Not Like CSI: Overcoming Resistance to Science in Law Enforcement

Failed Evidence: Why Law Enforcement Resists Science
David A. Harris
October 25, 2012

By David A. Harris, Distinguished Faculty Scholar and Professor of Law, Associate Dean for Research, University of Pittsburgh School of Law

The news everywhere today is full of headlines like “DNA Cracks Cold Case.” Popular culture is topped by television programs like CSI, in which police are more likely to use test tubes and high-tech gadgetry than guns and handcuffs to solve crimes. The bad guys better watch out: science is now the handmaiden of law enforcement.  And with that kind of partnership, criminals don’t stand a chance. 

But my new book, Failed Evidence: Why Law Enforcement Resists Science (NYU Press, 2012), exposes this picture for the myth that it is. Aside from DNA and analytical chemistry, police and prosecutors usually resist science, sometimes very vocally.  This scientific work concerns the basics of how police gather the evidence that prosecutors use in court every day: eyewitness identifications, suspect interrogations, and basic kinds of forensic evidence, like fingerprint analysis, and hair and fiber identification.  This science has been peer reviewed and published and replicated for years – sometimes for decades.  It tells us not only what the problems are in these basic areas of investigation, but how to fix them.  And yet, there is resistance to re-calibrating our police and prosecutorial practices so that they are consistent with the best of what science can teach us.  The question at the heart of Failed Evidence is why.  If we understand where that resistance comes from, we can find ways to overcome it, so that we can stop convicting the innocent, and get the real guilty parties off the street.

Here’s a sample of what science tells us about three of the most common types of evidence, and how we could avoid errors that have resulted in 300 DNA-based exonerations to date.

In the traditional kind of eyewitness identification, a witness views six or eight people or photos at a time.  This is called a simultaneous lineup: the witness sees the people or the photos together and asks himself “which of these looks the most like the man who assaulted me?” This is a process of relative judgment, and introduces errors. Of the 300 exoneration cases, 75 percent have contained a faulty identification. But the science shows us that if witnesses see the people or photos not together but one at a time – sequentially – those errors largely disappear.

Suspect interrogations sometimes produce false confessions or false statements of guilt; 27 percent of the exoneration cases have a false confession, without any evidence of physical abuse by police or mental infirmity. The research shows that these false confessions come from the extreme pressures brought to bear on the suspects by the police: by threats of the death penalty or other harsh punishment, or by lies about the results of scientific testing (e.g., “we found you fingerprint on the knife”). Electronically recording all interrogations, and prohibiting lies about scientific testing, would eliminate much of the problem.

Faulty forensic science shows up in half of the 300 exonerations. According to the National Academy of Sciences 2009 report, Strengthening Forensic Science in the United States, this is because forensic science is not really science at all: it is a craft based on human interpretation, not based on data.  Also, most forensic labs do not require proficiency testing, certification or quality control, and do not use common laboratory procedures used to screen out human cognitive biases.

(The percentages add up to more than one hundred percent because many cases feature more than one cause of error.)

Why the resistance to what is generally regarded as straightforward, well-accepted science?  First, there are cognitive barriers.  Forces like cognitive dissonance (“I’m one of the good guys” vs. “basic police procedures that I use sometimes produce terrible miscarriages of justice”), group polarization (in which a group with a strong identity tends toward the more extreme views in the group when they talk to each other),  loss aversion (we fear giving up what we might lose by giving up what we have now more than we like the idea of gaining something by change) and status quo bias make change in this area very difficult.  Second, there are strong institutional barriers against change among police and prosecutors.  For example, anything that seems like it might make it more difficult to arrest and convict will encounter massive resistance.  And the “us versus them” culture of police departments makes any suggestion coming from the outside highly suspect.

What can be done?  Failed Evidence makes many concrete suggestions, but two, which sometimes seem counterintuitive are: 1) to put police and prosecutors in the lead of reform efforts, and 2) to seek allies on the political right. There is now a small but growing cadre of police leaders and prosecutors who have substantial experience recording their interrogations or using state of the art eyewitness procedures. They are in the unique position of being able to tell their brothers and sisters in law enforcement, “I’ve done this for years. It’s good for the police to record interrogations – and the sky hasn’t fallen.”  As for seeking allies on the right, this seems backwards to some people, because the natural allies for reform usually come from the left.  But leaders on the left are often hesitant to lead efforts at criminal justice reform. They fear being called “soft on crime” during the next election. Leaders on the right have usually advocated tougher positions on criminal justice issues over their careers, and therefore aren’t as easily intimidated. They have proven to be clutch players on these issues in states as diverse as North Carolina, New Jersey and Ohio, and they’ve been key to successful reforms in each.

Failed Evidencet ells us that the problems with our use of eyewitness identification evidence, suspect interrogation, and forensics are real – and they can also be solved.  The goal is not a perfect justice system; anyone looking for that is na├»ve. But we are obligated to do as much as our intelligence and our scientific work can lead us to do. That way, we minimize the tragedies of wrongful convictions -- and we don’t leave the real predators on the street.