A computerized learning approach that enables a user to improve their performance on multiple-choice exams is disclosed. Although the learning approach includes various aspects and implementations, in general, the invention forces test-takers to practice their examination skills and subject matter knowledge in a systematic way. The systematic way forces the users to follow a proven problem-solving approach designed to improve their performance. The learning approach provides important performance feedback to user's, for example, elapsed time per question and percentage of correctly answered questions. The invention can also pinpoint for the user the substantive areas of the exam which the user is either weak or strong. Further, the learning approach is able to assist the user in predicting his/her eventual score.