Reports on an evaluation of answer changing on multiple choice tests. States that any change a student is inclined to make, no matter how low the confidence, was found to lead to a significant gain. Concludes that students who change answers are likely to benefit despite the widely held belief that "first impressions are best." (GEA)