The psychometric properties of the 18-item circadian type inventory (CTI) were examined in a non-shiftwork sample. Inter-item correlations for each scale were low to moderate. Confirmatory factor analysis (CFA) via structural equation modelling did not support the posited structure ofthe CTI. Principal components analysis and reliability analyses suggested the CTI is better represented by 11-items. The revised five item FR scale accounted for 27% of the variance, had stronger factor loadings and increased Cronbach alpha. The psychometric properties of the six item LV scale were marginally improved and explained 21% of the variance. The two-factor 11-item model was supported by CFA. Compared to the 18-item model, the 11-item model showed a marked improvement on several incremental fit indices and achieved a more parsimonious model fit. There was some indication of gender differences but a multi-group CFA indicated model parsimony was best for the invariant model. Gender difference needs to be further examined in large samples using a similar gender balance. It is concluded that the 11-item CTI is psychometrically superior to the original. However its predictive efficacy needs to be examined in longitudinal studies.
Funding
Category 1 - Australian Competitive Grants (this includes ARC, NHMRC)
History
Volume
36
Issue
8
Start Page
1953
End Page
1964
Number of Pages
12
ISSN
1873-3549
Location
Amsterdam
Publisher
Pergamon
Language
en-aus
Peer Reviewed
Yes
Open Access
No
External Author Affiliations
Faculty of Business and Law; TBA Research Institute; University of Wales;