Scores Based on Dangerous Responses to Multiple-Choice Items

Abstract

Scores based on the number of correct answers were compared with scores based on dangerous responses to items in the same multiple-choice test. The dangerous options were identified by content experts as representing responses to items that could cause significant harm to a physician's patients. The results of the study showed evidence for construct validity of both sets of scores. However, the total test scores and the dangerous option scores were completely redundant measures when evaluated by a correlation coefficient, corrected for unreliability. Therefore, unless more reliable dangerous options scores can be developed from afuture test, and unless these scores can be shown not to be redundant with the total test scores, the total test scores must remain the preferred choice for evaluating certification candidates.