Mitigating Biases
by Jim Bruce
… When Hiring Staff
Last week’s Tuesday Reading focused on cognitive biases, forces that can influence an “individual’s personal construction of his or her social reality.” This personal construction and not the objective input received from your senses may dictate your behavior in the social world. As a result, cognitive biases may sometimes lead to perceptual distortion, inaccurate judgment, or illogical interpretation. In a NeuroLeadership Institute paper, Breaking Bias,1 Leberman, Rock, and Cox have a helpful, more concise definition: “Cognitive biases are the unconscious drivers that influence our judgment and decision making.”
We ended last week’s essay by referring to Heidi Grant Halvorson and David Rock’s paper “Beyond Bias.”2 Halvorson and Rock note that our biases can be helpful and adaptive. They often enable people to make quick efficient judgments and decisions with minimal cognitive effort. However, they can also blind a person to new information or inhibit them from considering valuable options.
In their paper Halvorson and Rock grouped the very long list of common biases into five categories based upon their underlying cognitive nature – similarity, expedience, experience, distance, and safety. (Their research group, the NeuroLeadership Institute, has named this model SEEDSTM.) Each of the five categories can be described by a short description that helps us remember the model and what that category is all about:
Similarity – “People like me are better than others.” We tend to favor other individuals who are similar to ourselves.
Expedience – “If it feels right, it must be true.” We jump to conclusions without seeing all the evidence.
Experience – “My perceptions are accurate.” We believe our own perspectives most accurately reflect reality.
Distance – “Near is stronger than far.” We see what is near in space, time, or perceived ownership as more valuable.
Safety – “Bad is stronger than good.” We significantly over emphasize what is bad compared to what is good.
Today’s essay examines how these cognitive biases may effect the several stages in the hiring process. And, since most organizations today need and seek a more diverse and inclusive workforce, we want to understand the impact our biases have on who we hire. Since we cannot simply “outlaw” biases, we do need to also ask how their impact can be mitigated.
One way to think about the hiring process is to exam each of its elements:
- Codifying the position description and how the open position will be publicized.
- Reviewing resumes and selecting candidates to interview.
- Interviewing.
- Choosing the finalist.
- On-boarding.
Codifying the Positing
We begin by asking exactly what is the position we are trying to fill. And, too often, we are in a hurry – think expedience bias – and we take the last similar job description, make a few changes, quickly write a descriptive paragraph, and think we’re done. If we do this, we short change ourselves in at least two ways. First, the hiring manager needs to think carefully about the position’s “must haves” – the behaviors, skills, knowledge or other qualities that applicants need to demonstrate to be seriously considered for the position – as well as the “need not haves.” This information needs to be included in the job description and the posting’s materials as well as be available for use by the resume reviewers as the key factors in deciding who to be interviewed. (And they will be needed by the interviewers as well.)
And, second, we need to pay careful attention to the words we use to describe the position. Rebecca Knight, quoting Professor Francesca Gino, Harvard Business School professor, notes that “even subtle word choices can have a strong impact on the application pool.”3 Knight goes on to say that “research4 shows that masculine language, including adjectives like ‘competitive’ and ‘determined,’ results in women ‘perceiving that they would not belong in the work environment.’ On the other hand, words like ‘collaborative’ and ‘cooperative,’ tend to draw more women than men.”
So, be careful to clearly indicate what skills and aptitudes you want the applicants to have and not use gender biased language.
Reviewing Resumes and Selecting Candidates to Interview
You’ve posted your position description and the resumes begin to arrive. Your expedience bias is still running at full throttle so you are anxious to review the resumes and choose which individuals to interview.
Several words of caution are appropriate here. First, expedience biases impact our thinking when we are in a hurry to meet tight deadlines and when we have other priorities we are addressing at the same time. These biases make being objective very hard and can distort our first impressions, assumptions, and expectations about each of the individuals who have applied for the position and particularly those who will go on to be interviewed. In addition, the authors of “Select Better”5 suggest that some factors on a given resume may lead a reviewer to give either too much (halo effects) or too little (horn effects) weight to that piece of information.
