Comparison of Final Examination Formats in a Numerical Methods Course

摘要

With decreasing budgets for teaching assistants, large class sizes, and increased teaching loads, it is becoming ever more important to effectively utilize resources without sacrificing best practices of assessment. The objective of this study is to evaluate a hybrid multiple-choice final examination with optional partial credit (MC+PC) as a replacement for the same examination in constructed response (CR) or strict multiple-choice (MC) formats. In the hybrid MC+PC format, students were given multiple-choice options and were also allowed to submit constructed responses that would be graded for partial credit. The three examination formats were utilized once each in three offerings of a Numerical Methods course at the University of South Florida. Multiple linear regression and item analysis of student responses demonstrate that students approach the MC+PC format similarly to a CR exam, and the administrative requirements of the test were significantly reduced. This study finds the hybrid MC+PC format to be equally reliable and appropriate for a comprehensive final examination.