Abstract
dc:description.abstractWe introduce and develop a new statistical method for exploring latent structures: Frame Analysis. Frame Analysis drops the one-to-one correspondence between factor dimensionality and vector space representation found in Factor Analysis. This minor change obviates factor rotations, simple structure, and provides equal status for cross-loaded items. We show that in Frame Analysis, manifest items are defined by only one frame loading and are uniquely characterized as a linear combination of latent variables: a frame vector. Through a series of simulations, we characterize Frame Analysis performance in three scenarios: Exploratory, Constrained, and Partially-Constrained. Finally, we apply Frame Analysis to archival five-factor personality data and provide evidence that hierarchical personality models are disguised frame vectors.
Degree
thesis:*- Level thesis:degree_level
- Masters
- Department dc:contributor.department
- Psychology
- Grantor dc:publisher
- Georgia Institute of Technology
- Year dc:date.issued
- 2021
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Sanchez, Ryan C.
- Advisor dc:contributor.advisor
-
- Hunter, Michael D.
- Committee members dc:contributor.committeemember
-
- Roberts, James S.
- Fletcher, Keaton
Subjects
dc:subject × 5Rights
- Language dc:language.iso
- en_US
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- http://hdl.handle.net/1853/65998