DNA exonerates 300 but can't clear most innocents

The Washington Post's coverage of the 300th DNA exoneration nationally - that of Damon Thibodeaux, who was released last week from Louisiana's death row - made an astute observation about the limits of DNA evidence to rectify false convictions ("Louisiana death row inmate Damon Thibodeau exonerated with DNA evidence," Sept. 28):

The Thibodeaux case marks a dramatic mathematical milestone in the use of DNA in law enforcement, but it also signals the opening of a new, more complex phase in the use of such material in attempts to right the course of justice.

When DNA testing was first introduced in the late 1980s, the revolutionary new techniques shattered a widely held view in law enforcement and the public that American courts rarely convicted the innocent. Since then, high-profile exonerations and the increasingly common reliance on such testing have led many to believe that DNA can resolve doubts about almost any questionable conviction.

It’s now clear, however, that there is no DNA evidence in the vast majority of cases. In the first 15 years of DNA testing, almost all exonerations fit a basic pattern in which the defendant was accused of rape, or both rape and murder — because sexual assaults are the crimes in which DNA is most likely to be recovered. Between 1989 and the end of 2007, a total of 214 people were cleared using DNA evidence. In all but 14 cases — more than 93 percent — the alleged crime involved a sexual assault of some kind, according to a review by The Washington Post.

In hindsight, those straightforward, obvious miscarriages of justice were the low-hanging fruit of DNA exonerations. Now their numbers are declining. In their place are convictions such as Thibodeaux’s, in which serious doubts have been raised but little clear DNA or other scientific forensic evidence exists to conclusively prove guilt or innocence. In Thibodeaux’s case, the absence of any incriminating DNA evidence became as powerful an argument for his innocence as any other element of the case.

Of 83 exonerations in the past five years, more than 15 percent didn’t involve rape. As many as a quarter of the cases involved a false confession, in which one or more defendants admitted to the crime under interrogation.

Samuel Gross, an author of a report by the recently created National Registry of Exonerations at the University of Michigan, calculated that based on the proven rate of exonerations among death-row prisoners in the past two decades, U.S. courts appear to have an error rate in capital cases of between 2.5 percent and 4 percent. In June, researchers examining biological evidence from hundreds of Virginia rape convictions between 1973 and 1987 determined that new DNA testing appeared to exonerate convicted defendants in 8 percent to 15 percent of cases.

Applied against the approximately 140,000 prisoners on death row or serving life sentences in the United States, the findings suggest that many thousands of innocent individuals could be in prison for crimes they didn’t commit.

But the odds that many of those convicts will ever be able to prove their innocence through the existing systems of appeals are remote, given the lack of DNA evidence in the majority of cases.

Those observations are spot on, highlighting concerns this blog has been raising now for several years that focusing on DNA exonerations ignores most innocence cases. There have been several efforts to estimate rates of false convictions, ranging from a low of .75% (from an officer at the national District Attorneys association) up to 1.5%, to 2.3%, to 3.3% or even higher. A 2003 survey (pdf) of criminal justice practitioners estimating the proportion of "false positives" found that "Prosecutors and police perceive the least error (1/2-1%), while defense attorneys perceive the most error (4-5%). Judges perceive the national error rate to be between 1 and 3 percent."

If your correspondent were pressed on the point, my own guesstimate at the right proportion would fall between one and two percent - probably on the upper end of that range. Higher estimates tend to be based on capital murder and rape convictions, which for a variety of reasons likely see greater proportions of false convictions because of tremendous pressure to clear cases and lower standards of evidence required to convict, particularly for sexual assault, where a single eyewitness can generally be sufficient to secure a conviction. OTOH, most innocence estimates fail to take into account wrongful convictions arising the drug war, which in Texas, just taking into account the Tulia, Hearne and the Dallas fake-drug scandals, actually outnumber DNA exonerations.

Obviously, even the lowest of those estimates far exceeds the number of DNA exonerations. If just .75% of inmates in Texas prisons are actually innocent, that would translate to nearly 1,200 people currently incarcerated in Texas prisons for crimes they did not commit, though the real number could obviously be even higher. In the overwhelming majority of those cases there is and never will be DNA evidence available to prove their innocence.

For that reason, DNA exonerations should be viewed as a sample, the way pollsters judge public opinion by sampling the views of a much smaller group. Like opinion polls, conclusions drawn from such a sample may suffer from a margin of error, but in general they allow us a rare window into the causes of false convictions that was not previously available.

All that to say, while the 300th DNA exoneration is cause for celebration - particularly for Mr. Thibodeaux - it also reminds us that the underlying causes of false convictions remain and cannot be remedied solely by DNA.