
[3] E.H.Simpson, The interpretation of interaction in contingency tables, J. Roy. Statist. Soc. B 13,1951, 238-241.
[4] W.G. Cochran, Some methods for strengthening the common chi-square tests, Biometrics 24, 1954, 315-327.
[5] N. Mantel, and W.Haenszel, Statistical aspects of the analysis of data from retrospective studies of disease, J. Nat. Cancer Inst. 22,
1959, 719-748.
[6] L.A. Goodman, Simple methods for analyzing three-factor interaction in contingency tables, J Amer. Statist. Assoc. 59, 1964, 319-
352.
[7] L.A. Goodman, On partitioning
2
 and detecting partial association in three-way contingency tables, J. Roy. Statist. Soc. B 31,
1969, 486-498.
[8] C.R. Blyth, On Simpson’s paradox and the sure thing principle. J Amer. Statist. Assoc. 67, 1972, 364-366.
[9] A.Agresti, Categorical Data Analysis, (New York: Wiley & Sons 1990) pp 51-54
[10] A.Agresti, Exact inference for categorical data: recent advances and continuing controversies, Statistical Methods and Applications,
20, 1999, 2709-2722.
[11] A.Sergio Estay , I. PauletteI Naulin, Data analysis in forest sciences: why do we continue using null hypothesis significance tests?
BOSQUE 32(1),2011, 3-9.

More Related Content

What's hot

Estimation of Reliability in Multicomponent Stress-Strength Model Following B...
Estimation of Reliability in Multicomponent Stress-Strength Model Following B...Estimation of Reliability in Multicomponent Stress-Strength Model Following B...
Estimation of Reliability in Multicomponent Stress-Strength Model Following B...IJERA Editor
 
Inferential Statistics - DAY 4 - B.Ed - AIOU
Inferential Statistics - DAY 4 - B.Ed - AIOUInferential Statistics - DAY 4 - B.Ed - AIOU
Inferential Statistics - DAY 4 - B.Ed - AIOUEqraBaig
 
Meaning of statistics
Meaning of statisticsMeaning of statistics
Meaning of statisticsSarfraz Ahmad
 
Basic statistics by Neeraj Bhandari ( Surkhet.Nepal )
Basic statistics by Neeraj Bhandari ( Surkhet.Nepal )Basic statistics by Neeraj Bhandari ( Surkhet.Nepal )
Basic statistics by Neeraj Bhandari ( Surkhet.Nepal )Neeraj Bhandari
 
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 anovanaranbatn
 
1 descriptive statistics
1 descriptive statistics1 descriptive statistics
1 descriptive statisticsSanu Kumar
 
Data presentation and interpretation I Quantitative Research
Data presentation and interpretation I Quantitative ResearchData presentation and interpretation I Quantitative Research
Data presentation and interpretation I Quantitative ResearchJimnaira Abanto
 
Meta analisis in health policy
Meta analisis in health policyMeta analisis in health policy
Meta analisis in health policyrsd kol abundjani
 
Application of ANOVA
Application of ANOVAApplication of ANOVA
Application of ANOVARohit Patidar
 
Inferential Statistics: Chi Square (X2) - DAY 6 - B.ED - 8614 - AIOU
Inferential Statistics: Chi Square (X2) - DAY 6 - B.ED - 8614 - AIOUInferential Statistics: Chi Square (X2) - DAY 6 - B.ED - 8614 - AIOU
Inferential Statistics: Chi Square (X2) - DAY 6 - B.ED - 8614 - AIOUEqraBaig
 
Basic Concepts of Statistics - Lecture Notes
Basic Concepts of Statistics - Lecture NotesBasic Concepts of Statistics - Lecture Notes
Basic Concepts of Statistics - Lecture NotesDr. Nirav Vyas
 
Calculation of covariance and simple linear models
Calculation of covariance and simple linear models Calculation of covariance and simple linear models
Calculation of covariance and simple linear models Thendral Kumaresan
 
Introduction to Statistics - Basic concepts
Introduction to Statistics - Basic conceptsIntroduction to Statistics - Basic concepts
Introduction to Statistics - Basic conceptsDocIbrahimAbdelmonaem
 

What's hot (19)

Estimation of Reliability in Multicomponent Stress-Strength Model Following B...
Estimation of Reliability in Multicomponent Stress-Strength Model Following B...Estimation of Reliability in Multicomponent Stress-Strength Model Following B...
Estimation of Reliability in Multicomponent Stress-Strength Model Following B...
 
