Most people get their understanding of forensic science from crime shows, such as CSI, or movies.
The impression that’s given is that forensic science is incredibly accurate and reliable. For instance, most people believe that fingerprinting analysis is a reliable way to determine whether someone touched an item. Or they believe that shoe print analysis can help determine whether someone walked in a particular place.
Such beliefs are fine in a crime drama, where no real lives are at stake.
But in a criminal case – in a courtroom – the belief that fingerprints and shoe prints are reliable forensic tools can be catastrophic.
One of the reasons why such analysis is allowed into a courtroom is that judges are usually ill-equipped to determine whether a particular forensic science is reliable. So they defer to the people trained in that science.
And those people suffer from a separate problem. Having spent weeks, months, or years as a fingerprint analyst, a fingerprint analyst is unlikely to say that his field is unreliable.
Consequently, the very people who testify in a courtroom for the prosecution (and even people testifying for the defense) are same people who will say that their testimony is reliable. It’s a circular pattern that, unfortunately, permits juries – who usually know very little about these subjects – to hear about what is essentially unproven forensic science.
Take fingerprint analysis, which until the advent of DNA, was a major way (in addition to serology/blood typing) that investigators linked a defendant to a crime scene.
The National Academy of Sciences issued a report in 2009 pointing out that the evidence to support the reliability of many techniques, including most kinds of fingerprint analysis, is lacking.
As a result, there has been little rigorous research to investigate how accurately and reliably many forensic science disciplines can do what they purport to be able to do. In terms of a scientific basis, the disciplines based on biological or chemical analysis, such as toxicology and fiber analysis, generally hold an edge over fields based on subjective interpretation by experts, such as fingerprint and toolmark analysis. And there are variations within the latter group; for example, there is more available research and protocols for fingerprint analysis than for bitemarks.
In addition to investigating the limits of the techniques themselves, studies should also examine sources and rates of human error, the report says. As part of this effort, more research should be done on “contextual bias,” which occurs when the results of forensic analysis are influenced by an examiner’s knowledge about the suspect’s background or an investigator’s knowledge of a case. One study found that fingerprint examiners did not always agree even with their own past conclusions when the same evidence was presented in a different context.
So the next time you hear about fingerprint analysis, be very skeptical about whether the analysis shows what it purports to show.