Item Response Theory Analysis of Psychometric Properties of Subjective Vitality Scale in French University Students
Résumé
Previous validation studies using classical test theory (CTT) methods have proposed multiple versions of the Subjective Vitality Scale (SVS) due to some relatively inadequate items. To resolve the discrepancies between these versions, we used item response theory (IRT) modeling to examine the psychometric properties of the scale's items and the latent trait, in a large sample of French university stu-dents. Factor analysis showed that the original 7-item SVS fits the data well. However, IRT modeling revealed that Items 2, 3, and 5 performed poorly in discriminating the latent trait. As a consequence, Items 1, 4, 6, and 7 were the most discriminant items. The new version (4-item SVS) demonstrates an excellent model fit, and improves better than the 7-item version. The two versions correlated, in the expected direction, with measures of life satisfaction and depression, respectively, but the new version displayed significantly higher strength of correlation with depression measure than the 7-item version. This study confirmed the advantages of IRT methods over the CTT ones, and demonstrated that the 4-item SVS was a reliable and valid alternative version for accurately measuring subjective vitality. \textcopyright 2020 Cises.