Introduction to statistics
Introduction to statisticsIntroduction to statistics
Introduction to statistics
 
Analysis of variance
Analysis of varianceAnalysis of variance
Analysis of variance
 
Chapter 11
Chapter 11Chapter 11
Chapter 11
 
Inferential Statistics - DAY 4 - B.Ed - AIOU
Inferential Statistics - DAY 4 - B.Ed - AIOUInferential Statistics - DAY 4 - B.Ed - AIOU
Inferential Statistics - DAY 4 - B.Ed - AIOU
 
Panel data content
Panel data contentPanel data content
Panel data content
 
Meaning of statistics
Meaning of statisticsMeaning of statistics
Meaning of statistics
 
Basic statistics by Neeraj Bhandari ( Surkhet.Nepal )
Basic statistics by Neeraj Bhandari ( Surkhet.Nepal )Basic statistics by Neeraj Bhandari ( Surkhet.Nepal )
Basic statistics by Neeraj Bhandari ( Surkhet.Nepal )
 
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
 
1 descriptive statistics
1 descriptive statistics1 descriptive statistics
1 descriptive statistics
 
Anova copy
Anova   copyAnova   copy
Anova copy
 
Data presentation and interpretation I Quantitative Research
Data presentation and interpretation I Quantitative ResearchData presentation and interpretation I Quantitative Research
Data presentation and interpretation I Quantitative Research
 
Meta analisis in health policy
Meta analisis in health policyMeta analisis in health policy
Meta analisis in health policy
 
Application of ANOVA
Application of ANOVAApplication of ANOVA
Application of ANOVA
 
Inferential Statistics: Chi Square (X2) - DAY 6 - B.ED - 8614 - AIOU
Inferential Statistics: Chi Square (X2) - DAY 6 - B.ED - 8614 - AIOUInferential Statistics: Chi Square (X2) - DAY 6 - B.ED - 8614 - AIOU
Inferential Statistics: Chi Square (X2) - DAY 6 - B.ED - 8614 - AIOU
 
Akhiesh maurya
Akhiesh mauryaAkhiesh maurya
Akhiesh maurya
 
Basic Concepts of Statistics - Lecture Notes
Basic Concepts of Statistics - Lecture NotesBasic Concepts of Statistics - Lecture Notes
Basic Concepts of Statistics - Lecture Notes
 
Calculation of covariance and simple linear models
Calculation of covariance and simple linear models Calculation of covariance and simple linear models
Calculation of covariance and simple linear models
 
Introduction to Statistics - Basic concepts
Introduction to Statistics - Basic conceptsIntroduction to Statistics - Basic concepts
Introduction to Statistics - Basic concepts
 

Viewers also liked

Whistler Animal Shelter - 10th Annual K9 Wine & Dine
Whistler Animal Shelter - 10th Annual K9 Wine & DineWhistler Animal Shelter - 10th Annual K9 Wine & Dine
Whistler Animal Shelter - 10th Annual K9 Wine & DineWhistler Animals Galore
 
A.chuyende tnsua lan
A.chuyende tnsua lanA.chuyende tnsua lan
A.chuyende tnsua lanphihoanhbmt
 
Presentazione city sound 2012msagimp2
Presentazione city sound 2012msagimp2Presentazione city sound 2012msagimp2
Presentazione city sound 2012msagimp2lucaminetti
 
Media pitch
Media pitchMedia pitch
Media pitchHMonan
 
7 Tips for Selling Expensive Collectibles On eBay
7 Tips for Selling Expensive Collectibles On eBay7 Tips for Selling Expensive Collectibles On eBay
7 Tips for Selling Expensive Collectibles On eBaybelieve52
 
CLUBNIK угадай-ка
CLUBNIK угадай-каCLUBNIK угадай-ка
CLUBNIK угадай-каMihail Nadymov
 
Use Email Marketing wisely; Stand out from Junk Mail
Use Email Marketing wisely; Stand out from Junk MailUse Email Marketing wisely; Stand out from Junk Mail
Use Email Marketing wisely; Stand out from Junk Mailbelieve52
 
Issues in Sentiment analysis
Issues in Sentiment analysisIssues in Sentiment analysis
Issues in Sentiment analysisIOSR Journals
 
Literature Survey on Web Mining
Literature Survey on Web MiningLiterature Survey on Web Mining
Literature Survey on Web MiningIOSR Journals
 

Viewers also liked (20)

Final Year Project
Final Year ProjectFinal Year Project
Final Year Project
 
Whistler Animal Shelter - 10th Annual K9 Wine & Dine
Whistler Animal Shelter - 10th Annual K9 Wine & DineWhistler Animal Shelter - 10th Annual K9 Wine & Dine
Whistler Animal Shelter - 10th Annual K9 Wine & Dine
 
A.chuyende tnsua lan
A.chuyende tnsua lanA.chuyende tnsua lan
A.chuyende tnsua lan
 
Decision Trees
Decision TreesDecision Trees
Decision Trees
 
Presentazione city sound 2012msagimp2
Presentazione city sound 2012msagimp2Presentazione city sound 2012msagimp2
Presentazione city sound 2012msagimp2
 
Q6
Q6Q6
Q6
 
Ostorozhno doroga
Ostorozhno dorogaOstorozhno doroga
Ostorozhno doroga
 
A0610105
A0610105A0610105
A0610105
 
Media pitch
Media pitchMedia pitch
Media pitch
 
7 Tips for Selling Expensive Collectibles On eBay
7 Tips for Selling Expensive Collectibles On eBay7 Tips for Selling Expensive Collectibles On eBay
7 Tips for Selling Expensive Collectibles On eBay
 
CLUBNIK угадай-ка
CLUBNIK угадай-каCLUBNIK угадай-ка
CLUBNIK угадай-ка
 
Use Email Marketing wisely; Stand out from Junk Mail
Use Email Marketing wisely; Stand out from Junk MailUse Email Marketing wisely; Stand out from Junk Mail
Use Email Marketing wisely; Stand out from Junk Mail
 
