RACIAL BIAS AND IN-GROUP BIAS: Evidence from virtual reality courtrooms

A new study investigates whether the decision about the guilt or innocence of an individual, who is being accused of a crime, is race-blind. The research by Samantha Bielen, Wim Marneffe and Naci Mocan, to be presented at the 2020 European Economic Association Meetings, also asks whether the extent of the punishment depends on race.

Do these decisions depend on whether or not the defendant and the evaluator are of the same race? If so, are these influences stronger in case of white or minority evaluators? If there exist race effects on these decisions, are they impacted by whether the evaluator believes terrorism is a major problem in the country? 

To answer these questions, the authors shot 3D Virtual Reality (VR) videos of six criminal trials in Belgium, prosecuted by actual prosecutors and defended by actual defence attorneys in an actual courtroom.

Only the defendants in the courtroom are actors. The prosecutors and the defence attorneys are given the case files one week before the shooting of the trials so that they could do the background work and prepare their case and their defence. The prosecutors and the defence attorneys presented their cases orally in the courtroom, as they would normally do.

The VR technology made it possible to replace white defendants in the courtroom with individuals who have Middle Eastern or North African descent. This allowed the researchers to alter only the race of the defendants in these trials, holding all activity in the courtroom constant, including every word spoken by the prosecutor and the defence attorney, and all the body language in the courtroom.

A short clip of two videos can be seen here http://proficient.ninja/splitscreen/. This study is the first one that uses 3D Virtual Reality technology that uses actual people, rather than computer-animated scenes. Scrolling down at the link http://proficient.ninja/uhasselt/ allows one to observe scenes from all six trials. 

A total of 153 Master’s degree law students and undergraduate and Master’s degree economics students are randomly assigned to watch, with VR headsets, these trials. They made decisions on conviction as well as prison sentence and fine in accordance with the guidelines provided by the relevant law.

Half of the evaluators saw a white defendant, the other half saw a minority defendant in each case. The evaluators didn’t know the names of the defendants; thus names could not be used as signals of minority status. Background information obtained from evaluators allowed the researchers to identify their cultural heritage to determine if evaluators are part of the minority groups in Belgium. 

The study finds that both white and minority evaluators treat white defendants favourably relative to minority defendants during the conviction stage. These effects produce overall racial bias against minorities in conviction decision: minority defendants are 16% more likely to get convicted in comparison to white defendants even though they are tried for the same exact case and even though everything that went on during the trial is the same. 

In the sentencing phase defendants receive favourable treatment from evaluators of their own race. That is, convicted defendants receive shorter prison sentences and lower fines if the evaluator is of the same race as the defendant. Partly because of the fact that the numerical majority of evaluators are white, this pattern of behaviour leads to prison terms that are 1.3 months longer for minorities and to fines that are double that of assigned to whites. 

Evaluators’ concerns about terrorism do not affect the racial biases in these decisions. The researchers repeat the experiment with 36 practicing attorneys and find the same pattern of racial bias (in-group bias could not be analysed with the attorneys because all were white). Adding a small number of prosecutors and judges to the sample of attorneys generates similar results as those obtained from the attorney sample. 


Racial Bias and In-group Bias in Virtual Reality Courtrooms 

Samantha Bielen, Hasselt University 

Wim Marneffe, Hasselt University 

Naci Mocan, Louisiana State University, NBER and IZA 

Paper to be presented at the 2020 European Economic Association Meetings. 

For a copy of the paper email to mocan@lsu.ed