Implicit Bias

Implicit bias: a definition

Implicit bias (التحيز الضمني), also called unconscious bias, is a tendency of which the person who holds it is not aware, and which affects their judgement of other people according to the group those people belong to rather than according to what they have delivered.

The definition sets two possible bases for a judgement side by side. One is what a person has actually delivered: their work, their results, the answers they gave. The other is the group to which the person belongs. Implicit bias is present when the second basis enters a judgement that is meant to rest on the first, and when the person making the judgement does not know that it has.

Two features mark implicit bias. The first is that it is not intended. The person who holds it has not decided to judge anyone by their group, and has not chosen the tendency that does so. The second is that the person would reject it if it were put to them openly. Asked directly whether a candidate’s group should count for or against that candidate, they would say that it should not, and they would mean it.

Where implicit bias works in the hiring cycle

Implicit bias can enter the hiring cycle at the points below, and each gives it a different way into the decision:

• CV screening. The candidate’s name, university, year of graduation or place of residence can enter the decision without ever being given as a reason for it.
• The interview. The interviewer can favour the candidate who resembles the interviewer in background or in interests, and the preference is later explained as the candidate being the best fit for the team.
• Promotion. Nominations can go to the people who are most visible to management rather than to the people who have achieved the most.

At the screening stage, the name, the university, the year of graduation and the place of residence are details that sit on the CV beside the candidate’s experience and qualifications. None of them is entered in the record as the reason a CV was kept or set aside, and yet each of them can affect which CVs are kept and which are set aside. The decision is recorded on other grounds, while one of these details has shaped it.

At the interview stage, the resemblance in question is a resemblance in background or in interests, and it is a resemblance to the interviewer. A candidate who shares something with the person conducting the interview can be favoured for that reason. The explanation given afterwards does not mention the resemblance. It describes the candidate as the best fit for the team, and in that description the preference has become an assessment.

At the promotion stage, the comparison lies between visibility and achievement. An employee who is more visible to management is not, for that reason, an employee who has achieved more, and a nomination that follows visibility can pass over an employee who has achieved more but been seen less.

Why intention is not enough against implicit bias

Intention is not enough because implicit bias is, by definition, outside awareness.

Someone who says “I assess fairly” is telling the truth about their intention, and saying nothing about their behaviour. The statement is sincere. They do intend to assess fairly, and they would reject any suggestion that a candidate’s group should play a part. What the statement cannot report is whether a tendency of which they are unaware has entered their assessment, because a person can only report on what they are aware of.

For this reason implicit bias is not addressed by statements, but by procedures that reduce the room left for unconstrained judgement. The less of the decision that is left to judgement nobody has constrained, the less room the tendency has in which to operate.

What reduces implicit bias in practice

The measures below reduce implicit bias by procedure. Each one narrows the room for unconstrained judgement at a particular point:

• Blind screening. It hides the name, the photograph and the university during the first screening of applications.
• Criteria written before the candidates are seen. Criteria written after the candidates have been seen are easy to tailor to them.
• A structured interview. It puts the same questions, in the same order, to every candidate, with a defined rating scale.
• More than one interviewer, with different backgrounds.
• A review of the results in aggregate. It asks whether a pattern recurs among the people who are excluded. A pattern reveals what an individual case does not.

Blind screening acts at the first screening, where the name and the university were named above as details that can enter the decision. A detail the screener does not see cannot affect what the screener decides.

Writing the criteria before the candidates are seen fixes what the decision is meant to rest on before anyone is in a position to shape it around a particular person. Once the candidates are known, criteria can be written to match the candidate who is already preferred.

The structured interview fixes the questions, their order and the rating scale. Every candidate answers the same questions in the same sequence, and every answer is scored against the same defined scale. What is asked, in what order and against which scale is settled before the interview begins, and less is left to unconstrained judgement during it.

Having more than one interviewer, with different backgrounds, means that the decision does not rest on the judgement of one person, and that the people judging do not share a single background.

The review in aggregate looks at the results of decisions taken together rather than at one decision at a time. A single exclusion can be explained on the grounds recorded for it. A pattern in who is excluded, recurring across the decisions reviewed, reveals what no single case can show.

How implicit bias differs from discrimination

Discrimination is a behaviour that can be observed in a decision. Implicit bias is a tendency that may lead to that behaviour without being intended.

For that reason the two are addressed with different tools. Discrimination is addressed through policy and accountability, which apply to the behaviour that can be observed in a decision. Implicit bias is addressed through the design of the procedure, which limits the room the tendency has to affect the decision in the first place.

This is an explanation of the concept, not legal advice.