Depending on what the circumstance in which you are given the assessment, and the pool of developers who you will be sending them to, you may want to be concerned about cheating. If you are concerned (and not everyone should be), Qualified provides a number of features to help you determine if a candidate or student is trying to complete a coding challenge without doing the work themselves.
Code Similarity (Premium Feature)
Code Similarity is a feature which will compare a submitted solution to all other submitted solutions in the system for the same challenge. If a challenge is copied from the Qualified library and still linked (unmodified), then all solutions, even those on other teams, will be used to determine if the solution is similar to another.
This feature is powerful, as it compares not just the exact code itself, but the structure of the code. For example, changing variable names, function names or rearranging code will still result in the code being determined to be the same structure as other code that is otherwise the same.
This feature does not ever indicate that a code solution is plagiarized, because there is no way for the platform to know the context in which the code is being written. It may simply be that a certain type of challenge results in similar solutions being submitted. However there is a risk level associated with solutions that system finds to be closely similar to others. This risk level will either be High, Medium or Low. You can use these risk indicators as guides for when you should look closer at a solution, to see if it is in fact something to be concerned about.
Qualified provides the ability to playback an entire coding session. You can see the developer's key strokes in real-time, or speed it up to get a quick glance of how the solution came to be.
We recommend using this feature not only to get a view into how the developer thinks, but also to identify any red flags. The playback timeline will indicate any major code paste events, making it even easier for you to identify important events that you should pay attention to. Check out what was pasted in and see if it makes sense. Our system will only flag code pasted from an external source, so you won't get a bunch of noisy paste events that were caused by normal solution development.
This isn't a way to detect cheating, but to prevent it. Since the Qualified platform utilizes unit tests to test code, you can write randomized tests if you prefer. These randomized tests make it impossible for a candidate to hardcode their solution to pass specific test values. Many of our challenges in our library already contain randomized tests. However we should note, that within high-stakes situations such as job selection, candidates very rarely (< 1%) cheat in a way that randomized testing is necessary. However in cases of education with large classrooms, randomized testing may be something you want to consider.