These expedience biases can be significantly mitigated by two steps that can be taken: First, make use of the “must have” list mentioned earlier as the basis used to objectively and explicitly score each of the resumes. For example, you might use the list of “must haves” as a scorecard that each reviewer uses when reviewing a resume. If each reviewer gives a numeric score, say one to five on each of the requirements, then it will be easy to arrive at a numerical score that will permit you to easily rank the candidates.
And, second, as the resumes arrive, you might have someone who will not be part of the interview and decision process remove the identities from the resumes, for example, replacing names with unique numbers. Knight4 tells us that “you need to ‘level the playing field’ by taking steps so that the reviewers are focused on the candidate’s specific qualifications and characteristics, and not on their demographics. Iris Bohnet, director of the Women and Public Policy Program at Harvard’s Kennedy School said in What Works,6 “The fact is that Latisha and Jamal do not get the same number of callbacks as Emily and Greg.” Professor Gino suggests that a blind systematic process for reviewing applications and resumes “will help you improve your chances of including the most relevant candidates in your interview pool… It is easy for bias to trickle in when such a process is not decided a priori.”
Once all the resumes have been reviewed, you select the candidates who most closely match the requirements for the position and move on to interviews.
Interviews
The interview stage of hiring is likely to be the most vulnerable to biases as it relies on human interactions which are dynamic, unpredictable, and constantly changing. All the interviews can be conducted by a single individual or by multiple interviewers in a single session or in a session for each interviewer. And, they can be structured or unstructured. Each approach has it’s advantages and disadvantages. For example, research has shown that biases exert the strongest effects on behavior in the context of unstructured interviews. And, having multiple resume readers and candidate interviewers can give you a more comprehensive view of each candidate.
No matter the structure of the interviews, the candidate and interviewer are both constantly changing in response to the other’s questions, comments, and body motions. As a result, there are many opportunities for biases to exert strong effects that lead to nonverbal behaviors signaling disinterest, discomfort, or threat through tone of voice, the lack of eye contact, and your body posture.
Similarity biases and expedience biases are the biases most likely to surface and have negative impacts on the interviewer’s judgement in the interview. We tend to like people who seem to be like us or who share common interests. The result is automatic favoritism (similarity bias) which can lead us to ignore more objective measures of the candidates fit for the position. The interviewer can mitigate his or her behavior by deliberately seeking to identify common ground with each candidate. If you can find this common ground, you will treat the individual as an in-group member, someone to be perceived in a positive light. You can also look for differences when you read the resume or interview a person who is otherwise similar to you. If you take both of these steps you will have a more balanced view of all of the candidates.
Expedience biases also show up in the interview process but in different ways from how they have shown up earlier. The first is in what’s called “confirmation bias” where we seek information that confirms and fail to look for information that disconfirms our pre-existing beliefs. Getting information from colleagues before the interview is one way in which confirmation bias can undermine your decision making and lead you (unconsciously) to interpret a candidate’s responses to confirm your colleague’s opinions. Another way confirmation bias enters the interview is by leading the interviewer to ask questions that yield evidence to confirm an impression that you already have.
The other way expediency biases enter is through what’s called “recency bias,” our tendency to have better recall for, and be more influenced by information we acquired most recently. Sip,5 et al suggest four ground rules for interviewers to follow within a structured process to reduce this bias:
- Create a “no spoiler” policy. Avoid getting opinions from anyone before interviewing a candidate.
- Know the questions you will ask each and every candidate in advance. And, ask each candidate the same questions in the same order.
- Develop a way of objectively ranking or assessing answers in a formalized manner.
- Capture your assessment of a candidate immediately after the interview.
Choosing a candidate
You are now nearing the end of the process. You have reduced bias in the resume review and the interview stages and you have a short list of people to choose from. You are ready to choose. Research suggests that three forms of bias may appear as you take this final step – Experience, Distance, and Safety biases.
Experience bias leads us to believe that we see things as they really are, that there is nothing else to see. Unfortunately, this may not be the case. Research suggests that almost 60% of the hiring managers have discovered something false or misleading on resumes. Experience bias is mitigated by gaining multiple perspectives. Do reference checks where you ask detailed questions and check specific facts. Use the “snowball” approach by asking each reference who else might you talk with about the candidate.
