Improving Exams & Online Learning (2024)

It happens at all levels of education. From the anxiety-inducing timed multiplication test in 3rd grade to high-stakes final exams in college.

It happens more often during online exams. A survey1found that students are about 50% more likely to cheat in online courses than in live classes.

It happens for many reasons. Sometimes students cheat because of competitiveness and pressure. Other times, they don’t feel prepared, or it may just be because they can.

It happens in different ways. Whether it’s sneaking a quick Google search on their cell phones, whispering questions to Siri and Alexa, or having entire essays written by AI chatbots, technology makes cheating quick and easy.

No instructor looks forward to talking to a student about cheating (we really hope not). And while these are the last conversations you want to have with your students (and they feel the same), they’re necessary—no matter who, what, when, where, why, or how cheating happens.

But to make these conversations effective, you need to approach them thoughtfully. Talking to a student about cheating requires careful planning, consistency, emotional intelligence, solid evidence from the online proctoring software, and a deep understanding of disabilities and accommodations, which are often overlooked.

Responding to cheating the right way helps build a culture of academic integrity among your students.

“Instructors at our partner institutions have told us that addressing suspected cheating has an impact beyond each individual case,” said Jordan Adair, VP of Product at Honorlock. “Providing consistent feedback to students on their proctored exam results helps reduce future incidents and sends a clear message to students and their peers that academic integrity is valued at their institution.”

• How do you start the conversation?
• What should you say, and what should you avoid saying?
• What evidence do you need to prove they cheated during the proctored exam?
• What if their story is believable, but you still have concerns about academic dishonesty?
• Have you considered how disabilities may play a role?
• Are your test rules as objective as they should be?

• How do you start the conversation?

• What should you say, and what should you avoid saying?

• What evidence do you need to prove they cheated during the proctored exam?

• What if their story is believable, but you still have concerns about academic dishonesty?

• Have you considered how disabilities may play a role?

• Are your test rules as objective as they should be?

EMAIL TIPS

Templated email for other conversation purposes

Review your test rules and instructions

Writing clear test rules and instructions that leave no room for subjectivity is tough. While they seem straightforward, any ambiguity can complicate discussions with students.

Two example test rules that aren’t as objective as they seem:

With those examples in mind, play devil’s advocate as you review. This helps prepare for the conversation and refine your test rules. If updates are needed, inform students via email and LMS tools like messages, forum posts, and announcements.

Consider disabilities when reviewing evidence from the proctoring software

Tips for instructors to increase awareness and reduce stigma related to accommodations:

Tips for instructors to increase awareness and reduce stigma related to accommodations:

There’s no way around it: talking to a student about suspected cheating is inherently awkward, challenging, and even emotionally draining for both of you.

While we provide practical strategies and scripts for various situations, it’s important to recognize that each case is unique and rarely straightforward. However, being direct and sticking to the facts—firm and factual—keeps emotions in check and the discussion focused.

Prevent opinions and biases from getting in the way by focusing on objective information:

• The evidence you’ve gathered from the online proctoring software

• Your course policy and the exam/assignment rules

• The institution’s academic dishonesty policy

Thank the student for meeting with you, then tell them that you’ll discuss your concerns, present evidence from the online proctoring solution, and review academic integrity policies for your course and/or institution.

You also need to tell them they’ll have an opportunity to share their side of the story and ask questions. After that, the conversation can be similar to the email you sent them.

Tell the student exactly what your concerns are. Be specific and concise.

Example: “During the [exam name], the test proctoring software flagged you for [specific behavior, such as using a cell phone off-screen], which is against academic integrity policies.”

Then, you can transition to presenting the evidence:

“I watched a recording of the behavior, which the remote proctoring software flagged and was reviewed by one of their trained proctors. We can watch the recording together right now if you’d like.”

If they want to watch the recording:

“I have the video ready. Feel free to explain your actions as we watch. Let me know if there are times you’d like to pause or parts you’d like to rewatch.”

Instructor tips:

• Have the video cued and ready

• Write down the timestamps of important parts to watch
  • e.g., The student’s cell phone is visible at :20, :33, :45

• Ask open-ended questions instead of leading questions
  • For example, instead of asking a leading question like, “Were you looking at your cell phone here?” ask, “Can you describe what you were doing here?”

