Final Paper pdf 171KB

Authenticated Testing During Blended Delivery: Impacts On Assessment Scores Within An Engineering Undergraduate Core Course

Sustaining sufficient levels of academic integrity has been a challenge within various online delivery modalities. In this paper, the impact of requiring authenticated high-integrity testing is measured in a case study of a required undergraduate mechanical engineering course that is delivered via a combination of online components and weekly face-to-face meetings. Measurements of students’ assessment scores with the course are contrasted when delivered both with and without proctored lockdown assessment. Herein, proctored quizzes are compared to un-proctored quizzes in a mixed-mode undergraduate engineering course (Engineering Analysis: Dynamics) with 264 students enrolled who underwent alternating occurrences of assessment methods. The proctored quizzes took place with a lockdown computer-based assessment (CBA) at the Evaluation and Proficiency Center (EPC), which is an on-campus testing and tutoring facility that delivers quizzes authored in the Canvas Learning Management System (LMS). During the semester, four quizzes were administered, two of which were proctored and the other two un-proctored. The class was partitioned into two groups: Groups A and B. When Group A was proctored, Group B was not and vice-versa rendering fairness to all students. Results indicated a significant difference between for proctored quizzes and the un-proctored quizzes. The least difference in mean scores between one of the proctored quizzes (the fourth quiz) and the corresponding un-proctored one was 41%. This high difference could be attributed to the amount of possible cheating the students do when un-proctored, even for a timed quiz. Un-proctored online assessment is seen to greatly inflate the students’ grades, while keeping the students learning achievements at a lower level not corresponding to these grades. The student survey results at the end of the course corroborated the findings. A t-test resulted with p=0.0223 that take-home quizzes inflate grades and the students agreed that the EPC is effective in monitoring cheating activities (p=0.0145).