|
Activity Number:
|
32
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Health Policy Statistics
|
| Abstract - #305690 |
|
Title:
|
Factor Analysis with Categorical Data: a Methodological Illustration with the GAZA Child Health Survey Data
|
|
Author(s):
|
Dongguang Li*+ and John D. Pringle and Julio Arboleda-Florez and Heather Stuart
|
|
Companies:
|
National Cancer Institute of Canada and Queen's University and Queen's University and Queen's University
|
|
Address:
|
10 Stuart Street, Kingston, ON, K7L3N6, Canada
|
|
Keywords:
|
factor analysis ; categorical data ; war-trauma
|
|
Abstract:
|
Factor analysis is designed to identify the latent factors from a large set of variables using certain mathematics models. In contrast to the analysis for continuous data, the methods for categorical data factor analysis are relatively dubious. With the Gaza Child Health Survey data, this work conducts factor analysis using the multiple categorical war-trauma data and the transformed dichotomous data. The results are compared. The influences of different correlation coefficients are tested. The study evaluates the effects of the rotate options and the threshold for number of factors based on eigenvalues and factor loadings. The relevant methodological issues are discussed. The analysis finds the dichotomous transformation makes more appropriate evaluations of the latent factors and the tetrachoric correlation is the best index in implementing a factor analysis with dichotomous data.
|