If they don’t want to watch the recording:

• Acknowledge their decision
  “I understand you’d rather not watch the video of the proctored exam, and that’s perfectly okay. We can discuss the flagged behavior based on what I observed.”

• Be highly descriptive (and accurate) when describing the flagged behavior

Always be direct and concise, but give as much detail as possible when describing the flagged behavior, like in the examples below.“During the exam, the proctoring platform flagged several instances where you…”

• • covered your mouth and asked Siri for answers, which can be heard in the video. I’ll read them from the transcript provided by the Voice Detection software within the proctoring platform, [read transcript].

• • used a cell phone, which the AI detected was nearby, to search for answers to test questions.

Whether they choose to watch the recording or not, ask open-ended questions to get their side of the story, such as:

• “Can you help me understand what was happening during when [specific behavior]?”

• “Can you describe what you were doing when [specific behavior]?

Responding when the explanation resolves the situation

“Thank you for explaining what happened. Based on your explanation, it seems this situation can be resolved. I appreciate your honesty and cooperation. Let’s continue to follow the guidelines to avoid any future misunderstandings.”

Responding when you’d like to seek guidance from another expert

“Thank you for your explanation. Hearing your perspective helps me understand the situation better. To ensure fairness and handle the situation appropriately, I’m going to consult with [specific expert, department, etc.] to better understand [specific topic]. I’ll follow up with you once I have more information.”

Responding when students accept responsibility for cheating

The conversation can be difficult even when the student admits to cheating and accepts responsibility. Your response may depend on the institution’s academic integrity policy, which could require you to report it, or it may be left up to you.

“I appreciate your honesty in admitting to cheating. However, academic integrity is fundamental, and I’m required to follow the academic integrity policy, which means reporting this incident. The appropriate office will determine the consequences. I understand this isn’t ideal, but it’s necessary to maintain fairness for all students.”

Responding when students deny cheating

The best response when students deny cheating (even with clear evidence in the video recording) is to use neutral language, stick to the facts, and ask clarifying questions.

“I understand that you have a different perspective, but the video shows repeated instances where you [state the observed behaviors]. We can address each instance individually. Can you explain why you were [first observable action].”

If there’s little progress, explain the next steps:

“Given the evidence and your explanation, I need to follow the university’s academic integrity policy. This involves reporting the incident to the academic integrity office for further review. They’ll evaluate the evidence, and you’ll have an opportunity to share your perspective and ask questions.”

Unless their explanation resolves the situation, you should review your institution’s policies for addressing acts of academic dishonesty, the next steps in the investigation and/or process, and the potential consequences for violations.

Instructor note: continue using neutral language, regardless of what you learned in the conversation.

For example, instead of saying, “Since you cheated on the exam, you’ll be [insert consequence],” say, “Violations of this policy can lead to these consequences…”

Talking to students about suspected cheating isn’t easy, but it’s key to maintaining academic integrity and protecting your reputation.

Take the time to prepare: review your course policies, the institution’s academic integrity policies, the evidence from the online proctoring platform, the objectivity of your own test rules, and consider disabilities.

Consider the student’s perspective: these conversations can be stressful for students, so remember to give them the opportunity to share their side of the story while actively listening.

Every word matters: from the first email you send to request a meeting to the conversation itself, stay objective, focus on the evidence, and have a real dialogue with your student to keep the process fair and constructive.

Proctor online exams effectively: use the right proctoring solution to deter and prevent cheating while keeping the test experience fair, customizing exam settings to accommodate students’ needs, and making it as stress-free as possible.

Blog Sign Up

Blog Sign Up

¹ Newton, P.M., Essex, K. How Common is Cheating in Online Exams and did it Increase During the COVID-19 Pandemic? A Systematic Review.J Acad Ethics22, 323–343 (2024). https://doi.org/10.1007/s10805-023-09485-5

² Adam, T., & Warner-Griffin, C. (n.d.). Use of Supports among Students with Disabilities and Special Needs in College. Data Point. NCES 2022-071. https://eric.ed.gov/?id=ED618477

³ Driving Toward a Degree 2023: Awareness, Belonging, and Coordination – Tyton Partners. (2023, November 16). Tyton Partners. https://tytonpartners.com/driving-toward-a-degree-2023-awareness-belonging-and-coordination/

⁴ Herrick, S. J., Lu, W., & Bullock, D. (2020). Postsecondary Students With Disabilities: Predictors of Adaptation to College. Journal of College Student Retention, 24(2), 603–624. https://doi.org/10.1177/1521025120941011

