(But whose perception is being distorted?)


By Prof. William Harmening

Washington University in St. Louis


The concept of “perceptual distortion” has been offered up to the law enforcement community by pro-police researchers as a psychological get out of jail free card. It is a nice little tool used in nearly every questionable deadly force case to alter reality in favor of the officer’s narrative, and it is typically presented to judge and jury very eloquently by a pro-police expert witness with little or no background in the behavioral sciences. It allows an officer to testify that in the heat of the moment they failed to see something that was right in front of them, or actually did see something that was not even there. And with a pro-police expert witness ready to pull out the latest and greatest scientific research―typically carried out by pro-police researchers―to support such a phenomenon, the opposition truly has an uphill battle on their hands.


But the fact is, there is no latest and greatest scientific research, and there never has been. A deadly force encounter, and the stress that accompanies it, is a psychological exchange between two or more people that cannot be replicated in the laboratory. There is no independent variable that can be manipulated, and no scientific method that can be applied. We are left with a body of studies that have little ecological validity, or anecdotal surveys of small non-representative samples.


There are two studies in particular that are used in almost every questionable deadly force case to help justify an officer’s actions or inaction. While both might rightly be considered academic research, they both fall way short of the standard for scientific research. Both studies use samples so small that generalizing the results of either to a larger population is impossible. And finally, both suffer from an obvious and fatal flaw that renders them invalid, a circumstance that should also render them inadmissible in the courtroom.    




Before discussing the two studies in question, it is important to understand an important statistic, and that is the number of police deadly forces cases that occur each year. While this number has been difficult to come by in years past, the Washington Post now maintains an online and easily accessible database of all police killing in the U.S. since January 1, 2015. They have reported, with identifying data, 995 deadly police shootings in 2015; 958 in 2016; and 737 thus far in 2017 (October 1st). It is reasonable to assume that these numbers are relatively stable over time. And thus we can say that since the year the earliest of the two studies was begun (1994), there have been roughly 20-25,000 police shootings in the U.S. So let’s look at these two studies with those numbers in mind.


The first study is that of Klinger and Brunson, and is titled Police officers' perceptual distortions during lethal force situations: Informing the reasonableness standard.[1] In this study the authors interviewed just 80 officers who had been in a deadly force encounter. The second study is by Alexis Artwohl, and is titled Perceptual and memory distortions during officer involved shootings.[2] In her study, data was collected from just 157 officers over a six-year period. Klinger is a former police officer. Artwohl continues to be highly involved in police training.




The results of neither study offer much useful data that can be used to reach a determination of reasonableness, or lack of, in a police shooting. First of all, the majority of their conclusions only validate the obvious, that in a moment of high stress, activation of the sympathetic nervous system will cause a person’s perception of time and distance, as well as their auditory experience, to be changed. This is well researched and documented. But there are two problems with these studies. First, they both fail to quantify the amount of distortion that occurred. And second, neither makes an effort to correlate the level of distortion to an officer’s training and experience. There is a great deal of research that supports the idea that with more training and experience, the officer will feel more in control of a situation and better able to anticipate the outcome. This in turn leads to less stress, less activation of the sympathetic nervous system, and consequently, less perceptual distortion. Without this correlational data, neither study has much utility.


The bigger problem is when lawyers and pro-police experts attempt to use this data to construct a whole new reality, such as arguing that because of perceptual distortion, the officer saw a gun that in reality wasn’t there. Klinger and Brunson avoided this issue in their study, and thus offered no conclusions beyond those relating to distortions of speed of motion, auditory experience, and visual acuity. They did include a category called “other distortion,” which was reported by just 10 officers. Artwohl on the other hand does address the issue, and included a category called “memory distortion: saw, heard, or experienced something that didn’t really happen.” Her results were not much better. Only 32 officers reported experiencing this type of distortion. Like Klinger & Brunson, she offers no quantitative data on the level of distortion, nor the training and experience of those who reported this type of distortion. And neither distinguishes those officers who actually argued perceptual distortion in defense of a questionable use of force, as opposed to those who were clearly and unquestionably justified in their use of deadly force.




