New Study: Anger Sets The Stage For Seeing Threats Where None Exist

Print Friendly, PDF & Email

If you are angry when you confront a suspect, are you more likely to mistake a cell phone or other nonthreatening object in his hand for a gun?

Recent findings from university-based research suggest that indeed is the case. Through a series of time-pressured experiments, a behavioral science team at Northeastern University in Boston discovered, among other things, that:

  • Angry participants were significantly “more likely to misidentify an object as a gun” than were test subjects experiencing “happy” emotions or even other “negative” emotions, such as sadness or disgust;
  • Anger changed participants’ expectations, so that when angry they had a higher anticipation that they would “encounter more guns than neutral objects,” compared to people who were not angry;
  • Angry subjects “set a much lower threshold for saying that a target is holding a gun” —in other words, they “required much less information before they are willing to claim a target individual is threatening.”

“This appears to be well done, empirically solid work that supports the psychoanalytic concept that angry people project their anger onto others and see a more threatening world around them than actually exists,” Dr. Bill Lewinski, executive director of the Force Science Institute, told Force Science News.

However, Lewinski cautions, “this study should not be interpreted too broadly when it comes to its applicability to officer-involved shootings.” More on his concerns and reservations in a moment.

The research was led by Dr. David DeSteno, associate professor of psychology at Northeastern, and Jolie Baumann, a fourth-year graduate student who, like DeSteno, is interested in how emotions impact cognition and behavior. A full report of their study (“Emotion Guided Threat Detection: Expecting Guns Where There Are None”), prepared for the Journal of Personality and Social Psychology, Click here to read it free of charge.

As background for their study, the researchers cite 2 infamous cases of threat misinterpretation: the confrontation in which New York City police fatally shot Amadou Diallo, who they thought was reaching for a gun when he was actually reaching for his wallet and ID, and the incident in Baghdad where an Army helicopter crew killed a wire service photographer when they thought his camera with a telephoto lens was a weapon.

Threat assessment “can frequently occur in the presence of heightened emotional states,” DeSteno and Baumann note. They sought to determine whether strong emotions might “constitute a primary influence on threat detection,” and, in fact, might “push individuals’ judgments toward [mistakenly] favoring the existence of a threat, thereby leading them to actively aggress against others who may not have posed an actual danger.”


Using male and female undergraduates as test subjects, DeSteno and Baumann constructed this core approach:

Believing that they were participating in a memory test, the students were asked to write a 7-minute account of their daily routine (considered an emotionally neutral assignment), of an event that made they happy, or of an event that made them angry. They were not to detail any circumstances directly involving violence and/or guns. After the writing, the researchers used a questionnaire to verify that the subjects were emotionally neutral or sufficiently aroused to be classified as happy or angry. Roughly an equal number of students ended up in each category.

Seated in front of computers in individual cubicles, volunteers in each of the 3 groups were then briefly shown random background images of various urban and suburban scenes, including a park, a train station, and a street corner. At certain unexpected points, 10 different “target individuals” appeared one at a time “embedded” in the background for just a flash—750 milliseconds.

“[E]ach appeared twice holding a gun and twice holding a neutral object of similar size and color (i.e., camera, wallet, soda can, cell phone), resulting in a total of 40 targets,” DeSteno and Baumann explain.

In the fraction of a second that each individual was visible, the test subjects had to assess whether the person was holding a gun or a neutral object and push a corresponding computer key. No feedback about accuracy was given.

All the target individuals were of “apparent white European ancestry,” to diminish the risk of racial bias or stereotyping influencing the responses. And the test subjects had simply to identify the object being held as a gun or something “neutral;” they did not have to decide whether to shoot the target. “[W]e wanted the primary decision to be one of identification as opposed to one of aggressive behavior,” the researchers write.


Participants in all 3 groups made mistakes in identifying what was held, although the error rate was lowest among those in a “neutral” emotional condition. But the nature of the misinterpretations is revealing.

The unemotional participants as a group more often identified a gun as a nonthreatening object than they identified a neutral object as a gun. Happy subjects overall were virtually balanced in their erroneous identifications, miscalling a gun and a neutral object with about the same frequency.

But angry subjects were far more likely to mistakenly identify a harmless object as a gun than vice versa, and they made this error at a significantly higher rate than subjects in the other 2 categories.