Distance bias has to do with the way that distance in either time or space influences our decisions. This bias can disadvantage a candidate who needs time to develop in the position (even though you think he or she would be the better candidate in the long run) from someone who can “hit the ground running.” Distance bias also comes into play in terms of a candidate having to relocate geographically and in terms of when he or she would become available. You minimize distance bias by asking yourself what really matters and focusing on that when you decide.
Safety bias has to do with the human tendency to overweigh the negative. We tend to assign greater weight to the negative than to the positive, to loses than to gains. This also requires an accurate assessment of the risk. Unfortunately, the safety bias leads us to avoid, ignore, or draw negative conclusions from anything that makes a candidate look “non-traditional” or “risky.” Unfortunately, this bias can lead to lack of diversity.
You may be able to mitigate this bias by assuming that you have made a decision, the candidate has accepted, and has arrived. Do you regret your decision? Taking this post-hiring perspective will let you be more objective in your analysis. Another mitigation approach is to think of your decision as advice that you would give to a close colleague. What advice would you give if the choice did not affect you?
There is, of course, a lot more to the hiring process that I have outlined here. For example, I haven’t written much about who reviews the resumes – a team or a single individual – or who interviews the candidates – again, a team or a single individual such as the hiring manager. I haven’t talked about on-boarding, a critical process for a new hire’s success. What I hope I have done is to provide some solid foundation blocks that you can use to create a robust, low-bias process for hiring.
I trust that you’ll make this week a great one for you and your team.
. . . . jim
Jim Bruce is a Senior Fellow and Executive Coach at MOR Associates, and Professor of Electrical Engineering, Emeritus, and CIO, Emeritus, at the Massachusetts Institute of Technology, Cambridge, MA.
References:
- Matthew Lieberman, David Rock, and Christine Cox, Breaking Bias, NeuroLeadership Journal, Volume 5, May 2014.
- Heidi Grant Halvorson and David Rock, Beyond Bias, strategy+business, July 2015.
- Rebecca Knight, 7 Practical Ways to Reduce Bias in Your Hiring Process, Harvard Business Review, June 12, 2017.
- Danielle Gaucher, Justin Friesen, and Aaron Kay, Evidence That Gendered Wording in Job Advertisements Exists and Sustains Gender Inequality, Journal of Personality and Social Psychology, 2011, Volume 101, No. 1, pages 109-128.
- Kamila Sip, Jay VanBavel, Tessa West, Josh Davis, David Rock, and Heidi Grant, Select Better: How Managers Can Reduce Bias in Hiring, NeuroLeadership Journal, Volume 7, April 2017.
- Iris Bohnet, What Works: Gender Equality by Design, Harvard University Press, March 2016.
- November 2024 (3)
- October 2024 (5)
- September 2024 (4)
- August 2024 (4)
- July 2024 (5)
- June 2024 (4)
- May 2024 (4)
- April 2024 (5)
- March 2024 (4)
- February 2024 (4)
- January 2024 (5)
- December 2023 (3)
- November 2023 (4)
- October 2023 (5)
- September 2023 (4)
- August 2023 (4)
- July 2023 (4)
- June 2023 (4)
- May 2023 (5)
- April 2023 (4)
- March 2023 (1)
- January 2023 (4)
- December 2022 (3)
- November 2022 (5)
- October 2022 (4)
- September 2022 (4)
- August 2022 (5)
- July 2022 (4)
- June 2022 (4)
- May 2022 (5)
- April 2022 (4)
- March 2022 (5)
- February 2022 (4)
