Model Equation (Examining the Effect of Proctoring on Online Test Scores)

A linear mixed-effects model was used to evaluate the data. This is due to a realistic assumption that can be made––responses from tests taken by the same student or students with the same instructor may be related. A mixed-effects model allows for those relationships to be accurately reflected. 

Arcadia College

A 2017 study by Alessio et al. found evidence for the prevalence of cheating in online distance learning. The study examined 144 undergraduate students over the course of a full semester. Exams were administered online in one of two over-arching conditions: proctored and un-proctored. The results showed a significant grade disparity among the two conditions, with proctored students scoring an average of 17 points (two letter grades) lower per exam. Additionally, proctored students used significantly less test-taking time compared to their non-proctored peers. On average, proctored students used 30.5% less of the allocated exam time. These results were validated between-subjects as well as within-subjects. Likely, the higher exam grades and longer testing times observed in non-proctored students are due to time spent using restricted materials (Google, textbooks, notes, etc.). Given these results, the use of proctoring services seems to lessen the prevalence of cheating in online exams. However, additional research is needed in the field before assertions are made. 

Academic Dishonesty in Online Distance Learning

Alessio, H., Malay, N., Maurer, K., Bailer, A., & Rubin, B. (2017). Examining the effect of proctoring on online test scores. Online Learning, 21(1). https://doi.org/10.24059/olj.v21i1.885

Alessio et al. (2017) Examining the Effect of Proctoring on Online Test Scores

The purpose of this study was to compare online test results from proctored versus un-proctored online exams. While many studies have addressed the cheating angle, the overall data has been inconclusive on the prevalence of cheating in un-proctored vs. proctored exams. 

Examining the Effect of Proctoring on Online Test Scores: Purpose

Participants: 144 undergraduate students attending Miami University (Ohio), enrolled in Medical Terminology. Data collection ran through the entire semester, with measures taken to assure student anonymity. 


Conditions (3): Un-proctored, un-proctored and lockdown only, and proctored with video monitoring (Software Secure or Respondus Monitor)


**Of the nine available sections, four used proctoring software; the other five sections used no proctoring software on any exams. Students in the proctored conditions were not told which exams would be monitored (4 exams, random selection). 


ALL students, regardless of condition, were reminded that the exams were meant to be taken alone, with no peers or other resources aiding. 

Examining the Effect of Proctoring on Online Test Scores: Methods

Examining the Effect of Proctoring on Online Test Scores: Data Analysis

The results showed significant differences in proctored and un-proctored exams. First, proctored students scored an average of 17 points (two letter grades) lower per exam. The average test score for proctored exams was 74.3% and 89.4% for un-proctored exams. ​Additionally, proctored students used significantly less test-taking time compared to their non-proctored peers.
​The average percentage of time taken on proctored exams was 20.4% and 41.2% for un-proctored exams (showing that students took approximately half the amount of time when taking proctored exams compared to their un-proctored peers). 

Examining the Effect of Proctoring on Online Test Scores: Results

The results of this study show clear grade disparities between proctored and un-proctored exams. One explanation for the higher exam grades and longer testing times observed in non-proctored students may be time spent using restricted materials (Google, textbooks, notes, etc.). Un-proctored students may have taken advantage of the online format and time available to look up answers. 

Given these results, the use of proctoring services seems to lessen the prevalence of cheating in online exams. After controlling for the effects of test difficulty and student and teacher differences, students taking proctored online exams scored ~17 points lower compared to un-proctored students.

Examining the Effect of Proctoring on Online Test Scores: Discussion

While the results of this study were statistically significant, there were a few limitations present that may decrease its generalizability. First, the class relied heavily on memorization; all the exams were multiple choice and close-ended. Therefore, an effect this large may not be found in other class types. Second, the study only looked at two proctoring services: Software Secure or Respondus Monitor. Third, this study used a population of traditional students from a Midwestern university. Larger studies should be conducted with varying samples to better understand any potential effects of proctoring on exam scores. 

Learn Before

Related

Learn After