Creating An Unbiased Interview Process
(This is a draft that is still evolving based on research and feedback)
One of the biggest issues in the tech world is the lack of diversity. The problem is not just pipeline but the issues of getting people through the interviewing process to an offer. One of the areas that can be improved is how candidates are evaluated.
The goal of this document is to create a template and guidelines for designing an interview process that brings out the strengths in a candidate and aids in filtering out biases to create a more diverse culture in the workplace.
Pre-Interview Process
Creating a standardized process for each candidate helps ensures a fair interview. That begins with initially planning out the interviews in a structured way and giving each candidate the same interview.
1) Reducing Biases In Initial Screening
First, we should remove any information that might cause a bias about a who to interview for a position. Details such as name, past schools, past employers, and zip code should be removed. When deciding on who to give the opportunity to interview, focus only on information that is of merit value.
During the actual interview, information such as their zip code, where they went to school should potentially still be kept hidden.
2) Rank Your Required Skills For A Candidate
When interviewing a candidate, there is often a list of required skills for the position. While some skills are essential for the job, others can be learned on the job or nice to have. Each candidate will have different strengths and weakness that relate to the required skills. To help make equal assessments, consider the following:
- Rank the skills in the job description from most to least important.
- Decide what skills are nice to have vs. absolutely required
- Create a rubric, ie scale of 1–10, to evaluate how strong a candidate is in each skill
Later, this rubric can be used to assess a candidates strengths in relation to the position and help remove biases when ranking against other candidates.
3) Design Good Interview Questions
Develop a standard set of questions derived from the required skills for the job. The questions should dive deep into the candidate’s experiences. It is important to be cognizant of how the questions are phrased and the results they produce.
Examples of Bad vs Good Questions
Bad: Tell us about your skill in XYZ at job ABC
Better: At what jobs did you have experience in skill XYZ
Great: What positions, both professional and non-professionally, did you utilize skill XYZ and what were some of the greatest challenges you had to overcome
Per the examples above, often interviewers will ask questions that do not bring out the candidates strength. It is important to realize that candidates have other areas of professional experience ranging from side-projects to unexplored expertise from previous jobs that can be missed by bad questioning. Create questions in a way that you might be surprised by the answer.
4) Personal Questions
Focusing on professional skills in relation to a position does not tell a comprehensive story about a candidate. A candidate can often have hidden value, assets, and latent potential they can bring to the organization. Be open to asking them questions such as:
- What are a few things you are most proud of professionally that we might not have discussed already? It’s ok if we have previously discussed it.
- What is something you are proud of outside of work? For example, did you run a Marathon?
- What motivates you to come to work every morning?
- If offered the position, what do you feel you can bring to the organization?
Competency Testing
Almost every job has a test to assess how knowledgeable a candidate is in a given area. Tests, while applicable as a qualifier, can also filter great candidates if are administered incorrectly. Also not all tests have a correlation to job performance. Below are different perspectives to consider when giving a competency test.
Give Multiple Short Tests
One test should not be used to qualify a candidate, nor should passing every test always be required. Wording, phrasing, and understanding of the problem can influence how well someone performs. Take for example these two questions given to computer engineers:
Example 1: You are given an array of n integers and a number k. Determine whether there is a pair of elements in the array that sums to exactly k. For example, given the array [1, 3, 7] and k = 8, the answer is “yes,” but given k = 6 the answer is “no.”
Example 2: A group of n kids sit in a circle. Given a number m, start counting from one of the kids around the circle. The kid who counter is the number m goes out of the circle. Continue from the next kid until there is only one kid left. Write a function to find who is the last kid.
Which example is easier for you to understand, even if you do not fully interpret the problem? On interviews, testing different ways a person thinks can yield different results from the same candidate.
Take Home vs. Timed Tests
Candidates perform differently in different environments. Some candidates fail under timed tests from the pressure while others excel. Some candidates may get lazy on take-home test and not complete them. Consider how you are testing the candidates and if the test relates to the work environment and the candidates’ ability to perform at the job.
Avoid Unreasonably Long Tests
Be respectful of the candidates time. Candidates have lives, schedules, families and full-time jobs — and the time taken to take a test should be reflective of that. For example, requiring a single mother to take an unpaid 10-hour is not reasonable and good candidates will walk from tests that do not respect their time.
Theoretical Knowledge vs Working Knowledge
Theoretical knowledge is referring to the understanding of how an element works in theory. For example, someone might ask how many muscles are in the human arm or what is the run time of a hashmap. Working knowledge is the ability to execute a specific task in a real environment.
The two are not necessarily correlated in assessing a person’s ability to perform a task. For example, you are hiring a musician. Some people can “play by ear”, replay a song on an instrument just by hearing it and have no clue how to read a note. Other people that can compose songs and not play an instrument. If a candidate is not required to read notes, then a filtering process focused on reading notes will miss great musicians.
Test That Reflects The Interviewer Strengths
The last area of testing to consider how well the test relates to job performance and not be based on the strengths of the interviewer. For example, if an interviewer is great at remembering definitions, the test might end up being a quiz of definitions, erroneously filtering out candidates that are bad at definitions. Therefore test should be compromised by multiple interviewers who have different strengths and weaknesses.
Another approach to consider is asking a candidate what areas they are strong and weak in. Taking those areas into consideration, give them a test that allows them to show their strengths if their weaknesses are not something that is going to cause a problem on the job.
Post Interview Assessment
Either during or immediately after an interview, the candidate’s notes should be recorded on a universal rubric. Delays in writing this information down can cause important facts to be missed.
Multiple Interviewers
If an extra staff member is available, considering having multiple people sit in an interview. After the interview is over, have them submit their reports independently of each other without discussing the candidate. This practice can create different viewpoints of a candidate.
Consider 3rd Party Blind Review
3rd Party Blind Reviews are when a candidate assessment minus any personal identification information such name, gender or schooling, is given to someone else to review and compare candidates. This practice can help with weeding out affinity bias towards a candidate.
Culture Fit vs Culture Add
When thinking about how a candidate will do within the organization when working with others, the way the interviewer frames the “culture fit” question in their minds can influence hiring decisions. Take for example:
Examples:
Will candidate X get along with employee Y vs. can candidate X compliment employee Y in tasks
Candidate X reminds me of Bob vs. candidate X has strengths and expertise Bob does not have
Making a conscious effort to think of candidates in terms of value-add and not in likeness can influence how we see people. When thinking about diversity, it is essential to consider the spectrum it encompasses beyond race. Think of differences such as:
1) Age
2) Race/Culture
3) Gender
4) Sexual Identity
5) Education
6) Social Economic Status
7) Education
Understanding your organization and what diversity means from multiple perspectives can help hire people who are different and valuable. Diverse environments help with more expansive problem solving and the fighting of groupthink. We must take the time to dig deeper and find people strengths if we truly want to find the best candidates.
