You may have noticed that the process of interviewing software engineers has no standard across companies. There exists this vague hodgepodge of "leetcode-style" questions, coding, system design, etc. but you never know really what to expect.

Having a non-standard interviewing process is not necessarily a bad thing! Every company has its own unique, specific needs, and they are evaluating for those needs accordingly. For example, one company may favor high level systems design more than algorithms. This may even be broken down even further; e.g. a business may favor knowledge of technologies on the full stack more than business and finance considerations. Having an interviewing process which is not standardized enables a company to maximize the odds that they are selecting someone who will excel in their role.

The key problem is when a business has an inconsistent process internally for the same role.

The reason this is a problem is because that means different candidates will get different experiences based on hidden variables and factors. This is dangerous from a business perspective because it means you are increasing your false negatives or false positives, wasting time and money. Furthermore, this can quickly become unethical and immoral, because this is precisely where hidden bias slips in.

Hidden bias--or unconscious bias--occurs when we judge someone and then alter our behavior based on this. All of us are susceptible to this, and it is imperative that we remain vigilant to these issues. If we interview someone who reminds us of ourselves, we can very easily decide that they are somehow more qualified than someone who we feel has more differences from ourselves.

It isn't enough that we just talk about this issue. We have to put safeguards in place to mitigate our human shortcomings. One step in this direction is ensuring that our interviews are standardized internally.

This may sound rather hand-wavy, so here are some concrete scenarios of where unconscious bias can creep in, an example of a poorly-conducted interview, and an example of the interview which mitigates bias.

Example 1

The candidate seems to be struggling to reason through a problem for quite some time.

Poorly-conducted Interview

We get a feeling that we should start expounding more than ideally would be necessary. We may also give some basic code examples that come to mind.

Bias-mitigating interview

We refer to clearly defined interviewer documentation within our company. This document specifies:

  • How long we should wait before providing guidance
  • What kind of guidance to provide the candidate and how much to escalate
  • How much guidance affects the interview result

Spot the differences

In the poorly-conducted interview, we rely on gut feeling in order to provide hand-holding for the candidate.

Depending on our biased perceptions of the candidate, our interpretations are open to distortions, resulting in multiple interpretations of the same interview, e.g.

  • "The candidate is probably unqualified, inexperienced, and needs hand-holding", or
  • "The candidate is probably qualified, experienced, and obviously just needed a little nudge"

Example 2

The candidate performed quite well in the interview. We're at time now and it is time to conclude, but we are in the middle of a very fascinating, stimulating discussion.

Poorly-conducted interview

Ask the candidate if they have a hard-stop or are free to keep talking a bit more.

Bias-mitigating interview

Inform the candidate that it is time to wrap up and thank them for their time.

Spot the differences

When we allow certain candidates more time than others based on how we feel, we are contaminating the process with exposure to unconscious bias. Even when a candidate is doing particularly well, we must not allow special exceptions.

Conclusion

Interviewing is a very complicated process, which is currently does not have a one-size-fits-all panacea solution. While our interviews should be iterative and open to improvement, it is imperative that we are being consistent with our candidates and providing everyone an equal experience and equal opportunity.

Bias in Interviews