Search About Donate
A New York Police Department Sergeant demonstrates how to operate a body camera during a press conference in December.

Should Cops Get to Review the Video Before They Report?

Sorry, Mr. Bratton. Science says no.

The scenario is all too familiar. A police officer with a dash-cam or body camera stops an individual, the situation escalates, the individual is apprehended, a charge is made and the individual is arrested. The question is whether prior to being questioned or even prior to writing a report, should the officer be permitted to view the recorded footage?

Philip Eure, the Department of Investigation’s Inspector General for the NYPD, recently recommended that police officers be prevented from viewing recorded footage before giving a statement to investigators. Quick to respond, New York’s City Police Commissioner, Bill Bratton called this “one of the recommendations of the I.G. that we strongly, strongly disagree with and will not support under any circumstance.” His concern was enhancing the integrity of police officers; “I am not intending to use the cameras to play a game of gotcha with the cops.”

This is a complex issue, but one for which cognitive science research provides a clear answer. If the purpose of any investigation is to get the most complete, accurate information possible, then it could be argued that the officer should view the footage, probably multiple times, prior to being questioned and prior to testifying. Human memory is notoriously flawed, but we can consider recorded footage to be “ground truth.” So according to this argument, bolstering the officer’s account by having him view the recorded footage effectively serves to enhance the accuracy of the officer’s report. And it does.

The problem is that in so doing, two independent lines of evidence – the officer’s eyewitness memory and the recorded footage – are no longer two independent lines of evidence. That is, the eyewitness memory of the officer has been tainted by viewing the recorded footage. If in the prosecution of the case the officer is to serve as an eyewitness, and his memory is to be preserved untainted, then it is critical that the officer not view the footage.

Imagine another situation. Assume that a fingerprint expert is told that a suspect’s DNA was found at the crime scene. The expert is then asked if the fingerprint left at the crime scene matches that person’s fingerprint. Although this too is common practice, good sense would tell most people that this procedure is biased because the fingerprint evidence has been tainted by knowledge of the DNA evidence.

Here’s a way to test the utility of having an officer view recorded footage prior to testifying. Try this: (1) let an officer view the recorded footage for an arrest that he recorded, (2) randomly select a civilian who was not present at the incident and have him view the same recorded footage, (3) now interview both individuals about the incident and compare the accuracy and detail of the two accounts. How well does the information reported by each match the information in the recording? There is no reason to think that the officer’s account would be superior to the civilian’s. The real forensic value the officer brings to the case is his independent, unchecked experience of the event.

If the purpose of any investigation is to get the most complete and accurate information possible, and if the recorded footage is “ground truth,” then the testimony of the officer is not needed at all. He did his job by recording the event. The recording is the only evidence that is needed. Because of natural flaws in the memory process, the officer’s account of what transpired will almost always be less accurate than the recorded footage.

But often the officer’s perception of what occurred is as or more important than what actually occurred. Cognitive science research has amply demonstrated that perceptions and memories are not literal representations of reality, and our behavior is affected by our perceptions of reality not necessarily reality itself.

If — in light of an officer’s vantage point, his heightened level of stress, multiple distracting events, etc., — he perceives an individual to be more threatening than he actually is, his subsequent description of an incident may not match the events in the recorded footage. This does not necessarily undermines the integrity of the officer’s account. (In fact, if the officer’s account matched the video footage too well, this should raise questions about the source of his “memory.”) His account reflects his perception of the event, and it is this perception that is likely to have governed his behavior. It is thus important to preserve the officer’s perception of the event and not taint his memory by letting him view the recorded footage.

It may be argued that no prosecutor is going to put on a police witness without showing him the video first – to eliminate discrepancies between the officer’s testimony and the video, and avoid the “game of gotcha” that is Bill Bratton’s concern. But for an experienced defense attorney, the absence of discrepancies — testimony that seems suspiciously pat — is an opening for a devastating cross-examination. At every step in the process, the credibility of the officer’s account will only be enhanced if he has not viewed the video footage. (And, by the way, if the officer is entitled to review the footage, why not the suspect, for whom the consequences of the “game of gotcha” are even more dire? Would the police permit a suspect to refresh his memory by viewing the video before being interrogated?)

We know from cognitive science research that memory is vulnerable to the suggestive influence of what is called “post-event information.” An officer’s memory of what happened will be suggestively influenced by viewing video footage. And, it is important to know that such distortions in memory are both permanent, and they occur without conscious awareness.

This same logic applies to lay witnesses being permitted to view video footage prior to giving a statement, making an identification, or testifying. Frequently witnesses are shown a surveillance video from the crime scene and are subsequently questioned about the incident or presented with a lineup. Once any witness views video footage, his memory will be biased by the video, a salient source of “post-event information.” And if a case goes to trial, jurors are more likely to be swayed hearing what they consider to be two corroborating lines of evidence, when in fact they are hearing the video footage twice — once on screen and once through the eyes of the witness who watched it.

If it is important to know exactly what happened, then viewing the video footage will always be more accurate than the account of an officer, in which case the officer does not need to see the footage. If it is important to know an officer’s perception of an event, then it is important to preserve his memory untainted by viewing the video footage. In neither case is viewing the video footage recommended. Philip Eure seems to have understood the delicate nature of human memory, and his recommendations are well founded.

Kathy Pezdek chairs the Cognitive Science Program at Claremont Graduate University.