Transport1
Transport1Transport1
Transport1
 
Shakira 4 1
Shakira 4 1Shakira 4 1
Shakira 4 1
 
F0443041
F0443041F0443041
F0443041
 
C0541529
C0541529C0541529
C0541529
 
Designing Interactions at the Nexus of the Physical and Virtual World
Designing Interactions at the Nexus of the Physical and Virtual WorldDesigning Interactions at the Nexus of the Physical and Virtual World
Designing Interactions at the Nexus of the Physical and Virtual World
 
Issues in Sentiment analysis
Issues in Sentiment analysisIssues in Sentiment analysis
Issues in Sentiment analysis
 
A0150106
A0150106A0150106
A0150106
 
Literature Survey on Web Mining
Literature Survey on Web MiningLiterature Survey on Web Mining
Literature Survey on Web Mining
 

Similar to F0562023

Ratio and Product Type Estimators Using Stratified Ranked Set Sampling
Ratio and Product Type Estimators Using Stratified Ranked Set SamplingRatio and Product Type Estimators Using Stratified Ranked Set Sampling
Ratio and Product Type Estimators Using Stratified Ranked Set Samplinginventionjournals
 
IBM SPSS Statistics Algorithms.pdf
IBM SPSS Statistics Algorithms.pdfIBM SPSS Statistics Algorithms.pdf
IBM SPSS Statistics Algorithms.pdfNorafizah Samawi
 
2.4 Scatterplots, correlation, and regression
2.4 Scatterplots, correlation, and regression2.4 Scatterplots, correlation, and regression
2.4 Scatterplots, correlation, and regressionLong Beach City College
 
Multiple Linear Regression Model with Two Parameter Doubly Truncated New Symm...
Multiple Linear Regression Model with Two Parameter Doubly Truncated New Symm...Multiple Linear Regression Model with Two Parameter Doubly Truncated New Symm...
Multiple Linear Regression Model with Two Parameter Doubly Truncated New Symm...theijes
 
Week 5 Lecture 14 The Chi Square TestQuite often, patterns of .docx
Week 5 Lecture 14 The Chi Square TestQuite often, patterns of .docxWeek 5 Lecture 14 The Chi Square TestQuite often, patterns of .docx
Week 5 Lecture 14 The Chi Square TestQuite often, patterns of .docxcockekeshia
 
QNT 275 Exceptional Education - snaptutorial.com
QNT 275   Exceptional Education - snaptutorial.comQNT 275   Exceptional Education - snaptutorial.com
QNT 275 Exceptional Education - snaptutorial.comDavisMurphyB22
 
Chi square and t tests, Neelam zafar & group
Chi square and t tests, Neelam zafar & groupChi square and t tests, Neelam zafar & group
Chi square and t tests, Neelam zafar & groupNeelam Zafar
 
STATISTICAL TOOLS USED IN ANALYTICAL CHEMISTRY
STATISTICAL TOOLS USED IN ANALYTICAL CHEMISTRYSTATISTICAL TOOLS USED IN ANALYTICAL CHEMISTRY
STATISTICAL TOOLS USED IN ANALYTICAL CHEMISTRYkeerthana151
 
Quatitative Data Analysis
Quatitative Data Analysis Quatitative Data Analysis
Quatitative Data Analysis maneesh mani
 
Simple Correspondence Analysis
Simple Correspondence AnalysisSimple Correspondence Analysis
Simple Correspondence AnalysisVarun Gangadhar
 
Qnt 275 Enhance teaching / snaptutorial.com
Qnt 275 Enhance teaching / snaptutorial.comQnt 275 Enhance teaching / snaptutorial.com
Qnt 275 Enhance teaching / snaptutorial.comBaileya33
 
Swanetal.JournalofSedimentaryPetrology1978.pdf
Swanetal.JournalofSedimentaryPetrology1978.pdfSwanetal.JournalofSedimentaryPetrology1978.pdf
Swanetal.JournalofSedimentaryPetrology1978.pdfDrBaliReddy
 
2013jsm,Proceedings,DSweitzer,26sep
2013jsm,Proceedings,DSweitzer,26sep2013jsm,Proceedings,DSweitzer,26sep
2013jsm,Proceedings,DSweitzer,26sepDennis Sweitzer
 
QNT 275 Inspiring Innovation / tutorialrank.com
QNT 275 Inspiring Innovation / tutorialrank.comQNT 275 Inspiring Innovation / tutorialrank.com
QNT 275 Inspiring Innovation / tutorialrank.comBromleyz33
 
Z and t_tests
Z and t_testsZ and t_tests
Z and t_testseducation
 
A Critique Of Anscombe S Work On Statistical Analysis Using Graphs (2013 Home...
A Critique Of Anscombe S Work On Statistical Analysis Using Graphs (2013 Home...A Critique Of Anscombe S Work On Statistical Analysis Using Graphs (2013 Home...
A Critique Of Anscombe S Work On Statistical Analysis Using Graphs (2013 Home...Simar Neasy
 

