Rethinking Seniority in Coding Assessments

If you are using a coding assessment platform, you’ll need to ensure the test matches the seniority you are hiring for. Sure, most technical testing sites put a snazzy tag on their tests with a seniority level – junior, intermediate, or senior. Of course, it is not difficult to choose the “correct” seniority level, but mistakes can be made, and there is often more to each developer than meets the seniority-boxed eye. Believe it or not, the better way to approach seniority in testing is to have each test accommodate all seniorities. Why is this so? Allow us to explain.

The Reality of Developer Work

The everyday work developers do, just as with all professions, contains elements of all difficulty levels. Testing should reflect this reality. Let’s put this into context: Say a senior developer aces the difficult parts of a test but misses the easy things; this is an immediate red flag. As an employer, the last thing you want to be doing is constantly worrying about your employee’s inattentiveness and having to micromanage their tasking to only work on tasks of a specific difficulty or nature. In contrast, with a junior, why give them a junior-level coding test when, with a balanced test, you can see if they will overachieve and surprise you? Codeaid can help you do just that! I will tell you if a candidate scores in line with their seniority or if they under or overachieve.

Verify, Verify, Verify:
Uncovering the True Potential

If the results don’t align with the expected seniority, it’s time to take a closer look. Longer coding tests with balanced difficulty profiles are particularly well-suited to do this. For example, if a senior is sloppy, give them a chance to show that their lack of attention to detail is not part of their character. Conversely, perhaps you have a very motivated junior candidate that scored well, but you think they can do even better if given some feedback and time to test that theory.

Codeaid’s Feedback & Fix feature is excellent for such situations as it allows you to give automatically-generated, limited feedback to a candidate. With this, you can see if a concerning or promising aspect of a candidate’s initial performance continues.

The Calibration Balancing Act

As we mentioned above, longer coding tests support this comprehensive approach to seniority in testing. If a test is only 1-2 hours in length, it’s difficult to have meaningful elements of all difficulty levels. However, creating such a test requires careful calibration. Prior to publishing a Codeaid test challenge, we calibrate and fine-tune it using developers of known seniority. Once a coding test is public, the seniority scores are monitored to ensure proper separation and spread and can be adjusted by changing unit test weightings and by adding new content.

However, calibration is not only about scores. It is also about testing time. Juniors shouldn’t take 10x longer than seniors, but rather a 3x difference would be more reasonable. Keep in mind that it can take time for the calibration to settle into equilibrium. Calibration may look good at first glance, but as more test samples come in, problems can emerge, so long-term monitoring is essential.

The Takeaway: Don’t Box Candidates In

The bottom line is that a properly designed coding assessment should have richness, duration, and calibration accuracy that allows candidates to surprise both on the positive and negative sides. Don’t put candidates in a box before you even hire them, get a truer view of all their capabilities within various levels with Codeaid.