13 Ways to Stop Cheating on Online Tests

1. Detecting cell phones and other devices
2. Blocking AI like ChatGPT
3. Catching remote access software in contract cheating
4. Finding leaked test questions on the Internet automatically
5. Locking the test browser down
6. Listening for voice commands to activate voice assistants
7. Implementing hybrid virtual proctoring (AI + humans)
8. Verifying ID and monitoring behavior with a webcam
9. Creating explicit test rules & instructions
10. Reducing test anxiety
11. Prioritizing flexbility and accessibility
12. Addressing students suspected of cheating
13. Spotting trends in academic dishonesty through exam recordings and reports

Most people have cell phones or other devices, like smartwatches and laptops, which could be used to look up answers during online exams.

In fact, 71% of online proctored exam violations involve cell phones or other secondary devices.

Most exam proctoring services rely on a proctor to see a phone in real time, which is unreliable when a proctor is watching a nearly dozen test takers at once.

But some solution’s cell phone detection technology identifies when test takers attempt to use their phones and other devices to look up answers and when Apple devices are present in the testing area.

ChatGPT, Claude, and Google Gemini—all generative AI chatbots—are as controversial in online education as they are popular.

You tell it what to write about and it generates that content in a few seconds. They write pretty well overall, but they can be overly proper and generic sometimes. But with the right instructions, they can write like we (humans) do, and they’re really difficult for AI writing detection software to catch.

How to detect AI writing

Can plagiarism checkers help?

Nope. AI chatbots don’t plagiarize content. They create “fresh” text based on what billions and billions of resources they’ve been trained on.

Do AI writing detection tools work?

Not really. They’re helpful as a high-level gut check, but studies show that AI writing detection tools struggle when the AI-generated text is manually edited with a few small word swaps and paraphrasing. And, as we mentioned earlier, using specific prompts generates human-like content.

Can remote proctoring help?

Yes. Remote proctoring systems can prevent the use of unauthorized AI tools during online exams and even during essays and other written assignments by:

• Blocking access to other browsers and applications so AI tools can’t be used

• Listening for commands that activate voice assistants, which could be used to navigate AI

• Preventing test takers from pasting pre-copied text into exams and assignments

Has a technician ever remotely accessed and fixed your computer? That’s remote access software in action.

The downside is that it’s also used for cheating on online exams. The person getting credit pays a service for an expert to secretly take the test for them while they appear on video as if they’re completing the test.

How to block remote access software during online exams:

• Show locations by IP address: use proctoring software to show exam locations by IP address. If exams are taken in countries without known test takers, it could mean that proxy test-taking services were used and further investigation is needed.

• Record screens and require keyboard commands: before the tests or written assignments begin, ask learners to use specific keyboard commands that display all active applications on their device.

• Blocking applications: various settings within proctoring software can be used to block remote access applications, such as the browser lock.

Have you Googled your test questions? You should, because they’re often leaked on forums like Reddit, as well as sites that pretend to help with test prep and homework, but they’re really just repositories of test content alongside other avenues of cheating, like their “Expert Q&A” which is just hundreds of their “experts” answering questions 24/7 via chat.

The problem is that searching for your test content takes time… a lot of time.

The good news is that for your leaked test content in a few minutes and, if it finds any, gives you a one-click option to send a takedown request.

Use voice detection software that listens for specific keywords or phrases that activate voice assistants, such as “Hey Siri” or “OK Google,” to identify test takers who may be attempting to gain an unfair advantage. It then alerts a live remote proctor in real-time to review the situation and intervene if necessary.

This approach to makes proctored online exams less intimidating and non-invasive because a proctor will only intervene if the AI detects potential misconduct.

Hybrid proctoring solutions, which combine AI with human proctors and browser lockdown software, is highly effective to prevent cheating. But even with the same type of proctoring, the experience can vary widely depending on the company’s approach, for example:

• Company 1: a remote proctor, visible in a box on students’ screens, watches around a dozen test takers at once while AI monitors their behavior.

• Company 2: AI monitors behavior and pauses the exam for potential misconduct (even for innocent actions like sneezing), forcing a live proctor to step in unnecessarily.
• Company 3: behavior is monitored by AI, but instead of pausing the exam for potential misconduct, the AI alerts the proctor, who reviews the situation and only intervenes if necessary.

