Psychometric properties of the hope scale: A confirmatory factor analysis

Confirmatory factor analysis was employed to test several psychometric hypotheses regarding the Hope Scale. Across four large samples of college students, a two-factor (agency and pathways) model of hope reproduced the observed data consistently better than did a one-factor model. Support also was found for the tenability of a higher-order latent construct overarching these two factors. Neither the assumption of parallel nor tau-equivalent measures were met, however, suggesting that the items within a given factor are not interchangeable. Reliability estimates of (1) the items as indicators of the first-order construct, and (2) the first-order constructs as indicators of the higher-order latent variable also are presented. Support for the higher-order two-factor model of hope also was found for both men and women. Some support also emerged for invariance of first- and second-order loadings, though this was not a consistent finding. Implications for the quantitative application of the Hope Scale are discussed. © 1993 Academic Press, Inc.

Duke Scholars

Author Michael Alan Babyak Psychiatry & Behavioral Sciences, Behavioral Medicine & Neur ...