- January 2022 (4)
- December 2021 (3)
- November 2021 (4)
- October 2021 (3)
- September 2021 (4)
- August 2021 (4)
- July 2021 (4)
- June 2021 (5)
- May 2021 (4)
- April 2021 (4)
- March 2021 (5)
- February 2021 (4)
- January 2021 (4)
- December 2020 (4)
- November 2020 (4)
- October 2020 (6)
- September 2020 (5)
- August 2020 (4)
- July 2020 (7)
- June 2020 (7)
- May 2020 (5)
- April 2020 (4)
- March 2020 (5)
- February 2020 (4)
- January 2020 (4)
- December 2019 (2)
- November 2019 (4)
- October 2019 (4)
- September 2019 (3)
- August 2019 (3)
- July 2019 (2)
- June 2019 (4)
- May 2019 (3)
- April 2019 (5)
- March 2019 (4)
- February 2019 (3)
- January 2019 (5)
- December 2018 (2)
- November 2018 (4)
- October 2018 (5)
- September 2018 (3)
- August 2018 (3)
- July 2018 (4)
- June 2018 (4)
- May 2018 (5)
- April 2018 (4)
- March 2018 (5)
- February 2018 (5)
- January 2018 (3)
- December 2017 (3)
- November 2017 (4)
- October 2017 (5)
- September 2017 (3)
- August 2017 (5)
- July 2017 (3)
- June 2017 (8)
- May 2017 (5)
- April 2017 (4)
- March 2017 (4)
- February 2017 (4)
- January 2017 (4)
- December 2016 (2)
- November 2016 (7)
- October 2016 (5)
- September 2016 (8)
- August 2016 (5)
- July 2016 (4)
- June 2016 (12)
- May 2016 (5)
- April 2016 (4)
- March 2016 (7)
- February 2016 (4)
- January 2016 (10)
- December 2015 (4)
- November 2015 (6)
- October 2015 (4)
- September 2015 (7)
- August 2015 (5)
- July 2015 (6)
- June 2015 (12)
- May 2015 (4)
- April 2015 (6)
- March 2015 (10)
- February 2015 (4)
- January 2015 (4)
- December 2014 (3)
- November 2014 (5)
- October 2014 (4)
- September 2014 (6)
- August 2014 (4)
- July 2014 (4)
- June 2014 (4)
- May 2014 (5)
- April 2014 (5)
- March 2014 (5)
- February 2014 (4)
- January 2014 (5)
- December 2013 (5)
- November 2013 (5)
- October 2013 (10)
- September 2013 (4)
- August 2013 (5)
- July 2013 (8)
- June 2013 (6)
- May 2013 (4)
- April 2013 (5)
- March 2013 (4)
- February 2013 (4)
- January 2013 (5)
- December 2012 (3)
- November 2012 (4)
- October 2012 (5)
- September 2012 (4)
- August 2012 (4)
- July 2012 (5)
- June 2012 (4)
- May 2012 (5)
- April 2012 (4)
- March 2012 (4)
- February 2012 (4)
- January 2012 (4)
- December 2011 (3)
- November 2011 (5)
- October 2011 (4)
- September 2011 (4)
- August 2011 (4)
- July 2011 (4)
- June 2011 (5)
- May 2011 (5)
- April 2011 (3)
- March 2011 (4)
- February 2011 (4)
- January 2011 (4)
- December 2010 (3)
- November 2010 (4)
- October 2010 (4)
- September 2010 (3)
- August 2010 (5)
- July 2010 (4)
- June 2010 (5)
- May 2010 (4)
- April 2010 (3)
- March 2010 (2)
- February 2010 (4)
- January 2010 (4)
- December 2009 (4)
- November 2009 (4)
- October 2009 (4)
- September 2009 (4)
- August 2009 (3)
- July 2009 (3)
- June 2009 (3)
- May 2009 (4)
- April 2009 (4)
- March 2009 (2)
- February 2009 (3)
- January 2009 (3)
- December 2008 (3)
- November 2008 (3)
- October 2008 (3)
- August 2008 (3)
- July 2008 (4)
- May 2008 (2)
- April 2008 (2)
- March 2008 (2)
- February 2008 (1)
- January 2008 (1)
- December 2007 (3)
- November 2007 (3)
- October 2007 (3)
- September 2007 (1)
- August 2007 (2)
- July 2007 (4)
- June 2007 (2)
- May 2007 (3)
- April 2007 (1)
- March 2007 (2)
- February 2007 (2)
- January 2007 (3)
- December 2006 (1)
- November 2006 (1)
- October 2006 (1)
- September 2006 (3)
- August 2006 (1)
- June 2006 (2)
- April 2006 (1)
- March 2006 (1)
- February 2006 (1)
- January 2006 (1)
- December 2005 (1)
- November 2005 (2)
- October 2005 (1)
- August 2005 (1)
- July 2005 (1)
- April 2005 (2)
- March 2005 (4)
- February 2005 (2)
- December 2004 (1)