Course on analysis of multivariate categorical data

2008

Updated April, 2008

During the course we will use to different programs:

SCD/DIGRAM.

A zip file with version 1.70, the user guide and some examples may be downloaded here:

scd-digram 1-70.zip

Important:

There is a new version of DIGRAM

LEMWIN

Which may be downloaded here:

www.uvt.nl/faculteiten/fsw/organisatie/departementen/mto/software2.html

GMGM

Is here GMGM.exe

With user guide here:

http://pubhealth.ku.dk/upload/application/pdf/f7ddaf56/Research_report_07-15.pdf

Data to be used for demonstration and exercises may be found in the following ZIP files

DemoExamples.zip

Course program

February 22: Ordinal categorical data

Lectures 9.15 – 12.00 in CSS 5-1-34 Intro Lecture 1 Lecture 2

The gamma coefficient

The partial gamma coefficient

Exact conditional tests

Repeated Monte Carlo tests

The pseudo gamma coefficient for nominal and partially ordered variables

Exercises 13.00 – 16.00 in CSS-2-2-12 Exercise 1.zip

Introduction to DIGRAM

The user interface

Commands for analysis of contingency tables:

TABULATE

HYPOTHESIS

EXACT/REPEATED/SEQUENTIAL

TEST

PGAMMA

COLLAPS

Texts

Agresti, A (1984) :Analysis of Ordinal Categorical Data. John Wiley & Sons. Chapters 9 & 10

Kreiner, S (1987): Analysis of multidimensional contingency tables by exact conditional tests: techniques and Strategies. Scand. Journ. of Stat., 1987, 97-112

Dixon, WJ (1988) BMDP Statistical Software Manual. Appendix B11 544-555

Besag, J and Clifford, P. (1991) Sequential Monte Carlo p-values. Biometrika, 78, 301-304

Kreiner, S. (2003): Introduction to DIGRAM. Research report 03/10, Dept. of Biostatistics

Siersma, V. & Kreiner, S. (2007) A coefficient of association between categorical variables with partial or tentative ordering of categories. Research report 07/10, Dept. of Biostatistics

Christensen, KB & Kreiner, S (2007) A Monte Carlo Approach to Unidimensionality Testing in Polytomous Rasch Models. Applied Psychological Measurement, 31, 20-30

(pages 25-28 are particular important)

February 29: Graphical models

Lectures 9.15 – 12.00 in CSS 5-1-34 Lecture 2 Lecture 4

Exercises 13.00 – 16.00 in CSS-2-2-12 Exercise 2.zip

Definition of graphical models

See pages 57-60 + Appendix 1 of the user guide to DIGRAM

Model based statistical inference

See pages 61, 62 , 72-86 of the user guide

Texts

Whittaker, J 1993): Graphical Interaction Models: A New Approach for Statistical Modelling. In Dean, K (ed): Population Health Research, 160-180, SagePublications.

Wermuth, N (1993): Association Structures with Few Variables: Characteristics and Examples. In Dean, K (ed): Population Health Research, 181-203, SagePublications.

Kreiner, S. (1996) An informal introduction to graphical modelling. In Knudsen HC et.al. (eds): Mental health Evaluation, 156-175, Cambridge University Press

The introduction to graphical models (pages 1-14) in the Introduction to DIGRAM and the note on Graphical models and Global Markov properties.doc is also relevant for the introduction to graphical models

March 7: Model search

Lectures 9.15 – 12.00 in CSS 5-1-34 Lecture 5 - Model search

Exercises 12.30 – 14.00 in CSS-2-2-12 exercise 3.zip

March 14: Loglinear models

Lectures 9.15 – 12.00 in CSS 5-1-34 Lecture 7 - Loglinear models

Exercises 13.00 – 16.00 in CSS-2-2-18

April 25: Repetition of results on graphical and loglinear models

Lectures 9.15 – 12.00 in CSS 5-1-28

Exercises 13.00 – 16.00 in CSS-2-2-12 Data for exercise 5