|
Activity Number:
|
31
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, July 29, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistical Education
|
| Abstract - #309888 |
|
Title:
|
The Central Limit Theorem and Structural Validity in Factor Analysis
|
|
Author(s):
|
Landon Poppleton*+ and Mitchell Harris and Michael K. Lauritzen and Robert Bubb and Bruce Brown
|
|
Companies:
|
Brigham Young University and Brigham Young University and Brigham Young University and Brigham Young University and Brigham Young University
|
|
Address:
|
1001 SWKT, Provo, UT, 84602,
|
|
Keywords:
|
factor analysis ; optimization ; central limit theorem
|
|
Abstract:
|
Factoring methods such as principal components and factor analysis are often used for data simplification. Over the past century a number of strategies have evolved for optimizing factor analysis methods. The received view consists of recommendations for various extraction or rotational methods, recommended sample size, etc. An alternative view is proposed. It is argued that the single most important consideration in producing structurally valid factor solutions is the reliability of individual variables, and that an approach based upon the central limit theorem is particularly effective in increasing data reliability. Using a Monte Carlo simulation approach it is demonstrated that the central limit theorem averaging method is substantially more effective in producing structurally valid analyses than the various recommendations of the received view.
|
