“I remember his face very clearly. It's like a photograph in my memory.”

—Penny Beerntsen

About two decades before the Teresa Halbach murder, Steven Avery was accused of brutally attacking a woman named Penny Beerntsen. The truth, however, was that Avery never touched her. Actually, he had never even met her. In fact, the first time he ever saw her was inside the Manitowoc County Courthouse, well after the state had criminally charged him for the assault.

EYEWITNESS IDENTIFICATION TESTIMONY

Eyewitness identification testimony was the centerpiece of Avery's first trial in Making a Murderer. Beerntsen testified that Avery attacked her, and the jury convicted him on all counts—even though he denied committing the crime, the state had no physical evidence linking him to the crime, and sixteen alibi witnesses testified that he was nowhere near the scene of the crime.1

Eighteen years later, however, we learned that victim-eyewitness Beerntsen was mistaken. And we learned that the jury was wrong. Thanks to advances in DNA testing, Avery's defense team was eventually able to prove that serial rapist Gregory Allen, not Avery, was the guilty party.

Avery's first jury trial—a complete failure of the criminal justice system if ever there was one—perfectly illustrates the two problems with eyewitness identification testimony like Beerntsen's: first, such evidence is highly unreliable; and second, despite its unreliability, such evidence is highly persuasive.2

These two flaws have long been conspiring to convict innocent defendants. Even the Supreme Court of Wisconsin recently acknowledged that “eyewitness testimony is often hopelessly unreliable,” and, worse yet, erroneous in-court identifications “are the single greatest cause of wrongful convictions in the United States, and are responsible for more wrongful convictions than all other causes combined.”³

But why did Beerntsen testify that it was Avery who attacked her? And why was her testimony so persuasive to the jury? To answer these questions we need to travel back in time to the day Beerntsen was assaulted. On that day the Manitowoc Sheriff's Department set off a chain of events that, ultimately, caused Beerntsen to misidentify Avery at trial.

THE SHOW-UP IDENTIFICATION PROCEDURE

Making a Murderer presented evidence that, within hours after Beerntsen reported being attacked, the Manitowoc Sheriff's Department conducted what is known as a “show-up” procedure using a drawing of Steven Avery. In a show-up, unlike a lineup or photo array, the cops present a single suspect (or a photograph or drawing of that single suspect) to the victim, and then ask her to confirm that the suspect is, in fact, the perpetrator.⁴

On the day Beerntsen was attacked, the Manitowoc Sheriff's Department was already familiar with Avery from his previous arrests. They were also investigating him for a recent road rage incident involving a deputy's wife. Given this, they immediately targeted him as their prime (and sole) suspect in the Beerntsen case. It was even alleged that, after hearing Beerntsen describe her attacker, Deputy Judy Dvorak proclaimed, “That sounds like Steven Avery.”⁵

In hindsight, of course, we know that the sheriff's department's laser-like focus on Avery was misplaced. But even back then there were major differences between Beerntsen's description of the attacker and Avery. For example, Beerntsen was adamant that she got a good, close look at her assailant and his eyes were brown. Avery's eyes, however, were bright blue.6 Yet despite this and other discrepancies, the sheriffs were convinced that Avery was “the guy,” so they retrieved his booking photo from a previous arrest. They then had a sketch artist, Deputy Gene Kusche, produce a drawing of Avery based on that booking photo.⁷ Afterward, they presented the drawing to Beerntsen and asked her to confirm that Avery was, in fact, her attacker. And she complied.

This initial show-up procedure—where the sheriffs presented Beerntsen with only one suspect, Avery—is incredibly important. While eyewitness identifications are in general unreliable, show-up identifications are the most unreliable of all. In fact, show-ups are “the most grossly suggestive identification procedure now or ever used by the police.”⁸

In a show-up, the crime victim naturally thinks the cops are presenting her with only one option because they know who committed the crime.⁹ Why else would they be focusing on one particular suspect? This, in turn, makes the victim eager to positively identify the suspect presented to her. And the victim's eagerness is exacerbated if the cops explicitly tell her—as Dvorak allegedly told Beerntsen—that the victim's initial description of the attacker “sounds like” the suspect. The end result is that the victim will positively identify the suspect in the show-up, regardless of whether he is really the attacker. The suspect will then be arrested, prosecuted, and, in many cases, convicted.

