A psychometric evaluation of hierarchical and oblique versions of five variants of the Posttraumatic Growth Inventory. (30th December 2016)

• Record Type: Journal Article
• Authors: Horswill, Samantha C.; Desgagné, Gabrielle; Parkerson, Holly A.; Carleton, R. Nicholas; Asmundson, Gordon J.G.
• Is Part Of: Psychiatry research. Volume 246(2016)
• Page Start: 438
• Page End: 446
• Publication Date: 2016-12-30
• Subjects: PTGI -- Confirmatory factor analysis -- Assessment -- Measurement -- Trauma; Psychiatry -- Periodicals; Psychiatry -- periodicals; Psychiatrie -- Périodiques; 616.89
• DOI: 10.1016/j.psychres.2016.10.027
• Languages: English
• ISSNs: 0165-1781

Abstract

Abstract: The Posttraumatic Growth Inventory (PTGI; Tedeschi and Calhoun, 1996) is the most commonly used measure of posttraumatic growth. Although the original five factor structure has extensive psychometric support, evidence exists for alternate PTGI models. The current study assessed the validity of oblique and hierarchical factor structures of five PTGI models. Confirmatory factor analyses were performed among a heterogeneous community sample with a diverse trauma history. The oblique models provided a better fit to the data compared with their respective hierarchical models. Three oblique variants provided good fit to the data on two fit indices and all five oblique variants met the recommended criteria for at least one fit index. The 10-item model demonstrated the lowest Expected Cross-Validation Index (ECVI) values and is a brief and useful measure when examining PTGI total scores; however, current results suggest that consideration of the specific subscales may be more meaningful than total scores. Researchers and clinicians interested in subscale scores should consider using the 18- or 21-item variants due to a higher number of items per factor and, therefore, greater factor stability. Future directions in refining and measuring posttraumatic growth are discussed. Highlights: Five variants of the Posttraumatic Growth Inventory were evaluated. Oblique models provided a better fit to the data than hierarchical models. Based on its ECVI value, the 10-item oblique variant has the highest likelihood of replicability. The 18- or 21-item oblique models may be more reliable when using subscale scores. Qualitative and item analyses are needed to refine the posttraumatic growth construct.