Based on these examples, it’s important to dig into how different companies approach proctoring, as it can seriously impact the entire test-taking experience and exam integrity.

Writing test rules and instructions can be tricky because they need to be clear and concise, yet comprehensive and objective. Even when rules are direct and clear, there could be room to make them more comprehensive and adaptable, like in this example:

• Original test rule: Do not talk during the test.

• Improved test rule example: Do not communicate with anyone during the test, whether verbal, non-verbal, or electronic.

It’s important to understand what causes test anxiety before taking steps to help reduce it. A student survey indicated that many feel anxious before an exam because they don’t know what to expect and they have technology concerns.

Two tips to help test anxiety:

1. Provide frequent practice tests to help students understand what to expect and ensure that their technology works correctly.
2. Use online proctoring software that combines AI with human proctors to help support students during the exam.

Learn more about how to reduce test anxiety using online learning technology.

Talking to a student suspected of cheating isn’t fun but, but it’s necessary to protect academic integrity now and in the future.When cheating is handled right—easier said than done—you set the tone for your students and build a culture of academic integrity.

Discussing cheating with students might seem simple, but it takes a lot of preparation and planning to make the conversation effective and to be ready for any situation that comes your way.

Remote proctoring software collects extensive data, offering insights into how students approach exams through actionable reports and time-stamped recordings directly in the LMS. Ideally, the reports will be easy-to-read and include timestamped recordings of important behavior.

Here’s how a lecturer from the University of Florida used remote proctoring to identify trends and anomalies in test scores:

“After the second quiz in the third week of class, I had a ceiling effect that looked like a ski jump, with 80% of my students getting 100% on tests. I knew there was something seriously wrong. I began looking closely at who had missed which questions over the two quizzes. That’s when I realized I needed a proctoring solution of some kind. When I initially attempted to address the issue, I really didn’t even know enough to ask the right questions to get help… It was more than a tool to guard or block students from using inappropriate information. It was also a means to detect and determine many different ways that students approach the exams.” – Ryan P. Mears, PhD, Lecturer, University of Florida

4 common reasons for cheating on exams

1. Pressure: whether it's keeping a scholarship or passing a certification for a raise, pressure drives many to cheat on exams.
2. Opportunity: surveys tell us that people are more likely to cheat on online exams than they are in a classroom or testing center.
3. Competitiveness: school and work are competitive, which can drive some to use questionable tactics when they feel they can't compete fairly.
4. Lack of preparation: sometimes cheating happens because someone just isn't ready—maybe due to long work hours, extracurriculars, health issues, or family time.

Want to see Honorlock in action? Schedule a demo.

AI detection and plagiarism detection are similar, but different. Both aim to detect unoriginal content, but their focus is different.

• AI detection looks for writing patterns, like word choice and sentence structure, to identify AI-generated text.

• Plagiarism detectioncompares text against huge databases of content to identify copied or paraphrased content from other sources.

Whether it’s a 1,500-word research paper or a series of short answer responses, Honorlock’s exam proctoring system makes it simple to secure essays and other written assignments, whether they’re typed or handwritten.

To sum up the process, Honorlock proctors essays just like online exams, meaning learners WON’T be able to do the following while writing:

• Access other websites, applications, books, or their notes, unless permitted by the instructor

• Paste content they copied before starting the essay

• Use generative AI to answer questions or write content

• Look up information on their cell phones

• Ask Siri, Alexa, and other voice assistants for help

Honorlock’s AI proctoring technology listens for “Alexa” or “Hey Siri” and other specific commands that activate voice assistants. If any suspicious activity occurs, it alerts a live proctor to review it, records what was said, and provides a transcript for review.

This allows learners to talk through their essays without being flagged for irrelevant sounds like doorbells or coughing.

But what if a learner accidentally triggers our voice detection by saying a phrase that’s also a command for a voice assistant, such as “Okay Google”?

Our AI would notify a live proctor, but the proctor doesn’t immediately intervene. Instead, they review the situation in an analysis window, and if they determine no cheating occurred, they won’t intervene or disrupt the learner.

The instructor turns on Scratch Paper Allowed within Honorlock’s proctoring settings, and the AI won’t flag learners for looking down at their paper.

The learner simply shows the blank paper before they start writing and uploads the written essay after they complete it.

Providing the proctor with instructions that allow handwritten essays and rules for learners helps keep the review process organized.