The danger of the show-up procedure was well known, even decades before the Beerntsen case. In 1967, the US Supreme Court wrote that “the practice of showing suspects singly to persons for the purpose of identification, and not as part of a lineup, has been widely condemned.”¹⁰ In fact, show-ups are so suggestive, and pose such a high risk of misidentification, that even the Supreme Court of Wisconsin eventually banned their use by law enforcement—with limited exceptions.¹¹

But in one of the most interesting aspects of Making a Murderer, the sheriffs actually denied that they conducted a show-up procedure using the drawing of Avery. Rather, they claim they first asked Beerntsen to describe her attacker. Then, based on her description, Kusche sketched a drawing, and the drawing looked like Avery because Avery was, in fact, the perpetrator. Beerntsen then viewed the drawing and confirmed that it accurately depicted her assailant. In other words, it was Beerntsen's verbal description, with Kusche being “just the pencil,” that produced the drawing.12

For eighteen years after Avery's conviction, this explanation seemed plausible, and even probable. But now we know that Avery did not rape Beerntsen. Therefore, when Beerntsen described her attacker, it is incredibly unlikely that she would have described Steven Avery—a man who was nowhere near the crime scene when she was attacked, and a man whom she had probably never even seen before.

Given what we now know, a better explanation for why the drawing precisely matched Avery's booking photo (rather than the real perpetrator) is that when Beerntsen described her attacker, the sheriffs wanted it to be Avery—the man they had already decided was guilty. Kusche therefore sketched a drawing based on Avery's booking photo rather than on Beerntsen's description of the attacker. The sheriffs then presented that single drawing to Beerntsen and asked her to confirm that Avery was, in fact, her assailant.¹³ Thinking that the sheriffs knew what they were doing, she agreed.

In sum, Making a Murderer presented a persuasive case that the sheriffs did conduct a highly suggestive show-up procedure, and that is why Beerntsen initially misidentified Avery on the day that she was attacked. But didn't Beerntsen later identify Avery in a photo array and a lineup? And wouldn't those subsequent identification procedures cure the problem caused by the earlier, improper show-up procedure?

THE PHOTO ARRAY AND LINEUP

After the show-up, the sheriffs did conduct a more traditional photo array procedure, and then a lineup procedure, for Beerntsen's viewing. Each identification procedure forced Beerntsen to pick her attacker out from a group of people. And in each procedure, she picked Avery.

It is true that photo arrays and lineups are more reliable than show-ups. In a photo array or lineup, the victim is asked to identify the perpetrator among several—often a total of six—individuals. One of the individuals is the person that law enforcement believes is guilty of the crime, and the others will be fillers. These fillers are people who should, to some extent, resemble the suspect,14 but are known by law enforcement to be innocent—often because they were already incarcerated on unrelated cases at the time the victim was attacked. (As Avery learned after his trial where he presented sixteen alibi witnesses, the only surefire alibi is being in jail at the time of the crime.)

Unlike show-ups, when photo arrays and lineups are properly constructed and administered, the victim will not know which person law enforcement wants her to identify. Further, the risk of misidentification is spread among the suspect and the fillers, rather than falling entirely on the suspect, as it does in a show-up.¹⁵ Unfortunately, however, even a proper photo array or lineup cannot sanitize the witness's memory of an earlier show-up procedure—something that likely taints all subsequent identifications.

More specifically, because Beerntsen first identified Avery in the highly suggestive show-up procedure, and then identified him in the photo array and lineup, there was no way to know whether those subsequent identifications were based on her memory of him being the real perpetrator—which we now know he was not—or her memory of seeing him in the earlier show-up procedure.

Avery's appellate counsel raised this issue in his first appeal. As the appellate court put it, Avery “contends that the lineup was suggestive because he appears to have been the only person who was in both the [earlier] photo array and the lineup….”¹⁶ (It would also follow, of course, that the photo array was suggestive because Avery also appeared in the earlier show-up. This, however, was not raised on appeal; at that time, it was not even known that the sheriffs likely conducted a show-up procedure using the drawing of Avery.)

