2025 ASEE Annual Conference & Exposition

In an introductory transportation engineering course, four versions of four tests were developed as a measure to deter cheating. These tests included a total of 45 true-false questions. Of those, five questions were written as true statements (true only), and seven questions were written as false statements (false only). The remaining 33 questions were written as a true statement (true variant) and a corresponding false statement (false variant) and assigned to the different test versions. This approach raised three concerns: 1) that the student performance on the different versions of a test would differ; 2) that true-false statements that were written as a true statement and corresponding false statement would provide some hint to students; and 3) that student performance on the true variant and false variant questions would differ. The test responses from 53 students were analyzed. For each of the four tests, the performance on the true-false questions on the four versions was found to be comparable. In addition, the performance on the true variant questions and true only questions was found comparable as was the performance on the false variant and false only questions. Lastly, the performance on the true variant questions was found to be equal to or greater than the performance on false variant questions. The conclusion is that assigning true and false variants across versions of a test does not introduce a bias when the proportion of true to false questions is consistent.