Confirmatory Factor Analysis of the Inventory of Personality Organization-Reality Testing Subscale

Dagnall, NA, Denovan, AM, Parker, A, Drinkwater, KG and Walsh, RS (2018) Confirmatory Factor Analysis of the Inventory of Personality Organization-Reality Testing Subscale. Frontiers in Psychology, 9 (1116). pp. 1-12. ISSN 1664-1078

Preview

Published Version Available under License Creative Commons Attribution. Download (889kB) | Preview

Abstract

The reality testing dimension of the Inventory of Personality Organization, the IPO-RT, has emerged as an important index of proneness to reality testing deficits. However, to date few studies have examined the factorial structure of the IPO-RT in isolation. This is an important and necessary development because studies use the IPO-RT as a discrete measure. Additionally, psychometric evaluation of the IPO suggests alternative factorial solutions. Specifically, recent work supports multidimensionality, whereas initial IPO assessment evinced a unidimensional structure. Accordingly, this study, using a heterogeneous sample (N = 652), tested the fit of several factorial models (one-factor, four-factor oblique, second-order, and bifactor) via maximum likelihood with bootstrapping due to multivariate non-normality. Analysis revealed superior fit for the bifactor solution (correlated errors) (CFI = 0.965, SRMR = 0.036, RMSEA = 0.042). This model comprised a general reality testing dimension alongside four subfactors (auditory and visual hallucinations, delusional thinking, social deficits, and confusion). Inter-factor correlations were in the moderate range. Item loadings and omega reliability supported the notion that the IPO-RT emphasizes a single latent construct. The model demonstrated invariance across gender and partial age invariance. Overall, from a psychometric perspective, the IPO-RT functioned effectively at both global and, to an extent, factorial levels. Findings recommend that the IPO-RT should be scored as a total scale, and rather than treat subscales independently, future studies should consider examining factor variance alongside overall scale scores.