The reliability and validity of the EHAS in civil psychiatric patients

This study examined the reliability and validity of a relatively new measure of hostile attributions, the External Hostile Attributions Scale (EHAS). Participants were 77 civil psychiatric patients discharged from a psychiatric ward of a general hospital. Results showed that the EHAS had good internal consistency at baseline and first follow-up. Generalizability (G) theory analyses suggest that EHAS scores were moderately stable over time, and that changes in item scores across participants and time accounted for most of this variance. Results from univariate analyses demonstrated that the EHAS had good convergent validity with measures of thought disorder, criminal attitudes, and treatment compliance. It was not, however, related to ratings on risk assessment instruments. The EHAS also demonstrated good divergent validity in that it was unrelated to measures of substance use, social support, and impulsivity. Finally, EHAS scores were not related to violence, victimization, and suicide-related behaviour retrospectively or prospectively.