Fascinating piece in the NY Times:

 . . . according to recent research, interrogation recording may in fact be too vivid and persuasive. Even seemingly neutral recordings still require interpretation. As advertisers and Hollywood directors know well, camera angles, close-ups, lenses and dozens of other techniques shape our perception of what we see without our being aware of it.

In a series of experiments led by the psychologist G. Daniel Lassiter of Ohio University, mock juries were shown exactly the same interrogation, but some saw only the defendant, while others had a wider-angle view that included the interrogator. When the interrogator isn’t shown on camera, jurors are significantly less likely to find an interrogation coercive, and more likely to believe in the truth and accuracy of the confession that they hear — even when the interrogator explicitly threatens the defendant.

Professor Lassiter and other psychologists have consistently shown this “camera perspective bias” across a substantial series of experiments, finding in one study that even professionals like judges and police interrogators are not immune.

Experiments like these feed a larger concern: whether the police, prosecutors, defense lawyers, judges or jurors can actually tell the difference between true and false confessions, even with the more complete record of interactions that recorded interrogations provide . . .

We know that false confessions really do occur, even in very serious crimes, and probably more frequently than most people expect. But why? We know something about certain interrogation techniques, as well as defendant vulnerabilities like youth or mental disability, that may create heightened risks for false confessions. But we don’t yet know enough about the psychology of false confessions to be able to accurately “diagnose” the reliability of a given confession just by watching it.

It took a long time to persuade police agencies to start recording interrogations and confessions. That many now do is an important win for transparency. But it’s also important to remember that recording isn’t a panacea. Things can still go wrong.

There’s an interesting question here in what we do with this sort of research. The criminal justice system has long been slow to recognize and account for cognitive bias. For example, scientific research dating back more than a century has shown that eyewitness testimony is heavily susceptible to suggestion, misinformation and other errors. A few simple steps police and prosecutors can take drastically improve eyewitness testimony, at little to no added cost. But implementing those reforms will almost certainly mean fewer eyewitnesses testifying in court, which would make it more difficult to win convictions. Consequently, it’s been a chore getting police agencies and DA offices to adopt them. The U.S. Supreme Court hasn’t helped. In 2012, the court declined to make the reforms mandatory.

Even in an ideal court system that took note of studies like Lassiter’s, there will always be the challenge of knowing when there’s enough of a scientific consensus on a topic to account of the new research in the criminal process. There’s also the challenge of knowing exactly how to account for it.

But we’re nowhere near the point where we can even worry about those problems. The problems with eyewitness testimony are well-documented, well-researched and, as noted, have been known for decades. The solutions are also well-known and proven. Yet we still can’t get the criminal justice system to require police and prosecutors to implement them. Taking the time to discuss how we should handle emerging research, then, seems impractically conjectural.