Study: How race, clothes and demeanor influence police officer behavior

A research team compared how behavior and visible characteristics influence whether officers escalate or de-escalate street confrontations


Reprinted with permission from Force Science News 320

It’s all about a subject’s demeanor – not about race, ethnicity, or attire – when encounters with police escalate to violence, according to a new study from Washington State U.

Activist groups and mainstream media, of course, tend to insist otherwise. But a research team that conducted the first controlled laboratory study comparing how behavior and visible characteristics influence whether officers escalate or de-escalate street confrontations has found that appearance bias is not a dominant factor.

The way subjects act is what makes the difference. (Photo/PoliceOne)
The way subjects act is what makes the difference. (Photo/PoliceOne)

The way subjects act is what makes the difference.

“These findings offer important new insight into how fairly officers interact with people during routine encounters that have the potential to turn violent, and what this means for perceptions of police legitimacy, procedural justice, and allegations of racial bias,” writes the study’s lead author, Dr. Lois James.

She’s an assistant professor and researcher who works with the university’s Sleep and Performance Research Center on a number of police-related studies that have been reported in Force Science News.

While her latest findings are encouraging for law enforcement’s public image, her team uncovered some troubling evidence that she describes as “rather shocking.” Justifiably so!

TEST SCENARIOS

Participants in her new study were 50 officers randomly selected from a list of qualified volunteers from the patrol division of a mid-size metropolitan PD. All but a handful were white males, with an average of nearly 16 years on the job.

Armed with training-modified Glock 22s, they were exposed to a series of video scenarios, depicting police-citizen interactions in five situations: a vehicle stop, a welfare check, an investigation of “suspicious circumstances,” a disturbance of the peace and a community meeting.

Six versions of each scenario were filmed, so that the same action could feature key role players who differed as to race (white, black, or Hispanic) and attire (“business” dress, consisting of suit or slacks with a button-down shirt, or “street” clothes, including jeans, sneakers, and hoodie).

“Gender, age, body type of the suspect, and the environment for the interaction were held constant,” James said.

Half the scenarios within each racial/ethnic category depicted individuals who were confrontational from the onset, while the other half featured non-confrontational individuals, she says. Confrontational subjects, while not behaving criminally or displaying any pre-assault indicators, “acted with hostility, antagonism, contempt, or belligerence...being rude, disrespectful, and mocking.” Non-confrontational individuals were “friendly, polite, and respectful.”

James points out: “It is important to note that this variable was not dictated by [initial] actions of the officer. That behavior was apparent from the very start of the encounter, regardless of how the officer approached the scenario or initiated contact.”

Officers were told to “respond as they would in a routine police-citizen encounter,” interacting with people on the screen, trying to “resolve problems peaceably,” and de-escalating “where possible.”

DEPENDENT BRANCHING

Depending on what the participating officers did during the encounter, each scenario was “branched” in one of these ways:

  1. To a “positive track,” where the subject ultimately cooperates and ends up “visibly pleased” or at least “neutral” regarding the outcome, or
  2. To a “negative track,” which was initiated if an officer failed to display a professional attitude or dialogue, including disrespecting, patronizing, or insulting the subject, or pointed a gun at him/her “unnecessarily.”

Once the action branched negative, the subject “became visibly upset” or angry. Then the officer could initiate a “repair track” by trying to de-escalate these reactions.

If he failed to attempt de-escalation, however, the action escalated to the “deadly” level. The subject “became enraged, rapidly presented a weapon, and started shooting,” James explains.

Research findings

James’ team sought to answer two research questions:

  1. Did officers differ in how they treated on-screen individuals based on race/ethnicity, attire, or demeanor?
  2. If the negative track was initiated, did officers’ de-escalation attempts differ based on race/ethnicity, attire, and demeanor of the person they were dealing with?

Here’s what the researchers found:

  • “[O]fficers did not treat white, black, or Hispanic suspects significantly differently,” James writes. “[R]oughly equivalent percentages of scenarios featuring white, black, and Hispanic individuals resulted in cooperative, neutral, and deadly outcomes [indicating] that officers were not influenced by individuals’ race/ethnicity during their interactions.”
  • “[O]fficers did not treat street-dressed individuals differently [than] business-dressed individuals. [A]ttire did not predict the likelihood of a cooperative outcome...a neutral outcome...or a deadly outcome.”
  • “The sole significant result was demeanor.... [S]cenarios featuring confrontational individuals were significantly less likely to result in a cooperative outcome...and significantly more likely to result in deadly outcomes.... [O]fficers treated people better and avoided escalation when the on-screen individuals were friendly, respectful, and polite....[Officers] responded similarly to confrontational individuals regardless of their race/ethnicity or how they were dressed.”
  • Numerically, officers did attempt de-escalation (as evidenced by activation of the repair track) less frequently in scenarios with black individuals and with subjects in street garb. “But the difference was not statistically significant,” James says.
  • Again, “the sole significant variable was demeanor; officers were significantly more likely to attempt de-escalation when the individual was non-confrontational.”

 “Collectively, these results suggest that individual characteristics did not influence how officers treated people in the simulator,” James writes. “[B]eing confrontational was the sole significant predictor of a deadly outcome.”

MORE TO DO

The research findings suggest that police were impartial regarding race and attire, which speaks well for law enforcement in these troubled times. Nonetheless, James does point out a major dark element among the study’s discoveries.

Even in scenarios in which individuals were not rude or disrespectful at the outset, officers often “acted in ways that did not lend to the peaceful resolution of encounters,” she writes.

Indeed, 52% of the scenarios that began with non-confrontational subjects ended up with deadly force (compared to 63% of scenarios with hostile individuals). In the final split-second that officers are confronted with a lethal weapon, James notes, shooting may well be fully justified. But “when considering the ebbs and flows of the entire dynamic encounter” that led to that point, “the appropriateness of police actions [along the way] is less clear.”

Given her study’s set-up, where “any attempt to de-escalate the encounter would have resulted in its peaceful resolution,” the 52% could be “classified as unnecessary force,” James writes – a “rather shocking” result that shows “we still have much work to do.”

James’ study, “Testing the impact of citizen characteristics and demeanor on police officer behavior in potentially violent encounters,” appears in Policing: An International Journal of Police Strategies & Management and can be accessed in full for a fee by clicking here. Her colleagues in this research were Dr. Stephen James and Dr. Bryan Vila of Washington State U.

Dr. Lois James can be reached at lois_james@wsu.edu.

Recommended for you

Join the discussion

Copyright © 2019 PoliceOne.com. All rights reserved.