To the researchers, this response makes “great sense” in terms of human behavior. They point out that emotions are “informational and motivational.” Anger is “an emotional state that signals the presence of potentially violent or aggressive threats.” To innately “favor one’s own safety and survival” in that situation, it is more useful “to mistakenly harm a nonthreatening (unarmed) person than to risk being harmed or even killed by a threatening person; survival-wise, it is better to be safe than sorry when one’s emotions are signaling the presence of potential dangers.”

In other words, an angry person, feeling an exaggerated emotional need for self-protection, is biased toward finding threats where none exist, and for good reason from a preservation standpoint.

To test the uniqueness of anger as an error stimulus, the researchers altered the experiment with a new pool of volunteers to include disgust and sadness as negative emotional conditions, along with anger and a neutral state of mind.

Again, “angry participants made significantly more errors in calling a neutral object a gun than vice versa,” whereas the other groups showed a more balanced mix in their misidentifications.


In a third experiment, DeSteno and Baumann sought to determine whether anger causes subjects to “expect to encounter more guns than neutral objects” in their environment, which in turn would lead them “to adopt a more lenient decision threshold” for claiming a gun is actually present.

Here, they had fresh participants “identify” stimuli as neutral items or guns without really showing them any object at all. The subjects were told that they would be “subliminally exposed to images containing either guns or wallets” and they were to guess which stimulus was shown. In reality they really were exposed to a 50-millisecond flash of black and white dots clustered in the hands of target individuals, with the image “very quickly covered” with a gray oval as a mask “to give the impression” they had actually seen an object.

It seems likely that “wallets are a more common occurrence than guns in daily experience” and the collection of dots flashed on the screen was arguably more similar in shape to a wallet than a gun, the researchers note. Yet “the presence of anger resulted in a significant increase in guessing that the stimulus behind the gray mask was a gun…. In essence, it appears that anger leads participants to form heightened expectancies of encountering violence-related threats relative to nonthreats.”


In a final experiment, DeSteno and Baumann explored the extent to which angry participants realized their mistakes. In this test, angry subjects were allowed to reflect on their decision immediately after making each gun-or-neutral-object call. They were not allowed to see the image any longer, but they were not restricted on the time allowed for second-guessing afterward.

Under these conditions, the researchers found that as a group the angry participants revised their identification decisions to the extent that their error rate was essentially the same as that of subjects with neutral emotions. This convinced DeSteno and Baumann that the angry subjects were aware almost immediately of having made mistakes under severe time pressure and would correct them given the opportunity to do so.


Even “brief amounts of additional processing time” can drastically reduce errors, the researchers note, but in the police and military worlds “allowing such additional time is not always an option, as some decisions must be made very rapidly,” the researchers write.

“In such instances, training people to be aware of their emotions as well as the potential influence of those emotions on assessments of threat may represent one possibility for enhancing accuracy.” They cite a study by other behavioral scientists suggesting that people who “typically attend to their feelings” are less likely to allow their emotions to influence judgment of risk than those who “do not typically attend to their feelings.” That is, awareness of emotion may moderate its influence.

“Accordingly,” write DeSteno and Baumann, “programs designed to improve the accuracy of individuals who must make rapid decisions (e.g., police officers) might find the most success in eliminating emotion-based biases by training individuals to be aware of their emotions and the sources of those emotions.”

Research on that premise is necessary, they conclude, along with additional investigation of the role of perceptual distortions and fear is the misidentification of threat.


DeSteno’s and Baumann’s current study, which did not involve LEOs as test subjects, “is going to be accepted by the academic community and subsequently by society as a whole as an explanation for officer behavior in controversial ‘mistake-of-fact’ shootings,” Lewinski predicts. And that concerns him, he says, because it is “dangerous to automatically assume that an officer is angry at the time he or she misidentifies a benign object as a threat.”

Mistake-of-fact shootings, he explains, “often occur in highly ambiguous circumstances with apparently defiant and resistant subjects who are engaging in dynamic, rapidly evolving behavior that appears highly threatening.”

Lewinski has personally interviewed well over 1,000 officers who have survived shooting encounters. His investigations have convinced him that rather than feeling angry in mistaken shootings (as well as in other types of lethal encounters), “officers usually appear to be emotionally intense, very concerned, and even fearful about a situation that appears to be rapidly deteriorating and over which they have little control. They know that if they don’t act quickly, they could die.

“Having moderately angry civilian research subjects judge an object in a flash picture on a computer screen is interesting and may have some useful training considerations. But it does not duplicate the complex decision processes of an officer engaged in a real-world encounter.

“Unfortunately, there are many people who will infer that it does and the law enforcement community needs to be aware of the pitfalls of attaching too much relevancy to research at this early stage of development.”

Leave a Reply