A review of research on prejudice reduction was recently conducted by Elizabeth Levy Paluck of Princeton University, with colleagues from Hebrew University and Columbia University. While Paluck and Green’s original review on this subject focused on theoretical approaches, this latest review focuses on experimental studies, where the effects of interventions are properly tested.  

By taking this approach, the study looks set to become a pioneering piece of research on how to successfully reduce workplace prejudice and discrimination. Based on a meta-analysis* of 418 experiments, it is the most comprehensive and up-to-date study of its kind. The focus on experiments is welcome, as it shows the impacts that can genuinely be attributed to interventions.  

The authors highlight several large-scale studies in real-world settings. These should encourage future researchers to collect meaningful behavioural evidence with transparency and data openness. The research offers a number of important insights, which we summarise below. 

1. Face-to-face intergroup contact can reduce discriminatory behaviour 

Several studies found that intergroup contact can be effective in reducing prejudice, particularly among different faiths and castes. Members of Indian and Iraqi sports teams who worked in cross-caste or interfaith teams showed higher levels of trust and friendship and were more likely to train with their teammates months later or nominate them for an award, compared with those in the control group. These behavioural changes occurred even if attitudes were not affected. Another study showed similar findings in education, where Christian and Muslim applicants were randomly assigned to computer classes. Those who experienced intergroup contact were much less likely to discriminate against outgroups. 

2. Conversation and ‘perspective taking’ help people understand others 

In a series of field studies, voters were visited by volunteers for political causes – in this case, support for transgender and immigrant individuals. Through coaxing people into taking the perspectives of these groups, the researchers found their attitudes became less prejudiced over a period of months. In the case of transgender support, these conversations were effective irrespective of whether the volunteers were themselves transgender. 

3. Diversity training can be more effective for marginalised groups 

One trial in a multinational corporation showed that online diversity training led to reduced prejudice in specific workplace decisions, such as who gets nominated for an award or who is offered to be mentored. The main behavioural effects were found not among white men, but rather among women and minority groups. This study provides an excellent template for other training studies and leaves open the question of whether in-person training may have different effects. 

4. ‘Calling out’ unacceptable behaviour can reduce prejudice 

Social sanctioning, or calling people out for their prejudice and discrimination, can be an effective method of curbing their behaviour – but is this dependent on who is making that call? To test this, one study randomly assigned Twitter bots to reproach and track male users who publicly and repeatedly used a racial slur to harass other users. The research found that sanctioning only worked when it was seen to be from a perceived ‘credible’ member of the ingroup – that is, bots set up as accounts of white people with many followers. Even this effect weakened over two months. 

5. Emerging insights from smaller studies 

Paluck et al.’s review also describes an array of other potential interventions that look promising from early research. These include:  

• Cognitive and emotional training. Individuals can be trained to use thinking and emotion-regulation strategies to combat their personal implicit or explicit prejudice. Typically, this is what unconscious bias training (UBT) focuses on.
• Value consistency and self-worth interventions. This draws on the human desire to maintain a consistent and positive image of the self and the group, for example by reminding people of their preferences for equality and encouraging introspection about their beliefs and prejudices. 
• Peer influence and dialogue interventions. People who share identities can strongly influence one another’s ideas on the correct way of behaving. Some interventions use peers as messengers for a cause, such as leaders spreading the message of greater openness to other groups, while others focus on social norms, encouraging individuals to align their beliefs with their peers.
• Social categorisation. It’s important to encourage participants to rethink group boundaries and identify with outgroups. This often involves using approaches to highlight the similarities between the ingroup and the outgroup, as well as challenging negative stereotypes and perceptions of the outgroup.
• Entertainment and the arts. Telling stories about marginalised groups, made by and for their members, has been an important development in the last decade in film, music and television. This can be used to overcome human tendencies to counterargue messages or resist persuasion – it is hoped people will ‘get lost’ in the power of the narrative and forget their desire to critique. 
• Face-to-face contact. Interaction between different groups can work in a similar way to perspective-taking, as members begin to perceive common interests between each other. 
• Extended and imagined contact with outgroups. Fictional friends or characters in books and films can be used when face-to-face contact is deemed difficult or impossible to test whether the character’s contact with an outgroup member reduces prejudice.   

These types of interventions all show some promise, but there are caveats to note. Many of these smaller studies were conducted quickly and inexpensively, so the longer-term impact is unclear. There are also signs of publication bias, which means that studies with positive results are more likely to be published. The authors thus suggest caution when it comes to ‘small studies with optimistic conclusions’. Indeed, while many studies with small samples report very positive findings, those with larger samples tend to be more reserved. There are also notable weaknesses in the evidence base – for example, the authors found insufficient evidence to support popular interventions centred on ‘implicit association’ or UBTs.  

6. Key takeaways

This cutting-edge review of the best science provides excellent guidance on the best bets to help reduce prejudice and foster inclusion and diversity in the workplace. Our top five takeaways are:  

• Encourage diverse groups to work together, to talk to each other and to gain an appreciation of each other’s circumstances.  
• Integrating different groups will allow them to establish both professional and personal relationships to harness their potential.  
• Focus on empathy, rather than asking people to confront their biases, to reduce prejudicial attitudes and behaviours among ingroup members. 
• Recognise that diversity training may be more beneficial for marginalised groups. 
• Calling people out on their prejudice and discrimination can be effective for curbing unwanted behaviour.  

* What is a meta-analysis? Systematic reviews collate and appraise all the relevant scientific studies on a particular topic. Meta-analyses crunch the statistical data in these studies to give an overall ‘effect size’. For example, they tell us what the best research evidence says about the impact of an intervention. Systematic reviews and meta-analyses are thus crucial tools for understanding the body of knowledge.