Unfortunately, the appellate court dealt with this problem simply by ignoring it, and upheld Avery's conviction without even addressing the issue of successive identifications. But since that 1987 decision, even Wisconsin's attorney general has warned that conducting multiple, successive identification procedures is dangerous and should be avoided:

An eyewitness viewing a second procedure with the same suspect may believe that the suspect's presence in both procedures suggests that authorities believe the suspect is the perpetrator. Or, an eyewitness may become confused and identify the suspect based on recognizing him/her from the prior procedure rather than from remembering the suspect's presence at the crime. In either case, the suggestiveness of the second procedure may taint the eyewitness.17

Taking the specific identification procedures in reverse order, then, Beerntsen's identification of Avery in the lineup was based on her memory of him from the photo array. And Beerntsen's identification of Avery in the photo array was based on her memory of him from the show-up. And we now know that Beerntsen identified him in the show-up not because he was the real perpetrator, but because she believed the sheriffs when they implied (or possibly even explicitly told her) that the person in the show-up, Avery, was “the guy.”

In chronological order, her misidentification of Avery in the show-up led to her misidentification in the photo array, which in turn led to her misidentification in the lineup. At this point, it was all over but the shouting.

IT'S SHOWTIME!

Ultimately, Beerntsen had to identify Avery in court, at trial, for the jury. In some ways, a witness's in-court identification of a defendant is the ultimate in highly suggestive identification procedures. First, the victim-eyewitness will reasonably assume that the defendant will be present in the courtroom at the time of the trial. Second, the victim-eyewitness will also assume—again, quite reasonably—that the defendant will not be seated on the judge's bench, in the jury box, or at the prosecutor's table. This usually leaves only two people at the defense table, and the person wearing the suit and doing the talking will obviously be the defense lawyer. This, in turn, leaves only one person—the defendant—for the witness to identify as the perpetrator.

But even aside from the suggestive nature of the courtroom setup, Beerntsen's positive, in-court identification of Avery was easy to predict. After seeing him in the three out-of-court identification procedures, pointing the finger at him in court, during the trial, was a foregone conclusion. That is, the cumulative impact of those prior, out-of-court identifications “made it all but inevitable that the witness would identify the defendant [in court] whether or not he was in fact the man” who attacked her.18

The only question left was whether the jury would believe Beerntsen's identification or Avery's sixteen alibi witnesses. They chose Beerntsen's identification. And this demonstrates the ultimate problem with eyewitness identification testimony in court: “[T]here is almost nothing more convincing than a live human being who takes the stand, points a finger at the defendant, and says, ‘That's the one!’”¹⁹ And this is not just speculation. It has even been demonstrated in controlled studies:

This phenomenon is best illustrated by a classic psychological study, in which three sets of mock jurors were presented with identical evidence in a mock trial. The only difference was that each set of jurors was presented with different eyewitness evidence. One group was told that no eyewitness existed, and only 18% voted to convict. Another group heard a store clerk testify that he saw the defendant commit the crime, and 72% voted to convict, despite the defense lawyer's argument that the clerk was mistaken. Amazingly, when the third group learned that the store clerk “was legally blind and not wearing his glasses at the time” of the crime, an incredible 68% of the jurors still voted to convict. This demonstrates how willing jurors are to simply accept eyewitness testimony without any critical evaluation whatsoever.²⁰

To make matters worse, in Avery's trial the jury's willingness to accept Beerntsen's testimony was exacerbated by two things: first, Beerntsen testified she was “absolutely positive” that Avery was the perpetrator; and second, to support this claim, she insisted that “she had ample opportunity to view” her attacker at the time of the assault.²¹ These two things, of course, made her identification of Avery even more convincing. However, these two things did not make her identification more reliable.

Although counter-intuitive, controlled studies demonstrate that a witness's level of confidence in identification is not correlated with the accuracy of that identification.22 In other words, a confident witness can be mistaken just as a cautious or tentative witness can be mistaken. As we know from our experience in everyday life, just because someone is more confident in remembering things doesn't mean that person's self-confidence is justified. Beerntsen simply failed to realize that her memory was a false one, and had been contaminated by the Manitowoc Sheriff's Department's multiple, pretrial identification procedures.