Essay rules can include things like:

• Only use a black or dark blue pen and blank/lined white paper
• You can write on both sides of the paper
• Before you begin writing, hold the blank paper up to the camera and show both sides
• Write the question number in the top left corner of the paper
• Hold the paper up to the camera after you finish writing your response to each question

Honorlock’s hybrid proctoring software and services combine AI with live proctors and 24/7 live support.

Unlike other online proctoring services, Honorlock makes the entire experience smooth and hassle-free. Remotely proctored exams can be taken 24/7 without the need for scheduling, and learners can get technical support from a real person. During proctored essays and assessments, learners won’t struggle with old, slow software or feel uncomfortable because a live proctor—a stranger—is watching them the entire time.

Text Emotion AI

Analyzes written language to identify the sentiment and emotional tone of the content.

Example uses: analyzing written responses, like forum posts and course evaluations, to identify, understand, and address emotional states.

Voice Emotion AI

Identifies emotions based on vocal characteristics like volume, tone, pitch, and speed.

Example uses: monitoring learners’ spoken responses during online classes and virtual presentations to detect nuances in learner interest or distress.

Video Emotion AI

Observes body language, facial expressions, and gestures to determine emotional states.

Example uses: observing facial expressions and body language during video conversations and online exams to understand learners’ confusion and levels of interest or disinterest.

Even if you’re a native English speaker, you’ve probably heard this or something similar from voice assistants. They’re about 95% accurate, sometimes more or less depending on which one you’re using, which is pretty good, right?

Sure, but that still means every 20th word is wrong, which is the exact length of the sentence you’re reading.

However, the accuracy varies depending on your accent… even if you’re a native English speaker.

The research also found that people with non-native accents experienced 30% more inaccuracies when using voice assistants, withSpanish and Chinese accents being the least accurate. Imagine how frustrating that would be.

But there’s good news: accent recognition AI tools are available that are trained on extremely diverse data that allows it to better understand accents. They analyze and interpret speech patterns, intonations, and pronunciations specific to different accents.

Accent recognition AI can help improve the operability of voice-controlled technologies and generate more accurate live captioning and transcriptions.

What’s an easy way to stifle a sense of inclusion and belonging? Mispronouncing someone’s name.

Even if it’s an innocent mistake, it’s probably something that person encounters daily. But there’s a solution that can help: name pronunciation AI.

• Integrates throughout your online courses within the LMS, SIS, and other platforms to ensure that names are accurately pronounced.

• Uses databases of audio name pronunciations and algorithms that recommend correct pronunciations.

• Learners voice-record their names, and the recording is available throughout the platforms.

• Online class discussions
• One-on-one advising sessions
• Virtual information sessions
• Recruitment conversations
• Student support

3. Automatically finding leaked test content

Have you ever found your test questions leaked on the internet?

“Homework help” sites like Chegg and discussion forums like Reddit and Quora make it easy to find and share your test questions and answers.

You have 3 ways to tackle leaked test content:

That’s it. No more leaked content concerns.

4. Large Language Models can help review college admissions essays without bias

Completely eliminating biases isn’t realistic, but they can and should be acknowledged and addressed, and AI can help, even though it can be biased too.

How can AI be biased?

AI reflects the biases of the people using it, the data it’s trained on, and the ways it’s used.

Which would you pick?

Large Langage Models (LLM)

What are large language models?

Large Language Models are a type of AI that can understand, interpret, and generate human language by analyzing and learning from extensive datasets.

How do LLMs work?

RoBERTa vs. ChatGPT

The University of Pennsylvania used a LLM, RoBERTa to review college admissions essays for personal qualities

Research results and takeaways

“An AI approach to measuring personal qualities warrants both optimism and caution… We recommend AI be used to augment, not replace, human judgment. No algorithm can decide what the goals of a university’s admissions process should be or what personal qualities matter most for that community.”

5. Preventing remote access software contract cheating

How can you stop remote access cheating?

Honorlock’s remote proctoring platform has a few ways to help:

6. Using AI for on-demand tutoring

Intelligent Tutoring Systems (ITS)

Chatbots

AI writing assistants

Regardless of which AI tools you use, make sure that they’re used purposefully, ethically, and with full transparency. Always keep in mind that AI tools are just that—tools. They aren’t a replacement for the people using them.

Sign up for more online learning resources​