In academic circles and public interest groups, the reliability of eyewitness testimony has been under fire for some time. Psychologists have been studying the procedures used to identify suspects and the overall reliability of eyewitness accounts; the most prominent of these is Gary Wells of Iowa State University. The research indicates that individuals generally are poor observers; for example, eyewitnesses tend to be more certain about identification at trial than at initial interview. Procedures have a large impact on the reliability of eyewitness identification—e.g. , sequential identification is more reliable than the traditional all-at-once line-up; line-ups work better when the officer in charge does not know who the suspect is. The unreliability of this type of testimony has led several states to abolish or alter their use in cases involving the death penalty. For example, in Maryland, the state legislature changed their penal code to require DNA evidence or video of the crime or confession before allowing prosecutors to seek the death penalty. New Jersey and Illinois simply abolished capital punishment.
Clearly, the stakes are higher in a capital case than any other criminal prosecution. However if eyewitness testimony, based upon psychological understanding of the process of identification, is so questionable, why is it generally discussed only in the context of capital cases? If a procedure or method is deemed capricious or arbitrary, isn’t the due process clause triggered for all criminal defendants accused or convicted based upon eyewitness testimony?
The Supreme Court and the criminal justice system have long recognized that additional safeguards are necessary in capital cases (recall Furman v. GA 408 U.S. 238 (1972) and Gregg v. GA 428 U.S. 153 (1976)). The Court, though, has also recognized the deprivation or the potential deprivation of liberty as triggering other fundamental rights of due process, such as the right to counsel throughout the criminal justice process (see Duncan v. LA 391 U.S. 145 (1968); Argersinger v. Hamlin 407 U.S. 25 (1972), In re Gault 387 U.S. 1 (1967), Hamilton v. AL 368 U.S. 52 (1961), US v. Tucker 404 U.S. 443 (1972), Miranda v. AZ (384 U.S. 436 (1966)), among others.)
“Do the due process protections against unreliable identification evidence apply to all identifications made under suggestive circumstances or only when the suggestive circumstances were orchestrated by the police?”
This issue, though, does not address the fundamental question regarding eyewitness testimony. If this type of testimony is unreliable under the best of circumstances, if the use of eyewitness testimony can lead to arbitrary or capricious results, why are we only concerned with capital cases or cases with potential police misconduct? The costs of loss of liberty to an innocent person are also high and false conviction can haunt an individual well after a sentence is served. The consequences of false conviction can include possible loss of the right to vote, difficulty in obtaining employment, etc. What level of tolerance does or should the criminal justice system have for error? Wells found that an 18% error rate in line-up identification via computer when all suspects are shown at once; the error rate drops to 12% when the photos are shown one at a time. If a six percent reduction in error is enough to mandate a change in state criminal justice procedures, are we comfortable with a 12% error rate in eyewitness testimony? What does due process require of the criminal justice system now that we are aware of the deficiencies of this type of testimony?
 The recent controversy over the Troy Davis’ execution lead to some media coverage of this issue.
 The Court of Appeals for the Seventh Circuit accepted Wells as an expert witness along with his methodology in an appeal in 2003 (319 F.3d 301).