Prejudice. How do we overcome it?

This article reviews 418 experiments (conducted between 2007-2019) on reducing prejudice (based on race, ability, nationality, and other factors).

It builds on an earlier (2009) review by the same authors.

This new review highlights seven interventions that seem to work (many experiments testing the intervention done, showing sizeable effects on reducing prejudice):

1. Antibias, multicultural, and moral education (e.g. teaching health care providers strategies to combat mental illness stigma). Still, more field experiments may be needed.

2. Cognitive and emotional training. Including cognitive conditioning (e.g. pairing representations of stigmatised groups with positive stimuli); cognitive-emotional interventions (e.g. training to reduce negative emotions when encountering situations related to outgroups); and perspective taking (e.g. taking the "perspective of the elderly by climbing up a staircase while wearing restrictive knee pads").

3. Peer influence, discussion, and dialogue (e.g. using peers as messengers for a cause). Experiments show slightly smaller effects than other interventions.

4. Social categorization (e.g. modifying existing boundaries between groups, by asking research participants to list characteristics shared between groups).

5. Entertainment (e.g. incorporating "educational messages about prejudice into an entertaining storyline of a soap opera or film").

6. Face-to-face contact (e.g. assigning soldiers of minority group and majority group to room with one another during basic training).

7. Extended and imagined contact (e.g. reading children stories in which able-bodied children befriend children with disabilities). But larger experiments show much smaller effects (in reducing prejudice).

The review also highlights two interventions with mixed evidence / need more evidence:

1. Diversity training (programmes that "self-identify as a diversity training" or related e.g. sensitivity / cultural competence). Among the studies, only a small number of experiments conducted. Further, experiments with larger sample sizes find quite small effects (i.e. diversity training reduces prejudice very slightly).

2. Value consistency and self-worth (e.g. reminding someone of their (or their group’s) egalitarian preferences). Many lab and online experiments have shown sizeable effects (in reducing prejudice). But no field experiments have been done.

The review emphasises three overall limitations in the body of research (the 418 experiments it reviewed).

1. The majority (76%) of the studies only evaluate light touch interventions, for which the long-term impact remains unclear.

2. Publication bias may be exaggerating effects. The review found that experiments with larger sample sizes systematically showed larger effects (in reducing prejudice), compared to experiments with smaller sample sizes (which suggests small sample size experiments showing smaller effects have not been published).

Landmark studies often find limited effects (in reducing prejudice).

https://pubmed.ncbi.nlm.nih.gov/32928061/

This article argues that we can reduce exclusionary attitudes through interpersonal conversation. The researchers conducted three randomised field experiments:

1. The first experiment studied attitudes towards unauthorised immigrants. Canvassers (persons making the intervention) and voters (experiment participants) exchanged narratives about: (i) a previous personal experience with immigrants, and (ii) “a time when someone showed them compassion when they really needed it.”

2. The second experiment was about transphobia. In one condition, canvassers and voters exchanged narratives (like in experiment one), AND were shown a video narrative (where an unknown transgender woman shared her experience). In another condition, voters were ONLY shown the video.

3. The third experiment was also about transphobia. Here, voters were called by phone (instead of engaged in-person), and the caller and voter exchanged narratives.
   

Before the interventions (the exchange / video / call), seemingly unrelated surveys were carried out to recruit participants and collect baseline data.After the interventions, seemingly unrelated surveys were again conducted to see if the voters attitudes were less exclusionary (compared to the control groups). These surveys were conducted 1 week, 1 month, and 3-6 months after the intervention.

The results? In short, all interventions had durable effects (lasting up to the 3-6 months) in reducing exclusionary attitudes. The voters were more likely to support inclusive policies and express less prejudiced beliefs.

Other points:

• There seems two be three requirements for these interventions to work. The conversations must be: (i) interpersonal, (ii) non-judgmental, and (ii) involve exchanging narratives.

• The interventions' effect sizes may seem small (though statistically significant). This could be due to lack of experience of the canvassers (many of whom were non-profit staff and volunteers). (Also, sometimes small margins matter in policies/politics, and change can be incremental)

The experiments were all US-based, which has a particular culture of canvassing. The same interventions may not be suitable for other contexts.

https://www.cambridge.org/core/journals/american-political-science-review/article/reducing-exclusionary-attitudes-through-interpersonal-conversation-evidence-from-three-field-experiments/4AA5B97806A4CAFBAB0651F5DAD8F223

Here's a systematic review article from 2009, which highlights six interventions that work to reduce prejudice.

The article identified 985 existing studies on reducing prejudice, and synthesised the main findings. It's a little outdated, but is a comprehensive overview of "what works", as of 2009 (there is also an updated 2021 review by the same authors).

The six interventions that seem to work (supported by experimental evidence from field and laboratory):

1. Cooperative learning (e.g. students collectively assembling a puzzle)

2. Entertainment (e.g. children reading stories about "contact between children from different groups")

3. Discussion and peer influence (e.g. acquaintances "speaking out against biased jokes")

4. Contact (e.g. going on a mixed-ethnicity camping trip)

5. Value consistency and self-worth (e.g. telling someone – who sees themselves as valuing equality – that people who value equality are more likely to support the civil rights movement)

6. Cultural training (e.g. "teaching people how to interpret behaviours of different cultural and/or racial groups")

(To note: for the interventions that work, many may work only under certain conditions. E.g. contact interventions may only work under conditions of "equal status, shared goals, cooperation, and sanction by authority".)

Next, interventions that may work (supported mostly by experimental laboratory evidence – but lack experimental field evidence):

• Social categorisation, consisting:

  1. Decategorisation (e.g. instructing people to focus on individual – not group – identity)

  2. Recategorisation (e.g. encouraging people to see themselves and other groups as belonging to one overarching group – and not separate groups)

  3. Crossed categorisation (e.g. making two groups aware they also belong to a third common group)

  4. Integration (e.g. highlighting to two groups that they also belong to one overarching group)

• Cognitive training (e.g. stereotype retraining for children)

Finally, interventions that may or may not work (lacks robust experimental evidence one way or the other):

• Diversity training (e.g. sessions involving videos, role-plays, discussions)

• Multicultural, antibias, moral education (e.g. standalone instructional lectures or workshops)

• Sensitivity, cultural competence trainings in health and law enforcement professions

• Conflict resolution (for ordinary citizens, as opposed to elite negotiators)

A main conclusion of this paper is that more research – especially field experiments – needs to be done to better understand what works.

https://pubmed.ncbi.nlm.nih.gov/18851685/