Abstract

Recent studies have indicated that under the proper circumstances factor anaylisis may be accurately performed in samples as small as N = 9. However, all of these studies have extracted a pre-known number of factors, leaving an examination of determining the proper number of factors to future studies. The current study uses examines the following methods for determining the proper number of factors: Monte Carlo data to examine the performance of common versions of the Kaiser Rule, minimum average partial, parallel analysis and salient loading criteria under the conditions created by all possible combinations of method, model strength, overdetermination and sample size. Method performance was compared for overall accuracy (percent correct), and average discrepancy (mean difference from correct). ANOVA revealed that item level methods, including salient loading criteria and MAP procedures, maintain accuracy when model strength is at least moderate and overdetermiantion is high. Use of selected empirical methods for determining the number of factors is possible in small sample sizes only when overdetermination and model strength are adequately high, larger sample sizes should be preferred when possible.

LLU Discipline

Clinical Psychology

Department

Psychology

School

School of Behavioral Health

First Advisor

Boyd, Kendal C.

Second Advisor

Distelberg, Brian

Third Advisor

Herbozo, Sylvia M.

Fourth Advisor

Morrell, Holly E. R.

Degree Name

Doctor of Philosophy (PhD)

Degree Level

Ph.D.

Year Degree Awarded

2015

Date (Title Page)

9-2015

Language

English

Library of Congress/MESH Subject Headings

Factor Analysis; Quantitative Research; Psychometrics; Variability (Psychometrics); Number Concept

Subject - Local

Monte Carlo Data; Kaiser Rule; Minimum Average Partial; Parallel Analysis; Salient Loading Criteria

Type

Dissertation

Page Count

82

Digital Format

PDF

Digital Publisher

Loma Linda University Libraries

Usage Rights

This title appears here courtesy of the author, who has granted Loma Linda University a limited, non-exclusive right to make this publication available to the public. The author retains all other copyrights.

Collection

Loma Linda University Electronic Theses and Dissertations

Collection Website

http://scholarsrepository.llu.edu/etd/

Repository

Loma Linda University. Del E. Webb Memorial Library. University Archives

Download

Share

COinS