Reflections on ‘Innocence After Guilt’ and the Role of Forensic Evidence in Overturned Convictions

A few years ago, I had the opportunity to participate in a video series produced by the International Symposium on Human Identification titled "Innocence After Guilt." The series examines the role forensic DNA testing has played in post-conviction work over the past three decades, both in exonerating the innocent and in reaffirming the convictions of the guilty.

I was asked to appear in multiple episodes on the common causes of wrongful conviction, alongside DNA consultant Tiffany Roy, forensic scientist Peter Valentin, and Mike Ware, Director of the Innocence Project of Texas. The experience gave me reason to think again, and more carefully, about a set of problems I have spent much of my career trying to address.

In particular, the release of my newest book, Valid Comparisons, addressed some of the complexities inherent to our criminal justice system’s use and evaluation of scientific evidence.

The numbers alone are sobering, even with appropriate caveats. The National Registry of Exonerations, which has drawn legitimate criticism over the consistency and accuracy of its data, reports 3,337 exonerations in the United States since 1989, accounting for more than 29,500 years lost to imprisonment.

I raise those figures with a qualification of my own: an exoneration is a legal outcome, not a formal finding of factual innocence. Convictions are set aside for a range of reasons, and the population of exonerated individuals is not uniform in what it tells us about actual guilt or innocence in each case.

What the figures do reliably illustrate, whatever their precision, is that convictions can move through our courts with every appearance of legitimacy and still prove to be wrong. The causes are by now well documented: mistaken eyewitness identification, false confessions, testimony from jailhouse informants with an incentive to lie, prosecutorial or state misconduct, and perjured testimony.

What is less often discussed, and what came up repeatedly in our conversation, is the extent to which many of these errors take root not in bad faith but in overwhelmed jurisdictions that lack the resources, personnel, and budget to do the work properly. Wrongful convictions are rarely the product of a single villain. More often they are the product of a system operating under strain, where shortcuts become habits and habits become outcomes.

Forensic evidence, when properly developed and interpreted, is accurate the vast majority of the time. That was a point I wanted to make clearly in the episode, because it is sometimes lost in public discussion of wrongful convictions. The deeper issue is not usually the science itself but how that science is communicated to a jury of laypeople who are being asked to weigh technical evidence against the liberty of another human being.

In the segment in which I was featured most prominently, Tiffany and I discussed some of the common mistakes we see analysts make on the witness stand: overstating certainty, using jargon without translation, or failing to clearly distinguish between what the evidence shows and what it merely suggests. Peter made a related point that has stayed with me since we filmed: analysts must resist the temptation to stray into areas of speculation, because a jury may have no way of knowing where the established science ends and the witness's personal opinion begins. That distinction is often the difference between a conviction that holds up and one that does not.

Mike Ware's contribution to the episode grounded all of this in a single case. He described the overturned conviction of Thomas McGowan, who served nearly twenty-three years in a Texas prison based in part on faulty eyewitness testimony before DNA evidence proved his innocence. Cases like this one matter more than any aggregate figure could. A single name, a single sentence, and twenty-three years taken from one life say more about what is at stake than a statistic ever will

I have spent a good part of my career presenting workshops on expert witness testimony, with the strong belief that scientific accuracy is only half the job. The other half is communication: making sure that what an analyst knows to be true is what a jury actually understands. Participating in this series was a chance to make that case to a wider audience, alongside colleagues who have each seen the consequences of getting it wrong.

If the series accomplishes anything, I hope it is this: a clearer understanding that the integrity of forensic science depends not only on the laboratory, but on everyone who carries that evidence into a courtroom and puts it into words a jury can trust.