Key points are not available for this paper at this time.
A mental status rating was developed incorporating two major scoring techniques: the linear scale; and categorical judgements on yes-no or multiple choices. Using this form, 50 chronic hospitalized patients (over 2 years in the State Hospital) were evaluated independently but simultaneously by 3 Board-certified psychiatrists, one of whom served as interviewer, and the other two as non-participant observers. The role of interviewer was rotated, each psychiatrist interviewing approximately one-third of the patients. After a 6-week interval, 30 of the patients were re-evaluated by the same psychiatrists, each patient having a different interviewer than before. The obtained agreements on ratings of each item, as well as comparison or ratings of each psychiatrist with himself on first and second interviews (consistency scores), were calculated. The significance of the agreement and consistency scores was evaluated using the Pearson product moment correlation coefficient (r.) for scale items, and by comparing against chance expectations by Chi squared (X2) for categorical items. Several thousand calculations were performed using the IBM 650. The data breakdown was designed to demonstrate the effects upon reliability of differences between interviewers, differences between raters, changes in patients over time, and the limitations of the rating methods.
Rosenzweig et al. (Thu,) studied this question.