Skip to main content

Verified by Psychology Today


"I Saw It with My Own Eyes." But Did You Really?

Studies continue to show that eyewitness testimony can be completely unreliable.

Key points

  • Inaccurate eyewitness testimony has led to the false convictions of hundreds of individuals.
  • Studies have shown that the memory recall process is malleable and highly susceptible to manipulation.
  • Researchers have suggested important protocols to help improve the accuracy of eyewitness testimony.

In 2010, Aaron Scheerhoorn was stabbed to death outside of a Houston nightclub. The attack was merciless, with Scheerhoorn crying out "Help me! He's killing me!" as he was stabbed numerous times. A crowd witnessed the attack, and upon questioning by detectives, six different eyewitnesses identified Lydell Grant as the murderer.

At trial, the six eyewitnesses testified that they had seen Grant murder Scheerhoorn. For violent crimes, any eyewitness is a luxury for the prosecution; six fingering the same defendant is extraordinarily rare. Despite having an alibi, and despite the prosecution having no physical evidence tying him to the crime (which was unusual given the bloody nature of the attack), Grant was quickly found guilty and ultimately sentenced to life in prison. As any lawyer knows, few things sway a jury like eyewitness testimony.

Since the day he was arrested, Grant steadfastly maintained his innocence. From his maximum-security prison in Gatesville, Texas, he sent dozens of letters to defense attorneys in an attempt to have his case reevaluated, but most went unanswered. Years passed; soon Grant found himself imprisoned for nearly a decade.

Everything changed in 2019 when a new DNA testing method analyzed samples from underneath Scheerhoorn's fingernails, clearing Grant and instead implicating another man, Jermarico Carter, who later confessed to the murder. A writ of habeas corpus was filed bringing to light the new DNA evidence, and Grant was quickly released from prison.

Unfortunately, Grant's story is not unique. Data from the Innocence Project (an organization committed to exonerating individuals wrongly convicted of crimes) show that more than 375 individuals have been exonerated with new DNA evidence. Twenty-one of those individuals had been sentenced to death and served time on death row. Importantly, the overwhelming majority of convictions overturned through DNA testing were originally based on eyewitness testimony.

Eyewitness Testimony Is Often Unreliable

How did this happen? How did six different people all think they saw Grant murder Scheerhoorn? One might assume that Grant was the unlucky doppelgänger of Carter (the actual murderer), but as seen in their mugshots below, this isn't the case.

Houston Police Department
Mugshots of Lydell Grant (left) and Jermarico Carter (right)
Source: Houston Police Department

The reality is that numerous studies (in addition to real-world cases) have pointed to the reality that eyewitness testimony is often unreliable. This is of grave concern, given that surveys show that eyewitness testimony is among the most convincing forms of evidence presented in criminal trials (e.g. Benton et al., 2006).

Beginning in the 1970s, psychologist Elizabeth Loftus discovered the once-surprising (but now accepted) fact that memory is malleable and subject to manipulation. Several experimental studies have shown that participants can be primed to believe that they have witnessed any number of things, even when they did not. For example, individuals have been primed to believe that they saw a Stop sign, when they actually saw a Yield sign (Loftus, Miller, & Burns, 1978). Another seminal study, conducted by Loftus & Pickrell (1995), provides an illustrative example of the ease of creating "false memories" in participants. Researchers gave participants written accounts of four different events, three of which they had actually experienced. The fourth story wasn't real, and involved the participant getting lost as a child in a public place (for example a mall or amusement park). A family member was used to provide realistic details for the false story. Around one-third of participants claimed to have remembered the false story. This change in memory recall that arises after manipulation has been termed the "misinformation effect."

Why Is Eyewitness Testimony Unreliable?

Many people believe that human memory works like a video camera: We perceive an event, our brain records the experience, and then we later recall it in the same way that a film plays on a screen. But this isn't the case. Instead, memory recall is more akin to piecing a puzzle back together. As the pieces are reconstructed, one by one, our brains become highly susceptible to influence.

