VALIDATION OF A PSYCHIATRIC PROGRAM EVALUATION SCALE
Abstract
Trends within the field of mental health indicate a need for accurate program evaluation measures. Psychiatric program evaluation instruments that have been validated are scarce. The purpose of this study was two fold. The first phase was concerned with the validation of the experimental scale (Psychiatric Program Evaluation Scale) which was constructed prior to the study. The second phase examined the use of this experimental scale in program evaluation. The patients (N = 130) used in this study were diagnosed as clinically depressed by psychiatrists. The validation process consisted of determining the concurrent, construct, and content validity of this experimental scale. Reliability estimates were also obtained. Concurrent validity involved comparing 130 patient scale scores to 130 treatment staff scale scores (criterion) by means of a Pearson Product-Moment correlation coefficient method with significant results being realized. In addition, a random sample of the patient ratings (N = 65) were compared to the ratings of nondepressed individuals (N = 65) and analyzed using t-tests to measure the significance of difference between the means. A significant difference was realized (p < .001). Construct validity involved the use of two factor-analytic methods to examine the underlying construct of the scale. Both solutions yielded similar factors accounting for less than 50% of the common variance. Content validity was determined by a panel of 10 expert judges who found the scale to be presentable, suitable, and clinically appropriate. Reliability measures involving interrater correlations and estimates of internal consistency were found to be significant (p < .001). The second phase of the study involved program evaluation applications by using this experimental scale on a pretest and posttest basis to reflect treatment outcome for the patients (N = 112) over a 10-day treatment period. The scale was filled out by both patients and assigned treatment staff with the results being analyzed using paired t-tests. The findings demonstrated statistically significant treatment changes according to patient and staff scores. Results are discussed along with recommendations for future research.
This paper has been withdrawn.