Overcoming Workplace Inequality: Interventions to Reduce Social Category Biases and Increase Inclusion in the Workplace

The intervention is an online training that can be completed in under one hour in one sitting.

The intervention follows the best practices outlined by “wise psychological interventions” and uses a number of established psychological principles to maximize its effectiveness as a means of overcoming bias and addressing inequality. These principles include: (1) establishing the norm that people have biases, thus making it easier for people to acknowledge their own biases, (2) providing personalized feedback about individuals’ own biases to motivate them to change their behavior, and (3) providing research-backed strategies for overcoming bias and the opportunity to practice using these strategies, as (a) deliberate practice can make new behaviors stick, and (b) having people convince themselves of the usefulness of these strategies is more effective than being persuaded by others. Access to the actual intervention materials (with partner-specific information redacted) will be available upon request.

In the treatment conditions, participants are provided with information about stereotyping and social category biases. This information includes a summary of research findings and accounts of others’ personal experiences with stereotypes and biases. This content is delivered via video and audio presentations from faculty members and Ph.D.-level experts in relevant academic disciplines (e.g., neuroscience and decision sciences), infographics, and text (often accompanied by figures or other visual elements). In the control condition, participants are presented with information about what predicts success in teams via similar content mediums as in the treatment conditions.

The content of the training is divided into five sections. In the treatment conditions, the first section introduces the basic psychological processes (pattern recognition, categorization) that underlie stereotyping, explains what stereotypes are (with examples), and discusses how stereotypes can result in unconscious bias and undesired outcomes in the workplace. The aim of this section is to introduce the topic of bias in the workplace in a non-threatening way in order to reduce defensive reactions that can make people unwilling to acknowledge that they are biased. Specifically, we emphasize that (1) stereotypes are a function of generally adaptive psychological processes, (2) stereotypes can develop without conscious awareness, (3) everyone has stereotypes, and (4) having stereotypes does not mean that one is consciously hostile towards a group (or “prejudiced). In the control condition, this section will similarly provide a general introduction to the topic of what predicts success in teams.

The second section in the treatment conditions presents data from the lab, field, and real world that illustrate how bias manifests in the workplace. For example, participants learn that even when applicants have identical credentials, they are viewed as less competent if they are female (rather than male) in the gender-specific treatment or obese (rather than non-obese) in the general treatment. In addition to providing scientific evidence of the effect of stereotypes on women and other groups that face bias in the workplace, we complement these findings by providing first-hand accounts that illustrate how these biases affect individuals. The goal of this section is to combine the strength of scientific evidence with compelling anecdotes in order to persuade participants that stereotypes and bias have real effects on workplace outcomes, as argument strength and compelling anecdotes have been shown to be influential to people’s attitudes and judgments. In the control condition, participants will learn about what factors contribute to team success in the workplace (such as collective intelligence).

In the third section, in order to capitalize on the finding that providing personalized feedback about one’s biases can motivate relevant behavior change, participants in the treatment conditions are provided with personalized feedback about their own potential biases by completing the Gender-Career Implicit Associations Test. In addition to learning how they performed on the test, participants will then be presented with a first-hand account of a female faculty member’s reaction to learning that she held implicit associations favoring men and an opportunity to provide their own reactions to the test. In the control condition, participants will receive personalized feedback about how they score on a relevant assessment for success in teams (the Active Empathetic Learning Scale).

The fourth section in the treatment conditions teaches participants about strategies that they can use to overcome stereotypes and biases in the workplace. We have developed a memorable framework for introducing four research-backed strategies, which we call Blinding, Individuating, Articulating, and Substituting (BIAS). The BIAS framework specifically introduces the following strategies for reducing bias:
• Blinding: Removing identifying information from evaluations
• Individuating: Learning more about a person to reduce reliance on stereotypes
• Articulating: Setting clear expectations so people know they can safely break from stereotypes
• Substituting: Checking to see if you would make the same decision about a person from a different identity group
Participants learn about these strategies by reading descriptions of the strategies and the research supporting them. Participants also evaluate case studies to determine which strategy could have worked best to overcome stereotypes and promote a more inclusive workplace in particular situations. In the control condition, participants learn about strategies to increase psychological safety in teams (encouraging learning, acknowledging fallibility, and reflecting interest; EAR).

The fifth section presents participants in both the treatment and control conditions with realistic workplace scenarios and asks them to share their ideas about how to promote inclusivity and overcome bias (responses will be coded to evaluate the extent to which the strategies provided would achieve those ends). Importantly, it is only in the treatment conditions in which this exercise provides an opportunity for participants to practice applying the BIAS strategies to relevant workplace situations, as participants in the control condition are presented with strategies not relevant to bias. The aim of this section in the treatment conditions is for participants to engage in deliberate practice to help make these new behaviors habitual and to convince themselves of the usefulness of these strategies by applying them to these scenarios. In the control condition, although participants have not been provided with strategies for overcoming bias, asking them about how to promote inclusivity should seem relevant given the previous sections’ focus on team success in the workplace.

2017-02-02

2017-06-16

1) Behavioral measures of bias-reducing, inclusive behaviors
2) Engagement with the issue of workplace bias
3) Attitudinal change regarding bias

Please see outcomes explanation document.

Participants will be randomized into one of three conditions: gender-specific, general, or control.

Participants were invited by their employer to take part in a training program with the goal of helping to create a more inclusive work environment. Participants were randomly assigned to one of our treatment conditions or a control condition. The primary treatment conditions vary whether participants were presented with information that is only about one form of social category bias (e.g., gender bias) or is about a number of different forms of social category bias (e.g., gender and racial bias). The control condition presented filler content that is relevant to the workplace (such as videos of talks about what predicts the success of individuals and groups in the workplace, and personalized feedback on a psychological scale) that matches the length and feel of the intervention in the treatment conditions. It takes participants about 60 minutes during one sitting to complete the intervention (or control) training and to complete initial dependent measures assessing their views on stereotypes, bias, and discrimination (including open-ended responses to workplace scenarios in which they share their ideas about how to promote inclusivity and overcome bias).

Recruitment for this study began in early February, 2017, and continued through mid-March. Approximately 10,000 employees were invited to participate in the study and over 2,000 enrolled. At the end of the training, participants filled out attitudinal and survey measures. For the 12 weeks following the training, participants were texted up to once a week as a measure of engagement. Roughly one week after the recruitment period ended, participants were sent the voluntary follow-up survey measure from the research team. In early April, participants were sent an email from the partner organization regarding the creation of a new mentorship program and were encouraged to sign-up to connect with colleagues. At the end of April, participants were sent an email from the partner organization asking them to recognize and nominate colleagues for an award. In mid-May, participants will be asked to speak with a new male or female hire.

Randomization is completed online via Qualtrics

Individual participants are the unit of randomization

No

n/a

We recruited as many participants as possible during our recruitment window of February 2, 2017 to March 16, 2017. We ended up recruiting over 2,000 participants.

The participants were randomly assigned to one of three conditions. Roughly one third was assigned to the control condition, one third to the gender-specific condition, and one third to the general condition.

Outcomes Explanation and Analysis Plan