In criminal investigations, detectives will sometimes drop small clues that witnesses will pick up on, even if only subconsciously. For example, while using the quintessential lineup tactic for suspect identification, a detective might smile, grunt, or nod when a suspect is chosen (e.g., Zimmerman et al., 2007). More insidiously, detectives and prosecutors can outright manipulate people into having false memories by providing them with suggestive clues. Tactful questioning by a lawyer during trial can also alter the testimony of a witness as pieces of their memory become combined with information included within the lawyer's questions. For example, imagine that a car accident has occurred on a cloudy day, but during the trial, a lawyer asks the following question to an eyewitness: "Given that the accident occurred just past noon, did the sunshine affect your ability to actually see the crash?" Such a witness might now recall that the crash occurred on a sunny day, even though that isn't true.

Indeed, returning to the topic of Lydell Grant, reports suggest that all of the eyewitnesses to Scheerhoorn's murder were manipulated in some way either before or after their identification of Grant. Three reported that the detective told them that they had picked the same person that other people had. Two other eyewitnesses discussed their selection with each another and confirmed each other’s memory. The last eyewitness claimed that the detective stated “good job” following their identification of Grant. This type of manipulation can lead not only to incorrect identifications, but also more confidence in eyewitnesses that their memory is correct. Importantly, the more confident an eyewitness seems during trial (e.g., "I am 100% certain that is the man!"), the more persuasive they seem to jurors (Smalarz & Wells, 2014).

How Do We Make Eyewitness Testimony More Reliable?

It's important to note that eyewitnesses can be reliable; in fact, one study has shown that, under the right circumstances, eyewitnesses can recall information correctly more than 90% of the time (Wixted et al., 2015). The key is to ensure that one's memory is not contaminated, much in the same way that fingerprints should not be disturbed.

To study the most accurate methods of achieving accurate eyewitness memory recall, a committee of psychologists and criminologists was organized in 2014 by the U.S. National Research Council. Their seminal report, Identifying the Culprit: Assessing Eyewitness Identification, set out several important recommendations, including (1) "double blind" lineups, in which neither the administrator nor the eyewitness knows who the suspect is, (2) standardized instructions for eyewitnesses that ensure they are not primed to select a particular individual, (3) statements of confidence from eyewitnesses immediately after making an identification, where an eyewitness provides a statement, in their own words, that articulates the level of confidence they have in their identification, and (4) recording the identification process, preferably with both audio and video.

As it stands today, 24 states have adopted these types of procedural reforms for law enforcement, including Texas in 2011—one year after Lydell Grant was convicted of murder.


Benton, T. R., Ross, D. F., Bradshaw, E., Thomas, W. N., & Bradshaw, G. S. (2006). Eyewitness memory is still not common sense: comparing jurors, judges and law enforcement to eyewitness experts. Applied Cognitive Psychology, 20, 115–129.

Loftus, E.F., Miller, D.G., & Burns, H.J. (1978). Semantic integration of verbal information into a visual memory. Journal of Experimental Psychology: Human Learning and Memory, 4, 19–31.

Loftus, E.F. and Pickrell, J.E. 1995. The formation of false memories. Psychiatr. Ann. 25:720 -725.

Zimmerman, D. M., Chorn, J. A., Rhead, L. M., Evelo, A. J., & Kovera, M. B. (2017). Memory strength and lineup presentation moderate effects of administrator influence on mistaken identifications. Journal of Experimental Psychology: Applied, 23(4), 460–473.

Smalarz, L., & Wells, G. L. (2014). Post-identification feedback to eyewitnesses impairs evaluators’ abilities to discriminate between accurate and mistaken testimony. Law and Human Behavior, 38(2), 194–202.

Wixted, J. T., Mickes, L., Clark, S. E., Gronlund, S. D., & Roediger, H. L. III. (2015). Initial eyewitness confidence reliably predicts eyewitness identification accuracy. American Psychologist, 70(6), 515–526.

More from Maclen Stanley JD, Ed.M.
More from Psychology Today