Similar to F0562023 (20)

Ratio and Product Type Estimators Using Stratified Ranked Set Sampling
Ratio and Product Type Estimators Using Stratified Ranked Set SamplingRatio and Product Type Estimators Using Stratified Ranked Set Sampling
Ratio and Product Type Estimators Using Stratified Ranked Set Sampling
 
3.1 Measures of center
3.1 Measures of center3.1 Measures of center
3.1 Measures of center
 
IBM SPSS Statistics Algorithms.pdf
IBM SPSS Statistics Algorithms.pdfIBM SPSS Statistics Algorithms.pdf
IBM SPSS Statistics Algorithms.pdf
 
2.4 Scatterplots, correlation, and regression
2.4 Scatterplots, correlation, and regression2.4 Scatterplots, correlation, and regression
2.4 Scatterplots, correlation, and regression
 
Chapter12
Chapter12Chapter12
Chapter12
 
Multiple Linear Regression Model with Two Parameter Doubly Truncated New Symm...
Multiple Linear Regression Model with Two Parameter Doubly Truncated New Symm...Multiple Linear Regression Model with Two Parameter Doubly Truncated New Symm...
Multiple Linear Regression Model with Two Parameter Doubly Truncated New Symm...
 
Week 5 Lecture 14 The Chi Square TestQuite often, patterns of .docx
Week 5 Lecture 14 The Chi Square TestQuite often, patterns of .docxWeek 5 Lecture 14 The Chi Square TestQuite often, patterns of .docx
Week 5 Lecture 14 The Chi Square TestQuite often, patterns of .docx
 
QNT 275 Exceptional Education - snaptutorial.com
QNT 275   Exceptional Education - snaptutorial.comQNT 275   Exceptional Education - snaptutorial.com
QNT 275 Exceptional Education - snaptutorial.com
 
Chi square and t tests, Neelam zafar & group
Chi square and t tests, Neelam zafar & groupChi square and t tests, Neelam zafar & group
Chi square and t tests, Neelam zafar & group
 
STATISTICAL TOOLS USED IN ANALYTICAL CHEMISTRY
STATISTICAL TOOLS USED IN ANALYTICAL CHEMISTRYSTATISTICAL TOOLS USED IN ANALYTICAL CHEMISTRY
STATISTICAL TOOLS USED IN ANALYTICAL CHEMISTRY
 
Quatitative Data Analysis
Quatitative Data Analysis Quatitative Data Analysis
Quatitative Data Analysis
 
Simple Correspondence Analysis
Simple Correspondence AnalysisSimple Correspondence Analysis
Simple Correspondence Analysis
 
Qnt 275 Enhance teaching / snaptutorial.com
Qnt 275 Enhance teaching / snaptutorial.comQnt 275 Enhance teaching / snaptutorial.com
Qnt 275 Enhance teaching / snaptutorial.com
 
Swanetal.JournalofSedimentaryPetrology1978.pdf
Swanetal.JournalofSedimentaryPetrology1978.pdfSwanetal.JournalofSedimentaryPetrology1978.pdf
Swanetal.JournalofSedimentaryPetrology1978.pdf
 
2013jsm,Proceedings,DSweitzer,26sep
2013jsm,Proceedings,DSweitzer,26sep2013jsm,Proceedings,DSweitzer,26sep
2013jsm,Proceedings,DSweitzer,26sep
 
elementary statistic
elementary statisticelementary statistic
elementary statistic
 
QNT 275 Inspiring Innovation / tutorialrank.com
QNT 275 Inspiring Innovation / tutorialrank.comQNT 275 Inspiring Innovation / tutorialrank.com
QNT 275 Inspiring Innovation / tutorialrank.com
 
Z and t_tests
Z and t_testsZ and t_tests
Z and t_tests
 
Chi square test
Chi square testChi square test
Chi square test
 
A Critique Of Anscombe S Work On Statistical Analysis Using Graphs (2013 Home...
A Critique Of Anscombe S Work On Statistical Analysis Using Graphs (2013 Home...A Critique Of Anscombe S Work On Statistical Analysis Using Graphs (2013 Home...
A Critique Of Anscombe S Work On Statistical Analysis Using Graphs (2013 Home...
 

More from IOSR Journals (20)

A011140104
A011140104A011140104
A011140104
 
M0111397100
M0111397100M0111397100
M0111397100
 
L011138596
L011138596L011138596
L011138596
 
K011138084
K011138084K011138084
K011138084
 
J011137479
J011137479J011137479
J011137479
 
I011136673
I011136673I011136673
I011136673
 
G011134454
G011134454G011134454
G011134454
 
H011135565
H011135565H011135565
H011135565
 
F011134043
F011134043F011134043
F011134043
 
E011133639
E011133639E011133639
E011133639
 
D011132635
D011132635D011132635
D011132635
 
C011131925
C011131925C011131925
C011131925
 
B011130918
B011130918B011130918
B011130918
 
A011130108
A011130108A011130108
A011130108
 
I011125160
I011125160I011125160
I011125160
 
H011124050
H011124050H011124050
H011124050
 
G011123539
G011123539G011123539
G011123539
 
F011123134
F011123134F011123134
F011123134
 
E011122530
E011122530E011122530
E011122530
 
D011121524
D011121524D011121524
D011121524
 

Recently uploaded

Assure Ecommerce and Retail Operations Uptime with ThousandEyes
Assure Ecommerce and Retail Operations Uptime with ThousandEyesAssure Ecommerce and Retail Operations Uptime with ThousandEyes
Assure Ecommerce and Retail Operations Uptime with ThousandEyesThousandEyes
 
