An ecological study of the reliability of clinical judgment in executive appraisal

This research was a study of the reliability of clinical judgment findings (multitrait) across three different information sources (psychometric tests, structured interview, and psychometric tests and interview). Subjects (N = 74) were middle and senior executives of Western Canadian technical companies; clinicians (N = 3) were trained and experienced industrial psychologists. The study investigated the similarity of clinical evaluation of personological characteristics (based on an 18-factor multitrait paradigm) across the three different information sources. Subjects were independently rated by a single clinician on 18 criterion factors in each of the three information source categories. Test information source categories required the administration of approximately 12 hrs of standardized psychological assessment questionnaires to each of the 74 subjects. Interview source category involved a 1.5-hr structured interview per subject. Combined condition pooled both test and interview conditions. Generalizability of the findings was maximized by the undertaking of the experiment in a natural situation thus increasing ecological validity. Statistical treatments used were designed to assess the similarity of a clinician's evaluation of a subject based on the different category of information available about that client. Convergence (intrarater reliability) indexes range from a high of .64 to a low of .05. Results indicate a varying degree of convergence of multitrait clinical ratings dependent on clinician and trait being rated. Results are discussed in terms of implications for practitioners involved in executive personnel selection.