Accuracy, stability, and discriminant validity evidence: a comparative study of exploratory and confirmatory methods for assessing statistical dimensionality

UNCG Author/Contributor (non-UNCG co-authors, if there are any, appear on document)
Alexandra Leslie Lay (Creator)
Institution
The University of North Carolina at Greensboro (UNCG )
Web Site: http://library.uncg.edu/
Advisor
John Willse

Abstract: Defining the dimensionality of an educational or psychological assessment is a complicated task that requires great consideration given its potential for far reaching validity implications. This three-paper series of real data analyses aims to investigate how various methods for assessing dimensionality can work in tandem to strengthen claims regarding dimensional structure and whether there are administrative conditions, specifically related to sample size, in which some methods are more appropriate than others. The first paper provides a comparative analysis of the accuracy and precision of three exploratory dimensionality analyses, exploratory DETECT, common factor analysis (FA), and principal components analysis (PCA). The second paper investigates dimensionality at a finer grain size than the first, evaluating the stability of the item-to-factor mapping provided by exploratory DETECT. Finally, paper three begins to bridge the gap between exploratory and confirmatory analyses by employing the DETECT-defined factor structure with confirmatory DIMTEST and confirmatory factor analysis (CFA).

Additional Information

Publication
Dissertation
Language: English
Date: 2020
Keywords
Dimensionality Analysis, Factor Analysis, Test Dimensionality
Subjects
Educational tests and measurements
Psychological tests
Dimensional analysis

Email this document to