Tampa BSides - The No BS SOC (slides from April 6, 2024 talk)
Tampa BSides - The No BS SOC (slides from April 6, 2024 talk)Tampa BSides - The No BS SOC (slides from April 6, 2024 talk)
Tampa BSides - The No BS SOC (slides from April 6, 2024 talk)Mark Simos
 
How to Effectively Monitor SD-WAN and SASE Environments with ThousandEyes
How to Effectively Monitor SD-WAN and SASE Environments with ThousandEyesHow to Effectively Monitor SD-WAN and SASE Environments with ThousandEyes
How to Effectively Monitor SD-WAN and SASE Environments with ThousandEyesThousandEyes
 
QCon London: Mastering long-running processes in modern architectures
QCon London: Mastering long-running processes in modern architecturesQCon London: Mastering long-running processes in modern architectures
QCon London: Mastering long-running processes in modern architecturesBernd Ruecker
 
MuleSoft Online Meetup Group - B2B Crash Course: Release SparkNotes
MuleSoft Online Meetup Group - B2B Crash Course: Release SparkNotesMuleSoft Online Meetup Group - B2B Crash Course: Release SparkNotes
MuleSoft Online Meetup Group - B2B Crash Course: Release SparkNotesManik S Magar
 
Glenn Lazarus- Why Your Observability Strategy Needs Security Observability
Glenn Lazarus- Why Your Observability Strategy Needs Security ObservabilityGlenn Lazarus- Why Your Observability Strategy Needs Security Observability
Glenn Lazarus- Why Your Observability Strategy Needs Security Observabilityitnewsafrica
 
Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...
Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...
Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...itnewsafrica
 
React Native vs Ionic - The Best Mobile App Framework
React Native vs Ionic - The Best Mobile App FrameworkReact Native vs Ionic - The Best Mobile App Framework
React Native vs Ionic - The Best Mobile App FrameworkPixlogix Infotech
 
Abdul Kader Baba- Managing Cybersecurity Risks and Compliance Requirements i...
Abdul Kader Baba- Managing Cybersecurity Risks  and Compliance Requirements i...Abdul Kader Baba- Managing Cybersecurity Risks  and Compliance Requirements i...
Abdul Kader Baba- Managing Cybersecurity Risks and Compliance Requirements i...itnewsafrica
 
Kuma Meshes Part I - The basics - A tutorial
Kuma Meshes Part I - The basics - A tutorialKuma Meshes Part I - The basics - A tutorial
Kuma Meshes Part I - The basics - A tutorialJoão Esperancinha
 
Accelerating Enterprise Software Engineering with Platformless
Accelerating Enterprise Software Engineering with PlatformlessAccelerating Enterprise Software Engineering with Platformless
Accelerating Enterprise Software Engineering with PlatformlessWSO2
 
Generative AI - Gitex v1Generative AI - Gitex v1.pptx
Generative AI - Gitex v1Generative AI - Gitex v1.pptxGenerative AI - Gitex v1Generative AI - Gitex v1.pptx
Generative AI - Gitex v1Generative AI - Gitex v1.pptxfnnc6jmgwh
 
Infrared simulation and processing on Nvidia platforms
Infrared simulation and processing on Nvidia platformsInfrared simulation and processing on Nvidia platforms
Infrared simulation and processing on Nvidia platformsYoss Cohen
 
Connecting the Dots for Information Discovery.pdf
Connecting the Dots for Information Discovery.pdfConnecting the Dots for Information Discovery.pdf
Connecting the Dots for Information Discovery.pdfNeo4j
 
A Glance At The Java Performance Toolbox
A Glance At The Java Performance ToolboxA Glance At The Java Performance Toolbox
A Glance At The Java Performance ToolboxAna-Maria Mihalceanu
 
UiPath Community: Communication Mining from Zero to Hero
UiPath Community: Communication Mining from Zero to HeroUiPath Community: Communication Mining from Zero to Hero
UiPath Community: Communication Mining from Zero to HeroUiPathCommunity
 
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...Wes McKinney
 
Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...
Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...
Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...Nikki Chapple
 
[Webinar] SpiraTest - Setting New Standards in Quality Assurance
[Webinar] SpiraTest - Setting New Standards in Quality Assurance[Webinar] SpiraTest - Setting New Standards in Quality Assurance
[Webinar] SpiraTest - Setting New Standards in Quality AssuranceInflectra
 
Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...
Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...
Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...Nikki Chapple
 

Recently uploaded (20)

Assure Ecommerce and Retail Operations Uptime with ThousandEyes
Assure Ecommerce and Retail Operations Uptime with ThousandEyesAssure Ecommerce and Retail Operations Uptime with ThousandEyes
Assure Ecommerce and Retail Operations Uptime with ThousandEyes
 
