Measurement of External Shame: An Inside View

Abstract

The aims of this study were to investigate the construct validity of the Other as Shamer scale (OAS) using confirmatory factor analysis (CFA) and to examine the psychometric properties of its Italian version in a sample of 687 nonclinical individuals. The CFA results indicated that the hypothesized hierarchical model (with 1 higher order factor and 3 first-order factors) was the best fitting solution. Cronbach's alpha indexes, as well as test–retest stability, provided satisfactory results. Correlations of the OAS total score and its subscales with the Beck Depression Inventory–II (rs = .30–.48) and the Teate Depression Inventory (rs = .32–.45) were both substantial and significant (p < .01). Receiver operating characteristic curves were constructed to indicate sensitivity and specificity of the OAS and its subscales when determining those nonclinical subjects who met clinical thresholds for depression symptoms. A series of cutoff scores for the OAS scale and its subscales was developed, with sensitivity values between .70 and .62, and specificity values between .71 and .62, indicating good to fair discrimination between the 2 groups (depressed vs. nondepressed). The theoretical and practical implications of these results were discussed.

Notes

¹ Modification indexes (MIs) are often used to alter models to achieve better fit, but this must be done carefully and with theoretical justification. That is, blind use of MIs runs the risk of capitalization of chance and model adjustments that make no substantive sense (see Silvia & MacCallum, 1988). In MIs, improvement in fit is measured by a reduction in chi-square (recall a finding of chi-square significance corresponds to rejecting the model as one that fits the data).