Standardized Work-Sample Tests in Software Engineering Interviews

woman-in-black-and-white-polka-dots-dress-using-silver-3783824
Photo by Andrea Piacquadio from Pexels

As technology marches on, so should the interview process. The next step for software engineering is simple: standardized work-sample tests in coding, algorithms, and other such essential skills. These tests genuinely represent what it will be like to work for your company – a definite benefit both for yourself and your prospective employee! The standardization also makes the rubric clear for both you and the candidate and means everyone performs to precisely the same standard.

How to use these tests

Consider the tasks your new employee will need to do and include it when you’re creating your comprehensive programming and engineering test. You’ll get a sample of their actual work rather than their ability to answer exam questions! This process offers an objective evaluation of skills in a way that represents what they’ll need to be doing. It also limits bias towards one candidate or the other by measuring purely by this rubric.

Some facts about interview tests:

  • They have the best predictive performance when compared to other interviews
  • General Mental Ability (GMA) tests also rank high
  • Structured interviews perform well too, but most tech companies do not have a standardized structure for their interviews. The engineer running it often chooses the questions, giving a loose approach to the whole thing.
  • Ad hoc questioning is remarkably inconsistent, and candidates often are scored differently for the same answers by interviewers from the same company!
  • Both structured interviews and work-sample tests work best when used alongside a GMA test, though this is better off avoided in the US. The population tends to frown upon IQ tests as a rubric for intelligence culturally. As well, these can open companies up to discrimination lawsuits.
  • Graphology and age do not correlate with the success of an employee.
  • Surprisingly, job experience, reference checks, and education are not ranked highly in predictive performance!

The conclusion is simple, then: work-sample tests, along with some supplementation of other factors, are the best way forward.

Challenges

Designing the tests involves much work, which may be daunting to some – the kind of work you don’t need in an average interview process. As well, you must have a set time frame and absolute standardization while considering the needs of every candidate. Setting this gets more challenging the more people you interview.

On the other hand, you should consider the initial workload as an investment. It’s a one-time project (except for occasional updates) and saves time in the long run that would go towards coming up with ad hoc interview questions. As well, recruiting from a small pool of candidates or including set time frames in your interview invitations eliminates the second challenge.

Conclusion

Work-sample tests, at least for software engineering, are the way forward. They allow proper analysis of the candidate’s skills in an environment that matches them to their role. They help eliminate bias and inconsistency. Most importantly, standardization is more likely to get the exact candidate for the job who will be a boon to your company and help you grow.