Tampa BSides - The No BS SOC (slides from April 6, 2024 talk)
Tampa BSides - The No BS SOC (slides from April 6, 2024 talk)Tampa BSides - The No BS SOC (slides from April 6, 2024 talk)
Tampa BSides - The No BS SOC (slides from April 6, 2024 talk)
 
How to Effectively Monitor SD-WAN and SASE Environments with ThousandEyes
How to Effectively Monitor SD-WAN and SASE Environments with ThousandEyesHow to Effectively Monitor SD-WAN and SASE Environments with ThousandEyes
How to Effectively Monitor SD-WAN and SASE Environments with ThousandEyes
 
QCon London: Mastering long-running processes in modern architectures
QCon London: Mastering long-running processes in modern architecturesQCon London: Mastering long-running processes in modern architectures
QCon London: Mastering long-running processes in modern architectures
 
MuleSoft Online Meetup Group - B2B Crash Course: Release SparkNotes
MuleSoft Online Meetup Group - B2B Crash Course: Release SparkNotesMuleSoft Online Meetup Group - B2B Crash Course: Release SparkNotes
MuleSoft Online Meetup Group - B2B Crash Course: Release SparkNotes
 
Glenn Lazarus- Why Your Observability Strategy Needs Security Observability
Glenn Lazarus- Why Your Observability Strategy Needs Security ObservabilityGlenn Lazarus- Why Your Observability Strategy Needs Security Observability
Glenn Lazarus- Why Your Observability Strategy Needs Security Observability
 
Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...
Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...
Irene Moetsana-Moeng: Stakeholders in Cybersecurity: Collaborative Defence fo...
 
React Native vs Ionic - The Best Mobile App Framework
React Native vs Ionic - The Best Mobile App FrameworkReact Native vs Ionic - The Best Mobile App Framework
React Native vs Ionic - The Best Mobile App Framework
 
Abdul Kader Baba- Managing Cybersecurity Risks and Compliance Requirements i...
Abdul Kader Baba- Managing Cybersecurity Risks  and Compliance Requirements i...Abdul Kader Baba- Managing Cybersecurity Risks  and Compliance Requirements i...
Abdul Kader Baba- Managing Cybersecurity Risks and Compliance Requirements i...
 
Kuma Meshes Part I - The basics - A tutorial
Kuma Meshes Part I - The basics - A tutorialKuma Meshes Part I - The basics - A tutorial
Kuma Meshes Part I - The basics - A tutorial
 
Accelerating Enterprise Software Engineering with Platformless
Accelerating Enterprise Software Engineering with PlatformlessAccelerating Enterprise Software Engineering with Platformless
Accelerating Enterprise Software Engineering with Platformless
 
Generative AI - Gitex v1Generative AI - Gitex v1.pptx
Generative AI - Gitex v1Generative AI - Gitex v1.pptxGenerative AI - Gitex v1Generative AI - Gitex v1.pptx
Generative AI - Gitex v1Generative AI - Gitex v1.pptx
 
Infrared simulation and processing on Nvidia platforms
Infrared simulation and processing on Nvidia platformsInfrared simulation and processing on Nvidia platforms
Infrared simulation and processing on Nvidia platforms
 
Connecting the Dots for Information Discovery.pdf
Connecting the Dots for Information Discovery.pdfConnecting the Dots for Information Discovery.pdf
Connecting the Dots for Information Discovery.pdf
 
A Glance At The Java Performance Toolbox
A Glance At The Java Performance ToolboxA Glance At The Java Performance Toolbox
A Glance At The Java Performance Toolbox
 
UiPath Community: Communication Mining from Zero to Hero
UiPath Community: Communication Mining from Zero to HeroUiPath Community: Communication Mining from Zero to Hero
UiPath Community: Communication Mining from Zero to Hero
 
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...
The Future Roadmap for the Composable Data Stack - Wes McKinney - Data Counci...
 
Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...
Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...
Microsoft 365 Copilot: How to boost your productivity with AI – Part two: Dat...
 
[Webinar] SpiraTest - Setting New Standards in Quality Assurance
[Webinar] SpiraTest - Setting New Standards in Quality Assurance[Webinar] SpiraTest - Setting New Standards in Quality Assurance
[Webinar] SpiraTest - Setting New Standards in Quality Assurance
 
Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...
Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...
Microsoft 365 Copilot: How to boost your productivity with AI – Part one: Ado...
 

