15 Pages Posted: 29 Aug 2006
Date Written: August 30, 2006
In an ideal world, juries would always reach the correct result. In theory, we believe that the second best choice is to err on the side of acquitting the guilty rather than convicting the innocent. We say that it is better to acquit ten guilty men than convict one who is innocent. But I'm not sure that we really believe it. Would we really let ten child molesters walk the street to avoid convicting one innocent one? I have my doubts.
Although we think the system is tilted to protect defendants, it may not be. Juries may not really believe in the "presumption of innocence." Furthermore the prosecutor usually has far more resources than the defense. Searches have to be reasonable, but at least the government can conduct them. The defense cannot. More generally, the prosecution has a professional police force investigating for it, and greater access to forensic testing. If the prosecutor wishes to frame a suspect (which fortunately is not the norm) it may not be all that difficult.
It's hard to know which way juries in fact err more. Some high profile cases (John Hinckley, O.J. Simpson) may be instances in which a guilty man was acquitted. But recent studies have identified a huge number of wrongful convictions. I'm not persuaded that we really are acquitting more guilty people than convicting those who are innocent. There is probably no value in redefining reasonable doubt, but we can examine and try to minimize the causes of wrongful conviction.
False identification is one of the largest problems. Of course, we can't disallow eye-witness identification. But we can alter procedures to make them more reliable. One is to show pictures and suspects individually so that the witness will not assume that the array includes the culprit. Another is to have the lineup conducted by someone who doesn't know who the suspect is to avoid intentional or unintentional tipping of the witness.
False confessions are another problem. Perhaps the best remedy is to videotape all confessions so that defense counsel and the court (jury?) can see how the confession was obtained, and perhaps defense counsel can expose its probable inaccuracy.
Bad forensics can be either intentionally or unintentionally done. In the former case, severe sanctions should be available, and, of course, the conviction so obtained should be set aside. With unintentional botched forensics, defense counsel simply need to be trained to spot the problem and counter the bad forensics with their own better forensics.
Perhaps the most pernicious source of false convictions are so-called jailhouse confessions. These occur when an inmate claims to have overheard the defendant describe and confess to the crime. The informant is usually paid in the form of cash, dropped charges, and/or a reduced sentence. Obviously the probability of perjury is extremely high in such a circumstance. I would seriously consider disallowing this type of testimony entirely because of the high likelihood that it is false. Short of that, I'd at least require a cautionary instruction to the jury to be wary of the probability of falsity.
I conclude with four suggestions that are predicated on the reality that wrongful convictions happen. (1) There should be innocence commissions set up similar to the British model. (2) Defenses should not be artificially limited. For example, pending the outcome of a U.S. Supreme Court case, a State can (and some do) deny the defendant the opportunity to present evidence that somebody else committed the crime. (3) As long as we know there are mistakes, capital punishment should be abolished (as most of the civilized world has). And (4) parole should not be contingent on a person's admitting his guilt. This presents an untenable dilemma for an innocent person, and may actually cause him to spend more time in prison than a similarly-situated guilty one.
Suggested Citation: Suggested Citation
Loewy, Arnold H., Systemic Changes that Could Reduce the Conviction of the Innocent (August 30, 2006). UNC Legal Studies Research Paper No. 927223. Available at SSRN: https://ssrn.com/abstract=927223 or http://dx.doi.org/10.2139/ssrn.927223