Reclaiming Assessment Validity in the Age of Online Learning

Quiz 1 (Chart 1 below) produced a compressed distribution. Scores at both institutions clustered tightly at the maximum value. Density plots showed minimal spread. Most students earned near-perfect results.

That pattern appeared at both institutions. When a large group of students all score at the top, it becomes difficult to determine actual differences in mastery. The test loses resolution and validity. The key point: the distribution was not normal. It showed a strong ceiling effect, with most students achieving the maximum or near-maximum score.

When proctoring was introduced using Proctorio in Quiz 2 (Chart 2), the distribution changed immediately. Scores spread across the full grading scale, and variance returned. The density curve resembled a normal distribution rather than a spike at the maximum. In other words, the ceiling effect disappeared.

Statistically, paired t-tests comparing Quiz 1 and Quiz 2 showed p-values less than .001 at both institutions. That means the change was highly unlikely to be random, and that the two score distributions were significantly different from each other, removing the ceiling effect. Grades did not collapse. They showed normal and expected distributions.

To determine whether this was temporary, Quiz 2 was compared to Quiz 3 (both proctored).

As you can see from below (Chart 3), there was no statistically significant difference. A p-value greater than .05 indicates that there is no statistically significant difference between the two distributions. In other words, the ceiling effect observed earlier was successfully removed, resulting in normally distributed scores where the online asynchronous quiz performed similarly to the in-person, proctored quiz.

The largest shift occurred when proctoring was introduced. After that, performance patterns remained consistent.

This pattern suggests that expectations changed quickly. When monitoring was introduced, the assessment environment shifted. Students adjusted. The system recalibrated.

Dr. Basil described it clearly when “the goal was not to catch students. It was to ensure that grades reflect actual work.”

Proctoring, in this context, serves as a means of learning in an online environment, not as punishment.

The most important comparison may be modality.

One course was taught in person and online at the same time (Chart 4). Same instructor. Same materials. Same grading standards.

This confirms that proctoring technology effectively restores the integrity of the assessment, as it replicates the score distribution of a physically supervised classroom and eliminates the artificial ceiling effect seen in unproctored online environments.