F0562023

  • 1. IOSR Journal of Mathematics (IOSR-JM) e-ISSN: 2278-5728.Volume 5, Issue 6 (Mar. - Apr. 2013), PP 20-23 www.iosrjournals.org www.iosrjournals.org 20 | Page Post Component Analysis of Categorical data U. Sangeetha1 , M. Subbiah2 , M.R. Srinivasan3 1( Department of Management Studies, SSN College of Engineering, Chennai, India) 2 (Department of Mathematics, L. N. Government College, Ponneri, India) 3 (Department of Statistics, University of Madras, Chennai, India) Abstract: Categorical data are often stratified into two dimensional I × J tables in order to test the independence of attributes. Agresti (1999) has discussed testing methods with partitioning property of chi square to extract components that describe certain aspects of the overall association in a table. In this work an attempt has been made to study the pattern in which sub-tables exhibit a sign of reverse association when compared to a significant association of attributes with regard to the original I × J table. Simulation studies and subsequent results of vote counting method indicate that 2 × 2 tables have the reversal component association when compared to higher order sub-tables; interestingly, in more than 90% cases. The computationally extensive and exhaustive procedure provide a better tool to understand the association between the categories that focus on the strongest differences among all comparisons, and could be practically and other aspects in experimental studies such as clinical trials. Keywords: Contingency tables, Chi-Square test, Component Analysis. I. INTRODUCTION Categorical data analysis has found a rapid development since 1960’s which might be due to the methodological sophistication of the social and biomedical sciences. It is not uncommon in social sciences for measuring attitudes and opinions on various issues and demographic characteristics such as gender race and social class. Categorical scales have been used extensively in biomedical sciences to measure many factors, such as severity of injury, degree of recovery from surgery, and stage of a disease etc. This progress has been found as an outcome of stimulus ties between social scientists and / or biomedical scientists and Statisticians. However, one of the most influential contributions by Karl Pearson and Yule on association between categorical variables has also been used extensively in many research works and by the practitioners in this large data driven era. The data for this analysis are displayed in a table with I rows and J columns and many software perform this analysis to provide a test of significance using a chi-square analysis testing the null hypothesis that there is no association between the two variables. If the chi-square test results in a p-value of 0.05or smaller, the results are deemed significant, the null hypothesis is rejected. In such cases when there are only two groups being compared (a 2 × 2 table) it is easy to observe the association between them; however when there are three or more groups, there has been a need to identify the significant/non-significant association between pairs of groups. An extensive literature such as Berkson, (1938), Norton, (1945), Simpson, (1951), Cochran, (1954), Mantel and Haenszel, (1959), Goodman, (1964), Goodman, (1969), Blyth, 1972, Agresti, (1990), Agresti (1999), Sergio and Paulette (2011) provides a method based on portioning of chi-square it has laid a practical limitation to the users, in that partitioning the given I × J table satisfying the basic assumptions of the procedure. This paper has identified a need based analysis between the practice and the theory and has provided a procedure based on the post component analysis which is similar to the multiple comparisons in ANOVA for continuous response variables. A primitive approach and subsequent analysis issues have been discussed with limited size of the underlying classification table; this paper includes a simulation study to understand the decisions on the Reversal Association Pattern (RAP) in a I × J table. Also, this attempt has laid a foundation to visualize this post component analysis under various model settings and with large contingency tables. II. METHOD A bivariate relationship has been defined by the joint distribution of the two associated variables. If X and Y denote two categorical response variables, X having I levels and Y having J levels then the IJ possible combinations of classification could be displayed in a contingency table, where the cells contains frequency counts of outcomes. Such a contingency table is referred to as an I × J table. In a two-way contingency tables, the null hypothesis of statistical independence has been tested using Pearson chi-square statistic (Agresti 1990)
  • 2. Post Component Analysis of Categorical data www.iosrjournals.org 21 | Page χ2 =ΣΣ nij-mij mij 2 where nij and mij are observed and expected frequencies of (i, j) cell ( i=1,2,…,I; j=1,2,..,J). Interestingly mij used in χ2 depend on row and column marginal totals, but not on the order in which the rows and columns are listed. Chi square does not change under permutations of rows or columns and a treats both classifications as nominal scales. To understand the notion that the components represent certain aspects of the associations, the present attempt has relaxed the assumptions and include all possible a x b sub-tables. (2 ≤ a ≤ I and 2 ≤ b ≤ J). Chi square statistic have been based on this exhaustive and extensive list of sub-tables [(2I – I – 1) (2J – J – 1)-1] to portray the association pattern compared to the original table. In particular interest lies on the I × J tables which have significant over all association. By fixing size of the tables, the cell counts are simulated using resampling technique by considering arbitrary yet meaningful limits for lower and upper limits and suitable probability limits. A long sequence of tables of required size has been generated to study the nature of associations in sub- tables compared to the original I × J table. Decisions are made based on vote counting procedures; by calculating the proportions of sub-tables exhibiting RAP to the number of possible sub-tables. While in such calculations it has been emphasized for classifying the sub-tables as basic 2 × 2 categories and rest of them as higher order through the original I × J table has not been considered as