The Rorschach Oral Dependency Scale (e.g., Bornstein, 2002). Thus, it was unexpected that the present study failed to find gender differences on either the Thonzonium (bromide) custom synthesis implicit or self-report dependency instruments. It is also noteworthy that no gender differences emerged on self-report dependency scales in another recent study (Cogswell et al., 2006), which is again inconsistent with the prior literature. Dependency and GGTI298 web Depression In a more exploratory vein, the relative associations between self-reported dependency versus implicit dependency and depression were compared. Not surprisingly, self-reported dependency was most closely associated with self-reported concurrent depressive symptoms, as all three self-report indices were significantly correlated with BDI scores. TheJ Pers Assess. Author manuscript; available in PMC 2011 February 21.Cogswell et al.Pageself-report measures also had some predictive utility in relating to self-reported past depressive symptoms, with the IDI correlated with Criterion A symptoms, and all three measures significantly correlated with Criterion B symptoms. Notably, implicit dependency scores were significantly correlated with self-reported past depressive Criterion A and B symptoms, but not with concurrent BDI scores. Although the reason for the disparity here is unknown, prior researchers (e.g., Bornstein, 2002; McClelland et al., 1989) would oppose the idea that current affective states are more closely tied to self-report measures than to more indirect measures, as they argued for precisely the opposite conclusion. Also interesting were findings relevant to predicted distinctions between neediness and connectedness. Connectedness actually was more strongly related to the various depression indices than was neediness, which is counter to the theoretical contrast between those two constructs. The literature, as discussed earlier, has consistently found neediness to be more pathological than connectedness; nevertheless, it is important to avoid references to connectedness such as “mature” or “adaptive,” given its relations to depressive symptomatology in the present study, and also in the extant literature (e.g., McBride, Zuroff, Bacchiochi, Bagby, 2006). To further tease apart the contributions of the different types of dependency measures in predicting depression, analyses were conducted that controlled for the effects of other predictors. First, as depicted in Table 4, when all predictors of depressive symptoms were entered simultaneously into multiple regression equations, only implicit dependency remained significant for prediction of self-reported Criterion A symptoms. Use of a backward elimination regression analysis retained both implicit dependency and connectedness for prediction of self-reported Criterion B symptoms. Additionally, and as shown in Table 5, implicit dependency remained a significant predictor of self-reported past major depressive episodes after controlling for the effects of both self-report dependency measures. This pattern of results indicates that whereas all selected dependency measures had some predictive utility for self-reported depression, only the implicit measure was incrementally useful in predicting past symptoms and episodes. This provides further evidence of the implicit dependency measure’s validity, as well as a recommendation for its future use in accounting for unique variance beyond that contributed by self-report instruments. As the present data cannot address what i.The Rorschach Oral Dependency Scale (e.g., Bornstein, 2002). Thus, it was unexpected that the present study failed to find gender differences on either the implicit or self-report dependency instruments. It is also noteworthy that no gender differences emerged on self-report dependency scales in another recent study (Cogswell et al., 2006), which is again inconsistent with the prior literature. Dependency and Depression In a more exploratory vein, the relative associations between self-reported dependency versus implicit dependency and depression were compared. Not surprisingly, self-reported dependency was most closely associated with self-reported concurrent depressive symptoms, as all three self-report indices were significantly correlated with BDI scores. TheJ Pers Assess. Author manuscript; available in PMC 2011 February 21.Cogswell et al.Pageself-report measures also had some predictive utility in relating to self-reported past depressive symptoms, with the IDI correlated with Criterion A symptoms, and all three measures significantly correlated with Criterion B symptoms. Notably, implicit dependency scores were significantly correlated with self-reported past depressive Criterion A and B symptoms, but not with concurrent BDI scores. Although the reason for the disparity here is unknown, prior researchers (e.g., Bornstein, 2002; McClelland et al., 1989) would oppose the idea that current affective states are more closely tied to self-report measures than to more indirect measures, as they argued for precisely the opposite conclusion. Also interesting were findings relevant to predicted distinctions between neediness and connectedness. Connectedness actually was more strongly related to the various depression indices than was neediness, which is counter to the theoretical contrast between those two constructs. The literature, as discussed earlier, has consistently found neediness to be more pathological than connectedness; nevertheless, it is important to avoid references to connectedness such as “mature” or “adaptive,” given its relations to depressive symptomatology in the present study, and also in the extant literature (e.g., McBride, Zuroff, Bacchiochi, Bagby, 2006). To further tease apart the contributions of the different types of dependency measures in predicting depression, analyses were conducted that controlled for the effects of other predictors. First, as depicted in Table 4, when all predictors of depressive symptoms were entered simultaneously into multiple regression equations, only implicit dependency remained significant for prediction of self-reported Criterion A symptoms. Use of a backward elimination regression analysis retained both implicit dependency and connectedness for prediction of self-reported Criterion B symptoms. Additionally, and as shown in Table 5, implicit dependency remained a significant predictor of self-reported past major depressive episodes after controlling for the effects of both self-report dependency measures. This pattern of results indicates that whereas all selected dependency measures had some predictive utility for self-reported depression, only the implicit measure was incrementally useful in predicting past symptoms and episodes. This provides further evidence of the implicit dependency measure’s validity, as well as a recommendation for its future use in accounting for unique variance beyond that contributed by self-report instruments. As the present data cannot address what i.