The use of client surrogates in determining the reliability of a standardized assessment instrument in multiple, diverse settings

Since its introduction in 1980, the widely-used Addiction Severity Index has been subjected to extensive validity and reliability studies. Reliability procedures have commonly entailed test–retest and multiple judge video reviews, all within controlled environments. The following study uses a method that differs substantially from conventional reliability assessment by employing trained, professional actors as client surrogates to examine the reliability of the instrument when administered in the field, in naturalistic, diverse agency settings. This article details this method, addressing the question, to what extent did this method produce consistent stimuli to enable the measurement of reliability?