a sub table ( as opposed to the notion of subsets) III. RESULTS The RAP routine calculates the necessary values for the post component analysis and outputs a standard table of results. This includes the number of sub-tables with RAP, corresponding row and column number together with the conclusion of chi square test. The required proportions are calculated based on these numbers and conclusions are drawn there in. It further provides the distribution of cell counts in the given I × J table so as to understand the spread of the data that would be helpful in conjugating the new hypothesis pertain to RAP. However, due to paucity of space, this current simulation study has been limited to size 3 ×4 and 4 x 4. Figure 1 provides the box plot for the distribution of proportions of 2 × 2 sub-tables which have shown RAP; it could be observed that the pattern has been consistently high varies over the range of 60 % to 80 % in all the three cases considered and no extreme observation has been recorded. This pattern has a consistent behavior over the designs of the study. . . . . . . . . . (a)
  • 3. Post Component Analysis of Categorical data www.iosrjournals.org 22 | Page Figure 1: Box plot for the proportions of 2 x 2 tables which have RAP based on the simulation of three types of contingency tables (a) 3 × 4 (1 – 80) (b) 3 × 4 (101 – 160) (c) 4 × 4 (0 – 40) Further, the proportion of sub-tables that exhibit the characteristic of reversal association has been summarized in Table 1. This includes the simulation parameters together with some summary measures to understand the notion of RAP among the three contingency tables considered for the study. Numerical values for the mean as well as minimum and maximum for the proportions pertain to 2 × 2 sub-tables show that the dominance in terms of reversal associations. Also, it could be observed that the mean proportions of other higher order sub–tables that have RAP, have been systematically less than that of corresponding 2 × 2 sub-tables. However, maximum proportion does not differ as much as minimum proportions do as small as 0%, which indicates that in some cases could have higher order sub-tables without RAP though there could be no sign of such pattern observed in 2 × 2 sub-tables Table 1: Summary measures for the proportions of sub-tables which exhibit the Reversal Association Pattern (RAP) based on the simulation study Table Size Resample Limits Summary Measure Size of sub-tables with RAP Proportion of sub- tables with RAP Lower Upper 2 x 2 Other 3 x 4 101 160 Mean 0.7222 0.5091 0.5983 S E 0.0095 0.0142 0.0118 Minimum 0.3889 0.1200 0.2558 Maximum 0.8889 0.8800 0.8372 3 x 4 1 80 Mean 0.4592 0.1894 0.3023 S E 0.0095 0.0097 0.0094 1.0 .9 .8 .7 .6 .5 .4 .3 (c) 1.0 .9 .8 .7 .6 .5 .4 .3 (b)
  • 4. Post Component Analysis of Categorical data www.iosrjournals.org 23 | Page Minimum 0.1667 0.0000 0.0930 Maximum 0.8889 0.6800 0.7674 4 x 4 0 40 Mean 0.6506 0.3589 0.4464 S E 0.0095 0.0128 0.0115 Minimum 0.3056 0.0476 0.1333 Maximum 0.9167 0.7500 0.8000 IV. CONCLUSION In general a multi comparison procedure that has been applied in many statistical methods such as ANOVA could share a common aim. If a significant main effect or interaction is found, then there could be a significant difference amongst the levels of attributes somewhere. Yet still this significant effect has to be isolated exactly where the differences lie. However, in the case of categorical data, an attempt has been to understand the pattern in which such analysis could be performed; understanding the extensive possibilities of multiple comparisons between different levels of categories, this study has attempted an initial investigation to describe the pattern of reversal association compared to the original I × J table. The study used a non - parametric simulation procedure to generate the required contingency tables that has varied cell frequencies including zeros in few cells. It could be observed that in more than 90% cases 2 × 2 tables have the reversal component association when compared to higher order sub-tables and interestingly, higher order tables that have this pattern have strong relationships with their 2 × 2 counterparts. Though, this approach might lack some theoretical exactness but in terms of consistency and ease and breadth of application this would enable to understand and draw general conclusions regarding the general purpose of any multi comparison procedures. However, similar attempts could be made with higher order contingency tables from possible parametric simulations over a range of values that would yield low or high cell frequencies, so that the method of Post Component could be applied in many practical experimental studies such as clinical trials to design a drug efficacy study and the associations between different levels of categories could be explicitly analyzed to have cost effective yet pragmatic decisions. References [1] J.Berkson,. Some difficulties of interpretation encountered in the application of the chi-square test, Journal of the American Statistical Association 33, 1938, 526-536. [2] H.W,Norton, Calculation of chi-square for complex contingency tables, J. Amer. Statist. Assoc. 40, 1945, 251-258. [3] E.H.Simpson, The interpretation of interaction in contingency tables, J. Roy. Statist. Soc. B 13,1951, 238-241. [4] W.G. Cochran, Some methods for strengthening the common chi-square tests, Biometrics 24, 1954, 315-327. [5] N. Mantel, and W.Haenszel, Statistical aspects of the analysis of data from retrospective studies of disease, J. Nat. Cancer Inst. 22, 1959, 719-748. [6] L.A. Goodman, Simple methods for analyzing three-factor interaction in contingency tables, J Amer. Statist. Assoc. 59, 1964, 319- 352. [7] L.A. Goodman, On partitioning 2  and detecting partial association in three-way contingency tables, J. Roy. Statist. Soc. B 31, 1969, 486-498. [8] C.R. Blyth, On Simpson’s paradox and the sure thing principle. J Amer. Statist. Assoc. 67, 1972, 364-366. [9] A.Agresti, Categorical Data Analysis, (New York: Wiley & Sons 1990) pp 51-54 [10] A.Agresti, Exact inference for categorical data: recent advances and continuing controversies, Statistical Methods and Applications, 20, 1999, 2709-2722. [11] A.Sergio Estay , I. PauletteI Naulin, Data analysis in forest sciences: why do we continue using null hypothesis significance tests? BOSQUE 32(1),2011, 3-9.