Beyond the Score: The Importance of Feedback in Coding Assessments
Coding tests can tell you many things about a candidate, but what most don’t do is tell you how they respond to feedback.
Read MoreThis is super important because nobody wants cheaters to slip through the cracks. But the words used to describe these checks, like “plagiarism” or “similarity,” suggest that there’s more to this story. If a submitted code is a dead-on match or pretty darn close to a previous submission, it might mean something nefarious is going on, especially in longer and more complicated coding tests. However, lower levels of code similarity may not necessarily be good or bad, and that’s where the real debate begins.
Browser-based vs Take Home Tests
In the real world, developers make ample use of open-source code. It’s a way of working smarter instead of harder. But how do open-source code and coding tests intermix? Most online coding tests happen in one sitting and only last a couple of hours, which makes it tough to use outside sources. These systems usually block or make it hard to navigate away from the test browser window, so searching for open-source code is a no-go. Some might say this is a good thing, but it doesn’t really reflect the way developers actually work. These tests just don’t give developers the time or freedom to do what they’d normally do on the job.
In contrast, “take home” tests, like Codeaid, give developers the freedom to work in their own environment and use open-source code just like they would on a real project. Codeaid’s tests are 10 times longer than the industry standard of 1-2 hours, giving candidates several days to complete the test. With more time to complete the test, it’s less likely that two submissions will be too similar. However, with more time and freedom, using open-source code could result in more similar submissions. So, the question remains: what’s the ideal level of originality when it comes to code submissions?
What is the Optimal Level of Authenticity?
If a candidate scores higher and finishes quicker than average but has a slightly elevated code similarity level due to the use of open-source code, is that a good or bad result overall? Sure, code authenticity is important, but it’s just one piece of the puzzle. There’s no one right answer when it comes to the optimal level of authenticity; sometimes a little open-source use can actually improve a candidate’s score. But, at the same time, relying too heavily on open-source code can lower productivity and quality. The key is to use open source wisely and make sure you’re following the right licenses.
And there’s also a new wrinkle to consider with AI-generated code, like what you get from tools like ChatGPT. This is still a new development, so it’s tough to say yet what role these tools will play in coding tests. But it’s definitely something to keep an eye on.
It’s All About the Balance
The most effective coding assessment platforms will give candidates the freedom to use their creativity and judgment while also presenting a complete picture to the evaluator. This means taking into account not only code authenticity but also factors such as code quality, productivity, and problem-solving skills. Balancing the use of open-source code with code authenticity can be challenging, but it’s essential to evaluate the candidate’s skills and abilities in a fair and comprehensive way. And this is just the thing that Codeaid strives to do.
