Who gets an interview? Who gets the assignment that will matter at promotion time? Whose objection gets a serious hearing before the team commits to a decision? Workplace inclusion takes shape in choices like these. When bias awareness becomes the centerpiece of an inclusion strategy, the decisions that distribute opportunities can go on much as before. Behavioral science gives employers a way to see how those choices are made and test what happens when the process changes.

What does inclusion look like at work?

Diversity describes who is in the room. In Shore and colleagues’ (2011) model, inclusion depends on whether people feel they belong to the group and are valued for what makes them distinct. Someone who earns acceptance by adopting the group’s ways of working experiences what the authors call assimilation. They call it differentiation when a team values someone’s particular expertise while leaving that person out of decisions that affect their work.

Hiring records tell you who joined the company. How people get information, what work they’re trusted with, and what happens when they disagree tell you something about the experience of working there. A manager can give everyone time to speak. What the manager does with a disagreement tells people whether speaking up was worth the risk. That’s where inclusion intersects with psychological safety.

Where do hiring disparities begin?

In a field experiment overseen by Dares, the French labor ministry’s research and statistics directorate, researchers sent comparable fictitious applications to real job openings. In the 2021 report, applications with names that signaled a North African (Maghrebi) background were, on average, 31.5% less likely to receive a callback than applications with names that signaled a French background. The disparity appeared before an interview could take place. For an HR team looking for a place to start, the finding points to the criteria recruiters use to decide who gets a first conversation.

How reviewers see the candidates matters, too. In a lab experiment, Bohnet, van Geen, and Bazerman (2016) found that evaluators gave more weight to individual performance when they compared candidates side by side. Gender stereotypes carried more weight when evaluators considered candidates one at a time. A hiring panel can apply that insight by agreeing on what counts as evidence for the role, recording each reviewer’s assessment independently, and then comparing candidates against the same criteria. The panel has something concrete to discuss when someone says a candidate just seems like a better fit.

Who gets the work that leads to a promotion?

By the time a promotion committee meets, managers have already made decisions that shape what each employee can point to. In experiments by Babcock et al. (2017), women were more likely to volunteer for tasks that rarely help people get promoted. They were also more likely to be asked and to say yes. The researchers found evidence that expectations help sustain the pattern. People tend to ask whoever they think will agree.

Imagine a consulting team preparing for a client presentation. One project lead puts the deck together, coordinates reviews, and writes up the decisions. Her colleague presents the recommendations to the client’s executive team. Both do work the project needs, and he comes away with a chance to show he can defend a recommendation in front of senior leaders. If those roles fall to the same people on each project, the promotion committee ends up comparing employees who had different chances to demonstrate what they could do. Looking back at who was considered, who was asked, and who got the assignment helps managers find where those opportunities diverged.

Even equivalent performance ratings can lead to different rewards. At the U.S. company Castilla (2015) studied, women, employees from racial and ethnic minority groups, and employees born outside the United States received lower performance-based rewards than U.S.-born white men with comparable ratings and job characteristics. The gap narrowed after the company made pay decisions more transparent and required decision-makers to account for their choices. The before-and-after study followed one company. It draws attention to the decision that turns a performance rating into a dollar amount.

When does bias training change a decision?

Chang et al. (2019) followed more than 3,000 employees at one company after a brief online diversity training. The training shifted measured attitudes in groups that had been relatively less supportive of women at work. The behavioral changes the researchers observed occurred primarily in groups whose attitudes were already strongly supportive. A post-training quiz or satisfaction survey would have missed whether employees did anything differently in the weeks that followed.

Arslan et al. (2025) tested another approach at a telecommunications and engineering company. Hiring managers assigned to the diversity training were invited to watch a short video on diversity and candidate assessment just before selecting people for interviews. Compared with the control group, the share of job openings with at least one woman on the interview shortlist rose by 3 percentage points. The estimated effect on hiring a woman was statistically uncertain. For HR, the distinction matters. A change in screening can widen access to interviews while leaving the outcome of the final hiring decision open.

How can employers apply behavioral science to inclusion?

“Give employees equitable access to high-visibility assignments” is a goal. To turn it into a practice, specify who will do what and when. Before assigning a presentation to the executive team, for instance, the project lead reviews the list of eligible employees. The COM-B model developed by Michie, van Stralen, and West (2011) helps explain why that review might be skipped. Does the lead know what experience the assignment calls for? Is the list available before the decision? When a project starts in a hurry, who comes to mind first?

UNREST’s MEET method moves from diagnosing the problem to testing a change. In Model, the team reconstructs recent staffing decisions to see how project leads made their choices. In Engineer, it designs a change around the obstacle it found. If the list typically arrives after the assignment has been made, the team can put the review on the kickoff agenda before anyone commits. In Experiment, it pilots the change across several assignments, measures against a baseline, and uses a comparable team or a staggered rollout when feasible. Transfer begins when the results justify expanding the practice. The organization builds the review into its staffing process and checks whether project leads still use it once pilot support ends.

