Choosing Among Online Exam Options: Finding the Best Value

Human-based proctoring is often the first solution institutions consider to assess online exams. After all, human proctors are what we’re used to, and we’ve relied on them in classroom settings for decades. But as online learning grows, finding enough qualified proctors can be hard, and the cost can be astronomical. Some institutions look to solutions like proctor farms in an effort to secure staff and minimize costs. Such companies may have the number of proctors you need, but lack of quality can lead to disastrous results for students and loss of trust in institutions.

While human insight can be extremely beneficial in online learning, the drawbacks of relying exclusively on human proctors are significant:

• High Cost. An exam proctored in real time by a person costs nearly double the amount for an exam proctored solely through AI.
• ‍Lack of Training. Institutions may turn to volunteer proctors or proctor farms to find cost savings, but lack of experience and training can compromise test results.
• ‍Bias. Live human proctors introduce human bias, which has been shown to negatively impact women, students of color, and students with disabilities.
• ‍Security. A human proctor typically uses a webcam to observe the examinee’s every behavior throughout the entire test, searching for suspicious behaviors. Many students find this deeply uncomfortable, augmenting test anxiety and potentially reducing test performance.
• ‍Logistics. Scheduling exams with a large number of human proctors and students dispersed across the globe can be challenging and limit your ability to stay flexible and responsive.
  

These disadvantages can seriously impact your ability to administer tests online, produce valid results, and maintain good relationships with students.

Fully automated proctoring tools can offer scalability and may reduce costs by eliminating human proctors altogether. But complete reliance on AI can cause more harm than good.

Many fully automated systems are overly sensitive and feature AI trained on limited datasets, compromising reliability. This can result in not only inaccurate flagging, but profoundly damaging algorithmic bias that disproportionately affects historically disadvantaged students. Additionally, poorly designed solutions may tempt students to circumvent security measures, particularly if the system does not continuously refine its knowledge.

Both fully human and fully automated solutions have damaged trust between students and institutions due to their propensity to perpetuate inequality. As we have seen in recent years, overly invasive and poorly designed automated solutions can also lead to extraordinary costs in the form of lawsuits and data breaches.

Finding a solution with the ability to proctor reliably, securely, and cost-effectively on a large scale is vital to providing high-quality educational experiences in an increasingly digital world. Rosalyn’s human-in-the-loop technology offers an innovative system at a cost institutions can afford. Balancing artificial intelligence and human judgment provides outstanding scalability and greater comfort for students—something many other proctoring solutions lack regardless of cost.

The Rosalyn real-time AI proctoring system:

• Uses AI to identify, in real time, potential offenses and sends these events to human-in-the-loop proctors to assess in real time. This affords immediate action when needed, minimizes potential errors, eliminates the need for continual monitoring by humans, and reduces the number of proctors needed for each exam.
• Continually integrates new information and adapts to changing conditions.
• Is developed with student input to increase comfort and reduce anxiety so students can perform at their best.
  

With Rosalyn, you don’t have to choose between quality and affordability. Our innovative technology offers great value for students, educators, and institutions, now and in the future.