Reducing Bias in your interviews

You're biased. Yes, you have your biases no matter how certain you are that you don't. Some of them are conscious, others unconscious–if you still don't believe me, take this testBiases cause a lack of diversity in companies. They make you miss out on great candidates. Keep reading to learn some of the techniques I employ to reduce them in my interviews.

Diversity Matters Diversity matters. Getting the best candidates matters–Photo by Sharon McCutcheon on Unsplash

Our brain loves jumping to conclusions. It's an evolutionary mechanism that has kept us alive for all this time. We rely on intuition for most of our decisions. So our primary job here is delaying our intuition. Making it more accurate and based on what matters for success. We want to be objective in our assessment.

Objective assessment: Key Criteria for success

It always surprises me the number of interviewers that go into an interview without knowing what they're assessing. They've been to other interviews and have a set of questions that they know to ask, and that's it.  They don't know what traits correlate to success within their company, and this is one of the main reasons people fail to choose suitable candidates.

So the first step in reducing bias in your hiring process is determining a list of Key Criteria you want to assess. Don't jump to a 15 bullet point list. Choose carefully and prioritize–ideally, go for 5 to 6 items. There are many different sources for this, but the main question you want to answer is: "What would make someone successful in this role?" If you can't answer that question, go back and check:

  1. The job description;
  2. The career ladder;
  3. Your experience and
  4. The experience of your peers

These criteria should vary for different positions. Once we have them and trust that they're consistent measures of success, our next job is gathering information to determine how the candidates score on each of these.

Information Gathering: Study the Resume, Profile, and past phases

For non-entry-level positions, generally, the Candidate's experience is a gold mine of information. It's hard to surpass the information in a range of experiences that person had in all the years they have worked with a 1-hour interview. So we should value that. When we see a Resume that shows probable good scores on what we are trying to measure but has a hard time gathering that information during the interview, at the very least, we should doubt our performance in the interview and revalidate.

Some people pamper up their resumes, and extracting info from Resumes is hard, but I find that it's generally worth the time. Looking at the Resume before the interview, for me, has two objectives:

  1. Placing "Resume Scores" on key success criteria: It's hard to score some of the more non-technical scores. Again, this will be flawed, and we have to approach it as information that has noise in it. But it also has a signal.
  2. Knowing what points I want to press harder on in the interview. What are the criteria I cannot evaluate at all? What could be some good cues for behavioral questions?

If before your interview the Candidate has done other tests or interviews, you must check them. The checklist is the following:

  • Resume
  • Past Interviews
  • Technical test
  • Github
  • Personal Site / Portfolio
  • Social network profiles

It takes time to comb all this and prepare a pre-interview scorecard, but it's worth it. So take time to prepare, save at least 30 minutes before each interview to study the Candidate's profile.

Information Gathering: The Interview

Remember, we're gathering information, so we should go for what's more critical first. There are two main tools I like to use to extract information during an interview:

  1. Evidence-based Behavioral questions
  2. Problem-solving

I usually use both in an interview, roughly saving 2/3 of the time for problem-solving and 1/3 to behavioral questions. But this varies with how much information I think I can extract from each method.

Behavioral questions:

"There are essentially two ways to do behavioral-based interviews-one in which you ask the candidate to describe situations in which they exhibited certain behaviors (eg tell me about a time in which you dealt with a conflict at work- how did you handle it); the other asks the candidate what they would do in certain situations (eg if this situation happened what would you do)." -- Kathryn Keeton

Build a set of questions that enable you to assess what you want from the Candidate's experience. An example:

"Tell me about a time in which you had to work with multiple different teams to solve a large Product or Engineering problem. What was your role in this dynamic, and how did the project benefit from having you?"

Lead the conversation to the main things you want to assess. Manage time to make sure you get to all the points you want to measure.

Problem Solving

"Problem Solving" consists of giving the Candidate a problem to solve during the interview. This should be something similar to the issues they will need to solve when they're in the role inside the company.

The problem we choose should be related to the information we want to gather more of. If we've managed to assemble enough on all criteria, we can add more trust to the information we gathered. We can only cover so much in the evidence-based questions, so seeing the person in action is an additional time to allow them to provide the data we need.

Different problems explore different areas, and I will never be exhaustive here with them. It's an infinite combination of things. But just being aware of what you're searching for will help a lot. Some general things that are good to remember:

  1. Do NOT provide all information upfront. See how the Candidate gathers contextual information. See if they know how to figure out what's important;
  2. Take notes on how the person breaks down the steps to solve the problem. There are different ways to reach a satisfactory solution, but some will invite more collaboration than others. Be on the watch for how well you could collaborate with this person if they approached the problem in their work. See how well the person makes trade-offs and communicates them;
  3. Try to at least once add a requirement that would force a change to the best solution. Does the Candidate go back and rethink critical parts of their answer, or do they patch something on top of their solutions?

REMEMBER, we are gathering information. We can follow up and lead the conversation towards the scores we're less sure of.

Making a decision

If we've done the last steps with care, we're now in possession of a lot of data that will enable us to make a better decision. Our next step is giving this person an extra score on one more criterion: "Close your eyes and imagine working with this person. How well will this person do at this job?"

"But isn't this just intuition?" You might rightfully ask. And you're right it is, but after you have taken the time to gather information and categorize that information across our measure of success. This tends to lower the likeability bias. In general, when you're interviewing, you are a person that has a good grasp of what success looks like, so we should heavily value your intuition and experience as well. The process isn't here to eliminate it. It is here to help you build it based on evidence rather than on a halo effect. So this intuition we should trust slightly more than general intuition. 😊

Compare the scores–all key criteria + extra–of different candidates and make your decision based on that and that alone. Don't tamper with the data. Communicate your rationale about the decision to other people in the process and use their input to re-evaluate your assessment. And that's it. Hopefully, doing this will reduce some of the bias and help you get the best people!


Summing up:

  1. Commit to reducing bias. Be curious about every person and give them a fair chance.
  2. Allocate at least half an hour for your interview preparation
  3. Understand the key success criteria for the position, ensuring they are significantly different from each other
  4. Make a list of evidence-based questions that will enable you to accurately assess each of the criteria, with a clear understanding of what a high scoring answer (e.g., 10) would be, a moderate scoring answer (e.g., 6), and a low scoring answer (e.g., 1)
  5. In the interview, methodically ask questions to gather information on each of the points. It is of utmost importance that you collect enough data to grade ALL of the scores. Doing so is very hard, but we should make our best effort.
  6. Add up the scores to determine the most highly rated Candidate.
  7. Avoid putting a candidate's likeability ahead of the criteria you have decided the most important for success in the job.