To help provide every candidate with a fair chance, we’ve included numerous anti-cheating measures.
We like to believe that all candidates will follow the guidelines. Yet there is always the potential for some to try and bend the rules. To prevent this, we include anti-cheating settings on all plans. This article is relevant to all users of all plans.
Approx. reading time: 6 minutes
| While we employ several anti-cheating measures to monitor a candidate’s activity, we cannot say for certain that a candidate has attempted to cheat. These measures are meant to highlight potential areas of concern, and should not be used on their own as grounds to disqualify or reject a candidate. |
In this article
1. Understanding Behavioral Tiers
Behavioral Tiers provide a structured way to interpret anti-cheating signals by grouping candidate actions into three risk levels: No Issues, Some Issues, and Major Issues.
Instead of looking at individual anti-cheating flags in isolation, Behavioral Tiers contextualize a candidate's actions. This system recognizes that cheating is a spectrum and aims to distinguish the depth, intensity, frequency, and variety of concerning or non-concerning behaviors.
Any candidate who completes an assessment is placed into one of three tiers based on the depth, intensity, and variety of their behavior.
The Three Tiers at a Glance
Tier |
Icon |
Interpretation |
Recommended Action |
|---|---|---|---|
No Issues |
Checkmark |
Lowest risk level. No significant suspicious activity detected. |
Proceed with the recruitment process as usual. |
Some Issues |
Orange Flag |
Intermediate risk. Minor signals detected, such as moderate Tab Switches or Full-screen Exits, or opening Developer Tools as the last recorded event in a test. |
Open the behavior log and skim the event counts for context. |
Major Issues |
Red Flag |
Highest risk level. Significant signals detected, such as frequent full-screen exits or the webcam being disabled. We recommend caution. |
Review the full behavior log carefully before making a hiring decision. |
| Note: TestGorilla does not label candidates as "cheaters" because proving intent is impossible, we provide these signals to help you make an informed, human judgment. |
2. Monitored anti-cheating behaviors
To protect the integrity of your assessments, TestGorilla monitors several specific indicators of candidate activity. It is important to note that these flags are symptoms of behavior, not outright proof of cheating.
An event being flagged is not automatically a bad thing. For example, a candidate might exit full-screen mode to close a system notification or switch tabs to check a technical requirement if they believe it is permitted. Our goal is to provide you with the data needed to make an informed decision, rather than to label candidates as "cheaters" based on circumstantial evidence.
The following behaviors are monitored:
Webcam Snapshots: The system captures images of the candidate every 30 seconds to verify their identity.
Full-screen Tracking: We monitor whether a candidate stays within the full-screen test window or exits the mode.
Tab Switching: We detect if a candidate switches browser tabs during the assessment.
Webcam Status: The system tracks if the camera is disabled or turned off.
Developer Tools: The system logs if a candidate opens browser developer tools.
IP-based Location: We record the approximate location and IP address to identify potential unauthorized attempts from the same user.
Paste Detection: By tracking when and where content is pasted, this tool helps you cross-examine candidate submissions for authenticity, ensuring that the work you're evaluating is truly their own.
3. Other anti-cheating measures
We implement the following anti-cheating measures to protect our users and the integrity of our tests:
- Cycled questions. Every test in our library has a question bank numbering in the hundreds. We have an algorithm in place that ensures that different questions are selected and presented every time a test is used in an assessment. This makes it highly unlikely that candidates will face the same questions again, even if taking the test for a different company.
- Questions have a maximum exposure count. To further protect our test content, all questions have a maximum exposure limit. After this, we retire the question and bring a fresh one in to replace it.
- Study prevention: We prevent candidates from registering as new customers so they can't practice tests ahead of your assessment. We only allow users with business emails to sign up for accounts — personal and educational email addresses are blocked from signup. This prevents candidates from practicing tests or using TestGorilla as a study aid.
- Questions and answers are hidden. We keep the questions and their correct answers hidden to protect the integrity of our tests. Not even our staff can access them! While this means our customers can’t see the information, it also means anybody who might bypass our account creation filters can’t either. Additionally, if this information was easily accessible, it would increase the risk of the test content finding its way onto the general internet. At that point, a simple Google search would provide your candidates with all the answers they need.
- Multiple-attempt prevention. We detect if a candidate is trying to register for another attempt, and block them from doing so — one attempt per email address is all we allow.
- Timed tests. Timing tests lets us compare all candidates equally, while also adding a bit of pressure. Our tests are designed to be completed based on knowledge candidates already have. Adding time pressure means the candidate risks being unable to complete a test should they try to search for answers they don’t know. Once a test starts, the timer will continue running — even if the candidate closes the test window.
- Honesty agreement. Before proceeding, candidates must sign the TestGorilla honesty agreement. Doing so confirms their commitment to completing the assessment without using AI tools, retaking the assessment multiple times, or engaging in any other behaviors prohibited by our Candidate Honesty Policy.
| Tip: If you wish to provide extra time for the tests, you can do so in Step 4 of the assessment creation process. See our guide on advanced assessment settings for help with this. |
4. Common questions
How do I know if someone cheated?
The anti-cheating monitors are designed to highlight things that might be worth looking into. We cannot say for certain that a candidate has attempted to cheat. For example, an IP location flag could be because they were taking the assessment on their mobile device while in a car.
If any flags are raised, we recommend contacting the candidate first as they may have a valid excuse for it. We do not recommend rejecting candidates based on this information alone.
How can I disable the anti-cheating settings?
The only setting that can be disabled is the snapshot feature. Our advanced assessment settings guide explains how you can do this.
Should I immediately disqualify someone who has triggered a flag?
None of the information provided in the anti-cheating monitor should be taken as conclusive proof of cheating or an attempt to cheat.
TestGorilla never recommends using the information from the anti-cheating monitor as grounds to immediately reject or disqualify a candidate.