There is a confounding variable in each study that renders them invalid. It is a variable that should be apparent to any attorney, and that is the statute of limitations on murder and federal civil rights violations. Every police officer involved in a deadly force encounter is acutely aware of the possibility of criminal, civil, or administrative action against them. It is a possibility that may take years to diminish. Take for example the case of St. Louis Metropolitan Police officer Jason Stockley. Stockley and his partner were involved in a vehicle pursuit of a suspected drug dealer in 2011. At the end of the pursuit, Stockley shot into the suspect’s crashed vehicle and killed him. At the time he was cleared of any wrongdoing (the suspect had a gun) by his own prosecuting attorney and the FBI. Five years later, and for reasons that remain a mystery, the prosecuting attorney presented the case to a St. Louis grand jury and Stockley was indicted for first degree murder. Not surprisingly, he was acquitted in a 2017 bench trial.


Following a deadly force encounter, police officers have little to risk by reporting perceptual distortion. It can help diminish their responsibility for what happened, and lessen the impact of any questionable judgment. But officers may place a great deal at risk by reporting that their thoughts were clear and unobstructed. In the case of Artwohl, she admitted to interviewing officers only days after their deadly force encounters. In most cases this would have been before the various state and federal investigations of the incident had been completed or even begun. It is little wonder that some of the officers reported perceptual distortion. The authors of both studies avoid any discussion of this problem, and their methodologies include no strategy for controlling for this confounding variable. Both studies are thus invalid as a means for determining the reasonableness of an officer’s actions.


Now it is true that under 42 USC 3789g federally funded research data is protected under a confidentiality requirement. Klinger and Brunson identify their study as in fact being federally funded. It is unclear in the case of Artwohl’s. But it is unlikely that for a police officer involved in a questionable use of force, as well as the attorneys who represent them, such an assurance from an academic will be enough to compel an honest response to a survey if such a response is perceived as exposing the officer to liability. It must also be pointed out that the disclosure prohibitions under this statute are binding only on federal employees and those who receive federal funding. Police officers deal daily with things such as Miranda and the fruit of the poison tree. They understand the impact of a suppression hearing, and how certain evidence gets to that point. They understand that there are workarounds to avoid these problems, and are often masters at using them. Police officers are naturally suspect of academics. Regardless of the confidentiality requirements of 42 USC 3789g, there is little chance a police officer will ever voluntarily expose themselves to potential criminal, civil, or administrative liability, and thus, in surveys such as these, it is expected that they will answer in ways that support their efforts to avoid such liability now and forever.


Finally, and as mentioned above, neither study points out how many of the cases involved in their survey involved an actual defense of perceptual distortion against a prosecutor’s allegation of unreasonableness, as opposed to the incident being unquestionably justified. Artwohl’s study also suffers from additional problems. She admits that the survey she used was not completed until after the officers involved had been informed of the facts of the case during required critical incident debriefings following the incident. She also admits that a full third of the surveys were administered by other people, including other police officers. She does not say if these other police officers were union representatives, internal affairs investigators, commanders, or just helpful friends. In the case of Klinger and Brunson, both of whom are criminologists as opposed to behavioral scientists, they have the additional problem of how their participants were selected. They were either acquaintances of the researchers, or acquaintances of those individuals. The potential for “demand characteristics” in such a scenario―the officers answering in ways they believe will validate the research―is greatly increased for a number of reasons.




The concept of perceptual distortion finds its way into many cases involving an officer’s questionable use of deadly force. It is typically misused to support the officer’s narrative and to reduce or eliminate responsibility for the officer’s own actions. And almost always the two studies discussed above will be offered to support the testimony of a pro-police expert. But both studies have major problems that attack their validity, and their use should be challenged. There is no doubt that changes in perception occur when the sympathetic nervous system is activated, but the degree of change is mitigated by a police officer’s training and experience. Without controlling for these variables, both studies are of limited utility beyond the small amount of anecdotal evidence they provide.

[1] Klinger, D., and Brunson, R. (2009). Police officers’ perceptual distortions during lethal force situations: Informing the reasonableness standard. Criminology and Public Policy, vol. 8(1), pgs. 117-140.

[2] Artwohl, A. (2002). Perceptual and memory distortions during officer involved shootings. FBI Law Enforcement Bulletin, vol. 71(10), pgs. 18-24.