And perhaps even more surprising, an eyewitness's “opportunity to view” the perpetrator at the time of the crime “does not translate into an eyewitness being able to accurately recall the suspect.”²³ There are at least two reasons for this: first, in hindsight, the witness can easily misstate her degree of attention during the crime; and second, the witness often fails to appreciate the impact that stress and anxiety have on her ability to focus during the crime and, later, to accurately recall the perpetrator in court.²⁴

In sum, Beerntsen's in-court identification of Steven Avery—though confident and convincing—was no more accurate than the highly suggestive show-up procedure on which it, and the state's entire case, was based.

DOUBLE THE DAMAGE

Misidentifications and wrongful convictions are a huge problem with eyewitness identification evidence—particularly when those identifications are the result of, or have been tainted by, a show-up procedure. But this is really only half of the story.

It is important to remember that whenever a witness misidentifies a suspect, that suspect will be arrested, prosecuted, and possibly convicted—all while the real perpetrator remains free to commit more crimes.²⁵ This point can be lost on the police, prosecutors, and judges who are often determined to convict someone—anyone—whenever a crime is committed.

As discussed earlier, the Supreme Court of Wisconsin eventually banned the use of show-up procedures in most circumstances, including situations where the cops plan to take the suspect into custody.26 The court reasoned that if the police are going to arrest the suspect, then he isn't going anywhere anytime soon, thus giving the police ample time to assemble a more reliable photo array or lineup for the victim's viewing. This, in turn, will reduce the risk that the victim will misidentify an innocent suspect.

Yet, despite the reasonableness of this rule, government agents still look for disingenuous ways to circumvent it. In several cases, Wisconsin prosecutors and trial judges have conspired to produce this line of reasoning: despite the ban on show-ups in situations where the police are going to arrest the suspect, a show-up procedure should still be permitted when the police plan to arrest the suspect for something other than the crime under investigation, such as a probation hold or arrest warrant in an unrelated case.

Fortunately, the appellate court eventually rejected this frivolous distinction and held that the reason for the suspect's arrest is irrelevant.²⁷ Instead, the point is that whenever the suspect is in custody, regardless of the reason, the police will have ample time to administer a more reliable identification procedure. Therefore, under such circumstances, the police have no justification for using a highly suggestive and error-prone show-up procedure.

For the short-sighted Wisconsin police, prosecutors, and judges who continue to look for ways around the general ban on show-ups, they should simply think of the Beerntsen case. They should remember that, when Beerntsen misidentified Avery in the show-up procedure, not only was an innocent man arrested, prosecuted, convicted, and imprisoned, but the real perpetrator—serial rapist Gregory Allen—remained free to commit more sex crimes and victimize more women before he was eventually caught. The lesson: show-up identification procedures double the damage and should rarely, if ever, be permitted in our criminal justice system.

ALIBIS VERSUS EYEWITNESSES

When the police assemble lineups and photo arrays, they need to locate five other people—known as fillers—to include along with the suspect. If the victim identifies one of the fillers as the perpetrator, the filler is not at any risk of being prosecuted. Instead, the police will know that the victim's identification was mistaken, as the filler has an airtight alibi: he was in jail on an unrelated matter at the time of the crime.

But is being in custody the only airtight alibi? Avery learned that his sixteen-witness alibi defense was far from airtight; the prosecutor easily defeated it with a single eyewitness. But in perhaps the most ironic eyewitness identification case of all time, a man named Donald Thomson learned that, sometimes, alibi defenses do work.

In 1975, a woman had been watching television when a man broke into her home, attacked and raped her, and left her unconscious. When she woke up she contacted the police and described Thomson as her attacker; the police were able to arrest him the next day. Fortunately for Thomson, he, just like the fillers described above, had an airtight alibi. But Thomson was not in jail at the time of the crime. Rather, he was far away from the crime scene and was being interviewed on a live television show. It turns out that the woman had been watching that same television show immediately before she was attacked, and confused Thomson's face (that she had seen on television) with the face of her assailant.28

That misidentification is, in one respect, similar to the Beerntsen-Avery misidentification; that is, it demonstrates how stress at the time of the event can impact the victim's future ability to accurately identify the suspect. But the irony is that Donald Thomson was a research psychologist and was invited on the television show to discuss the flaws in human memory and eyewitness identifications.²⁹

It's a good thing that Thomson's television show was live, rather than tape-delayed, or his airtight alibi would have vanished. Without it, he may have ended up, like Avery, a convicted criminal instead of an ironic (and iconic) example of the hopeless unreliability of eyewitness testimony.