A reorganization makes these choices easy to see. When managers have to staff new roles quickly, they’re likely to call people whose work they already know. The point at which new teams are formed gives leaders a chance to widen the pool of people they consider. The same approach applies when a hiring panel selects interviewees, a committee decides on promotions, or a manager responds to dissent in a meeting.

Decision Practice to test What to look for
Select candidates for interviews The recruiter assesses each application against the criteria agreed on before screening and records the basis for the decision. Share of job openings where the recruiter recorded the basis for each screening decision before sending interview invitations.
Assign high-visibility work The project lead reviews the list of eligible employees before choosing who will present to senior leadership. Share of assignments preceded by that review, and who received the assignments among eligible employees.
Decide on a promotion The committee ties each promotion decision to examples of the employee’s work and to the responsibilities of the new role. In a sample of decisions, look for concrete examples that support the stated rationale against the published criteria.
Make a decision in a meeting The meeting lead asks for a different view before making the decision and explains how they considered it. Which objections came up before the decision and how the meeting lead responded.

How do you measure whether inclusion is improving?

Start with the practice itself. How often did project leads review eligible employees before filling a high-visibility assignment? Count every assignment that falls under the new rule. Opening the list after the decision doesn’t count as a prior review. This behavioral indicator tells you whether managers changed how they made the decision. Who actually gets the assignments tells you whether those opportunities are still going to the same people.

Employees are the source for a different kind of evidence. A person can land a high-visibility assignment and still be left out of the conversations where decisions take shape. Ask employees whether they feel they belong and whether the team values what they bring to the work. Track those responses alongside the new practice and the distribution of assignments. Comparing the pilot with similar teams that still follow the usual process can help estimate what the intervention changed, especially if the teams were on similar trajectories before it began.

Which decision should you change first?

If the applicant pool has broadened and the interview pool has barely moved, look at how recruiters screen applications. If hiring has opened up and promotions are still concentrated, look upstream at who gets the work that qualifies people to move up. UNREST’s work on diversity and inclusion starts with gaps like these and helps the people who make the decisions choose a change to test. Leadership’s first move is to give the person responsible for that process the authority to change how it works.

Frequently asked questions

Does fair hiring require an identical assessment for every candidate?

Candidates for the same role should be evaluated on job-related criteria. A recruiter can adapt the format while assessing the same skill. If the exercise tests how well candidates analyze a file, a screen-reader-compatible version gives a candidate with low vision access to the same task.

How can you tell when a promotion criterion is too vague?

Ask two reviewers what they observed that led them to say someone “lacks executive presence.” If one points to leading a cross-functional project and the other to confidence in meetings, they’re judging different things. The committee can go back to the responsibilities of the role and agree on what evidence it expects for each.

Can employers measure workplace diversity in France?

Yes, subject to conditions. France’s data protection authority, the CNIL (2025), says participation in workplace diversity surveys must be voluntary and employees must be informed. It recommends favoring anonymous surveys and limiting the data collected. A trusted third party can administer the survey so the employer has no access to individual responses.

References

Dares, Institut des politiques publiques, & ISM Corum. (2021). Discrimination à l’embauche des personnes d’origine supposée maghrébine : quels enseignements d’une grande étude par testing ? Dares Analyses, 67.

Shore, L. M., Randel, A. E., Chung, B. G., Dean, M. A., Holcombe Ehrhart, K., & Singh, G. (2011). Inclusion and Diversity in Work Groups: A Review and Model for Future Research. Journal of Management, 37(4), 1262–1289. 10.1177/0149206310385943.

Bohnet, I., van Geen, A., & Bazerman, M. (2016). When Performance Trumps Gender Bias: Joint vs. Separate Evaluation. Management Science, 62(5), 1225–1234. 10.1287/mnsc.2015.2186.

Babcock, L., Recalde, M. P., Vesterlund, L., & Weingart, L. (2017). Gender Differences in Accepting and Receiving Requests for Tasks with Low Promotability. American Economic Review, 107(3), 714–747. 10.1257/aer.20141734.

Castilla, E. J. (2015). Accounting for the Gap: A Firm Study Manipulating Organizational Accountability and Transparency in Pay Decisions. Organization Science, 26(2), 311–333. 10.1287/orsc.2014.0950.

Chang, E. H., Milkman, K. L., Gromet, D. M., Rebele, R. W., Massey, C., Duckworth, A. L., & Grant, A. M. (2019). The Mixed Effects of Online Diversity Training. PNAS, 116(16), 7778–7783. 10.1073/pnas.1816076116.

Arslan, C., Chang, E. H., Chilazi, S., Bohnet, I., & Hauser, O. P. (2025). Behaviorally Designed Training Leads to More Diverse Hiring. Science, 387(6732), 364–366. 10.1126/science.ads5258.

Michie, S., van Stralen, M. M., & West, R. (2011). The Behaviour Change Wheel: A New Method for Characterising and Designing Behaviour Change Interventions. Implementation Science, 6, 42. 10.1186/1748-5908-6-42.

CNIL. (2025). Enquêtes de mesure de la diversité au travail : la CNIL publie ses recommandations.

About the author

Maxime Rabéchault is the founder and president of UNREST, a change enablement consultancy. He applies behavioral science to help organizations change how they work.