DETECTING ERRORS IN A SCORING PROGRAM - A METHOD OF DOUBLE DIAGNOSIS USING A COMPUTER-GENERATED SAMPLE

This paper discusses a new method for locating errors in diagnostic co mputer scoring programs for structured clinical interviews. It was pro posed as a test of the accuracy of the scoring program for the Composi te International Diagnostic Interview, version 1.1. The proposal was t o create an independent scoring program in a different computer langua ge but serving the same criteria. Both programs were then applied to t he same large set of valid (i.e., logically consistent) computer-gener ated test cases, and differences in diagnostic assignments reviewed. T he method described can identify the program steps that account for th e sources of the errors. Corrections can be made and the programs run again on new sets of test cases until discrepancy-free results are ach ieved. While this method cannot discover errors that are repeated in t he two programs, it does discover more of the errors in a scoring prog ram than we have previously been able to identify. This technique prov ides a systematic and rigorous approach to assuring the accuracy of sc oring programs based on established algorithms.